Skip to main content

Full text of "Intrinsic motivation : the effects of task choice, reward magnitude and reward choice"

See other formats



(3x UBMS 
wmbmw 







Digitized by the Internet Archive 
in 2019 with funding from 
University of Alberta Libraries 


https://archive.org/details/Hallschmid1977 











THE UNIVERSITY OF ALBERTA 


release FORM 

NAME OF AUTHOR: Claus A. Hallschmid 

TITLE OF THESIS: Intrinsic Motivation: The Effects 

of Task Choice, Reward Magnitude 
and Reward Choice 

DEGREE FOR WHICH THESIS WAS PRESENTED: Doctor of Philosophy 
YEAR THIS DEGREE GRANTED: 1977 

Permission is hereby granted to the UNIVERSITY OF 
ALBERTA LIBRARY to reproduce single copies of this 
thesis and to lend or sell such copies for private, 
scholarly or scientific research purposes only. 

The author reserves other publication rights, and 
neither the thesis nor extensive extracts from it may 
be printed or otherwise reproduced without the author* 
written permission. 













THE UNIVERSITY OF ALBERTA 

INTRINSIC MOTIVATION: THE EFFECTS 
OF TASK CHOICE, REWARD MAGNITUDE 
AND REWARD CHOICE 



by 


CLAUS A. HALLSCHMID 


A THESIS 

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES AND RESEARCH 
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE 

OF DOCTOR OF PHILOSOPHY 

DEPARTMENT OF EDUCATIONAL PSYCHOLOGI 


EDMONTON, ALEERTA 
FALL, 1977 













































UNIVERSITY OF ALBERTA 


FACULTY OF GRADUATE STUDIES AND RESEARCH 


The undersigned certify that they have read, and 
recommend to the Faculty of Graduate Studies and Research, for 
acceptance, a thesis entitled "INTRINSIC MOTIVATION: THE EFFECTS 
OF TASK CHOICE, REWARD MAGNITUDE AND REWARD CHOICE," submitted by 
CLAUS A. HALLSCHMID in partial fulfillment of the requirements 
for the degree of Doctor of Philosophy. 





ABSTRACT 


Attribution theory proposes that the larger an 
extrinsic monetary reward the more likely the subject's 
attention will become focused on that reward (Lepper £- 
Greene, 1976). A second proposition is that any externally 
imposed condition (i.e. deadline, reward choice) will also 
result in a perceived shift in locus of causality to the 
external condition. In the present study both the effects of 
a larger versus a smaller monetary reward and the effects of 
reward choice on subsequent intrinsic motivation were 
investigated. More speficicaily, the present study examined 
the effects of individually determined task choice, reward 
magnitude, and reward choice on subsequent intrinsic 
motivation. Three seperate studies were conducted. 

The first study was designed to (a) establish a 
behavioral norm of intrinsic motivation on five experimental 
task, ib) to allow comparison of this norm to high and low 
task motivation levels determined by means of a paired- 
comparison procedure, and (c) to test the relationship 
between additudinal measures and a behavioral measure of 
intrinsic motivation. The results indicated first that 
although high and low task choice did not differ 
significantly on the behavioral measure from the previously 
established norm the two motivational levels did differ 
significantly from each other in opposite directions from 
the norm. This established the validity of the paired- 
comparison procedure to provide individually determined 


IV 
















levels of intrinsic task motivation. Secondly, the various 
self-report measures showed only a moderately low 
correlation with task persistence. Furthermore, a factor 
analysis indicated that the attitude measures may not be 
part of the same factor structure as the behavioral measure. 

The second study explored the effects of pre-testing on 
both behavioral and attitudinal measures of intrinsic 
motivation. The results indicated that two of the attitude 
measures were significantly affected by pre-testing. 
However, the task persistence measure remained relatively 
immune to any pre-test effects. 

The third study tested the effects of reward magnitude 
and individually determined monetary reward choice on 
subsequent intrinsic motivation for a task by manipulating 
both intrinsic task motivation and extrinsic reward as 
independent variables. In two reward conditions subjects 
were given a high ($2,00) and a low (250) monetary reward 
for working on an experimental task of either high or low 
perceived intrinsic motivation for a 10-minute period. In a 
third condition subjects were allowed to choose the monetary 
reward at the end of this 10-minute period. Their subsequent 
intrinsic task motivation, defined as task persistence, was 
measured unobtrusively during a second 10-minute free-choice 
interval and compared to high and low motivation control 
groups. The results showed that reward magnitude was not, 
but that individually determined reward choice was, 
significantly related to subsequent impairment of intrinsic 


v 








.. 




'c - H I I I H 












task motivation. These findings were discussed in relation 
to both attribution theory and some alternative theoretical 
expla nations. 


vi 















































ACKNOWLEDGEMENTS 


I am grateful to the many persons who assisted me in 
completing this study. 

In particular, I wish to express my sincere 
appreciation to the following: 

Dr. George Fitzsimmons, my supervisor whose analytic 
ability, guidance, and patient understanding of my personal 
idiosyncracies made it possible to see the study to its 
conclusion. 

Dr. John Osborne, whose probing and critical attitude 
provided sharp theoretical understanding into various 
aspects of the study. 

Dr. Robert Mulcahy whose insights into the nature of 
the study increased the quality of the final product. 

Dr. Micheal Ross, Dept, of Psychology, University of 
Waterloo, whose insights and suggestions, both through 
personal contact and his publications, made valuable 
contributions to this study particularly during its final 
stages. 

Dr. Charles Lee, Chairman, Dept. of Organizational 
Analysis, Faculty of Business Administration and commerce. 
University of Alberta, whose cooperation and assistance made 
the completion of the study less difficult. 

Dr. Dong-Yul Lee and Ajit Mohanty, collegues and 

vii 
























■ 











































friends, who on many occasions provided the personal support 
that only friends can. 

My wife Erika and children, Michelle and Peter, whose 
love and understanding, motivated me throughout the years. 

Thank you, for providing an enriching experience! 


C.A.H. 


vm 
















' 


. . . 



TABLE OF CONTENTS 


Page 

I INTRODUCTION 1 

II REVIEW OF THE LITERATURE 6 

THEORITICAL FORMULATIONS 6 

Cognitive Evaluation Theory 7 

Attribution Theory 8 

The Concept of Behavioral Justification 10 

Inequity Theory 13 

The Competing Response Hypothesis 15 

EVALUATION OF THE THEORIES 17 

SELECTIVE REVIEW OF THE EMPIRICAL LITERATURE 20 

Early Studies Investigating the Hypothesized 

Shift in Locus of Causality 21 

Contingent Vs Non-Contingent Rewards 23 

Timing of Reward 23 

Expectation of Reward 25 

Task Performance and Task Intrinsic 

Motivation 28 

The Content-Consequence Distinction 29 

Negative Extrinsic Reward 29 

Intrinsic and Extrinsic Motivation as 

Independent Variables 30 

Reward Saliency 32 

Performance Contingent Rewards 33 

Non-Tangible Reward Conditions 35 

EVALUATION OF EMPIRICAL RESEARCH 36 

Reward Salience and Reward Magnitude 36 

Reward and Task Choice 43 

Dependent Measures of Intrinsic Motivation 45 

III RATIONALE, DESIGN, AND HYPOTHESIS 48 

RATIONALE 48 

DESIGN 52 

HYPOTHESIS 54 

Individual Task choice 55 

Reward Magnitude 56 

IV METHOD 58 

SUBJECTS 58 


IX 























■ 

. 






















Page 

PILOT STUDY 58 

STUDY 1 59 

Procedure 60 

Dependent Variable Measures 64 

STUDY 2 65 

Procedure 66 

Dependent Variable Measures 67 

STUDY 3 67 

Procedure 68 

Dependent Variable Measures 69 

V INDIVIDUAL TASK CHOICE: RESULT AND DISCUSSION 70 

RESULTS 70 

DISSUSSION 73 

VI TEST-RE-TEST EFFECTS: RESULTS AND DISCUSSION 76 

RESULTS 76 

DISCUSSION 78 

VII REWARD MAGNITUDE AND REWARD CHOICE: RESULTS 

AND DISCUSSION 80 

RESULTS 80 

DISCUSSION 88 

VIII CONCLUSIONS 100 

REFERENCES 107 

APPENDIX 1 112 

APPENDIX 2 116 

APPENDIX 3 122 

APPENDIX 4 131 

APPENDIX 5 139 

APPENDIX 6 148 

APPENDIX 7 157 


X 



























































LIST OF TABLES 


TABLE 1 


TAELE 2 


Page 


. MEAN TIME IN SECONDS OVER A 10-MINUTE 

FREE TIME PERIOD 83 

. SUMMARY ANCOVA TABLE FOR TASK PERSISTENCE 

DATA: 2 (GROUPS) x5 (LEVELS) 84 


xi 












































LIST OF FIGURES 


Page 


FIGURE 1. KEAN V-ALUES IN SECONDS FOR REWARD 


CONDITION GROUPS 


86 










































CHAPTER I 


INTRODUCTION 

Currently, we are witnessing an increasing interest in 
the theoretical issues surrounding the attributional 
analysis of motivation (Weiner, 1970; Weiner & Kukla, 1970). 
Attributional reformulations are generating research in a 
number of diverse areas ranging from skill-oriented 
performance tasks (Miller & Ross, 1975) and interpersonal 
influence studies (Beckman, 1973) to intrinsically motivated 
task performance (Kruglanski, 1975). With respect to 
intrinsically motivated task performance, de Charms (1968) 
originally concluded that different types of motivation are 
non-additive. That is, motivation to perform a task is not 
necessarily the greatest when an individual performs an 
activity both for its own sake and as a means to obtain a 
reward. This core hypothesis has generated considerable 
research on diverse aspects related to the hypothesized non¬ 
additive effects of extrinsic rewards on intrinsic 
motivation. 

Lepper and Greene (1976) in an important conceptual 
paper provide strong argument for the heuristic and 
theoretical value of the distinction between two broad 
classes of motivation: extrinsic and intrinsic motivation. 
At the base of this internal-external partition is the 
assumption that a person is intrinsically motivated if he 
perceives himself performing an activity for no apparent 


1 


















' 

, 





2 


reason except the activity itself (Deci, 1971). Extrinsic 
motivation on the other hand, refers to the performance of 
an activity because it leads to an external reward (e.g. 
status, money, verbal approval). Therefore intrinsic 
motivation is often equated with a uniqueness dimension for 
a particular individual, whereas external motivation is 
equated with non-unigueness (Kruglanski, 1975). Although the 
operationalization of intrinsic and extrinsic motivation 
poses a problem some important issues have been raised with 
respect to the hypothesized non-additive relationship 
between the two. 


Cognitive evaluation theory (de Charms, 1968; Deci, 
1971) suggests that this non-additive relationship can be 
explained in terms of the individuals perception of what 
causes his behavior (Locus of Causality) . That is, the 
association of an extrinsic reward to an already 
intrinsically satisfying task has the effect of changing the 
perceived locus of causality from internal to external and 
thereby reducing the motivation for the task. Therefore, the 
motivation to perform a task solely on the basis of its 
intrinsic task satisfaction will suffer if an extrinsic 
reward is introduced. The basic assumption is that we infer 
our beliefs from observation of our own behavior and the 
circumstances under which it takes place. The individual's 
perceptions of the causes of his behavior becomes the major 
unit of analysis in cognitive evaluation theory. It is 
assumed therefore, that if the person attributes his 




. 


























3 


behavior to intrinsic factors that he then behaves as if he 
is intrinsically motivated and indeed we can say he is 
intrinsically motivated. Therefore, self-perceptions are 
assumed to have important implications for an individual's 
behavioir. Whatever the objective cause of behavior, if an 
individual perceives himself performing an activity for its 
own sake, he is likely to behave as if he were intrinsically 
motivated and conversely, if he perceives himself performing 
an activity for the external consequences attached to the 
activity he is likely to behave as if he were extrinsicaily 
motivated (Ross, 1976). Major emphasis is placed on what 
causes these perceptions, what modifies them, and what the 
behavioral consequences are once they are assumed. 

In the empirical research most studies investigating 
this conceptualization use a similar experimental design -- 
an extrinsic reward is introduced to an already interesting 
task, and then one or more dependent variable measures of 
intrinsic motivation are assessed. Thus, a basal level of 
intrinsic motivation is determined, a treatment is given, 
and the post-experimental level of intrinsic motivation is 
measured. Measures of intrinsic motivation as dependent 
variable measures have included task persistence (Deci, 
1971,1972a), willingness to participate in a similar 
experiment (Calder & Staw, 1975b; Kruglanski, Riter, Amitai, 
Margolin, Shabtai, S Zaksh, 1975), reported task interest 
and enjoyment (Deci, 1971; Calder & Staw 1975b), and free¬ 
time activity (Deci, 1971; Ross, 1975). However to determine 








’ 


4 


the initial level of intrinsic motivation all of the studies 
to date have employed group validated task motivation level. 
That is, through pre-testing a task was equated on all 
dimensions except the affective dimension of perceived 
liking or enjoyment, which was manipulated as the 
independent variable. Within this paradigm Deci (1971) 
compared subjects who were paid to engage in an activity 
during the second of three work periods with control 
subjects who were not paid. Subjects who had been paid 
showed a greater decrease in performance from the first to 
the third period than control subjects. Apparantly, payment 
decreased the subjects' interest in performing the activity 
for its own sake. Calder and Staw (1975b) tested this 
hypothesized relationship more directly by manipulating both 
intrinsic and extrinsic motivation as independent variables 
and measuring the effect of their interaction on both task 
satisfaction and task persistence. The results revealed a 
significant decrease in intrinsic motivation in the presence 
of extrinsic reward when the initial level of intrinsic 
motivation was high. However this finding applied only to 
the task satisfaction measure not the persistence measure. 

With respect to variables mediating this relationship, 
other studies have reported that several variables influence 
the interaction between extrinsic and intrinsic motivation: 
expectancy of reward (Deci, 1972b; Greene & Lepper, 1974; 
Kruglanski, Alon, & Lewis, 1972; Lepper, Greene & Nisbett, 
1973), magnitude of task intrinsic reward (Kruglanski, 


. 
























5 


Riter, Arazi, Agassi, Montiqio, Peri & Peretz, 1975), 
contingent versus non contingent rewards (Deci, 1972b; 
Kruglanski, Freidman 8 Zeevi, 1971) , saliency of rewards 
(Ross, 1975a), perceptions of effective control (Karniol 8 
Ross, 1975), delay of gratification (Ross, et al, in Press), 
and success contingent rewards (Ross, 1975b). Although most 
of these studies provide fairly consistent support for the 
hypothesis that intrinsic and extrinsic motivation interact 
non-additively, several issues and questions still remain 
untested. 

The purpose of the present study is to investigate and 
test: (1) the effects of individually determined level of 
intrinsic motivation, rather than group validated motivation 
level, (2) establish a behavioral task norm of intrinsic 
motivation and determine whether individually determined 
task choice results in significant deviation from this 
behavioral norm, (3) determine the underlying factor 
structure of various measures of intrinsic motivation, and 
(h) test the effects of reward magnitude and reward choice 
(self-reward) on subsequent intrinsic motivation. Thereby, 
the present study addresses itself to a number of issues and 
untested predictions derived within an attributional 


framework 






















' 









* 










CHAPTER II 


REVIEW OF THE LITERATURE 

THEORET ICAL FORMULATIONS 

The theoretical distinction between intrinsically and 
extrinsically motivated behaviors has introduced a number of 
theoretical issues. Generally these issues have centered 
around the conceptual criteria of internality and 
externality in motivation. The identification of intrinsic 
motivation with factors internal to the individual and 
extrinsic motivation with factors external to the individual 
seems to have brought a certain degree of confusion into the 
conceptualization of intrinsic versus extrinsic motivation. 
The theoretical relationship between these two classes of 
motivation is largely dependent on how internality and 
externality is conceptualized. To depict the nature of this 
relationship a number of speculative frameworks have been 
advanced. Common to these is the notion that the perception 
of the causes of behavior (self-attribution) plays a central 
role in the motivation process since the allocation of 
causal responsibility for behavior is assumed to guide 
subsequent behaviors. For example, the belief that an event 
is over-justifiabiy caused (more than one valid cause) may 
lead to different attitudes and actions than the belief that 
the event is justifiably (one valid cause) or 

under justifiably caused (no valid cause). 


6 




: *||| 7 * 


7 


Cognitiv e Evaluation Theory 

de Charms (1968) has developed a phenomenological 

theory of personal causation predicated on the assumption 

that mastery of certain situations is valued for its own 

sake rather than the consequences it produces. Central to 

this conceptualization is the concept of personal causation. 

Calder and Staw (1975b) write: 

For the Charms, the crux of the distinction 
between intrinsic and extrinsic motivation 
stems from the feeling or perception of 
personal causation. Satisfaction derives from 
an activity which is perceived as intrinsically 
motivated because of a person's need to feel a 
sense of personal causation in his or her 
actions (p. 599). 

de Charms also utilizes Rotter's (1954) concept of locus of 
control. However, he stresses the notion of instrumental 
feedback inherent in the activity whereas Rotter stresses 
the idea of reward as a conseguence of the activity. Thus, 
the perception of the cause of behavior results in the 
feeling of being either an Origin or Pawn, depending on 
whether the locus of causality resides in the self or 
outside the self in the environment. When the locus of 
causality is perceived within the self, the individual 
considers himself to be intrinsically motivated (Origin) . 
When the locus of causality is perceived outside the self in 
the environment, the individual considers himself to be 
extrinsically motivated (Pawn). Within this conceptual 
framework, extrinsic reward is thought to interact with 
intrinsic motivation by affecting a change in the perceived 
locus of causality with a subsequent change in feelings from 













0} Lz-ita&j ,aeoofioi.-i -1 '&*-:>£? v ^ r t-t *tti* - -’•• 5 * s2 

' 






, 


8 


Origin to Pawn. 


Deci (1971,1972a,1972b,1973), extending de Charms' 
theory further stresses the need for self-determination and 
competence in intrinsically motivated behaviors. He suggests 
that there are two processes by which extrinsic rewards can 
affect intrinsic motivation: a change in perceived locus of 
causality, and a change in one's feelings of competence and 
self-determination. These two processes are initiated at 
different times since every reward has two aspects: (1) a 
"controlling" aspect, which initiates change in the 
perceived locus of causality, and (2) an "information" 
aspect which provides knowledge to the person about his 
competence and self-determination. Which process is invoked 
depends largely on which aspect of the reward is more 
salient. 

Att ri butio n Theor y 


Kruglanski, et al (1975b) have taken issue with the 

above conceptualization of intrinsic motivation: 

The latter formulation has several 

implications; (a) its identification of 
intrinsic motivation with a specific need 
(reminiscent of the postulation of intrinsic 
and extrinsic needs by some industrial 
psychologists) removes the analysis from the 
original attributional level impartial to the 
type of motive; (b) it seems to misidentify 
internal attributions of causality with the 
need to make internal attribution (or the need 
of self-determination); (c) it seems to 

suggests that motives outside that of self- 
determination could not motivate activities in 
an intrinsic way (p. 745). 


To deal with these issues Kruglanski, et al (1975b) adopt a 













v y 1 


* 



























- 


9 


purely attributional level of analysis. The crux of this 

theory is a content-consequence distinction. A person 

considers his behavior intrinsically motivated when he 

perceives the cause of his behavior to reside in the content 

of the tasK; a person considers himself extrinsically 

motivated when he perceives the cause for his behavior to 

reside in the consequences of a task. In terms of this 

content-consequence distinction, when an extrinsic reward is 

introduced to an already intrinsically rewarding activity, 

it is thought to change the perception of causality from 

content to consequence. Thus Kruglanski (1975a) redefines 

the internal-external dimensions: 

Let the term endogenous attribution of an 
action denote the case wherein an action is 
attributed to itself as a reason. In other 
words, an endogenousey attributed action is an 
end in itself .... Let the term exogenous 
attribution of an action denote the case where 
the action is not an end on itself but rather a 
means that mediates a further goal, one 
exogenous to it (p. 390). 

As Kruglanski states, the conceptual criterion for this 
exogenous-endogenous distinction is relatively 

straightforward, as it is based on a common differentiation 
between means and ends. In contrast, the previously defined 
internal-external distinction (de Charms, 1968; Deci, 1971) 
does not provide a theoretically consistent means or 
criterion for designating the contingent monetary reward as 
the external cause of action. 


Lepper (1976) and Calder and Staw (1975b), taking into 
account the effects of boundary variables , view the self- 








. 

. 




' 


10 


perception process in terms of positive, negative, or 
neutral affect associtated to various means and ends of an 
activity. Different self-perceptions, which subsequently 
guide behaviors, result depending on the nature of the 
affect associated to an activity or its consequences. 
Therefore, the individual considers himself extrinsicaliy 
motivated when the ends of an activity are positive while 
the means are perceived as negative or neutral. On the other 
hand, the individual considers himself intrinsically 
motivated when the ends of an activity are perceived as 
negative or neutral while the means are perceived as 
positive. When both means and ends are perceived as positive 
an ••unstable" attributional situation is created which may 
be resolved by a shift in the perceived liking or 
satisfaction attached to the task. This attribution process 
is hypothesized to be responsible for the reduced subsequent 
intrinsic task motivation. 

The Conc ept of B ehavioral Justific a tion 

Compliance research in the area of insufficient 
justification provides strong evidence that subjects will 
develop internal justification or control when there is 
insufficient external justification or control. Festinger 
and Carlsmith (1959), in their now classic study, found that 
subjects who received insufficient external reward ($1) grew 
to like the experimental task more than those who received 
sufficient rewards ($20). The insufficiently rewarded 
subjects developed internal justification, or a more 














* 






. 








. 


Jtt* 





' 


I 





11 


favourable attitude toward the task. 


On the other hand, in the cognitive evaluation studies, 
interesting experimental activities are employed allowing 
subjects to be sufficiently justified even when no rewards 
are given. Oversufficient justification occurs when a reward 
is given for an already interesting activity. Therefore, 
both reward and task justify the individual's behavior. As 
hypothesized in attribution theory, liking for the task and 
subsequent intrinsic motivation will decrease. Nisbett and 
Valins (1971) suggest that these two processes are not the 
same. Nonetheless, despite the fact that these two phenomena 
may have different underlying psychological processes, Deci 
(1972) suggests they still may be quite complementary. 


Gerard, Connolley, and Wilhelmy (1974) in this 

tradition present a single-factor theory with which all the 

forced compliance research in both insufficient and 

oversufficient conditions is to be understood: 

The framework, depicts the relationship between 
•'resultant justification" and both dissonance 
and "sweetness" effects on cognitive change. By 
resultant justification we mean the difference 
between all the various justifications for the 
complaint act and all the justifications 
(reasons) against performing it. Resultant 
justification is shown on the abscissa going 
from very low to very high levels. When the 
resultant justification for a given act (such 
as complying with the experimenter's request to 
taste bitter solutions) is very low, the person 
will not choose to engage in the act. For any 
given action alternative, there is some point 
at which resultant justification is just barely 
sufficient for the person to choose that 
alternative -- his "action" threshold. 
Dissonance-induced cognitive change will be 
maximum at this threshold. To the extent that 









12 


resultant justification is greater than the 
minimum required to elicit the decision, 
dissonance will be lower. As we move along the 
abscissa to the right, we finally reach the 
"comfort" threshold, a point at which the 
relatively small amount of inconsistency can be 
tolerated with no accompanying tension to 
reduce it. This tolerance level is the 
individual difference variable we referred to 
above. Following to the right, we reach a point 
of zero inconsistency where the negative 
behavioral consequences are just balanced by 
positive justification (positive anticipated 
behavioral consequences). As we move beyond the 
trough, the resultant justification is 
positive. The "sweetness" threshold is the 
point at which the amount of net positive 
justification begins to enhance or "sweeten" 
the behavior -- the so called incentive effect. 
By a secondary reinforcement -- like process, 
oversufficient justification colors the view 
the person has of his choice in such a way that 
its value is enhanced. In general, the 
resultant justification for a given alternative 
reflects the spread in choice 


This model is similar on the surface only to Calder and 
Staw*s (1975b) means-ends analysis conception since it is 
based on dissonance theory. Ail factors affecting attitude 
change do so through their effects on resultant 
justification for behavior. Similarly in attribution theory 
the affects attached to tasks and their consequences change 
attitudes towards the tasks through the hypothetical means- 
ends analysis. Thus, resultant justification plays a key 
role in both conceptions. The degree of resultant 
justification is determined from a combination of both 
intrinsic and extrinsic factors. Intrinsic justification may 
thus be regarded as synonomous to intrinsically motivated 
behavior. For example, when the individual perceives himself 
to be motivated by intrinsic factors (means) or similarly. 


1 






















13 


when he attributes locus of causality for his behavior to 
the interest he has for the task itself he is said to be 
intrinsicaliy motivated. However, this represents a 
theoretical departure from the forced compliance research 
which is based on cognitive dissonance theory and utilizes 
the concept of an underlying motivational tension state 
(Festinger, 1957) . 

A prediction derived directly from dissonance theory is 
that the larger a reward received for an activity the less 
the resultant cognitive change (Brehm & Cohen, 1962). On the 
other hand, both cognitive evaluation theory and attribution 
theory 'introduce the concept of an "unstable" attributional 
state, where behavior is perceived as over- justifiably 
caused (more than one valid cause). The prediction derived 
from this conceptualization is that the latger a reward 
received for an activity th e more l ikely that cognitive 
change will occur. Both within over- and under^justification 
research however, reward magnitude is considered to be an 
important variable although the specific predictions rest on 
fundamentally distinct processes. 

Ine qui ty T heory 

In essence Adam's theory of inequity (1963) states that 
individuals make comparisons of their behavioral efforts 
with those of others in similar situations. Equity exists 
when the perceived input (effort) to output (reward) ratio 
is equivalent to the ratio of others. Inequity results when 
these ratios are not equivalent. The perceived inequity in 

















. 




. 







14 


such situations results in a hypothesized tension state to 
restore equity. A number of hypothesized alternatives are 
available: (1) increase or decrease the outcomes (rewards), 
or (2) increase or decrease the inputs by varying effort. 
Most research employing predictions testing the resolution 
of inequity have had considerable methodological 
difficulties. However, in general there exists some support 
for the hypothesis that subjects increase their efforts to 
bring inputs and outputs into balance and thus reduce 
feelings of inequity (Goodman S Freidman, 1971). In the 
analysis of over-and-under sufficiency conditions ineguity 
theory would predict a reduction in perceived inequity as 
reward magnitude increased to an optimal point. Thereafter, 
inequity would again arise as a result of an 
over justification effect. Also, inequity theory is based on 
subjective comparisons of input-output ratios between self 
and other and therefore not on a purely self basis. This 
analysis suggests that individuals would reduce their 
behavioral efforts (outputs) in order to balance the self- 
other ratios when deprived of reward in relation to others. 
This also has particular relevance to the present study 
since inequity theory represents another tension-state 
theory raised as an alternative to the attribution 
hypothesis (Deci, 1972). 

However, there are as yet two unresolved methodological 
issues in inequity research: (1) the problem of devalued 
self-esteem, and (2) the problem of induction. The former. 






. 

* ew>? <«*>* i.leixfi »Mti Si §£ttlj| I lBP i ’ 

. 


) 










15 


Goodman 6 Freidman (1971) state, relates to the fact that 
most of the studies supporting the inequity hypothesis have 
not empirically determined whether self-devaluation or 
inequity feelings account for the observed variances in 
behavior. The latter, relates to the method of inducing the 
subject, to perceive that his outcomes are related to his 
inputs, which is necessary for inequity feelings to occur. 
Thus, when there does not exist a perceived connection 
between outputs and inputs no observed behavior changes 
should occur. 

Deci (1972) states that an internal standard or 
"compari'son other" may also be used within the framework of 
inequity theory. Based on Pritchard's (1969) work, he 
suggests that inequity feelings may arise from a 
disproportionate relationship between person's own inputs 
and his outputs. Deci explains that predictions from 
cognitive evaluation theory seem diametrically opposite to 
predictions based on equity theory. That is, "The cognitive 
evaluation theory prediction, is that he would be less 
likely to continue to work (p. 114)." Deci then proposed a 
conceptual integration of these two theories, and suggests 
that whether or not a person has feelings of inequity his 
intrinsic motivation is still reduced in the presence of an 
external reward. 

The Comp et ing Respo n se Hypothes is 

The fundamental proposition of this hypothesis, 
introduced by Reiss & Suschinsky (1975a # 1975b,) as a viable 




. 

' 

: -• 

' 





















16 


alternative to the attributional model, is that 
environmental stimuli (e.g. reward, distractors, etc) can 
impair or facilitate intrinsic motivation by eliciting more 
or less enjoyable ways of performing an activity relative to 
a control condition. Therefore "any responses that 
facilitate task enjoyment are termed competing responses (p. 
235)". The introduction of an expected reward in the 
presence of an already interesting activity can arouse many 
competing response effects (e.g. performance anxiety, 
frustrative delay of reward, distraction effects, hurried 
performance, etc). The greater the extent to which such 
competing responses are aroused during training, the greater 
is the subsequent loss in intrinsically motivated responses 
in the experimental activity. Keiss & Suschinsky (1975b) 
also postulate three major processes by which loss in 
intrinsic motivation can be accounted for: (1) the arousal 
of unpleasant affect may lead to aversive Pavlovian 
conditiong (especially under moderate levels of frustrative 
delay of reward), (2) the subject's recognition that he did 
not enjoy the activity during the experimental session may 
mediate a relatively persistant effect, and (3) arousal of 
competing ' responses could impair performance quality and 
this could mediate a relatively persistant effect. With 
respect to the major competing response effects Logan (1976) 
concludes that," the major factor determining the 
detrimental effects of delayed reward is the behavior that 
occurs during the delay interval ip. 105)." This is 
precisely what the competing response hypothesis postulates 








. 

■ 


I 

. 

' 


17 


lies at the basis of the observed decrements in the 
intrinsic motivation studies. Differential hypothesis 
generated by attribution theory and competing response 
theory, have been tested by Ross (in press) and Reiss & 
Suschinsky (1975b) and will be reviewed in the following 
chapter. 

EVA LUATION OF THE TH EORIES 

The theories reviewed are cognitive theories since they 
resort to cognitive mediational explanations to account for 
empirical data. However, an important distinction can be 
made in their conceptulization of the motivation construct. 
Cognitive dissonance theory (Festinger, 1956) and equity 
theory (Adams, 1963) rely on a postulated tension-reduction 
process to account for both behavioral and attitudinal 
displacements. On the other hand, cognitive evaluation 
theory (Deci, 1971) and attribution theory (Kruglanski, 
1975), strictly speaking, do not postulate an aversive 
motivational drive. That is, attribution theory accounts for 
behavioral displacement in terms of the causal attributions 
an individual makes regarding his behavior. Often these 
attributions may be based on observable and salient stimuli 
outside the individual himself. Bern (1967) in an important 
conceptual reanalysis of diverse dissonance research 
suggests that more often than not publicly observable 
stimuli and behaviors are at the basis of the attributions 
made. Thus, an individual’s attitudes and behaviors may be 
viewed as inferences from observation of his own behaviors 











. 




18 


and the accompanying stimulus variables. 

In terms of the motivational analysis of under-and 
over-justification phenomena related to the introduction of 
extrinsic rewards, dissonance theory makes the less obvious 
and more subtle predictions within the underjustification 
condition and attribution theory makes the less obvious and 
more subtle predictions in the over justification 
conditions. Dissonance theory predicts less cognitive change 
with increasing reward magnitude and low task interest, 
whereas attribution theory predicts more cognitive change 
with increasing reward and task interest. Although the 
theories' postulate distinct and different underlying 
psychological processes, they may still be considered as 
complementary since each model seems to have relevance to a 
particular segment of the justification curve. The important 
distinction lies in the hypothesized underlying process. In 
the case of perceived overjustification attribution theory 
hypothesizes a shift in the perceived locus of causality, 
with a subsequent decrease in intrinsic motivation. The 
larger and more salient the extrinsic reward the more likely 
this decrease. In the case of perceived under justification 
dissonance theory hypothesizes an increase in intrinsic 
motivation with a subsequent decrease in experienced 
dissonance (i.e. tension). The larger the rewa-rd or extefnal 
justification the less the experienced dissonance with less 
subsequent cognitive change. 


Adams (1963) proposed a theory of inequity wich also 










i- p j t i A ' - i ' f - 5l * 1 

■ 




19 


relies on a postulated underlying tension state. 
Interrestingly, inequity theory makes predictions opposite 
from those of attribution theory within the 
over justification conditions. However, inequity theory also 
poses serious menthodological problems that have made it 
difficult if not impossible to test. Despite this, Deci 
(1972) has attempted a conceptional integration of the two 
models. 

Finally, the competing response hypothesis, forwarded 
by Reiss and Suschinsky (1975,1976) is based on the delay of 
reinforcement concept. Competing response effects are 
thought to be reinforced responses that are incompatible 
with the originally reinforced responses during the delay 
period. This results in the competing responses to be more 
strongly reinforced than the intended responses. 
Interestingly, Logan (1976) qualifies the delay of 
reinforcement concept by adding that language can offset the 
detrimental effect of delayed reward through secondary 
reinforcement to help bridge the time of delay of actual 
reward. This is particularly relevant to the present study 
since reward is not made success contingent but is allocated 
on the basis of performing the task without regard to 
quality of performance or degree of success on the task. 

In sum, a number of alternative explanations have been 
forwarded in the literature to account for the negative 
effect of extrinsic rewards on intrinsic motivation. Because 
of these divergent explanations and theoretical issues the 






















. 


20 


concept of intrinsic motivation as a psychological construct 
remains conceptually somewhat unclear. However, as Calder 
and Staw (1975b) state it still remains a usefull concept 
since it addresses the interaction of extrinsic and 
intrinsic variables from the point of view of the individual 
himself. That is, intrinsic and extrinsic motivation are 
defined in terms of the individual's subjective perceptions 
of personal causation. This was the original concept 
introduced by de Charms (1968); and subsequently this core 
idea has been extended and added to by various formulations 
as the concept of intrinsic versus extrinsic motivation came 
under empirical scrutiny. For example, Bern's (1972) self¬ 
perception theory suggests that the individual infers his 
internal state by observing his own behavior in the context 
in which it occurs. This leads to the individual labelling 
his behavior as internally caused in some situations and 
externally caused in other situations. The central question 
then becomes what this does to subsequent attitudes and 
behaviors. 

This question has led to the empirical investigation of 
variables that play a major determining role in self¬ 
perception processes, which up to this point have not been 
investigated in other conceptual systems. 

SEL ECT IVE REVIEW OF THE EMP IR ICA L LITERATUR E 

The following section is a selective review of 
empirical studies that have investigated various aspects of 
















. 



















21 


the intrinsic-extrinsic motivation question. Only those 
studies of particular relevance to the present investigation 
are dealt with in detail. With the exception of one early 
field study and several more recent nursery school studies 
all of the experiments were conducted in laboratory settings 
employing an essentially similar paradigm. The experimental 
time was devided into training or work periods and free-time 
periods. During the free time periods subjects could choose 
or not choose to work on the experimental task, usually in 
the presence of various distractors. Both behavioral and 
attitudinal measures usually served as the dependent 
variable measures. The early studies were designed to 
investigate the basic hypothesis formulated originally by de 
Charms (1968). This led to various modifications in the 
theoretical explanations and investigation of what Calder 
and Staw (1975a) call boundary variables. These were defined 
as parameters of the task environment and reward or other 
variables that have a delimiting effect on the hypothesized 
relationship between intrinsic and extrinsic motivation. 
More recently, specific issues have dominated the 
literature. These have resulted in differential predictions 
derived from alternative models to the original cognitive 
evaluation model. 

Earl y Studies Inv estigat i ng the Hyp ot hesize d Shift in Loc us 
of Cau sality 

Deci (1971,1972a) conducted two laboratory and field 
experiments to investigate the effects of external reward on 









*q*M**z xig mods* u i u+amtlW** '*>* ^ - J *' 

■ 






■ 





22 


intrinsic task motivation. In each experiment, subjects were 
requested to perform an activity over three different time 
periods (periods 1,2,3). The degree of intrinsic motivation 
was the amount of time the subject spent on the puzzle 
during an 8-minute free-choice situations as well as the 
subject's rating of the degree to which they found the task 
interesting and enjoyable on a Likert-type scale. External 
rewards were given to the experimental subjects during the 
second period only, while the control subjects received no 
reward. The differences in the experimental groups' 
intrinsic motivation between Period 1 and Period 3 relative 
to the differences in the control's indicated that: (1) with 
money as the external reward, intrinsic motivation 
decreases, and (2) with verbal reinforcement in the form of 
positive feedback as external reward, intrinsic motivation 
increases. 

To account for these findings Deci (1971,1972b) 
forwarded the hypothesis that a change in perceived locus of 
causality from external to internal cuased the reduced task 
persistance. More specifically, he asserted that extrinsic 
rewards have more salience than intrinsic rewards and thus 
"co-opt" intrinsic motivation (Deci, 1975). Furthermore, he 
also forwarded the idea that individuals make choices about 
their behavior on the basis of their perception and will 
behave in accordance with these perceptions regarding the 
locus of causality. 


' ‘ I 








. 














23 


Cont ingent v ers us Non-Contingen t Re ward s 

Deci (1972b) in another study investigated the effects 
of extrinsic reward on intrinsic motivation as mediated by 
reward made contingent on performance versus reward made 
non-contingent on performance. As well r in this study he 
investigated the effects of verbal external reward. He 
reported that a person’s intrinsic motivation to perform an 
activity decreased when he received contingent monetary 
payment, threats of punishment for poor performance, or 
negative feedback about his performance. Non-contingent 
monetary payments left intrinsic motivation unchanged, and 
positive verbal reinforcements appeared to enhance intrinsic 
motivation. 

In all of Deci's early studies money was administered 
contingently (i.e. money was received by the subjects 
dependent on performance). Therefore, this study represented 
the first test of the contingent versus non-contingent 
question. To account for the results Deci (1972c) reasoned 
that under the non-contingency condition performance is not 
tied directly to rewards which makes it less likely that 
subjects will percieve the locus of causality outside 
themselves. He stressed the "less likely" "suggesting it is 
not impossible that non-contingent rewards decrease 
intrinsic motivation. 

Timin g of Rew ard 


Using the same experimental design as in his earlier 











. 















24 


studies., Dec! ['\972d) , also investigated the effects of the 
timing of extrinsic and intrinsic factors. The independent 
variable of extrinsic reward was varied so that subjects 
solved puzzles in one of the six conditions: 

(1) no reward, 

*2) Money before the free-choice period, 

(3) money after the free-choice period, 

(4) money and positive verbal reinforcement, 

*5) positive verbal reinforcemnt before the 

free-choice period, 

(6) positive verbal reinforcement after the 

free-choice period. 

\ 

As predicted, subjects who were rewarded with money 
were less intrinsically motivated than non-rewarded subjects 
and subjects paid before the free-choice period worked 
harder on the puzzles than those paid after the free-choice 
period. The effects of verbal reinforcement were in the 

predicted direction but did not reach statistical 

significance. Interestingly, subjects who received money and 
verbal reinforcement similtaneousiy, showed no significant 
difference in intrinsic motivation as compared to the 
control group. 

It is also interesting to note that when male and 
female subjects were considered seperately, Deci found 
verbal reinforcement significantly increased intrinsic 
motivation in males but not in females. This finding 

prompted Deci to make the distinction between the 






. 


























. 








25 


"controlling" aspect and an "information" or "feedback" 
aspect of extrinsic reward. When money is given "as an 
external reward," the controlling aspect is clearly the 
strongest and leads to a decrease in intrinsic motivation. 
When positive verbal reinforcement is given on the other 
hand, the information aspect is clearly the strongest and 
leads to an increase in intrinsic motivation. The 
differential male-female findings seem to indicate that 
positive feedback strengthened feelings of competence and 
self-determination in males, whereas for females it changed 
their perceived locus of causality. Deci (1972b) suggests 

that socialization of males versus females in our society is 

\ 

responsible for this 'effect. However, the major purpose of 
this study was to test a cognitive evaluation versus 
inequity hypothesis. Deci (1972d) reasoned that subjects 
paid prior to the free-time period would experience inequity 
and thus continue to work while subjects paid after the 
free-time period would not experience inequity and thus not 
continue to work. Therefore, timing of reward was 
hypothesized to determine whether or not the subject would 
experience inequity or cognitive reevaluation. Calder and 
Staw (1975) have criticized this study suggesting that 
Deci's conclusions cannot be validly derived from his data 
since the use of only one independent measure to tap two 
distinct underlying processes is not experimentally valid. 
Expectation of Re ward 


There is some evidence indicating that expectation 





»4> an a»*it -* *4£m*ifi-**** ! -’ v " ' a :v “‘ 








26 


versus no expectation of reward is related to intrinsic 
motivation. Lepper, Greene, and Nisbett (1973) using a group 
of kindergarten children varied the expectation as well as 
the level of extrinsic reward. In this study, preschool 
children were assigned to one of three treatment conditions 
(expected reward, unexpected reward, and no reward) on the 
basis of a criterion of intrinsic motivation for a drawing 
activity during baseline observation in the classrooms. The 
extrinsic reward was a "good player reward" certificate 
adorned with a gold seal and ribbon. Measures of intrinsic 
motivation were obtained one or two weeks after the 

experimental sessions and consisted of recording spontaneous 

\ 

classroom drawing activity. They found that children who 
expected a reward for an interesting task (playing with 
magic markers) subsequently showed less intrinsically 
motivated behavior in a free-time situation than subjects 
who did not expect the reward or who received no reward. 

Greene and Lepper (1975) replicated the above study and 
in addition measured the effects of low performance versus 
high performance demands, making the reward contingent on 
the level of performance. The results of this study provided 
a near perfect replication of the earlier study by Lepper et 
al (1973). However, the manipulation of the performance 
demand in this study produced no significant effect. Greene 
and Lepper suggest that the negative effects of expected 
rewards on intrinsic motivation are therefore not limited to 
whether rewards are presented contingent or non-contingent 


■ 


















27 


on performance. 

Kruglanski, Alon, and Leiws (1972) manipulated two 
independent variables: (1) the presence versus absence of 
prizes awarded to members of the winning teams in a series 
of game competitions conducted in classrooms, and (2) the 
time at which the measurement of the dependent variable was 
carried out: immediately following the experimental session 
and one.week after the experimental session. No significant 
difference between the groups was found on an open-ended 
question testing for causal attribution. However, on a 
close-ended question, a significant difference between the 
groups was found. Also, the prize group rated their task 
significantly less enjoyable than the no-prize group. 
Analysis of the scores of rated task enjoyment showed 
virtually no change over time. Kruglanski, Alon and Lewis 
concluded that intrinsic interest for an activity decreases 
when subjects receive an unexpected reward after performing 
an interesting activity. 

The data of these studies suggest that when extrinsic 
reward is expected the individual is more likely to perceive 
the reward as the cause of his behavior (Deci, 1975). 
However, this effect may not be as important as the 
empirical data suggests since unexpected reward may initiate 
retrospective evaluation (Deci, 1975). In the Kruglanski, et 
al (1972) study rewards were recieved by the children after 
the game activities were performed and were not expected, 
yet there was less subsequent intrinsic interest in the 





























28 


activities than in the no-reward controls. Deci (1975) 
suggests that reward expectation increases the salienly of 
the causal link between reward and behavior and thus 
mediates the reward-intrinsic motivation relationship. 

Task Performance and Magnitude of T ask Intrinsic Mo tivatio n 

Kruglanski, Freidman, and Zeevi (1971) demonstrated 
that the decline in liking of an activity due to extrinsic 
attributions is paralleled by a decline in the quality of 
performance. Subjects were tested under two experimental 
conditions on a variety of tasks (creativity, recall, and 
Zeigarnik measures). Subjects in the extrinsic-incentive 
condition' were promised a reward (guided tour) for their 
participation in the experiment. No reward was given to the 
no-incentive group. It was found that subjects in the 
incentive group as opposed to subjects in the non-incentive 
group exhibited superiority in creativity of performance and 
task recall. In addition, they manifested a stronger 
Zeigarnik effect, and reported greater enjoyment of the 
experiment. 

Kruslanski, Riter, Arazi, Agassi, Montiqui, Peri, and 
Peretz (1975a) investigated the relation between the 
magnitude of task intrinsic reward and the degree of 
intrinsic and extrinsic motivation towards the task. In one 
experimental condition (high intrinsic reward) the task was 
designed to be interesting and challenging. In the other 
condition (low intrinsic reward) it was designed to be dull. 
The results of three seperate experiments provided strong 








■ 






* 



29 


support for the hypotheses that: (1) 
motivation varies positively with 
intrinsic rewards and, (2) the 
motivation varies negatively with 
intrinsic rewards. 

The Cont ent-Con se quence D istinct i on 


the extent of intrinsic 
the magnitude of task- 
extent of extrinsic 
the magnitude of task- 


Krugianski, Biter, Amitai, Margolin, Shabtai, and Zaksh 
(1975b) studied the content-consequence aspect of external 
rewards. The data supported the hypotheses that when 
payments are inherent to the task's content their presence 
(versus absence) increases intrinsic motivation, whereas 
when they constitute the task's exogenous consequence their 
presence (versus absence) decreases intrinsic motivation 
toward the task. The findings were interpreted as consistent 
with the assumption that intrinsic motivation ensues 
whenever the actor causally attributes his performance of 
the task to the task's content and inconsistent with the 
proposal that intrinsic motivation be identified with 
internal (or self) attributions and extrinsic motivation 
with attributions only to the external environment. 

Neg ative Ext rin s ic R ewa rd 

Deci and Cascio (1972) studied the effects of negative 
feedback and threats of punishment, on intrinsic motivation. 
Subjects in this study solved puzzles during the first part 
of the experimental sessions, and then observations relevant 
to their intrinsic motivation were made. Subjects in the 
negative feedback condition were given very difficult 
























30 


puzzles to solve so that they failed on more puzzles than 
the control subjects who were given easier puzzles. Those in 
the high failure (negative feedback) condition showed less 
intrinsic motivation following their puzzle-solving session 
than did control subjects. Subjects in the threat condition 
received an aversive buzzer each time they were unable to 
solve a puzzle, while the control subjects did not receive 
the buzzer. Those subjects threatened with the buzzer showed 
less intrinsic motivation than control subjects. These 
results however only reached marginal significance. 

Intrinsi c and Ext rin sic Motivation as I nde p en d ent Variables 

Calder and Staw (1975b) designed a study to test the 
effects of extrinsic reward on intrinsic motivation more 
directly by manipulating both the intrinsic and extrinsic 
factors as independent variables. This study has particular 
relevance to the present study since the measure of 
intrinsic motivation was achieved by analysis of scores of 
rated task enjoyment by subjects. One aspect of the present 
study is designed to test the relationship between self- 
report measures of intrinsic motivation and a behavioral 
measure of task persistance. The measure of intrinsic 
motivation consisted of subjects rating the extent to which 
they found "the puzzle task itself" enjoyable on a 17-point 
Likert-type scale. Several other questions on a 11-point 
scale concerned the subjects* perception of the situation. 
Since the procedure did not employ the typical free-time 
measure (Deci, 1971, 1972a, 


1972b; Lepper et al, 1973), 
















. 




















31 


subjects were asked to volunteer for future experiments of a 
similar nature without payment (time for which subjects 
volunteered as coded in minutes). The independent measure of 
intrinsic motivation consisted of 15 jigsaw-type puzzles, 
one set with pictures (high intrinsic motivation) and the 
other set blank (low intrinsic motivation). Through pilot 
testing task behavior was eguated so that the tasks differed 
only on the affective dimension of intrinsic motivation. 
Results revealed a significant Blank-Picture x Money 
interaction for the task enjoyment variable but a non¬ 
significant interaction for the time volunteered variable. 
In addition, two other findings are reported; (1) the 
introduction of the extrinsic reward (money) apparently 
decreased the perception of task effort ("trying") on both 
the high and low intrinsically motivating tasks, and (2) 
subjects were only "aware" of the effects of external 
factors on their level of motivation and not internal 
factors. 

Within this paper Calder and Staw (1975b) also 
distinguish between a statistical and mot ivat ional 
int er ac tion. The term interaction with respect to 
motivational phenomena is meant to define their finding that 
total motivation (in a situation where extrinsic rewards are 
added to an already interesting activity) is equal to 
intrinsic plus extrinsic minus an interactio n between the 
two. Had these authors not found a statistical interaction 
they would still have found the motivational interaction 





( 











32 


effect because of the main effect of high rewards producing 
a decrease in subsequent intrinsic motivation. The 
intera ction hypothesis was predicted upon two effects: (1) 
an inverse effect (seif-perception) when the task is 
initially high in interest (overjustification phenomena), 
and (2) a direct (or reinforcement) effect when there is 
initially less interest in the task. Therefore, when the 
task is less interesting, self-perception phenomena is not 
expected to apply since no over-justification perceptions 
exist. This is also relevant to the present study since the 
prediction of a direct reinforcement effect is not expected 
to apply with duller tasks when individually chosen. In the 

t 

Calder and Staw (1975b) study the two tasks were group 
validated (see next section). 

Rew a r d S aliency 

Ross (1975) conducted three important experiments where 
pre-school children were shown either a salient-reward 
condition or a non-salient-reward condition and a cue 
reminding them of the reward to be administered at the end 
of the task. He found that both in immediate and delayed 
(free-time) play the children in the salient-reward 
condition showed significantly greater decrease in free-time 
play than non-salient or control children. 

In a second study (Ross, 1976), children were asked to 
think about an unrelated topic (distraction condition) or 
were not explicitly asked to think. Children in the 
distraction condition and in a non-reward control condition 





, ts»-; o?! -i '■ viXeiitfll 





































33 


displayed the most interest in the activity, adding weight 
to the findings of his first study. It is here that Ross 
suggests a delay of gratification hypotheses as an 
alternative to the attribution hypotheses to explain his 
results. 

In a third study children drew under one of three 
conditions: No reward (control), wait-contingent reward 
(rewarded for merely waiting for experimenter to return), 
and task-contingent reward (rewarded for drawing). The 
results indicated that the task-contingent condition 
resulted in both the greatest degree of reduced intrinsic 
task interest and quality of performance. Ross suggests that 
these results indicate that the frustration due to delay of 
reward effect is not the critical proces involved in reduced 
task interest. This study also has relevance to the present 
study since Ross utilized several play activities in the 
experimental room. 

Perf ormance C on ti ngent R ewards 

Most of the studies reviewed so far seem to indicate 
that salient and contingent rewards reduce intrinsic 
motivation _ for a task. Ross, Karniol, and Rothstein (in 
press) conducted two inportant studies which indicated that 
decrements in intrinsic motivation however may not be 
inevitable. Rewards have cue value when their attainment is 
a function of task success and their reception signifies 
competence at the task. In most of the previous studies 
rewards were dispensed contingent on performing the task 






, 




' 






34 


only rather than on success criteria. Thus, Ross (1976) 
suggests, that such rewards have little or no cue value and 
may have a particularly detrimental effect on intrinsic 
motivation since there is simply a means ends relationship 
between activity and reward. On the other hand, when rewards 
are performance contingent, reception of the reward provides 
a cue of successful performance and task competence. 
Therefore, performance contingent rewards should increase 
subsequent intrinsic motivation. 

In the first experiment, college students worked on a 
cooncept-j;or matron task under three conditions: (1) no 
reward (controls), (2) money contingent on undertaking task, 
and (3) maney contingent on successful solutions. Dependent 
measures of intrinsic interest were subjects evaluations of 
task and reported willingness to perform the task on 
subsequent occasions without reward. The results provided 
strong support that rewards reduced intrinsic interest only 
when not success-contingent. 

In a second study with an early school population a 
nonmonetary reward (marshmallows) and a behavioral measure 
of intrinsic motivation (persistance at a concept-formation 
task during, a 6-minute free-play period) was used. Also 
included in this experiment was a manipulation of the degree 
to which subjects succeeded at the activity. The results 
indicated that performance-contingent and control subjects 
played with the task significantly longer than did the 
noncontingent-reward subjects. These results therefore 






■ 






.0UIO4PD1O'. x J*t ai_-.0o^i«s p IH ^ 1 : *** 




35 


paralleled to those of the first study. 

Nont an gibl e Rew ard C onditions 

fiore recently, Amabile, Dejong, and Lepper (1976) 
designed a study to explore the effects of externally 
imposed deadlines on individuals task performance and 
subsequent interest in the task. In two deadline conditions 
subjects were given an explicit time limit for solving a 
series of initially interesting word games. In one 
condition, the deadline was left implicit. In two control 
conditions, subjects were asked to solve puzzles to work at 
their own pace, and as fast as possible. However, here no 
explicit time limit was presented. Unobstrusive behavioral 
and attitudinal measures of interest indicated that in the 
absence of external constraints subjects in the deadline 
conditions were less interested in the game than subjects in 
the non-deadline conditions. However, no significant 
difference in interest were found between subjects in the 
two deadline conditions (explicit and implicit). 

This study seems to demonstrate that an externally 
imposed deadline for task completion can result in 
significant decrements in intrinsic task interest. Therefore 
an extrinsic reward or goal need not be tangible to reduce 
intrinsic motivation. Unfortunately, these results could be 
easily explained using the competing response hypothesis at 
the explanatory model, since both the arousal of unpleasant 
affect and the subject*s recognition of a lack of enjoyment 
during the experimental phase under the deadline conditions 











$L U:».t ' , « >6 

' 

■ 





3b 


may have mediated a relatively persistant effect and become 
associated to the task. However, the authors found that 
performance quality and experimenter (not subject) 
evaluation of the task was the same in all groups. 

EVA LUAT ION OF EMPIRICAL RE SEAR CH 

The findings in the majority of empirical studies are 
at least relatively consistent with the derivations made 
from attribution and cognitive evaluation theory. The 
effects of expected, salient, and contingent rewards on 
intrinsic motivation have been replicated using different 
rewards, subject populations, ranging from pre school 
children to college students, and different dependent 
measures. However, there are still methodological and 
theoretical issues that remain to be investigated. Many of 
these issues are dealt with in recent and comprehensive 
reviews (Deci, 1975; Notz 1975; Calder & Staw, 1975a; Ross, 
1976; Reiss & Suschinsky, 1976). Only those issues of 
particular importance to the present study will be reviewed 
here. 

Rewa rd Sal ience and Rew ard Magnitud e 

One of the recurrent themes found in the empirical 
lirterature is an attempt to determine the conditions under 
which extrinsic rewards cause a decrease in subdeguent 
intrinsic motivation. Within the attributional and cognitive 
evaluation model reward saliency is regarded as a major 
variable related to the individuals perceptions of 
behaviorsal causality. That is, the greater the saliency of 










. 






















. 











37 


the external reward the more likely it will be perceived as 
the cause of behavior. However, two alternative explanations 
are found in the literature; one derived from cognitive 
evaluation theory (Deci, 1972d), and the other derived from 
a reinforcement model (Reiss & Suschinsky, 1975). 

Within the framework of cognitive evaluation theory, 
Deci (1972d) suggests that it is not reward saliency per se 
but the salience of the controlling component of extrinsic 
rewards that is the critical variable causing the subsequent 
shift in percieved locus of causality. Deci (1971,1972,1973) 
provides evidence that positive verbal feedback increases 
intrinsic motivation whereas monetary extrinsic rewards 
decrease intrinsic motivation. Thus, he asserts that it is 
the saliency of the controlling component of monetary 
extrinsic reward that accounts for these differential 
results. It seems however, that these conclusions on the 
basis of Deci's evidence alone remain somewhat, speculative. 
Within most of the studies reviewed in fact the extrinsin 
reward was designed to be highly salient to the subjects. 

Ross, et al (in press) on the other hand provides much 
stronger evidence that rewards that are success contingent 
do not reduce intrinsic motivation. To test this hypothesis 
he conducted two studies manipulating the cue value of 
rewards. Both studies, using different subject populations 
and dependent measures of intrinsic motivation, supported 
the hypothesis that only rewards of low cue value 
(performance contingent rewards) decrease subsequent 





. 










38 


intrinsic interest in a task. 

To test the effect of other saliency variables Boss 
(1975) also conducted two experiments with preschool 
children. In the first study saliency was manipulated with 
regards to reward visibility to the subjects while in the 
second study saliency was varied along an attentional 
dimension. Both studies provide strong evidence that reward 
salience in terms of visibility and attentional focus are 
critical variables and determining conditions that mediate 
the effects of extrinsic rewards on subsequent motivation. 

j 

In sum, within the attributional and cognitive 
evaluation models reward saliency is hypothesized to mediate 
the effects of extrinsic rewards in intrinsic motivation in 
three ways: (1) visibility of the reward, (2) attentional 
focus, and (3) cue value (controlling vs. informational 
component). 

A prediction derived from the saliency hypothesis is 
that reward magnitude is related to subsequent decrements in 
intrinsic motivation. That is, a larger and more desirable, 

and thus more salient reward is mor e likely to cause a shift 

\ 

in percieved locus of causality than a smaller, less 
desirable, and less salient reward. This prediction is made 
on the basis of two related effects: (1) a powerful 
extrinsic reward may lead individuals with an initial 
interest in an activity to reatribute their interest in that 
activity to the salient external reward, and (2) conversely 










■ o ' * ll 










39 


the presentation of a small and minimally salient but 
sufficient reward may induce in individuals* with less 
initial interest in a task, an internal attribution with 
increased intrinsic motivation (Lepper & Greene, 1976). 
Interestingly, within the overjustification literature, 
these predictions have remained untested as only the effects 
of a salient reward versus no-reward have been tested. This 
leaves a serious gap with respect to an important property 
of extrinsic monetary rewards. This is of singular 
importance to our understanding of the effects of rewards on 
intrinsic motivation since virtually nothing is known 
empirically of the nature of the relationship between reward 
size and subsequent intrinsic motivation. It is interesting 
to speculate that this relationship may not be a monotonic 
one. Although a very small reward may not be salient and 
powerful enough to be percieved as the cause of behavior, 
there may be other so far ignored factors, operative that 
could affect subsequent intrinsic motivation in unexpected 
ways. This point will be persued after a brief review of the 
competing response effects explanation. 

As stated before, a major alternative explanation of 
the intrinsic motivation phenomena, forwarded in recent 
years by Reiss & Suschinsky (1975,1975a), is the competing 
response hypothesis based on a reinforcement model. This 
alternative explanation represents the most serious 
challenge to the attributional and cognitive evaluation 
explanations. In fact, Reiss & Suschinsky (1976) suggest 





* 2 : * ** ♦¥****] *** 

' 












40 


that all research supporting the overjustification 
hypothesis is more appropriately accounted for in terms of 
the competing response hypothesis. The most fundamental 
proposition is that there exist different ways in which any 
particular activity can be performed. Presumably there is an 
underlying continuum of degrees of enjoyment associated to 
the manner in which a particular experimental task is 
performed. This hypothesis predicts that practice of an 
activity under a high degree of enjoyment relative to past 
experience or a control group leads to acquisition of 
intrinsic motivation. On the other hand, practice of an 
activity in a way less enjoyable relative to past experience 
or a control group impairs subsequent intrinsic motivation. 
Thus, Reiss and Suschinsky state, "That for any play 
activity there exists a set of responses that facilitate 
task enjoyment. Generally, these responses include a certain 
degree of attention to the activity ... performance of the 
activity in the absence of aversive affect, performance of 
the activity at a relaxed pace, successful performance ... 
so that any response that interferes with responses that 
facilitate task enjoyment, will be termed competing 
responses (p. 235)." Thus, it is suggested that the 
introduction of an expected reward can arouse many responses 
that compete with responses facilitating task enjoyment. 
"Moreover, simultaneous exposure to two reinforcing 
activities (reward and intrinsically motivated activity) 
could arouse attentional approach conflicts that are 
aversive (p. 214)." 


,. 

. 


41 


Reiss and Suschinsky suggest this hypothesis leads to a 
number of critical issues: First, the competing response 
hypothesis proposes that the greater the extent of arousal 
the greater the decrement in intrinsic motivation of the 
experimental activity. The second issue relates to Reiss and 
Suschinsky's use of the internal-external partition. That 
is, their definition of intrinsic motivation is largely 
dependent on the absence of any immediately discernable 
environmental reward. Therefor, any studies attempting to 
study intrinsically motivated behaviors are potentially open 
to criticism of whether the testing procedures evaluate 
extrinsic motivation, since it is difficult to control an 
environmental stimuli. This, it is suggested is of major 
consern and raises the issue of possible transfer of an 
expectation of reward from training to testing situations. 
Basically, Reiss and Suschinsky critique what is referred to 
as the "all or none" conception of intrinsic motivation as 
conceptualized in the attributional model. However, this is 
precisely where the opposing theoretical positions* 
divergence is critical since the attributional explanation 
is based on> an analysis of the individuals perceptions of 
what causes his behavior. Whatever the objective cause of 
behavior may be, attribution theory, self-perception theory, 
and cognitive evaluation theory take the individual*s 
subjective perceptions of behavioral causality as the 
central theoretical springboard to generate testable 
hypothesis. 










42 


The theoretical divergence is of particular interest to 
the present study since the concept of competing response 
effects draws attention to other important aspects of 
extrinsic rewards ignored in previous investigations. This 
is highlighted in the review of the empirical literature by 
several unresolved issues. For example, the findings 
surrounding the issues of contingent versus non-contingent 
and expected versus unexpected rewards are still eguivocal. 
The results from various studies are mixed and may be 
related to slight differences in environmental settings, 
task parameters, and rewards aspects. Each of these 
variables could create distinctly different psychological 
situations, with important attitudional and behavioral 
consequences. However, in terms of the reward magnitude 
variable, competing response effects are also influenced by 
the size of extrinsic reward, but the effect predicted is of 
a more continous nature. That is, decrements in subsequent 
intrinsic motivation are dependent on increments in reward 
magnitude inducing degrees of arousal from competing 
effects. Thus, a large reward is assumed to induce greater 
frustration during the delay period than a smaller reward 
with a subsequentally larger decrement in task interest. 

In sum, in terms of attributional processes the 
presence of a large reward leads to the individual's 
attention being focused outward away from the task in an 
•'all of none" fashion. When in the presence of a small and 
negligable reward the individual's attention is more 




, -41 <4 ‘ • * ; **!*•> 1 

' 








43 


internally focused away from the reward, which may even lead 
to an increase in intrinsic interest. On the other hand, in 
terms of competing response effects large rewards represent 
more powerful distractions and frustrators than smaller 
rewards, therefore leading to a greater decrease in 
intrinsic motivation. Thus, both theoretical explanations 
lead to a similar prediction: that reward magnitude is 
related to subsequent decrements in intrinsic task interest. 
However, the hypothesized underlying processes are 

distinctly different, with the competing response 

explanation perhaps drawing attention to a host of factors 
not taken into account within a pure attributional model. 
S ewa rd and Task Cho ice 

Another aspect that the preceding literature review 
points out is related to the level of reward and task 
interest as independent variables. In all of the previous 
studies level of reward (reward vs. no-reward) and level of 
task interest (high interest vs. low interest) was 
determined on a group validated basis without consideration 
for individual choice (self-choice). That is, the 
independent measure of level of intrinsic motivation was 
determined through pilot testing to achieve tasks that 
differed significantly only on the percieved intrinsic 
interest dimension but were equated on other cognitive and 
behavioral task dimension (if more than one interest level 
was employed). Calder and Staw (1975b) for example, utilized 
a series of semantic differential scales to equate a blank- 









44 


and picture-jigsaw puzzle on all but the affective (liking) 
dimension. However, even then subjects could still have 
differential task perceptions, not tapped by the equating 
procedure, that could influence subsequent dependent 
measures of task behavior in unexpected ways. 

With regards to rewards none of the previous studies 
have provided subjects with a reward choice. Recent findings 
have reported that self-reinforced responses may be more 
resistant to extinction than externally reinforced responses 
(Weiner & Dubanoski, 1975). A number of studies have shown 
greater 'resistance to extinction among self-reinforced 
groups compared to experimenter-reinforced groups on 
discrimination tasks (Johnson, 1970; Johnson & Martin, 1971; 
Bolstad & Johnson, 1972). Weiner and Dubanoski (1975) 
suggest four possible causal eplanations: (1) reinforcement 
recieved under external control (Rotter, 1966) is less 
likely to strenghten behavior than is reinforcement under 
internal control, (2) percieved control over the environment 
may be more reinforcing (Chaikin, 1971; Kanfer & Siedner, 
1973), and self-reinforced groups may percieve greater 
control over their environment, (3) differential 
reinforcement history may result in the establishment of 
self-evaluation as a conditioned reinforcer, and (4) under 
conditions of self-reinforcement subjects may chose a 
schedule most reinforcing to themselves and thus more 
resistant to extinction. While these findings raise 
interesting questions it seems that some important questions 


fv*to4q> * ii vh 110 iW to sa-. .j-ii^iu 


45 


regarding the effects of self-reinforcement on intrinsic 
motivation have so far been overlooked. How, for example 
does self-reward affect subsequent interest in a task? 
Although this question is not directed at the acquisition of 
a new response and subsequent extinction of that response as 
related to self-reinforcement it deals with the effects of 
self-reinforcement on responses already within the 
behavioral repertoire of the subject and with a high degree 
of associated interest. There is a crucial similarity 
however in both the learning and intrinsic interest 

situations; it can be said that in both situations the 

( 

subject works on the experimental task to achieve an end 
state or result (i. e. reward choice). Within the intrinsic 
motivation situation the emphasis however is on the extent 
to which reward choice leads to perceptions of extrinsically 
motivated behavior with an expected concomittant decrease in 
intrinsic interest for the task. 

Depe ndent Measu res of Intrinsic Mot i vation 

Although there are several distinct conceptualizations 
of the construct of intrinsic motivation its 
operationalization rests primeraly on observed behavior in 
the absence of any immediately identifiable rewards other 
than the activity itself. Lepper and Greene (1976) argue 
that the extrinsic versus intrinsic motivation distinction 
has proved to be a useful concept in generating new and 
testable hypothesis on two ways: (1) among experimental 
social psychologists the distinction between internally and 






tfiAt ip tfy -n*ti i ^ '.ue uni o^o^ra^ s 



















46 


externally governed behavior has long been regarded as 
fundamental (Kelman, 1958; Collins, 1973; de Charms, 1965; 
Bern, 1972), and t2) heuristically, the distinction has 
proved valuable in drawing attention to the fact that 
different processes may be involved in controlling behavior 
in situations were there are clearly identifiable external 
contingencies in contrast to those situations were there are 
no immediately idenfiable contingencies (Collins, 1973; 
Rosenhan, 1969). 

This theoretical conceptualization has led to the 
operationalization of intrinsic motivation according to a 
specific set of criteria; (1) subjects must display interest 
in an activity in the absence of any salient extrinsic 
contingencies, and (2) subjects must not expect any future 
reward condition to be implemented contingent on their 
activity on the task. All of the studies reviewed utilized 
experimental designs that incorporated measures of intrinsic 
motivation based on these criteria. However, the actual 
dependent measures employed have ranged from observable task 
persistence during a free-time period to various self-report 
measures or attitudinal measures. 

A fundamental assumption has been that both the 
attitudinal and behavioral measures tap the same underlying 
state of intrinsic motivation for an activity. Moreover, 
both the attitudinal and behavioral measures are thought to 
reflect intrinsic motivation as defined in attributional 
terms; that is, the subject’s perception of an internal 


























































locus of causality or a high degree of liking for the task 
in relation to external variables. The major assumption has 
been that there exists an innate relationship between 
attitude and behavior. However, Doob (1947) and Bandura 
(1969) long ago argued that there may not be a one to one 
relationship between the two. Basically, an attitude may be 
a learned predisposition to respond or a learned mediating 
response. Once the attitude is learned the appropriate 
behavioral response must also be learned. Thus, two 
individual's may hold the same attitude but also display 
different learned responses given certain other variables 

i 

are present. 

Although the majority of studies have utilized either 
behavioral or attitudinal measures (and more recent studies 
have utilized both measures to generalize their results) and 
found fairly consistent results with both measures a number 
of studies have also found mixed results demonstrating that 
the attitudinal and behavioral measures do not always 
correspond as directly as is believed. It is interesting to 
note that in none of the previous studies was a systematic 
attempt made to analyze the underlying factor structure of 
these two broad classes of intrinsic motivation. 










■ 














!£. 1 





CHAPTER III 


RATIONALE, DESIGN, AND HYPOTHESIS 

RATIONALE 

The preceding theoretical and empirical literature 
review has pointed out a number of specific areas and issues 
for further empirical investigation. The present study was 
designed to investigate some of the issues raised by the 
literature. Three major questions are addressed. The first 
question Investigated is the effect of extrinsic monetary 
reward on individually determined, rather than group 
validated task motivation. All of the studies to date have 
pre-tested experimental tasks to determine motivation level 
and to equate other (behavioral and cognitive) dimensions to 
control differential task perception (Calder & Staw, 1975). 
The present study investigates subjects' task motivation 
level based on individually determined task choices with 
behavioral and cognitive dimensions controlled through 
randomization,. If extrinsic reward undermines intrinsic task 
interest this would add more support to the attributional 
analysis of task motivation because of the initially strong 
individual task interest factor. Notz (1975) in a lucid 
review of the literature concludes that more theoretical and 
empirical work needs to be conducted on specifying 
properties of tasks related to intrinsic motivation. More 
specifically, the degree to which the individual perceives 


48 




















&* i" a '• Un*u* e : •***• P° *1 












' 












49 


personal freedom of choice and self-investment or commitment 
to the task may be related to personal causation. 
Individually determined task choice provides an opportunity 
to investigate this variable in the laboratory (p. 694). 

An area of research relevant to this question consists 
of th,e free-choice studies where cognitive dissonance theory 
is the major theoretical framework used to generate 
predictions. Subjects are asked to select an object from a 
set of objects. The dependent variable is the subjects' 
subsequent attitude ratings on the chosen and rejected 
alternatives. Dissonance theory predicts an attitude 
displacement in the positive direction on the previously 
chosen object and a negative attitude displacement on the 
rejected alternatives. Bern (1967) however has done an 
interesting re-analysis of this prediction in attribution 
terms. If it can be demonstrated that individually 
determined task choice results in the predicted differential 
levels ©f task persistence in a free-choice situation where 
a number of alternative activities are available, this would 
add credibility to the concept of intrinsic motivation as a 
psychological construct since it is defined as task 
persistence (1) in the absence of any immediately 
identifiable external contingencies and (2) on a self-chosen 
target activity. 

Also, the random distribution of tasks within the 
treatment groups (high and low intrinsic motivation) would 
ensure that any differential task perceptions are controlled 






. 




■ 








I 








50 


since both groups would be working on all of the tasks with 
only affect towards the task manipulated as the independent 
variable. 

The second major question investigated is the predicted 
negative relationship between the size of an extrinsic 

s 

monetary reward and subsequent task interest for both 
initially interesting and less interesting activities. More 
specifically, the literature suggests that the larger the 
reward the more likely (1) it will be percievea as the cause 
of behavior or (2) the greater its potential distracting and 
frustrative' effect. Conversely, a small reward is less 
likely to be percieved as the cause of behavior. However, 
this relationship may not be as clear cut conceptually as 
this. A very small reward may activate feelings of self¬ 
devaluation which could be related to observed decrements in 
task interest. T’o control for this effect rewards are 
allocated only on a performance-contingent basis in the 
present study. 

Finally, the Calder & Staw (1975b) finding of a direct 

reinforcement effect with a less interesting task is not 

\ 

expected to apply in the present study due to a predicted 
dissonance effect associated to negative task displacement 
on the rejected task alternatives. However, dissonance 
theory also predicts less cognitive change with increasing 
reward magnitude. As a result less decrease in task interest 
is expected in the presence of a large reward in contrast to 


a smaller reward. 






























. 




51 


The third major question examined was raised in 
relation to self-reinforcement. The findings on self¬ 
reinforcement effects suggests that self-reinforcement leads 
to greater resistance to extinction on discriminative tasks 
than experimenter-reinfcrcement. In relation to intrinsic 

i 

motivation this concept has remained unexplored to this 
point. Although within the attribution framework responses 
already established within the behavioral repertoire of the 
individual are investigated one commonality with the 
learning paradigm is, the fact that the subject works on the 
task to achieve an end result (i.e. choose an individually 
determined guantity of the reward). Although the self¬ 
reinforcement literature suggests this is an important 
variable, the rationale for this part of the present study 
also emerges out of attribution theory which predicts that 
reward self-choice leads to a shift in locus of causality 
perceptions with subsequent impairment in intrinsic 
motivation. 

Finally, the underlying factor structure of the two 

previously mentioned classes of measures (altitudinal and 

\ 

behavioral) is investigated. 

One other question investigated and indirectly related 
is the effect of a purely cognitive variable classified as a 
deprivation condition. Although no specific hypothesis are 
generated this variable relates to inequity theory since 
subjects will be deprived of rewards or outcomes. Inequity 








. 


















52 


theory predicts a subsequent drop in work behavior or task 
persistence due to feelings of inequity generated by the 
monetary deprivations in relation to other subjects. 
Moreover, the induction of inequity does not include 
possible self-devaluation since dubjects will be informed 
that the deprivation is the result of a random process. 
However, one real limitation to this procedure is that no 
provision is made to have the subject experience a 
connection between his inputs and behavioral outputs. 

DESIGN 

i 

In order to investigate the listed question, one pilot 
study, and two preliminary studies were conducted prior to 
the main investigation. The pilot study was designed to 
determine the five experimental tasks to be utilized as 
choice objects. In study one the main focus was given to 
establish a norm for a behavioral measure of intrinsic 
motivation (i.e. task persistance). This study consisted of 
two phases: 1 1) first, a behavioral norm was established on 
five experimental tasks, and then, (2) subjects' 
individually determined choice of high and low intrinsic 
task interest was compared to the previously established 
norm and to each other. Two seperate and independent samples 
of subjects were randomly assigned to each group. In 
addition, the intercorrelations and factor structure of the 
same set of dependent variable measures were compared to the 
results of study 3. In sum, study 1 and 3 were designed for 
the purpose of: (1) Determining the 


intercorrelations 











. 

V 


53 


between various measures of intrinsic motivation (task 
persistance and self-report measures - study 1 and 3), (2) 
determining whether the measures of intrinsic task 
persistence vary as expected from the previously established 
norm (study 1) and, (3) to determine whether individually 
determined task persistence results in significant 
differences betwen high and low task motivated groups (study 
1). The second study was designed to determine the effects 
of pre-testing on the behavioral measure of task persistence 
which is the primary dependent variable of intrinsic 
motivation in the main experiment. Finally, the main study 
(study 3) investigated the effects of reward magnitude, 
reward choice, and reward deprivation on intrinsic task 
interest. 

The main experimental design (study 3) consisted of 
five treatment groups with two levels of intrinsic task 
interest: High intrinsic task motivation and low intrinsic 
task motivation. Subjects were asked to choose by means of a 
paired-comparison procedure the task of the five previously 
established tasks, which they perceived as the most 
interesting and enjoyable. As established in study one, this 
enabled the experimenter to randomly assign subjects to one 
of the two levels of intrinsic motivation. Simultaneously, 
subjects were randomly assigned to work on their respective 
task under five treatment conditions during a 10-minute 
experimental time period: 

(1) A two dollar reward. 







. 








. 






54 


(2) A twenty-five cent reward, 

(3) A deprivation condition, 

(4) A self-choice condition (any quantity between 
zero cents and two dollars), 

(5) A no-reward condition (control group) , 


A number of 
utilized in stud 
chosen and presen 
correlate self-r 
measure of intrin 
the task) which 
IC-minute free-ti 


rating scales similar to those previously 
ies investigating intrinsic motivation were 
ted on a pre- and post-basis in order to 
eport measures with the primary dependent 
sic motivation (i. e. time spent working on 
was measured unobstrusively during a second 
me period. 


HYPOTHESIS 


THe hypothesis concerning the individual task choice 
experiment were tested in an attempt to ensure that this 
procedure leads to significantly different levels of 
intrinsic motivation as measured by task persistence. No 
hypothesis regarding the effects of pre-testing were 
specified and therefore this study may be regarded as 

exploratory. The hypothesis with respect to reward magnitude 

\ 

and reward choice were tested to evaluate the relationship 
between size of reward and reward self-choice as related to 
subsequent intrinsic motivation at both an interesting and 
dull motivational task level. 


Finally, because the methodological difficulties 
associated to inequity theory make it difficult to determine 








. 




■ 

\ 



55 

its functional significance to the intrinsic motivation 
construct, no specific predictions were made. Therefore, 
examination of this variable (deprivation) in the present 
study may also be considered as exploratory. 

Individu al Task Choice 
Hypothesis 1. 

The high task motivated group will display greater task 
persistence over a 10-minute free-time period than the low 
task motivated group. This prediction is advanced since it 

is argued that individually determined task choice, based on 

/ 

the subjects percieved liking and enjoyment for a task, 
would be directly related to the degree of observed task 
persistance in the absence of any external contingencies. 

Hypo t hes is 1. 1 

Both the high and low task motivated groups will 
display respectively greater and less task persistence 
during a 10-minute free-time period than a previously 
established behavioral norm. The rationale for this 
hypothesis stems from the assumption that both high and low 
percieved task liking should result in significant 
deviations in opposite directions from a behavioral norm 
which represents an average level of intrinsic motivation 
for a particular population (in this case college 
undergraduates). Previous research has utilized 

significantly different levels of intrinsic motivation 
through group validated pre-testing. The present 
methodology, should the hypothesis be confirmed, allow the 









. 






56 


investigation of not only individually determined task 
choice levels of intrinsic motivation but also the 
determination of whether a significantly higher and lower 
level of intrinsic motivation from a population average is 
being investigated. 

Reward M agnitud e 
Hypothesis 2. 

A two dollar reward group will display significantly 
less task persistence over a 10-minute free-time period than 
a twenty-five cents reward group or no-reward control group 

j 

on a high ' intrinsic interest task. This hypothesis is 
derived from the basic assumption of attribution theory and 
self-perception theory that a negative relationship exists 
between size of extrinsic monetary reward and subsequent 
intrinsic task interest. 

Hypo t hesis 2.1 

A two dollar reward group will display significantly 
more task persistence over a 10-minute free-time period than 
a twenty-five cents reward group or no-reward control group 

on a low intrinsic interest task. This hypothesis is derived 

\ 

from dissonance theory which predicts decreasing cognitive 
change with increasing reward magnitude. 

Hyp ot hesis 3. 

A self-reward (choice) group will display less task 
persistence over a 10-minute free-time period than a no¬ 
reward control group on a high intrinsic interest task. This 














. 


















. 





57 


hypothesis is based on the assumption that subjects will 
percieve self-reward as an extrinsically imposed reward 
condition with a resultant shift in locus of causality from 
internal to external. 

No hypothesis is formulated to predict the effects of 
self-reward on a less interesting task, since no self¬ 
perception phenomena is expected to occur in this group and 
the relationship of reward choice to dissonance theory is 
conceptually unclear. To summarize, these hypotheses are 

based on the following effects: (1) a negative relationship 

/ 

between size of monetary extrinsic reward and initially high 
task interest, (2) a positive relationship between size of 
monetary extrinsic reward and initially low task interest, 
and (3) a negative relationship between externally imposed 
monetary reward choice and initially high task interest. 








JG to i.oot £ 1 '"' 


t ; 




CHAPTER IV 


METHOD 

This chapter includes a detailed description of the 

s 

general experimental procedures. It also includes 
description of the experimental subjects, the apparatus and 
experimental tasks utilized, the rating instruments used for 
pre-post treatment purposes, and a description of the 
primary behavioral measure of intrinsic motivation and the 

procedure involved in its measurement. 

/ 

SUBJECTS 

The total sample of subjects consisted of 343 under¬ 
graduate students drawn from various faculties, but 
predominantly from the Faculty of Business Administration 
and Dept. Educational Psychology at the University of 
Alberta. Their ages ranged from 43 to 18 with a median of 24 
years. 

PILOT STUDY 

The subjects for this study were a total of 60 
undergraduate students at the University of Alberta. The 
purpose of the study was to determine five tasks that would 
utilized in the remaining studies. The procedure consisted 
of asking the subjects to choose from thirty tasks presented 
to them, the three tasks with the highest perceived interest 
and enjoyment component and the three tasks with the lowest 
perceived interest and enjoyment components. During this 


58 
















































59 


procedure thirty tasks were laid out on tables and subjects 
in groups of 10-15 were asked to walk around the tables 
familiarizing themselves with each task while the 
experimenter stressed that they could use as much time as 
needed to make their decisions. Subjects were also free to 
walk to and fro to reinspect a specific task should they so 
desire. The procedure did not require the subjects to work 
on the task, but they were free to manually manipulate the 
tasks. 

The results of this pilot study allowed the 

/ 

experimenter to determine 5 tasks with the highest 
discrimenatory component in terms of interest and perceived 
enjoyment for the tasks. A simple item analysis revealed 
that 5 out of the 30 tasks had a remarkably equal number of 
high and low choice scores. The description of the 5 tasks, 
and procedure of analysis is outlined in detail in Appendix. 

STUDY 1 

The subjects for this study were 75 undergraduate 
students drawn from the Faculty of Business Administration 
and the Dept, of Educational Psychology at the University of 
Alberta. Their ages ranged from 38 to 19 years, with a 
median of 23 years. There were 32 male students and 18 
female students. 

This preliminary study served to determine: (1) whether 
individually determined task choice and subsequent task 
persistence varied significantly from an independently 


' 













60 


established norm of task persistence across five tasks, (2) 
whether high and low individually determined task choice 
results in significant difference during a free-time period, 
and (3) to test the degree of relationship between self- 
report measures of task motivation and a behavioral measure 
of task motivation. The basic procedure parallels Deci's 
(1972) study with appropriate modifications. The study 
consisted of two phases. During phase one 25 subjects were 
assigned to the five experimental tasks without pre-testing 
or selection and measured on the amount of time spent 
manipulating the task (task persistence) during a 10-minute 
free-time period after they had worked on the task for a 
preliminary 10-minute time period. The mean time spent 
working on the task for this group of subjects represents 
the norm sample. Phase two consisted of randomly assigning a 
second group of 50 independently sampled subjects to high 
and low task motivation groups on the basis of self-choice 
(pair-comparison) and then comparing the means of these 
groups to the mean of the norm sample and each other. 
Procedur e 

The following represents the experimental procedure for 
all three groups with the exception of the norm smaple where 
no pre-testing or self-choice procedure was initiated. 

The experimenter escorted each subject to the 
experimental room and asked him to be seated at a table on 
which the five experimental tasks were located (jigsaw 
puzzle, watercolour set with paper and water, crossword 



m* - * 1 










61 


puzzle, block construction task with designs, and practice 
course in business writing). The tasks were then presented 
to the subject by the method of paired-comparisons, 
providing standardized instructions (See Appendix 2 for a 
detailed description of the instructions). The experimenter 
asked the subject to point to the task he perceived as the 
most interesting and enjoyable on the basis of his feelings 
towards the task at the moment. Each task was paired once 


with every 

other task 

across successive 

trials. The 

ten 

possible pairings of 

tasks 

were 

presented 

in random order to 

minimize 

position and order 

effects. 

Based on the choice 

score, the 

five tasks 

were 

rank 

ordered 

from high to 

low 

interest 

value. 

Th e 

experimenter 

was doing 

this 

simultaneously during 

the 

presentation 

of tasks to 

the 

subjects. 

In this 

way. 

the 

pre-experimental level 

of 

intrinsic 

task interest 

for 

a particular task 

was 

individually determined. 





The 

subjects 

were 

then 

randomly 

assigned to 

the 

treatment 

conditions. 

Half 

of the subjects worked on 

the 

task ranked as the 

least 

liked 

task and 

half worked on 

the 


task most liked. The experimenter randomized the choice of 
tasks in case of ties. Thus, two conditions of task interest 
were achieved. In this way, each experimental group 
consisted of a balanced distribution of all five tasks. This 
was necessary to ensure that groups varied only on the 
affective dimension of high-and-low-perceived intrinsic 
interest and not on behavioral and cognitive dimensions. 


--4T (t - J '■ S* 






















































62 


Following this, each subject was asked to relocate his 
position to an adjacent table and be seated. Here the 
subjects were told that they are taking part in an 

experiment on cognitive learning and that they will be asked 

\ 

to participate in performing a task. On the table in front 
of the subject the experimenter placed the experimental 
task. On this-' same table in the upper left hand corner a 
pile of recent magazines, consisting of a wide range of 
general interest topics, were located within easy reach of 
the subject. Directly to the subject's right on the table, 
were located the now regaining four tasks. To reach these 
the subject would have to stand up and reach over. Directly 
above and slightly to the rear of the sitting subject was 
located a speaker of the intercom system and a hanging 
microphone. 


The experimenter asked the subject to fill out the 
rating forms with respect to the task in front of him prior 
to doing the task (see Appendix 3 for rating forms). During 
this time the experimenter left the room and obsetved the 
subject behind a one-way mirror. When the subject had 
completed the rating forms the experimenter returned, 
removed the forms and indicated to the subject that he would 
go behind the mirror to observe the subject (pointing to the 
exact spot) and provide the instructions over the intercom. 
The experimenter then left, located himself behind the one¬ 
way mirror and read the instructions to the subject (see 
Appendic 2). The subject then worked on the appropriate task 


ait 

, 

. 









63 


for a period of 10 minutes. During this time two independent 
raters were recording unobtrusively the amount of eye and/or 
hand contact the subject made with the task. Pretesting 
indicated an inter-rater reliability of .99 (Pearson product 
moment correlation). At the end of this 10-minute period the 
experimenter indicated to the subject over the intercom that 
time was up and again read instructions. These instructions 
included that certain recordings were made and he (the 
experimenter)- must leave the process this data. He told the 
subject that he rnay do anything he wished, but that he 
remain in the room until the experimenter returned, since 
more questions need to be asked depending on the data 
processing. The amount of time spent by the subject on the 
task during this 10-minute time period was the measure of 
subsequent task interest. In order to control completion as 
a possible causal factor, each task was designed to make 
completion impossible during the 10-minute period. In the 
case of the water colour painting where more than one 
picture could be completed in 10 minutes, depending on the 
subjects style, the subject was informed during the 
instructions for the first 10 minutes that he may paint as 
many pictures as he wished. At the end of the second 10- 
minute period the experimenter entered the room and gave 
instructions for the completion of the post-rating 
questionnaires which the subject completed in the 
experimental room (see Appendix 2) . 


Finally in order to control cross-subject contamination 














- 

. - . ’ ft i 
































64 


subjects were asked not to divulge the contents of the 
experiment to anyone. This was stressed by the experimenter, 
and all subjects seemed very cooperative in this respect. 
Depe nd ent Va ria ble M easures 

\ 

(1) The primary dependent measure consisted of time 
spent wording on the task during the second 10- 
minute or free-choice period. The criteria for 
task persistence was eye and/or hand contact 
with any component part of the task in a 
sitting or standing position. 

(2) Secondary dependent measure consisted of a 
number of rating scales: 

(a) Semantic differential scales: These scales, 
developed by Caider and Staw (1975b), included 
affective, cognitive, and behavioral items. The 
overall scale consisted of 13 items (4 
affective, 4 behavioral and 5 cognitive). The 
only modification to the scales deemed 
necessary was that all items were arranged 
randomly with respet to both order and 
direction of the scales to control for any 
tendency errors. Each item consisted of a 7- 
point scale (see Appendix 3). 

(b) Perceived enjoyment and interest: Two 
Likert-type rating scales were utilized to 
measure perceived interest and enjoyment. One 





























- i 




















65 


scale consisted of two 7-point summated Likert- 
type scales presented on a test-tre-test basis. 

The other consisted of a single 11-point 
Likert-type scale presented only on a post¬ 
treatment basis. These scales represent 
modification of scales utilized in a number of 
studies (see Appendix 3) . 

(c) Perceived effort: This scale consisted of a 
single 11-point Likert-type scale measuring the 
degree of perceived effort by the subject on a 
post-treatment basis (see Appendix 3) . 

(d) Perceived pleasure of work: This was a 0-1 
data question where the subject indicated 
whether he perceived the task as work or 
pleasure, (see Appendix 3). 

(e) Willingness to colunteer: This was also a 
0-1 data scale where the subject answered 
either yes or no. (see Appendix 3) . 

STUDY 2 

The subjects for this study were 64 undergraduate 
students drawn from the Faculty of Business Administration 
and the Faculty of Education at the University of Alberta. 
The ages from 30 years to 18 years, with a median of 22 
years. There were 23 female students and 41 male students. 
The study was a Solomon four-group design to determine the 
effects of pre-testing on self-report post-tests and task 
performance. Performance is defined as the time spent on a 





tij iWB • ’flpgMf 







' 








66 


task by the subject in a free-time situation over a 10- 
minute interval. 

The design consisted of two levels of intrinsic 

motivation and four groups of test-re-test and treatment 

\ 

combinations. Subjects were randomized across all levels and 
groups. Therefore, the design was a level of intrinsic 
motivation by treatment factorial design with data collected 
as test-re-test and performance measures. Analysis consisted 
of two analyses of variance: First, a 4 x 2 analysis of 
variance of post-test scores across two levels of intrinsic 
motivation; second a 2 x 2 analysis of variance of 
performance scores across two levels of intrinsic 
motivation. 

Prod ecur e 

Experimental procedure paralleled that of Study 1 in 
every respect in the test-performance-re-test groups. 
Modifications were introduced in the remaining three groups 
to satisfy the criteria of a Solomon four-group design. 

In the two (high and low intrinsic motivation) test-re¬ 
test no-performance groups procedure paralleled that of 
Study 1 with the exception that the experimenter asked the 
subject to remain in the room until he (the experimenter) 
returned to ask more questions. Again, the subject was told 
that he could do anything he wished, except leave the room, 
under the pretense of asking further questions after the 
data processing. The experimenter as in Study 1 took the 


















■ 






■ 






• 1 



67 


completed pre-test rating form with him as he left. The 
procedure again paralleled that of Study 1 on entry of the 
experimenter after a 20-minute time period had elapsed, (see 
Appendic 2 for detailed instructions). In the treatment 
post-test only group the procedure again paralleled every 
aspect exactly as the test-re-test treatment procedure of 
the control group in Study 1. The only modification across 
both groups of high and low intrinsic motivation consisted 
of eliminating the pre-test procedure (see Appendix 2 for 
detailed instructions). 

Finally, in the post-test only group procedures were 
modified to include the post-test ratings after the subject 
waited for the 20-minute time interval as in the other group 
(see Appendix 2 for detailed instructions). 

Depe nd en t Variabl e Measu res 

All dependent variable measures of study 1 were 
included with the exception of the perceived effort and the 
perceived enjoyment and interest scales, and the two 0-1 
data scales since these were post-test instruments only and 

no reference to pre-test could be made. 

\ 

STUDY 3 

This study consisted of the major experimental 
investigation of five treatment groups across two levels of 
intrinsic motivation. Subjects were randomized across levels 
of intrinsic motivation by means of the previously (Study 1) 
defined paired-comparison procedure. Subjects were 






■ 







68 


simultaneously randomized 

across treatment 

levels. 

The 

sample consisted of 

144 

undergraduate 

students at 

the 

University of Alberta. 

Ages 

ranged from 43 

to 

18 years 

with 


a median of 25 years. There were 97 male subjects and 47 
female subjects. 

Procedure 

The experimental procedure was similar to the 
experimental procedure of Study 1. However, the treatment 
conditions required ' slight modifications in the both 
procedure and instruments. Subjects were informed over the 
intercom that they would receive their appropriate reward 
when the experimenter signalled that time has elapsed. At 
that point the subject would place the reward in an envelope 
and place the envelope in his pocket. Then the experimenter 
read the remaining instructions to the subject. The reward 
was located in an open drawer of the same table to the 
subject's left in full view of the subject at all times in 
order to ensure high saliency of reward. The drawer and 
evelope procedure was designed to control for a rater halo 
effect. However this was difficult to achieve as a number of 
subjects did not comply and placed their rewards into the 
envelope in full view of both raters. However, since the 
raters were naive and the hypothesized relationship was in 
the opposite direction of what common sense would dictate 
this did not seem to be a serious methodological problem 
(see Appendix 2 for detailed instructions). 







69 


De pe nd en t Va ria ble Measu r es 

Ail of the dependent variable measures as descreibed in 
Study 1 were also utilized in this study. The primary 
measure consisted of time spent by the subject manipulating 
a task during a 10-minute free-choice period. Again, the 
criteria for working on the task consisted of eye and/or 
hand contact with the experimental task in both a sitting or 
standing position. 








































CHAPTER V 


INDIVIDUAL TASK CHOICE: RESULTS AND DISCUSSION 

RESULTS 

To test for significant differences on the task 
persistence measure between high and low task interest 
groups (on the basis of individual task choice) and the norm 
sample, the data of study 1 were analyzed in a one-way 
analysis of variance (Aiiova) and a two-way analysis of 
variance with repeated measures. The first analysis tested 
the main effects of three levels of task motivation and the 
second analysis tested main and interaction effects of two 
levels of task motivation (A) and two rater measures of task 
persistence (B), treated as repeated measures. 


The results of the first analysis (see Appendix 4.1) 
revealed a significant main effect for level of task 
motivation; F(2,72)=5.78,p <.005. Scheffe multiple 
comparison revealed that neither the mean of the high 
(454.65) or low (253.83) motivation groups differed 
significantly from the mean of the norm sample (305.14); 
p=.06 and p=.71. However, the analysis also revealed that 
the means of the high and low motivation groups differed 
significantly from each other (p=.01). The results therefore 
confirm hypothesis 1 but not hypothesis 1.1 and are 
consistent with the basic derivation from attribution theory 
that perceived task interest as measured by the paired- 
comparison procedure is directly related to task persistance 


70 









' 









71 


in the absence of any external contingencies. Appendix 4.4 
provides descriptive data of task distributions within the 
high and low motivation groups. Appendix 4.2 provides the 
means, standard deviations, and variance ratios. 

The results of the second analysis (Appendix 4.3) also 
revealed a significant main effect for level of task 
motivation (A);F (1,48) = 10.53,p <.005. These results, as were 
the results of the first analysis, are again consistent with 
and confirms hypothesis 1 that individually determined task 
choice and task persistence are related in the absence of 
external contigencies. Appendix 4.5 provides the means, and 
standard deviations of the two groups for the two 
independent raters. 

In addition to these analysis inter-correlations of 
test-re-test measures of the high and low motivation groups 
were computed and then factor analyzed using a principle- 
axis rotation, varimax nethod (Kaiser, 1956). The varimax 
was used since it is both invariant under changes in the 
composition of a test series and satisfies the criteria of 
simple structure best. Thus, factors obtained by this method 
in a sample will have a greater likelihood of portraying the 
universe of varimax factors (Harman, 1967). 

The results of the inter-test correlations (all pre and 
post measures) indicates that only two self-report measures 
correlate significantly with task persistance (see Appendix 
4.6). Both the semantic differential scale-consisting of 






























u< - "• » ’■» jm r •* 









72 


four affective 7-point scales (Test 2) and the summated pre¬ 
rating scale consisting of two 7-point Likert-type scales 
(Test 8) are signigicantly correlated with task persistence. 
However, although these correlations are significant they 
are also moderately low. Inspection of Appendix 4.6 also 
indicates that the test-re-test reliability coefficients 
(over a 20 minute time period) reach significance (test 2 to 
9 inclusive). All of these correlations with the exception 
of the semantic differential consisting of cognitive scales 
are fairly high. Also, with respect to Appendix 4.6 two 
post-test only measures consisting of 0-1 data had too small 
a variance to compute a meaningful correlation coefficient 
(Test 12 and 13). Finally, the inter-rater reliability 
(computed seperately) indicates a significant correlation of 
.99 (p <.001). 

Appendix 4.7 presents the results of the factor- 
analysis of the 13 tests and indicates four factors. The 
four tests with the greatest loading on intrinsic task 
interest are the semantic differential pre- and post-scales 
consisting of the four affective scales, and the pre- and 
post-summated rating scales measuring self-reported task 
enjoyment and task-interest. Interestingly, task persistance 
only correlated at a moderately low level to this factor. 
The tests of the second factor are the semantic differential 
pre-test, the two 11-point Likert-type scales measuring 
preceived effort and perceived interest and enjoyment. 
However, the cognitive post-semantic differential scale also 
















. 








. < ' . 
























73 


correlated only moderately low with this factor. The third 
factor consisted mainly of the pre-and^post-semantic 
differential scales measuring perceived behavioral aspects 
of the task. Factor IV also seems to be a factor consisting 
mainly of the spre-and-post-semantic differential (cognitive 
scales), but also the subject volunteer test correlated 
quite highly with this factor. However, the interpretation 
of these factors because of the relative low sample size 
(N=50) necessitates a cautious bias. 

DISCUSSION 

The results of the analysis testing the prediction that 
individually determined task choice is a valid measure of 
free-time task persistance provide qualified support for the 
prediction derived from attribution theory that defines 
intrinsic motivation as (1) the perceived liking and 
enjoyment for a particular task, and (2) the amount of time 
a subject spends at an activity in the absence of any 
external contingencies. 

The purpose of this study was to test this relationship 
between individually determined task choice and task 
persistance according to the following set of criteria: (1) 
absence of any salient or immediately indentifiable external 
contingencies (reward, experimenter, etc), (2) target 
activities (tasks) are avialable in a free-time situation, 
and (3) subjects had no expectation of any extrinsic reward. 
To the degree that these conditions were met by the 
experimental design the criteria for measured intrinsic 







- 







- 




74 


motivation, as defined by attribution theory, have also been 
met. In conclusion the results provide strong support that 
individually determined task choice, as determined through a 
paired-comparison procedure, is a valid and reliable measure 
of level of s intrinsic motivation as measured by task 
persistance within a free-time situation. 

However, an interesting additional finding demonstrated 
that neither the high nor the low motivation groups differed 
significantly from a previously established norm group, 
although the results of the high compared to norm group were 
marginal. The major significance of this finding is relevant 
to the methodology of measuring intrinsic motivation. That 
is, the paired-comparison procedure did not result in a very 
low level of intrinsic motivation, at least not 
significantly lower than that of the norm sample. This may 
be of particular importance in the interpretation of results 
from study 3 since a "true" insufficiency condition is 
probably not being achieved by means of the present 
methodology. 

Finally, the results of the correlational analysis and 
factor analysis of the self-report measures and task 
persistance scores revealed that only two of the rating 
scales are significantly correlated to task persistence 
(semantic differential-affective scales and the pre-rating 
scale consisting of two summated 7-point Likert-type scales 
measuring perceived interest and enjoyment). Previous 
research seems to be based on the assumption that task 




' 

. ,, |.:« -t. «• O '-“ 9V * 1 “ 






,, .Mel •?- VI ‘ot'W -§ e ’ 








75 


persistance and perceived interest and enjoyment are more 
highly related than this. The factor-analysis revealed four 
factors of which Factor I could be called a perceived 
intrinsic motivation factor consisting of the semantic 
differential-effective scales and the pre- and post-summated 
scale (above) as these are most highly correlated with this 
factor. However, task persistance does not correlate 

substantially with this factor. Keeping in mind that this 

/ 

data must be interpreted with caution since the number of 
subjects sampled was only 50 the results would seem to 
indicate that a behavioral measure like task persistance may 
not be part of the same factor measured by pre- and post- 
self-report measures of intrinsic motivation, as least in 
the absence of extrinsic rewards. This suggests the need for 
further research into the factor structure of dependent 
measures of intrinsic motivation. 

In sum, the results of this study indicate that: (1) 
perceived liking and enjoyment for a task as measured by a 
paired-comparison procedure is a valid measure of task 
persistence during a 10-minute free-time period, and (2) 
that self-repont attitudinal measures are not part of the 
same underlying factor of intrinsic motivation as defined by 
task persistance. 




- 

- 


















































CHAPTER VI 


TEST-RE-TEST EFFECTS: RESULTS AND DISCUSSION 

RE SUL TS 

\ 

To test the effects of pre-testing on both attitudinal 
self-report measures and a behavioral measure of intrinsic 
motivation the data were analyzed in two analysis of 
variance, according to a Solomun Four-Group design (Campbell 
& Stanley, 1963). The first analysis consisted of a 4x2 
Anova of four self-report scales (Test 5,6,7 & 8) with four 
groups of test-re-test and performance combinations (A) and 
two levels of, intrinsic task motivation (B) . The second 
analysis consisted of a 2x2 Anova of task persistance data 
with two test performance groups (A) and two levels of task 
motivation (B). 

Appendix 5.1 and 5.4 inclusive provides the results of 
the first analysis. 

Four test-re-test main effects (A) were significant and 
one motivation level main effect was significant (B). Also 
one interaction effect was significant for the pre-semantic 
differential (affective) test. 

Sem an tic Diffe ren tial Measure ( Affectiv e Scales ): 

The Anova revealed significant effects for test-re¬ 
test-performance orders (A); F(3,56)=7.11, p <.001; for 
motivation levels (3); F(1,56)=60.79, p <.001; and for Order 
(A) x Motivation level t-B) interaction; F (1,56)=5.56, p 


76 




















. • -I " I H ] : .f 










77 


<.005. 

Semant ic D iffer en tia ]. Mea sure (Behavioral S cales) : 

For the data of this semantic differential scale the 
Anove revealed significant effects for motivation level (B); 
F (1,56)-18.78, p <.001; and Order (A) x Motivation level (B) 
interaction; F(1,56)= 2.68, p <.05. 

Sema ntic Differ e ntial Me asu re ( Cognitiv e Scales) : 

This measure revealed only a motivation level (B) 
significant effect; F (1,56)=11•40, p <. 005. 

Summated Batina S cale (Liking a nd Enjoyment Scales) : 

\ 

This measure also revealed only a significant 
motivation (B) level effect; F (1,56)=41 .44, p <.001. 

A Posterio ri Com parison s: 

Because of the significant order and interaction 
effects for two of the semantic differential (affective 
scales, behavioral scales) measures, Scheffe a posteriori 
multiple comparison tests (Kirk, 1969) were conducted to 
test for the specific test-re-test-performance (order) 
effects. Appendix 5.5 summarizes the results of this 
analysis. From a general inspection of this table it is 
evident that within the high task motivation groups all 
measures do not seem to be affected by pre-testing. However, 
in the low task motivated groups the semantic differential 
(affective and behavioral scales) show significant pre-test- 
performance effects. Within the high task motivated groups 
only one significant effect was revealed in relation to the 


















t«i leu** i <*i‘ w wHi" 















78 


semantic differential consisting of behavioral scales. 
Appendic 5.6 provides the means and standard deviations of 
the significant order effects. 

The results of the second Anova (see Appendix 5.7) 
testing the effects of pre-testing on task persistance 
showed only a significant motivation level (B) effect, 
F (1, 28)=23.05, p <.001; and no significant order (A) effect; 

F (1,28)=1.54, p <.05; nor Order (A) x Motivation level (B) 

/ 

interaction effect; F (1,28)=1.04, p <.25. The means and 
standard deviations of the significant motivation effects 
are listed in Appendix 5. 8. 

DISCUSSION 

The analysis revealed that pre-testing on attitude 
self-report measures does not significantly affect task 
persistence. However, two of the self-report measures 
(semantic-differential affective and behavioral scales) were 
significantly affected by pre-testing, performance, and pre- 
test-performance interaction. These results are therefore 
relevant to the interpretation of the findings in study 1 
and study 3 in two ways: First, the results provide 
additional support to the findings in study 1 that task 
persistance- and the various attitude measures do not tap the 
same inderlying factor, since task persistence is relatively 
independent of aftitudinal pre-test effects. Secondly, the 
significant test-re-test effects on two of the attitude 
measures necessitate a cautious interpretation in study 3 
with respect to these measures as dependent measures of 










■ 

















subsequent intrinsic motivation. 

















. 

s. 



















■ 

























, 














- 
























. 















CHAPTER VII 


REWARD MAGNITUDE AND REWARD CHOICE: 
RESULTS AND DISCUSSION 


RESULTS 


The major independent variables under 
this study were: (1) reward magnitude 

monetary reward) (2) reward choice, a 
deprivation. Initial level of intrinsic ta 
treated as an independent variable (high i 
interest) and determined by means of 
validated paired-comparison procedure. The 
measure of intrinsic movivation was task pe 
10-minute free-time period. Also, secon 
intrinsic motivation consisted of six self 
measures. The relative importance of 
dependent variable measures was pointed out 
1). The predictions are formulated primar 
to the behavioral measure of intrinsic moti 


consideration in 
(large and small 
nd (3) reward 
sk motivation was 
nterest and low 
the previously 
primary dependent 
rsistance over a 
dary measures of 
-report attitude 
both classes of 
earlier (study 
ly with reference 
vation. 


The data were analyzed in a 2x5 analysis of covariance 
(Ancova) covarying over test-performance-re-test scores. The 
analysis of covariance tests the significance of the 
differences among means after controlling initial 
differences between groups on the covariate measures, 
tehrefore taking the degree of relationship between the 
covariates and the dependent measures into account 


80 











- 

. 



iao re 


81 


(Kerlinger, 1973). This allows for g sore accurate 
reflection of dependent measure scores between experimental 
groups (Munally, 1975). In the present study individuals 
were assigned randomly to groups, pre-tests were 
administered, the experimental treatments were applied 
(reward conditions), and both post-test measures and a task 
persistance measure taken. However, random assignment to 
groups does not ensure that the groups have equal means on 
all pre-test measures. Therefore, part of the observed post 
test differences among groups may arise from chance 
differences in the pre-test; or, important differences in 
the post-test means may be masked by the chance differences 
in the pretest. 


To summarize, two sets of independent measures were 
collected: (1) a behavioral measure of task persistence over 
a 10-minute free-time period following the administration of 
reward conditions, and (2) attitudinal self-report measures 
collected at the end of the 10-minute free-time period. The 
data from each of these measures are considered seperately 
in the following sections. 

Behavioral Measure of Intrinsic Motivation: 


Of primary interest was the amount of time the subjects 
spent when they beleived their behavior was no longer being 
monitered. The mean time spent for each subject group is 
provided in table 1. The results of the Ancova are 
summarized in table 2. Only the reward conditions (B) 
reached significance; F (4, 130)=6.16, p <.001. 


Neither 










. 




82 


motivation level (A); F (1,130)=1.16, p >.25; nor Motivation 
(a) x Reward (B) ; F (4,130) =1 .48, p >.05, reached a 
significance level. Since, Scheffe multiple comparisons of 
treatment means reflect only differences among the average 
of two means for each treatment it was necessary to again do 
Scheffe a posteriori individual comparisons of means within 
each motivation level to determine which treatments resulted 
in significant change on the dependent variable measures 
from the control groups means and from each other. To 
calculate Scheffe a posteriori comparisons use of the 
adjusted means and MS error 2 was necessitated since the 
original analysis was an Ancova (Kirk, 1968, p. 482). 

To test the hypothesis that the high reward group will 
display less subsequent intrinsic task behavior than the low 
reward group a posteriori comparisons compared the two 
conditions whithin the initially high task interest group. 
This comparison failed to reach significance (F <1). 
Therefore the prediction that a larger reward will result in 
a greater decrement in subsequent intrinsic motivation than 
a smaller reward was not confirmed. Similarly, in the case 
of the initially dull task the hypothesis with respect to 

V 

the two reward conditions of differing magnitude failed to 
be confirmed as the difference between their means also did 
not reach significance (F <1). 

Interestingly, both the high and low reward conditions 
showed significantly less task persistence however than the 
no-reward control group within the initially high interest 


' 

. 


TABLE 1 


MEAN TIME IN SECONDS OVER A 
10-MINUTE FREE TIME PERIOD 

High Task Motivation Low Task Motivation 

Mean S.D Mean S.D 


TD 

239.14 

271.70 

236.27 

259.26 

TC 

222.55 

273.28 

165.30 

261.73 

DP 

440.25 

162.45 

249.47 

219.49 

CH 

300.00 

258.06 

182.02 

284.62 

CO 

543.50 

103.51 

299.38 

242.83 

Note . 

TD = Two 

dollar reward 

condition 



TC = Twenty-five cents reward condition 
DP = Deprivation reward condition 
CH = Choice reward condition 
CO = control reward condition 







































. 




























84 


TABLE 2 

SUMMARY ANCOVA TABLE FOR TASK PERSISTENCE 
DATA: 2 (GROUPS) x5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

57528.379 

1. 153 

P 

>.25 

Reward Conditions (b) 

4 

305960.750 

6.159 

P 

<.001 

Motivation x Reward 
(AxB) 

4 

73542.813 

1.480 

P 

>.05 

Motivation-Reward 
Within Groups 

130 

49671.824 





























































85 


task; F(4,150)=10,38, p <.001 (high reward); F(4,150)=15,41, 
p <.001 (low reward). 

With respect to reward choice, both initially 
interesting and less interesting task groups showed 
significantly less task persistance than their respective 
no-reward controls; F (4,180)=17.28, p <.001 (high task 
motivation); F (4, 180)=13.57, p <.001 (low task motivation). 

Finally, neither of the two deprivation conditions 
(high and low task motivation) revealed a significant 
difference from the no-reward control groups. These results 
are graphically displayed in figure 1. 

Atti t udina l Measures of Intrinsic Mo tivatio n 

The remaining dependent variables concern the subjects' 
perceptions of various aspects of intrinsic interest as 
measured by several self-report scales. A guestionaire 
including the previous pre-test measures and four additional 
post-test only questions was administered after the 10- 
minute free-time period. 

Appendix 6.1 to 6.6 inclusive provide the results of 
the Ancova of all secondary measures of intrinsic 
motivation^ Only the following means of the secondary 
measures showed a significant difference from the control 
group means: (1) post-test summated rating scale mean 
(4.63,S.D.=.71) and control mean (2.80,S.D.=1.38) for low 
task interest and high reward; F(df=4,150)=45.23, p <.001; 
(2) post-only rating scale mean (5.23,S.D.=1.83) and control 














' 

. 


) 





TASK PERSISTENCE 


86 


600.00 


500.00 


400.00 


300.00 


200.00 


100.00 


000.00 



TD TC CH DP CO 


REWARD CONDITIONS 


FIGURE 1 

MEAN VALUES IN SECONDS FOR REWARD CONDITION 
GROUPS (TD = TWO DOLLARS, TC = TWENTY-FIVE 
CENTS, CH = CHOICE, DP = DEPRIVATION, CO — 

CONTROLS,*-—HIGH TASK MOTIVATION,*-=■ 

LOW TASK MOTIVATION). 























































87 


mean (6.3b f S.D. = 1.57) for low task interest and low reward; 
F (df=4,150)=3.03 r p <.025; (3) post-only rating scale mean 
(5.10,S.D.=2.14) and control (6.36,S.D.=1.57) mean for low 
task interest and deprivation reward condition; 
F (df = 4, 150)=3.64, p <.005; (4) post-only rating scale mean 
(4.32,S.D.=1.36) and control mean (6.36,S.D.=1.57) for low 
task interest and choice reward condition, 
F (df = 4,180)=36.95, p <.001, (5) post-only rating scale mean 
(7.46,S.D.=1.29) and control mean (6.91,S.D.=1.23) for high 
task interest and choice reward condition; F(df=4,180)=3.23, 
p <.025; and (6) post-effort scale mean (8.24,S.D.=1.54) and 
control mean (6.67,S.D.=1.59) for high task interest and low 
reward condition; F (df4. 150)=2.92, p <.05. All of these 
results are in the opposite direction from that predicted. 

Finally, a second correlational and factor analysis was 
conducted on the data for all the dependent measures of 
intrinsic motivation (see Appendix 6.7 and 6.8). As in the 
first factor analysis the tests for Factor I (perceived 
intrinsic task interests) consist of tests 2 and 5 (pre- and 
post-semantic differential affective scales) and tests 8 and 
9 (pre- and post-summated rating scale consisting of two 7- 
point Likert-type scales measuring perceived task enjoyment 
and interest). Interestingly, inspection of Appendix 6.8 
shows a slight shift on Test 1 (task persistance) towards a 
lower correlation with Factor I. Factor III and IV were 
loaded by pre- and post-semantic differential measures 
(Factor III - behavioral scales and Factor IV - cognitive 


I • • • • • 

e ( ; . ino- .c (‘>i : ■■ , J. *( 2 ,»* ) 


















s 




, 

- 



88 


scales). However, an important shirt in the factor structure 
of Factor II occurred. This Factor was loaded with Test 9 
(post-summated enjoyment and interest scale), Test 10 (post- 
perceived enjoyment and interest scale). Test 11 (post- 
perceived effort scale), Test 12 (work-leisure question), 

N 

and Test 13 (volunteer again question). The condition common 
to all of these tests is that they were administered after 
the experimental treatment conditions were presented and 
therefore on a post-experimental basis only. Interestingly 
Test 1 (task persistance) correlated highest with this 
factor, although still only moderately low. 

DISCUSSION 

The data of the third study, designed to test the 
effects of four different reward conditions in both high and 
low intrinsic task motivated groups revealed strong main 
effects of extrinsic reward on intrinsic motivation as 
measured by task persistance. For the high intrinsically 
task motivated groups, task persistance during a free-choice 
period decreased significantly below that of the high 
motivation control group, with the introduction of a $2.00, 
250, and choice reward condition. However, the predicted 
effect of reward magnitude on subsequent intrinsic 
motivation was not confirmed. This suggests that a less 
desirable extrinsic reward has the same potency in under 
mining intrinsic motivation than a more desirable reward, 
assuming that $2.00 is more desirable than 250. This seems 
to be somewhat contradictory to the findings in the Lee, 








. 


3 At ipjM 





89 


Syrnyck and Hallschmid (1976) study where the desirability 
of extrinsic reward was found to significantly interact with 
level of task motivation. It also raises some interesting 
questions with regards to quantitative versus qualitative 

differences of reward and timing of reward choice. In the 

\ 

Lee r et al (1976) study subjects were randomly assigned to 
either high or low extrinsic reward conditions based on 
individually determined rather than group validated, reward 
choice. First of all, the rewards were varied on a set of 
qualitatively different extrinsic reward dimensions. 
Secondly, the choice of reward and therefore determination 
of reward desirability occurred prior to task manipulation 
by the subjects. This may in part be related to the 
significant Task x Reward interaction found since reward 
desirability was maximized but choice itself seperated from 
the total reward complex. This suggests that attributions 
were formed on the basis of perceived desirability of 
extrinsic reward in the Lee, et al (1976) study, and 
confirms the attributionally derived hypothesis that a more 
desirable reward and therefore more salient reward would 
focus the individual's attention outward away from the task. 
Similarly, .in terms of competing response effects the more 
desirable reward would act as a more potent distraction than 
the less desirable reward. This raises the question of why 
the small reward (twenty-five cents) condition in the 
present study resulted in virtually the same decrement in 
intrinsic motivation as the large (two dollar) reward, given 
the assumption that the smaller monetary reward was 


44 f»r- ■ xr)K * m 




jjt 






90 


perceived as less desirable by the subjects. Deci (1973) and 
Boss (197 ) have suggested that a small reward could provide 
negative feedback regarding task success. In this light a 
small or negligable reward could act as a status threat and 
mediate the effects of reward on task interest either by: 
(1) mediating a relatively persistant negative effect that 
becomes associated to the task or, (2) by providing negative 
feedback to the subject with respect to task competence. 
However, both of these effects occur only when rewards are 
allocated on a success contingency basis. The present 
experimental procedure was designed to reduce the likelihood 
of negative feedback regarding task success by allocating 
rewards on a performance contingency basis only. 

One possible explanation that has as jet not been 
suggested in the literature could be referred to as input 
saliency or forced attentiveness. A common criterion for the 
existence of cognitive dissonance for example is the length 
of time it takes a subject to make a decision. Jecker (1964) 
has found that as dissonance increases, decision time also 
increases. Similarly, the divergent perceptions by subjects 
within an overjustification situations could lead to the 
subject examining the situation and making decisions as to 
the appropriateness of the rewards and the desirability of 
the rewards. This heightened attentiveness may be a 
concomittant effect that subsequently heightens the saliency 
of stimulus cues. That is, under the effects of heightened 
attentiveness or vigilance the subject may notice more 


« 














91 


subtle cues increasing the saliency of rewards. Although 
this is only speculation and remains to be investigated 
empirically a heightened attentiveness effect may explain 
not only why a small and presumably less salient reward is 
as potent in decreasing subseguent intrinsic motivation as a 

s 

larger more salient reward, but also may have relevance in 
relation to the whole extrinsic reward issue. Heightened 
attentiveness does not need to imply that percieved changes 
in locus of causality will occur. The question of percieved 
change is largely dependent on what the subject becomes 
aware of which could change the psychological meaning of the 
situation for each subject. Even, if a stimulus is 
inconsequential, it may become more noticed under certain 
situations resulting in a different orientation towards the 
task. This may be particularly relevant to brief time 
periods $s in most of the intrinsic motivation studies. 

Finally, the high or low motivated tasks were 
individually assigned, and not group validated. Because 
subjects worked on the lowest task of their choice, it seems 
reasonable to suspect that a dissonance effect may have 
offset the usual incentive effect of extrinsic reward on the 
low task motivation condition (Calder f> Staw, 1975b) or, the 
mean of a sample of subjects in the low task motivation 
groups as in the validation study was not significantly 
different from the previously established behavioral norm 
and thus did not represent a truly low (dull) motivation 
level. This represents one of the major advantages of the 




. .pa*** l=ifs »f.a: ; i#* >.- ' 9ot > 

. 

- 







92 


methodology devised in the present study. More specifically, 
comparison to an independently established behavioral norm 
allows the determination of what could turn out to be a 
relevant task characteristic -- level of task interest in 
relation to the behavioral norm. This would allow empirical 
investigation not only across the entire justification curve 
but also of cross-sample comparisons to help determine if 
the observed relationship between intrinsic motivation and 
extrinsic reward is a more or less consistent. However, the 
finding remains that the negative effect of extrinsic reward 
on an intrinsically interesting activity is not dependent on 
reward magnitude. 

For the choice reward conditions, where both task and 
reward magnitude were individually determined, the results 
demonstrated that an externally imposed reward choice can 
result in subsequent decreases in intrinsic motivation. 
Compared with subjects in the no-reward control groups, 
subjects in the reward-choice conditions, whether they chose 
money or no-money, showed a decrement in intrinsic 
motivation. However, further analysis also revealed a 
significantly greater decrease for the money choice group 
versus no-money choice group (see appendix 7). 

The actual quantity chosen may have been determined by 
an individual difference factor (social desirability factor, 
or task relevance factor). Since subjects were assigned to 
these groups on the basis of their choice behavior no valid 
conclusions can be drawn with respect to cause-effect 


- 











93 


relationships- The results only suggest that either an 
individual difference factor or a shift in locus of 
causality based on reward-choice decreased subsequent 
intrinsic interest. However, the finding remains significant 
and important and awaits further empirical investigation. 

It is . difficult to reconcile these findings within 
either a pure attributional or competing response framework. 
Perhaps both processes are operative in the simultanious 
presence of task intrinsic and extrinsic rewards, and a 
conceptual framework that formulates these underlying 
processes on a complimentary basis would be a more fruitful 
position to take. Perhaps a more appropriate question is not 
whether one model should replace the other, but under what 
conditions the two processes compliment or counteract each 
other. On the basis of these findings and earlier research a 
justifiable assertion is that further isolation of important 
variables at the theoretical level can result in fruitful 
empirical investigation of the underlying psychological 
process involved in overjustification situation. At least to 
this investigator it seems premature to adopt a closed 
system perspective and more appropriate to attempt to 
isolate further important variables (e.g. situational) to 
generate experimentally testable hypothesis with a view to 
incorporate them into a more comprehensive model. 

The competing response hypothesis suggests that there 
are any number of possible effects (e.g. performance 
anxiety, delay of reward frustrations, approach-approach 











94 


attentional conflicts, etc.), so that control of one does 
not rule out the possibility of another operating in the 
experimental setting (Reiss & Suschinsky, 1975). Therefore, 
it might be argued that the choice of reward situation 
represents an approach-approach attentional conflict between 
two positive stimuli (Reiss & Suschinsky, 1975) which could 
be highly distracting and aversive. Granted that there was 
significant reduction in task presistence in the no-money 
group, it remains improbable that the money versus no-money 
groups would not be exposed to the same aversive stimuli 
were one to adopt this position. Also, none of the subjects 
on an open-ended post-experimental questionnaire made any 
reference to such a conflict or any concommitant anxiety. 
Indeed, most of the subjects made reference to a social 
desirability factor (e.g. "no one wants to look cheap"). 

i 

The question of transfer of environmental reward 
expectancy from training to testing situations does not seem 
to be an issue in this study either since the experimenter 
clearly indicated that: (1) the experiment is over, (2) that 
he would leave for a short period of time to process data, 

and (3) that he would return only to ask further questions. 

\ 

Thus the non-recurance of behavior in the rewarded groups 
and recurance of behavior in the control groups does not 
seem likely to be function of an expectation factor. 

Finally, the criticism that reduced task persistence is 
due to a performance anxiety effect (Reiss & Suschinsky, 
1975) again seems inconsistent with the highly probable 


. 

. 


95 


assumption that there would be little difference in the 
perceptions of the choice reward subjects to differentially 
regard the task as evaluative since the experimenter 
stressed in his instructions that the money should not be 
dependent on task performance. 

The important finding remain, that reward magnitude is 
not a critical factor in undermining intrinsic task 
persistence, but that choice of monetary reward is. 

With respect to reward magnitude a puzzling finding is 
the lack of any evidence for the predicted dissonance and 
reward effect. A possible alternative explanation is that a 
true under-justification condition did not exist since 
neither high nor low extrinsic reward affected subsequent 
intrinsic motivation significantly, but did show a non¬ 
significant decrease. 

Finally, the intercorrelations and factor-analysis of 
the same set of variables after reward treatments had been 
introduced in = 144) revealed an interesting shift in the 
factor structure. Task persistance now correlated to a 

r 

higher degree with a factor consisting of all of the post- 
experimental rating measures. This correlation was still 
relatively low however. Causal generalization of course is 
not warranted, however; task persistance and self-report 
measures seem to be less correlated in the absence than in 
the presence of extrinsic rewards. A more directly obvious 
explanation however for the observed factor shift lies in 




, 










96 


the properties of the tests themselves (Guilford, 1954). 
Factors obtained from pre- and post-treatment scales may not 
be identical to factors obtained from post-treatment only 
scales. This becomes even more likely when taking into 
account the results of the second preliminary study which 
indicated significant pre-test effects in the semantic- 
differential (affective and behavioral) scales. This in part 
at least could account for the slightly higher correlation 
of task persistance with Factor II (post-treatment only 
scales) since the post-test rating scales may be more immune 
from pre-test effects. However the results would suggest the 
need for further investigation into the underlying factors 
of intrinsic motivation measures and their hypothesized 
relationships under the influence of and in the absence of 
extrinsic rewards. 

One of the main limitations of this study is that no 
direct measure of attributions was devised. The main reason 
for this was to avoid cross-subject contamination because a 
number of subjects were drawn from the same faculties. No 
satisfactory unobtrusive rating scale or questionnaire was 
devised. This represents a problem for further 

s 

investigation. It would be valuable to replicate the choice 
reward condition using a greater number of subjects for the 
sample and devising an instrument to measure attributions 
directly. These results represent important findings in 
light of research that indicated that perceived control over 
the environment is reinforcing (Weiner & Dubanoski, 1975). 




. 

3 - *f I W* ( - 








97 


Presumably having a reward choice is conducive to perceiving 
greater control over the environment. 

A second limitation relates to the sampling procedure 
of the three studies. Comparison of control group means to 
the previously established behavior norm violates the 
assumption that both samples be drawn from the same 
population since the time span involved cuts across two 
semesters. This limitation does not allow a direct 
comparison of group means to the previously established 
behavior norn (study 1) to determine if a true 
underjustification condition existed in the low task 
motivation groups. As stated above this was hypothesized to 
occur since subjects worked on the least liked task as 
established by the paired-comparison procedure. In future 
research utilizing this methodology, a number of precautions 
could be observed: (1) tasks that have a very low intrinsic 
interest value could be added to the experimental task list, 
and (2) subjects could be assigned their second or third 
choice to ext.ablish three levels of task intrinsic 
motivation (high, low, and medium) . This would allow 
research across the entire justification curve to observe 
changes in the extrinsic-intrinsic motivation relationship 
and represents a new methodology for research on intrinsic 
motivation. 

The reward deprivation condition did not have any 
significant effect on task persistence during the free time 
period in high or low motivated task groups. To reiterate. 














98 


no specific predictions were made regarding the effects of 
this variable on intrinsically task motivated behavior. 
However, the variable has some relevance to Adam*s (1963) 
inequity theory in relation to self and other ratio of 
imputs and outputs comparisons. Since the subjects did not 
perceive, however, any relationship between their behavioral 
inputs and monetary outputs (reward) this cannot be 
considered to be a true inequity situation. Only the 
operation of a purely cognitive deprivation variable was 
observed. Interestingly, the methodology utilized could be 
modified to develop tests of the effects of inequity on high 
and low task motivation provided the induction variable is 
introduced to insure an inequity condition exists as 
outlined by the theory. 

The findings of this series of studies generally 
provide further support for the attributional analysis of 
the relationship between extrinsic and intrinsic motivation. 
However a number of interesting questions are also raised 
that require further investigation. 

The first finding relates to the underlying factor 
structure of intrinsic motivation. This underlying structure 
may be more complex than previously assumed. Perhaps a 
fruitful line of direction would be to consider task 
persistence and various types of self-report measures as 
measuring seperate factors. 


Secondly, the finding that reward, magnitude is not 


. 










99 


significantly related to the degree to which extrisic reward 
undermines task motivation raises further questions in light 
of the Lee, et al (1976) finding that desirability of reward 
is a significant variable. Future investigation could 
manipulate reward magnitude across a much larger range as 
was done in the dissonance literature. A second research 
direction would investigate quantitative versus- qualitative 
reward components (i.e. consummatory vs. monetary reward). 
Perhaps the factor structure of extrinsic reward components 
would provide some interesting research leads also. 

Third,\the question of self-choice of extrinsic reward 
magnitude could be further investigated by employing a 
larger sample of subjects, measuring attributions directly, 
and isolate the individual difference factor. 

As explicated above research across the entire 
justification curve would provide a more comprehensive 
analysis of the relationship between extrinsic and intrinsic 
motivation. 

Finally, the effects of reward deprivation would lead 
to an investigation of equity theory operating under low and 
high task motivation provided the induction factor is 
carefully considered. 




i ) :i 

. 


, : 





CHAPTER VIII 


CONCLUSIONS 

Most of the research on the effects of extrinsic 
rewards on intrinsic motivation have provided relatively 
consistent evidence in favour of attributional and self- 
perception phenomena. However, one distinguishing feature in 
this research is that the extrinsic reward has always been 
disigned to be highly salient to the experimental subjects. 
Most of the empirical evidence shows that when, and only 
when, the .extrinsic rewards are salient does subsequent 
intrinsic motivation decrease. This has raised the 
possibility of competing response effects as an alternative 
explanation to self-perception and attribution phenomena. 
The finding that expected salient, and tangible rewards 
decrease subsequent intrinsic motivation, whereas 
unexpected, social, and less salient rewards do not, is 
consistent with the competing response explanation. 
Therefore, at the present time all of the empirical findings 
reviewed can be understood in terms of both attributional 
and self-perception processes or competing response effects. 

The present study was designed to test several such 
hypothesis derived primarily from attribution theory, but 
explicable also in terms of competing response effects. 
However, the main rationale for the study is found in 
several issues that had remained unexplored up to this 
point. First, all of the previous empirical studies had 


100 























u.n|tj| wltllrii^ »nB] alaali^|U jflfB’H ’' 93 •* 

t 










101 


employed group validated experimental tasks. In the present 
study tasks were individually chosen from a group or set of 
tasks to maximize the feeling of self-investment or 
commitment to the task which is assumed to be intimately 
related to intrinsic task motivation. The results clearly 
indicate that individual task choice is directly related to 
task persistence during a free-time period. This provides 
strong evidence for the validity of this procedure to 
determine level of intrinsic motivation in the absence of 
external contingences. However, with respect to attitudinal 
(self-report) measures of intrinsic motivation it was found 
that task persistence did not correlate highly with these 
measures. Furthermore, a factor analysis showed that these 
measures are not part of the same underlying factor in the 
absence of extrinsic rewards. This is a surprising finding 
since attitude and behavioral measures are frequently used 
interchangeably as dependent variable measures of intrinsic 
motivation. However, one qualification to this is that on 
the introduction of external reward contingencies the post¬ 
test only measures correlated to a slightly higher degree 
with the behavioral measure, than in the absence of rewards. 
Both measures also loaded to a slightly higher degree on the 
same underlying factor in the presence of external rewards, 
than in their absence. 

These findings throw some doubt on the fundamental 
assumption that attitudinal and behavior measures of 
intrinsic motivation tap or measure , the same underlying 




. 








. 






102 


states at all times. Possibly, the Doob (1947) and Bandura 
(1969) argument that attitudes are learned predispositions 
to respond or learned mediating responses is a more 
appropriate way of conceptualizing the relationship between 
attitudinal and behavior measures of intrinsic motivation. 
Particularly, in the presence of individual task choice 
there may be as yet several undetermined processes that 
cause individuals to display different learned responses 
despite the similarities as reported on the attitudinal 
measures. However this remains highly speculative and 
suggests the need for further research into the underlying 
factor structure of measures of intrinsic motivation. 

The second major question investigated in the present 
study is related to the effect on intrinsic motivation of 
extrinsic monetary rewards of differing magnitudes. Ail of 
the previous empirical studies had contrasted only the 
effects of a salient reward with no reward. This left a gap 
in the empirical literture because of the untested 
attributional derivation that the larger the extrinsic 
reward the m ore l ikel y subsequent decrements in intrinsic 
motivation. Conversely, it left untested a second derivation 

s 

from attribution theory that a minimally salient or 
neglibable reward would not decrease intrinsic motivation. 
Given the all or none nature of the attributional process a 
minimal reward could even lead to subsequent increase in 
intrinsic motivation. These predictions were not confirmed 


as both a relatively larger reward and small or neglibable 















. 




103 


reward were found to have the same detrimental effect on 
subsequent intrinsic motivation. 

It might be argued that subjects were responding to the 
informational aspect of a small extrinsic monetary reward. 
In other words, a small extrinsic reward could convey 
negative feedback or have negative cue value. However, it 
will be remembered that the experimental design attempted to 
ensure that subjects would not percieve the rewards 
contingent on any success criteria. The data, therefore, 
would not seem to provide any evidence for this explanation, 
since both\ rewards (high and low) were contingent only on 
performing the task and not on task success. 

A possible explanation presented here for future 
investigation is what may be called a heightened 
attentiveness effect or forced attentiveness effect, which 
has only been explored in relation to dissonce phenomena so 
far (Jecker, 1964). Within an overjustification situation 
subjects may be responding to a number of divergent 
perceptions of various stimuli, examining the situation and 
making decisions as to reward appropriateness, experimenter 
motivation, and so forth. This may lead to heightened 
attentiveness which in turn raises the saliency of various 
stimulus cues. Thus, heightened attentiveness may lead to an 
increase in the saliency of even a small or initially 
minimally salient reward. Interestingly, this effect is 
consistent with the empirical evidence that reward saliency 
mediates the effects of extrinsic reward through reward 






. 

f 

% 


visibility or the subject’s attentional focus (Ross, et al, 
in Press, Ross 1975). However, this also remains highly 
speculative and awaits further empirical investigation. 

Finally, the effects of an externally imposed or 
experimenter imposed reward choice on subsequent intrinsic 
motivation was investigated. Experimental eveidence was 
cited indicating that self-reinforcement leads to greater 
resistence to extinction (Weiner & Dubanoski, 1975). 
However, in the intrinsic motivation literature the effects 
of self-reward had been left unexplored up to this point. 
Therefore tiie effects of externally imposed self-reward on 
subsequent intrinsic motivation were observed with two 
interesting findings: (1) externally imposed reward choice 
leads to reduced intrinsic motivation, and (2) the money 
choice versus no-money choice group showed a significantly 
greater decrease in subsequent intrinsic motivation. The 
reward choice again was also contingent only on performing 
the task and not on task success. Interestingly, both the 
money-choice and no-money choice groups showed a significant 
decrease in intrinsic motivation. This finding would seem to 
support the overjustification hypothesis since in either 

s 

case the subject would perceive himself undertaking the task 
as a means to achieve the reward choice. However, forced 
attentiveness could also act as an alternative explanation 
in this case since subjects may presumably be even more 
prone to examining the experimental situation and make 
decisions as to the appropriateness of their reward choice. 






105 


Finally, the finding that the money choice group showed 
the greatest decrease in intrinsic motivation suggests an 
area for further research. There are three possible 
explanations for this result: il) A possible shift in locus 
of causality as a result of the reward choice, (2) An 
individual difference factor, and (3) An interaction between 
the two. Because subjects were assigned on the basis of 
self-selection it is not possible to draw any valid 
conclusions as to causality. However, the finding remains 
interesting and important enough to warrant further 
research. 


In sum, whereas the data of the present study provides 
some significant support for the hypothesis that both 
monetary extrinsic rewards and monatary reward choice per se 
leads to subsequent decrements in intrinsic motivation, the 
results failed to reveal the expected difference between 
extrinsic rewards of different magnitudes. Also the findings 
with respect to measures of intrinsic motivation raise some 
doubt as to whether attitudinal and behavioral measures tap 
the same underlying state in all situations. 


However, the study also demonstrates the utility and 
validity of a paired-comparison procedure for future 
research across the entire justification curve. Also in 
terms of future research the findings point to a possible 
individual difference factor that could be investigated in 
relation to reward choice. Undoubtedly, there exist 


reward 










' 


















, 




* 
















106 


individual differences that have as yet not 
systematically explored. 


been 
































. 










REFERENCES 


Adams, J.S. "Inequity in Social Exchange." In Berkowitz 

(Ed.) Ad va nces in Experime n tal Soc ial Psychol ogy, 1965, 
2, 267-300. 

Amabile, T.M., Dejong, W. , and Lepper, M. "Effects of 

Externally Imposed Deadlines on Subsequent Intrinsic 
Motivation," J ournal of Personalit y and S ocial 
Psy chology . 1976, 34, 92-98. 

Bandura, A., Pri nciples of B ehavior Mod i vicatio n. New York: 
Holt, Rinehart 6 Winston, 1969. 

Bolstad, O.D., & Johnson, S.M. "Self-regulation in the 

Modification of Disruptive Classroom Behavior" Journa l 
of Appl ied Behavioral Analysis, 1972, 5, 443-454. 

Beckman, L.J. "Effects of Students* Performance on Teachers’ 
and Observers" Attribution of Causality,"J our n al of 
E duca t iona l Research , 1973, 6J, 76-82. 

Bern, D.J. "Self-perception: The dependent variable of human 
performance." Or gan iz ation al Beh avior and Hum an 
P erformance , 1967, 2, 105-121. 

Bern, D.J. "Self-perception Theory," M.L. Berkowitz. Advances 
in Ex perimenta l Social Psy c hol ogy, 6, New York: 

Academic Press, 1972. 

Bern, D.J. "Self-perception: An Alternative 
Interpretation of Cognitive Dissonance Phenomena," 

P sy chol ogi cal Review , 1967, 74 , 183-200. 

Berkowitz, L. "Social Motivation." M. Gardner and E. 

Aronson, The Ha ndb ook of Social Psychology (2nd Ed.), 
1968. 

Brehm, J.W., and Cohen, A.R. Expl orations in cogni t ive 
Dissonance . New York: Wiley, 1962. 

Calder, B.J. and Staw, B.M. "Self-perception of intrinsic 
and extrinsic motivation." Jou rnal of Personality and 
Social Psy ch ology , 1975, 3J, 76-80. 

Goodman, P.S. and Freidman, A. "An Examination of Adams* 

Theory of Inequity," Administrativ e Science Quarterly . 
1971, 16, 271-288. 

de Charms, R. Per sona l Ca u sation : Th e Internal Affecti ve 
D eter m inants of Behavi or. New York: Academic Press, 
1968. 


107 




























































































. • 




















108 


Deci, E.L. "Effects of externally mediated rewards on 
intrinsic motivation." Jou r nal of Personality and 
S ocia l Psych olog y, 1971, J 8, 105-115. 

Deci, E.L. and Cascio, W.F. "Change in intrinsic motivation 
as a function of negative feedback and threats." Paper 
presented at the Eastern Psychological Association 
Meeting, Boston, Mass., April, 1972a. 

Deci, E.L. "The effects of contingent and noncontingent 
rewards and controls on intrinsic motivation." 
O rganizati onal Behav ior an d Hu m an Perf ormance , 19 72b, 8 
, 217-229. 


Deci, E.L. "Intrinsic motivation, extrinsic reinforcement, 
and inequity." Jou rnal of Pers onality and Social 
Psychology . 1972c, 22, 113-120. 

Deci, E.L., Benwane, C. and Landy, D. "The attribution of 
motivation as a function of output and rewards." 

Jour nal of Personal ity, 1974, 42, 652-667. 

Deci, E.L., Cascio, W.F. and Krusell, J. "Cognitive 

evaluation theory and some comments on the Calder and 
Staw critique." Journal of Personality and Social 
P sycholog y, 19 75, 3_1, 81-85. 

Doob, L.W. "The Behavior of Attitudes." Psy cho l ogi cal Review 
, 1974, 54, 135-156. 

Festinger,-L.A. Theory of Co gnitiv e D issonanc e. Evanston, 
Illis Row, Peterson & Co., 1957. 

Festinger, L., and Carlsmith, J.M. "Cognitive Consequences 
of forced compliance." J ournal of Abnormal and Social 
Psychology , 1959, 58, 203-210. 

Gerard, H.B., Connolley, E.S. and Wilhelmy, R.A. In 

E xperi m ent al S ocial Psycholog y, New York: Academic 
Press, 1974 , 7, 231-232. 

Greene, D. and Lepper, M.R. "Effects of extrinsic rewards on 
children*s subsequent intrinsic interest." Chi Id 
Dev elopmen t, 1974, 45, 1141-1145. 

Guilford, J.P. P syc ho metric Methods , New York: McGraw-Hill, 
1954. 

Harman, H.H. Modern Facto r Analysis . Chicago, University of 
Chicago Press, 1967. 

Jecker, J.D. "The Cognitive Effects of Conflict and 

Dissonance." In L. Festinger (ed.) Confl ict, Decision, 
and D issonance . Stanford, Calif: Stanford University 
Press, 1964, 21-30. 

















































, 

- 













109 


Johnson, S.M. "Self-reinforcement Versus External 

Reinforcement in Behavior Modification with Children.” 
D evelo p men tal Ps ych o logy , 1970, 3, 147-148. 

Johnson, S.M. & Martin, S. "Developing Self-evaluation as a 
Conditioned Reinforcer", in B. Ashem & E.G. Poser 
(eds.), Be ha vior Modificaton with Childr en, New York: 
Pergamon Press, 1971. 

Kaiser, H.F. "The Varimax Method of Factor Analysis.” 

Unpublished Ph.D. Thesis, University of California, 
1956. 


Kirk, R.E. Experimen t al Desqn: Procedures for the B ehavior al 
S cien ce. Belmont, California, Brodes/Cole, 1969. 

Kerlinger, F.N. Fo undat ions of B eha v ioral R esearch . New 
York: Holt, Rinehart & Winston, Inc., 1973. 

Kruglanski, A.W., Friedman, I., and Zeevi, G. "The effects 
of extrinsic incentive on some qualitative aspects of 
task performance." J ournal of Personalit y. 1971, 39, 
606-617. 

Kruglanski, A.W., Alon, S. and Lewis, T. "Retrospective 
misattribution and task enjoyment." Journal of 
E xper i men tal S ocia l Ps ychol ogy, 1972, 8, 493-501. 

Kruglanski, A.W., Riten, A., Amitai, A., Margolin, B.S., 

Shabtai, L. and Zaksh, D. "Can money enhance intrinsic 
motivation?: A test of the content-consequence 
hypothesis." J ournal of Persna l ity and S ocial 
Ps y chology , 19 75, 3_1 , 744-750. 

Kruglanski, A.W., Piter, A., Arazi, D., Agassi, R., 

Montequio, J., Peri, I., and Peretz, M. "Effects of 
task-intrinsic rewards upon extrinsic and intrinsic- 
motivation." Journal of Personalit y and Social 
P sycholog y, 1975, 3J, 699-705. , 31., 699-705. 

Kruglanski, A.W. "The Exogenous-Endogenour Partition in 

Attribution Theory." Psycholog ical Review, 197 5, 82, 
387-406. 

Lee, D.Y., Syrnyk, R., and Hallschmid, C. "Self-Perception 
of Intrinsic and Extrinsic Motivation: Effects on 
Institutionanlized Mentally Retarded Adolescents." 
A meric an Jo urnal of Men tal Deficiency . 1976, 81., 331- 
337. 

Lepper, M.R., Greene, D. and Nisbett, R.E. "Undermining 

children*s intrinsic interest with extrinsic reward: A 
test of the overjustification hypothesis." Jou rnal of 
Persnality a nd Social P sychol ogy, 1973, 28, 129-137. 























































' 


































' 

■ 

, 










1 ^ ' • 

. . x ' • » % • ' * 








110 


Lepper, M.R. and Greene, D. "On Understanding 

Over justification: A Reply to Reiss and Suschinsky." 
Journal of P ersonal ity and Soc ial Psycholog y, 1976, 33, 
25-35. 

Miller, D.T., and Ross, M. "Self-Serving Biases in the 
Attribution of Causality: Fact or Fiction?" 
P sych ol ogi c al Bulle t in, 19 75, 82, 213-225. 

Mishel, W. , Ebbesen, E.B. and Zeiss, A.P. "Cognitive and 
attentional mechanisms in delay of gratification." 

J ournal of P ersona lity and Socia l Psycholog y, 1972, 2_1, 
204-218. 

Notz, W.W. "Work motivation and the negative effects of 
extrinsic rewards: A review with implications for 
theory and practice." Amer ican Psycholog ist. 1975, 884- 
891 . 

Nunally, J.C. Psycho metric Theory . New York: McGraw-Hill, 

1 967. 

Nunally, J.C. I ntrod u ction to S tatistic s Fo r Psychology and 
Educatio n. New York: McGraw-Hill, 1975. 

Pritchard, R.D. "Equity Theory: A Review and Critique." 

O rganizati o nal Behavio r and Human Per f ormance . 1969, 4, 
176-211. 

Reiss, S. and Suschinsky, L. "Overjustification, competing 
responses, and the acquistion of intrinsic interest." 

J ou rnal of P ersonality and Socia l Psychology . 19 75, 31 . 

1116-1125. 

Reiss, S. and Suschinsky, L.W. "The Competing Response 

Hypothesis of Decreased Play Effects: A Reply to Lepper 
and Greene." J our nal of Pers on alit y an d Socia l 
P sy ch ology , 1975, 33 , 233-244. 

Ross, M. "Salience of reward and intrinsic motivation." 

Jo urn al of P ersona lity a nd Social Psycho l ogy . 1975, 32, 
245-254. 

Ross, M., Rarniol, R. and Rothstein, M. "Reward Contingency 
and Intrinsic Motivation in Children: A Test of the 
Delay of Gratification Hypothesis." Jo urnal of 
Person alit y and Social Psycholog y(In Press). 

Rotter, J.B. Social L earnin g and Cl i nical Psyc holo gy. 

Englewood Cliffs, New Jersey: Prentice-Hall, 1954. 

Weiner, B. "Attribution Theory, Achievement Motivation, and 
Educational Process," Review of Educatio nal R esea rch. 
42, 203-215. 





















































. 

* • 


: 











Weiner, H.R. and Dubanoski, R. A. “Resistance to Extinction 
as a Function of Seif- or Externally Determined 
Schedules of Reinforcement." Journal of P ersonali ty an 
S ocia l Psy ch ology , 1975, 31_, 9 05-910. 

Weiner, B. and Kukla, A. “An Attributional Analysis of 

Achievement Motivation," J our nal of Personalit y an d 
Socia l Psy ch ology , 1970, 21., 1-20. 



















. . 













APPENDIX 1 


DESCRIPTION OF EXPERIMENTAL TASKS AND 


PILOT STUDY RATING FORMS 






113 


INSTRUCTIONS 

A. Pilot Study Instructions* 

This is a list of the various puzzles, games and tasks 
that you are asked to rate. The numbers in the left column 
^column 1) correspond to the task or the puzzle displayed. 
Please familiarize yourself with each of the tasks by 
reading the brief instructions above the tasks. You may also 
manipulate each task for this purpose. Once you have done 
this choose the three tasks you find the most interesting 
and enjoyable and then the three tasks you find the least 
interesting and enjoyable. Check off your choices in the 
boxes opposite to the task. 


Note: *The Headings A and B are for the benefit of the 

reader only and were not included on the original 
questionnaire in the pilot study. 














































114 


TEST RATING FORM 
E. Pilot Study Rating Form 

Task No. Most interesting Least interesting 

and enjoyable and enjoyable 


1 . 

2 . 

3. 

4. 

5. 

6 . 

7. 

8 . 
9. 

10 . 

11 . 

12 . 

13. 

14. 

15. 

16. 

17. 

18. 

19. 

20 . 
21 . 
22 . 

23. 

24. 

25. 

26. 

27. 

28. 

29. 

30. 


1 . 

2 . 

3. 

4. 

5. 

6 . 

7. 

8 . 

9. 

10 . 

11 . 

12 . 

13. 

14. 

15. 

16. 

17. 

18. 

19. 

20 . 

21 . 

22 . 

23. 

24. 

25. 

26. 

27. 

28. 

29. 

30. 



































































































115 


BRIEF DESCRIPTION OF EXPERIMENTAL TASKS 


1. Jigsaw Pizzle 


2. Water Colour Set 


3. Crossword Puzzle 


u. Block Design 


5. Business Writing 


This was a regular 500-piece jigsaw 
puzzle with a coloured pastural 
scene. 

This included seven cakes of colours 
including white, a stack of white and 
coloured paper, two seperate cups of 
water and two brushes. 

This task consisted of a regular 40- 
60 word crossword puzzle mounted on a 
white paper background. 

This task consisted of 100 coloured 
cubes (2 centimetres square) which 
subjects arranged according to 30 
three-dimensional designs. 

This task included a series of 
diverse writing tasks (e.g. letter 
writing, memo writing, spelling. 


etc. ) 




























- 























APPENDIX 2 


MAIN EXPERIMENTAL INSTRUCTIONS 


116 



EXPERIMENTAL INSTRUCTIONS 


S tudy One 

The subjects came individually to the testing room from 
a waiting area nearby. Upon entering the testing room, each 
subject was seated at a table upon which the five 
experimental tasks were located in a row directly in front 
of the subject. At this point the experimenter sat himself 
behind the table directly facing the subject and began: 

"My name is Mr. Hallschmid. Before we start I would 
like to present to you a series of two tasks, and in each 
case I would like you to point to the task, of each pair 
that I place in front of you, that you find the most 
interesting and enjoyable on the basis of how you feel at 

the moment." 

/ 

If the subject had no questions, the experimenter 
continued and presented the pairs of tasks and recorded 
choices on a worksheet (see Appendix 3) which was placed out 
of view of the subject. Once all possible randomized pairs 
had been presented the experimenter continued: 

"Thank you, now would you please sit over here 
(pointing and escorting subject to the other end of the 
table) . " 












- 














118 


The experimenter placed the appropriate task in front 
of the subject together with the rating scales: 

"Now would you please fill out these questions, as soon 
as you have done so I will return and give you the rest of 
the instructions.” 

If the subject had no questions the experimenter left 
and observed the subject through the one-way mirror until 
the subject had completed the ratings 

"Alright, thank you. Now I will go into that room 
(pointing to the appropriate spot) and give you your 
instructions over the intercom." 

Once the experimenter had reached the monitor room he 
continued and provided the instructions over the intercom: 

"Please look at the task in front of you. I will ask 
you to start working on this task in a minute. You may work 
on the task in any way you think is appropriate until time 

i 

is up. Are there any questions before we begin? Alright you 
may start now." 

Now was an audio cue for the two independent raters to 
begin timing. The subject worked on the task for a 10-minute 
interval at which point the experimenter wonld interrupt 
over the intercom: 

"Alright (rater cue) time is up. I have made some 
recordings while you were working on (the task) and have to 


-JO - 93. »il‘ bO'i avJfc;. orf* Au’fs'i Ulw f d* W** * vftd oa X 25 

. 

■ 


119 


leave for a while. You may do anything you like until I 
return, but please remain in the room since I will probably 
ask you some more questions when I return. Thank you." 
(Rater cue.) 

Experimenter left and walked down a long hallway. 
Raters now recorded task persistance during a 10-minute time 
interval which was the primary dependent measure. Within the 
experimental room the subject had access to the task he had 
worked on, the remaining four tasks, and distraction stimuli 
(magazines). At the end of the 10-minute interval the 
experimenter, entered the room (rater cue) and continued: 

"Thank you for waiting. Before you leave I would like 
you to fill out again these questions." 

Once the subject had completed the rating scales the 
experimenter terminated the experiment with a few questions 
and debriefed the subject to the extent possible at this 

time. All subjects were asked not to disclose the contents 

{ 

of the experiment to their classmates or friends at 
university. 

Stu d v Tw o 

Slight modifications had to be introduced in 
conformance with the parameters of this study. When subjects 
did not work on a task the experimenter indicated he would 
leave to process some data and asked the subjects to stay in 
the room as in the above instructions. These were the only 
















, 





120 


modifications necessary. Finally, in the performance post¬ 
test group subjects were asked to work on the appropriate 
task over the intercom without prior rating. 

Study Thr ee 

The only modifications necessary from the main 
experimental procedure of Study One revolved around the 
presentation of the monetary reward. The experimenter read 
the instructions over the intercom in accordance with each 
reward condition as follows: 

(1) High and low reward conditions 

"Please look at the task in front of you. I will ask 
you to start working on this task in a minute. You may work 
on the task in any way you think is appropriate until time 
is up. (At.that time I would like you to take the two 
dollars (Twenty-five cents) in the drawer to your right. The 
money will then be yours. Please note that receiving this 
money is not contingent on how you perform on the task. It 
is only contingent on you working on the task until I tell 
you time is up.)"* 

After the subject had worked on the task for 10-minutes 
the experimenter interrupted over the intercom: 

"All right time is up. (Please take the two dollars in 
the drawer, place the money in the envelope and place it in 
your pocket. The money now belongs to you.)" 













‘ 





























- 

* 





121 


* Eaters were unable to hear the bracketed statements. 

(2) Choice condition 

"Please look at. the task in front of you. T will ask 
you to start working on this task in a minute. You may work 
on the task in any way you think is appropriate until time 
is up. (At that time I would like you to take any amount of 
money that is in the drawer to your right. Choosing the 
money is only contingent on you working on the task until 
time is up, not how you perform on this task. You may take 
any amount you prefer, all of it, none of it, or some 
quantity in between.)" 

After 10-minutes had elapsed the experimenter 
interrupted over the intercom: 

"All right time is up. (Please take any quantity of 
money you prefer from the money in the drawer.") 

(3) Deprivation condition 

Procedure was exactly the same except that all subjects 
were told that they are not receiving any money and that 
this was decided on a purely random basis. Subjects were 
reminded of this again after the completion of the first 10- 
minute time interval. 

These instructions were given to all subjects with the 
minimum amount of deviation possible. 


- 

, 




APPENDIX 3 


EATING SCALES FOE INTEINSIC TASK INTEEEST 


122 






















123 


INSTRUCTIONS 


A. Pre-Treatment Instructions* 


This is a study in cognitive learning that requires the 
measurement of feelings towards a task. This is done by 
having various people judge these tasks on a series of 

scales. In taking this test, please make sure your 

judgements are on the basis of your feeling s about the task. 
On each of the following pages you will find a set of scales 
and you are to rate the task presented to you on each of the 
scales in their respective order. Here is how you are to use 
the scales: 

If you feel the task is v er y b orin g, you should place 
your check mark as follows: 

boring X :_:_:_:_:_ : interesting 

If you feel the task is ver y intere stin g, you should 
place your check mark as follows: 

boring_:__:_:_:_:_: X : interesting 


If you feel the task is f airl y boring or f airly 
inte re sting , you should place your check mark as follows: 

boring_: X :_:_:_:_:_: interesting 

OR 


boring 


X 


: interesting 































124 


If you feel the task is onl y slightly borin g or onl y 
slightl y in teresting , you should place your check mark as 
follows: 

boring_:_; X :_:_:__: interesting 

OR 

boring_:_:_:_: X :_:_: interesting 

If you consider the task to be neutral on the scale, or 
if the task is equally boring or i nte r esting you should 
place you check mark in the middle space. 

Please do not omit any item and never place more than 
one check mark in a single scale. Work at a fa irly h ig h 
speed thr ougho ut the guesti o nnair e. It is your immediate 
impression, the immediate "feeling" about the task that we 
want. Thank you for your cooperation. 




*N ot e: The headings A to F are for the benefit of the 
reader only and were not included on the 

questionnaires used in this study. 
















■ 










JIGSAW PUZZLE** 


B. Semantic 

Good 
Easy 
Slow 
Free 
Ordinary 
Rational 
Monotonous 
Passive 
Interesting 
Ambiguous 
Pleasurable 
Simple 
Rigied 


Differential Scales* 

_:_:_: ; Bad 

_:_:_:_: Hard 

_:_:_:_: Fast 

_:_:_:_: Constrained 

_:_:_:_: Novel 

_: :__: Intuitive 

_:_ :_: Exciting 

_:_:_:_: Active 

_:_:_:__: Boring 

_:_:_:_: Clear 

_:__:_:_: Painful 

_:__:_:_Complex 

: : : : Loose 


Note 


' * Developed by Calder and Staw (1975) 
** Titles appropriate for task 

















































- 



















126 


C. Pre-Treatment Summated Eating Scale 

On the basis of your own impressions, rate the 

following questions as you did the first set: 

1. To what extent do you find this task enjoyable? 

JIGSAW PUZZLE 

Very Very 

unenjoyable_: _:_:_:_:_: enjoyable 

2. To what extent do you find this task interesting? 

Very Very 

uninteresting_:_:_:_:_: :_: interesting 























. 






















127 


D. Post-Treatment Instructions 

New that you have completed the experiment please rate 
the task again on the basis of your fe eling s towards the 
task. On each of the following pages you will find a set of 
scales and you are to rate the task presented to you on each 
of the scales in their respective order. Here is how you are 
to use the scales: 

If you feel the task is very b orin g, you should place 
your check mark as follows: 

boring _X_ :_:_:__:_:_:_: interesting 

If you feel the task is ver y int eresting , you should 
place your check mark as follows: 

boring_:_:_:___:_: X : interesting 

If you feel the task is f airl y boring, or f airly 
interesti ng, you should place your check mark as follows: 

boring_: X :__:_: : :_: interesting 

' OR 

boring_:_:_:__: interesting 

If you feel the task is onlj slightly boring or only 
s lig htly int erestin g, you should place your check mark as 
follows: 

boring_:_: X :_:_:_ : interesting 

OR 

boring __:_:_:_:_X_:_:_: interesting 


If you consider the task to be neutral on the scale, or 






















UC 

^ __ 

- 















128 


if the task is equally boring or i nteresting you should 
place you check mark in the middle space. 












129 


E. Post-Treatment Pvating Scales 

Please rate the following questions as you did the 
first set: 

s 

1. To what extent do you find this task enjoyable? 

JIGSAW PUZZLE 

Very Very 

unenjoyable _:_:_:_:_:_:_: enjoyable 

2. To what extent do you find this task interesting? 

Very Very 

uninteresting_:_:_:_:_:_: : interesting 

3. To what extent did you work on the task because you 
found the task interesting and enjoyable? 

Very Not 

much_:_:_:_:__:_:__:_:_:_: at all 

4. How much effort did you put forth working on the task? 

Very Very 

much_:_:_:_:__:_:__:_:_:_: little 

5. Did the task seem like work or pleasure? 

Work_ Pleasure_ 

6. Would you volunteer for another similar experiment? 

Yes 


No 






















. 























F. Experimenter Pair Comparison Worksheet 
Randomized Pairs 


C-E 

A-D 

B-E 

C-D 

B-C A 3 C D E 

A-C 

B-D 

D-E 

A-B 

A-E 


N ote : 

A = Jigsaw Puzzle 
B '= Water Colour Painting 
C = Crossword Puzzle 
D = Block Design 
E = Business Writing 






























APPENDIX 4 


DATA SUMMARY TABLES OF STUDY 1 



131 
























132 


Source 


Between 
W it hin 


APPENDIX 4.1 

ANOVA SUMMARY TABLE FOR FREE CHOICE 
TASK PERSISTENCE OF HIGH, LOW, 

AND NORM SAMPLE GROUPS 

df MS F P 


Groups 2 280887.50 5.78 p <.005 

Groups 72 48602.05 












































133 


MEANS 


HIGH 


APPENDIX 4.2 

PROBABILITY MATRIX OF SCHEFFE MULTIPLY COMPARISON 
OF TASK PERSISTENCE MEANS OF HIGH, LOW, 

AND NORM SAMPLE GROUPS 

HIGH LOW NORM 

454.65 253.83 305.14 

S.D.=202.64 S.D.=236.48 S.D.=220.96 


.01 


.06 


LOW 


71 



































































134 


APPENDIX 4.3 


ANOVA WITH REPEATED MEASURES SUMMARY TABLE 
FOR FREE-CHOICE TASK PERSISTENCE DATA: 

2 (GROUPS) X 2(REPEATED MEASURES FOR RATERS) 


Source 

df 

MS 

F 


T 

Between Subjects 
Motivation (A) 

1 

1013701.20 

10.528 

P 

<.005 

Subjects Within 
Groups 

48 

96289.10 




Within Subjects 

Rater (B) 

1 

253.90 

0.868 

P 

<.25 

Motivation x Rater 
(A x B) 

1 

332.00 

1.135 

P 

<.25 

E x Subject Within 
Groups 

48 

292.60 
























































135 


APPENDIX 4.4 

TASK DISTRIBUTION WITHIN HIGH AND 
LOW TASK MOTIVATION GROUPS 



HIGH 

LOW 

JIGSAW PUZZLE 

5 

4 

WATER COLOUR PAINTING 

4 

6 

CROSSWORD PUZZLE 

5 

5 

ELOCK DESIGN 

6 

3 

BUSINESS WRITING 

5 

7 






















































136 


APPENDIX 4.5 

SIGNIFICANT MOTIVATION LEVEL DIFFERENCES 


Group 

High Task 

Inte rest 

Low Task 

Interest 


Rater 1 

Rater 2 

Rater 1 

Rater 2 

Mean 

454.45 

454.80 

256.67 

249.90 

Standard Deviation 

248.07 

24 7.53 

172.45 

171.00 























. 





































CORRELATION MATRIX OF INTRINSIC TASK INTEREST MEASURES (N=50) 


137 


p 











0 

0 

O 

0 

0 

0 

O 

0 

O 

O 

O 

p 












0 

O 

0 

0 

0 

O 

0 

O 

O 

O 










* 

* 

CO 

S3 

0 

CN- 

0 

C^ 

vn 

0 

p 

ON 

0 

VO 

-ct 

CM 

P 

P 

00 

p 

1 

SR 

p 

0~N 

vO 

CM 

* 

vO 

O 

t- 

* 

4t 

On 

C^. 

CRS 

MD 

p 

0 

p 

VO 

P 

c^- 

VO 

p 

00 

p 

O 

O 

CM 

r r ', 

p 

p 

1 

O' 

0 

* 

VO 

P'N 

* 

VA 

P 


P 











Eh 


* 



* 



* 



2 

CO 

0 

CM 

O 

vO 

VO 

VO 

00 

ON 

p 

* 

ON 

C^ 

vO 

0 

On 

P 

ON 

-t 

p 

* 

On 

£N- 



p 











Eh 





* 






< 

cc 

p 

-t - 



* 

it 

★ 

O 

0 




CO 

0 

CN, 


5 

p 

p 

vO 

Cn- 

CM 

VO 

CM 





0. 

o 

Q 

C/l 


o 

gp 

P-. 

c/i 


o 

Ph 

< 

Q 

c/l 


CP 

o 

Q 

CO 


CP 

CO 

Q 

CO 


CP 

< 

Q 

CO 


CP 

Eh 


10 

-P 

(0 

0 

+> 

(0 


0) 

-p 

c 


10 

aJ 

Eh 


(0 

G 

P 

M 

P 

C 


o 

o 


CM 

cm 


* 

^t 

cc 

CM 


P 

O 



rR 

p 

00 

P 

P 

CM 

P 

-t 

CM 


* 



* 


* 



* 

ON 

O 

ON 

On 

f—H 


CM 

CM 

^t 

C rN , 

CM 

-t- 

C^- 

O 

-q- 


* 




ON 

▼ 

rv 

-t - 


C^- 


00 

O 


CM 

00 

2h 

P 



0 

o 


p 

(0 

•H 

(0 

p 

0) 

CP 

** 

(0 

id 

Eh 


8? 


00 

CM 

it 


vO 


p 
P (0 
cd <D 
P Eh 
P I 
G Q) 

££ 


cd 


p 

(0 

0) 

Eh 

I 


<D 

<H 

Ch 


0 

> 

P 

P 

o 

a) 

<H 

<H 

< 


£ 

0) 

CO 


CM 


s e 
s* 

0 P 

<H CTj 
<H P 
P P 
Q O 
> 

o id 
P jo 
p 0 

q cn 

id 

£ 

0 

CO 


p 
p CO 
id 0 
P Eh 
P I 
G 0 

££ 
0 - 
<H 0 
<H > 
p P 
Q P 
P 
O G 
P bo 
P o 
a 


ss § 


p 

(0 

P 0 
cd Eh 
P I 
P P 
G (0 

(S 

0 - 

P 0 
P > 
P P 
O P 

o 

O 0 
p p 
p p 
< 


e 

0 

CO 


£ 

0 

CO 


vr> 


* 

o 

CM 


p 

(0 

0 

Eh 

P 

(0 

O 

P 


S(S B 0 : 


0 

P 

P 


P 

s 

£ 

0 

CO 


VO 


> 

id 

x. 



cd 

o 

co 

be 

G' 


P 

cS I 


oo 


p 
CO 
0 
Eh 
I 

T3 P 
0 0 
P O 
0 P 


ON 


0 >> 
p p 

g<§ 

w 

M 

w> c 

C P 
p p 




>> 

p 

c 

o 

bO 

•I 


C • 
O 
P 
P 
CO 
0 

(5 


p 
T 3 10 
0 O 
C P 

•fH '— y 

X 

£ 

c 

o 


cd 

o 

CO 

t 

<2 

P 

Crl 


>> 

P 


0 P 

0 I 
0 P 
0 10 

H <2 


G 

O 


I 

■fi 

O 


P G 
W P 
0 P 

<§ £ 

U P 

0 w 

0 O 
p p 
G w 
3 
P 
O 
> 


Note ; *p <.05 

**p <.01 

Decimal point omited 
























































































































































































































































FACTOR ANALYSIS OF INTRINSIC TASK INTEREST TEST SCORES (N-50) 

A VARIMAX ROATION 


138 



o 

ca 

00 


Hi 

A- 

cn 


o 

cc 

a- 


00 

O' 

a 

a 

ca 

(A 

O 

VO 

cvj 

ca 

o 

ON 

va 

va 

00 

vo 

cc 

00 

VA 

VO 

ca 


nl 

A 

A 


> P 


C 

So 

o 

u 


VA 

On 

O 


A 

CA 

O 


O 

Hi 

rH 

I 


CVJ 

rH 

VO 

On 

ca 


ca 

rH 

VA 

ON 

o 

ca 

O 

O 

va| 

rH 

rH 


I 


Hi 

C\i 

o 

I 






C^J 

o-| 

nt 1 

rH 

A 

00 

CA 

VO 

Hi 

00 

a 

rH 



rH 

rH 


2 


O 

P *H 

O' 

a 

VO 

o 

vo 

H 1 

CO 

CO 

ca 

o 

ca 


O' 

nt 

vO 

M > 

00 

A 

00 

o 

A 

00 

_h- 

ON 

3 

CA 

on 

CA 

00 

C\J 

vO 

VA 

M jjj 

CVJ 

CVJ 

A 

rv 

OsJ 

col 

3- 

rH 

o 

rH 

o 

o 

rH 

O 

£ 










1 

i 


1 

CVJ 

VA 

rH 



ca 


o 

VO 

On 

O 

va 

O' 

Hi 

tv 

rH 

A 

va 

vO 

O 

ON 

rH 

cH 

O 

rH 

cvj 

CVJ 

Hi 

vo| 


I 


VA P 

00 CVJ 

0 I rH 



I 


*A 

vO 

O 

cvj 


rH 

00 

00 


P 

(0 


0 

P 

C 


TJ 

<r> 

-p 

& 


M ^ 

CA 

vO I 

CVJ 

rH 

Ol 

o 

o 

col 

ON 

h! 

vo 

CA 

va 

C/3 

ON 

VO 

3 

VA 

00 

Hi 

Hi 

Hi 

CO 

ON 

Hi 

CA 

A 

e5 

CVJ 

001 

CA 

Cvj 

l 

ool 

CVJ 

o 

oo| 

aI 

CVJ 

rH 

CA 

O 

» 


& 


VO 

CVJ 


p 

to 

<D 

H 




/-—V 



S 


P 










r—v 

p 



p 


CO 

p 





G 




rH P 

rH U 

rH -P 

rH 

(0 


0 

P CO 

0 

0 

0 


O 




cd to 

as © 

cd (0 

cd 

0 

cd tH 

aj © 

i—1 

1—1 

P N 

/ —s 

P 

—V 



P <D 

P Eh 

-H 0 


P 

t 


cd 

cd 

ct) >5 

>> 

P 


> 


-p Eh 

P I 

-p h 

P 

1 

P 

P 

P 1 

o 

o 

o P 


(0 


G P 


C 1 

C 0) 

G I 

g 

P 

G 

(0 

d -P 

03 

03 

03 C 


0 

c 

O G 

<D 

O 

2 8 
(v £ 

2£ 

8 8 
0 P 

2 

to 

o 

2 

o 

CP 

2 S 

M 

bC 

O 

bC 

3 

o 

P O 

c 

0) 

Q) 

Ph 

a) 


© CL. 

c 

c 

d be 

bC 


bo 

(0 bl 

m 

<M 

«H rH 


<M 


<H 

rH 


p 

p ^ ' 

P c 

c 

0 

c 

0 G 

P 

tH <D 

tH rt 

«h 0 

cm 

0) 

P 

cd 

<H © 

p ^ 

p p 

p p 

0 P 

N 

p 

<§> +j 

(0 

P > 

*H U 


P 

> 

P 

u 

P > 

0 

cd 10 

o? a) 

Eh 


p p 

p 

p 

«H 

JO 

Q P 
-P 

Q O 

P 

Q P 
-P 

Q 

P 

P 

« 

o 

p 

O p 
p 

g<g 

in 

cd 

<8 

^ cS 

u 

o o 

O > 

O P 

O 

O 

o 

> 

O P 

Ti Eh 

T* 1 

Ti I 

03 1 

0 

l 

0 I 

CD 

P 0) 

p cd 

P p 

P 

a> 


cd 

P c 

0 1 

0 P 

0 P 

P 

£ 

P 

© -H 

Ph 

-H <H 

P -C 

-P bD 

P 

<M 

p 

r! 

p be 

P 0 

P CO 

C (0 

P w 

(0 

p m 


ssa 

C <D 

c o 

g 

<H 


£ 

c o 

$1 

0 O 

P o 

H O 

i 

o 

d o 

P2 

n) m 

cd o 


C 

Cd 

cd O 

e pl, 

fi pL. 

O Q-i 

a: 

LM 

3 pH 

(0 


e ^ 


£ 

_ _ ' 

g 

v 

g v_/ 

M '- / 

M '-- 


P ^ 

M 

-' 


cd 

0 

0) 

0 

(1) 


0) 


0 

P 

P 

o 

tH 

O 


O 

H 

03 

03 

03 

03 


03 


03 

03 

03 

a 

w 





CVJ 

(A 

Hi 

VA 


vO 


CN 

00 

ON 

o 

p 

CVJ 


CA 












p 

p 

p 


P 


0) 

o 

G 

cd 


1 


O 

§ 

■c 


cd 

> 

rH 

cd 

-P 

o 

<m 

o 




Note ! Decimal point omited 














































































. 






































APPENDIX 5 

DATA SUMMARY TABLES OF STUDY 2 


139 






140 


APPENDIX 5.1 

ANOVA SUMMARY TABLE FOR AFFECTIVE SEMANTIC DIFFERENTIAL 

DATA: 4(GROUPS) X 2 (LEVELS) 


Source 

df 

MS 

F 

P 

Order (A) 

3 

6.914 

7.1 1 

p <0.001 

Motivation Level (B) 

1 

59.097 

60.79 

p <0.001 

Order x Motivation 
(A x B) 

1 

5.412 

5.56 

p <0.005 

Within Order-Motivation 
Groups 

56 

0.972 










' 




































. 



















141 


APPENDIX 5.2 

ANOVA SUMMARY TABLE FOR BEHAVIORAL SEMANTIC DIFFERENTIAL 

DATA: 4 (GROUPS) X 2 (LEVELS) 


Source 

df 

MS 

F 


P 

Order (A) 

3 

1.293 

1.63 

P 

<0.05 

Motivation Level (B) 

1 

14.82 2 

18.78 

P 

< .001 

Order x Motivation 

Level (A x B) 

3 

2.117 

2.68 

P 

< .05 

Within Order-Motivation 
Groups 

56 

0.789 











































. 





























142 


APPENDIX 5.3 

ANOVA SUMMARY TABLE FOR COGNITIVE SEMANTIC DIFFERENTIAL 

DATA: 4(GROUPS) x 2 (LEVELS) 


Source 

df 

MS 

F 


P 

Order (A) 

3 

0.375 

1.098 

P 

<. 25 

Motivation Level (B) 

1 

3.900 

11.401 

P 

<.005 

Order x Motivation 

Level (A x B) 

3 

0.13 7 

0.401 

P 

<. 75 

Within Order-Motivation 
Groups 

56 

0.342 












































143 


APPENDIX 5.4 

ANOVA SUMMARY TABLE FOR SUMMATED RATING SCALE 
DATA: 4 (GROUPS) X 2 (LEVELS) 


Source 

df 

MS 

F 


P 

Order (A) 

3 

3.316 

2.058 

P 

in 

o 

• 

V 

Motivation Level (B) 

1 

41.441 

25.7 19 

P 

<.001 

Order x Motivation 
Level(A x B) 

3 

2.014 

1.2 50 

P 

<. 25 

Within Order-Motivation 
Level Groups 

56 

1.61 1 























































144 


APPENDIX 5.5 


SUMMARY OF SCHEFFE A POSTERIORI COMPARISONS 

OF MEANS FOR SEMANTIC DIFFERENTIAL 

AFFECTIVE AND BEHAVIOR SCALES 

High Task Low Task 
Motivation Motivation 

Test Comparisons -—-——--— 

df ( 1,56) F df (1,56) F 



A 1A2 

2.69 

1.42 

' 

A1 A3 

0.60 

45.42* 


A 1 A4 

0.06 

0.37 

S.D. (Affective) 

A2A3 

0.74 

30.77* 


A2A4 

1.92 

3.25 


A3A4 

0.28 

54.02* 


A1A2 

3.60 

3.97 


A1 A3 

1.85 

4. 90 


A 1 A4 

1.23 

0. 73 

S.D. (Behavioral) 

A2A3 

10.61 * 

0.05 


A2A4 

0.63 

8. 1 0* 


A3 Ah 

6.08 

9. 41* 


Note : * p < .05 

A1 = Test-Performance-Test 
A2 = Test Test 
A3 = Performance-Test 
A4 = Test 


















































































































































145 


APPENDIX 5.6 


SUMMARY TABLE OF MEAN AND STANDARD DEVIATIONS OF 
SIGNIFICANT TEST-PERFORMANCE-RE-TEST DIFFERENCES 


TEST 

SEMANTIC DIFFERENTIAL 
(AFFECTIVE) 


SEMANTIC DIFFERENTIAL 
(BEHAVIORAL) 

Al 

2.69 (S.D.=.41) 

A2 

4.94 (S.D.=.33)* 

A3 

5.12 IS .D . = .32) 

A3 

3.91 (S.D.=.24)* 

A2 

3.87 (S.D. = .28) 

A2 

3.87 (S.D.=.28) 

A3 

5. 12(S.D. = .10) 

A4 

2.97 (S.D.=.36) 

A3 

5. 12 (S .D. =. 10) 

A3 

5.12 (S.D.=.32) 

A 4 . 

2.47(S.D.=.13) 

A4 

2.97 (S.D.=.36) 


Note : * Low Task Motivation 

Al = Test-Performance-Re-Test 
A2 = Test Re-Test 
A3 = Performance-Re-Test 


A4 


Re-Test 



















. 1 













. . . ' 




.. \ . 






















146 


APPENDIX 5.7 

ANOVA SUMMARY TABLE FREE-TIME TASK 
PERSISTENCE DATA: 2 (GROUPS) x 2 (LEVELS) 
Source ' df MS F 


Order (A) 

Motivation Level (B) 

Order x Motivation 
Level (A x B) 


1 52600.244 1.538 
1 787849.219 23.029 
1 35617.139 1.041 


P 

P 

P 


P 


<.05 

<.001 

<.25 


Order-Motivation 
Within Groups 


28 


34209.692 





























- 





























147 


APPENDIX 5.8 

MEANS AND STANDARD DEVIATIONS ON TASK PERSISTENCE 
OF TEST-PERFORMANCE-TEST AND PERFORMANCE ONLY 

DATA IN SECONDS 


Group 

High Task 

Interest 

Low Task 

Interest 


Mean 

S.D. 

Mean 

S.D. 

Test- 

Performanee-Test 

534.35 

164.10 

153.81 

124.00 

Performance-Test 

386.54 

196.96 

139.44 

182.74 










































































APPENDIX 6 


SUMMARY ANCOVA TABLES OF ATTITUDE MEASURES 
OF INTRINSIC MOTIVATION 


148 






APPENDIX 6.1 


SUMMARY ANCOVA TABLE FOR SEMANTIC DIFFERENTIAL 
AFFECTIVE DATA: 2lGROUPS) x 5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

0.593 

0.642 

P 

>.25 

Reward Condition (B) 

4 

2.367 

2.562 

P 

<. 05 

Motivation x Reward 
(AxB) 

4 

1.277 

1.382 

P 

>. 05 

Motivation-Reward 
Within Groups 

130 

0.924 























. 
































APPENDIX 6.2 


SUMMARY ANCOVA TABLE FOR SEMANTIC DIFFERENTIAL 
BEHAVIORAL DATA: 2(GROUPS) x 5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

3.489 

5.832 

P 

>. 05 

Reward Condition (B) 

4 

1.003 

1.677 

P 

<.05 

Motivation x Reward 
(AxB) 

4 

0.884 

1.4 78 

P 

>.05 

Motivation-Reward 
Within Groups 

130 

0.598 





































































APPENDIX 6.3 


SUMMARY ANCOVA TABLE FOR SEMANTIC DIFFERENTIAL 
COGNITIVE DATA: 2 (GROUPS) x 5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

3.489 

5.833 

P 

>.05 

Reward Condition (B) 

4 

1.003 

1.677 

P 

in 

o 

• 

V 

Motivation x Reward 
(AxB) 

4 

0.884 

1.478 

P 

>. 05 

Motivation-Reward 
Within Groups 

130 

0.598 


























































APPENDIX 6.4 


SUMMARY ANCOVA TA3LE FOR SUMMATED-RATING SCALE 
DATA: 2 (GROUPS) x 5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

0.051 

0.029 

P 

>. 75 

Reward Condition (B) 

4 

4.829 

2.8 12 

P 

<.05 

Motivation x Reward 
(AxB) 

4 

0.843 

0.491 

P 

o 

in 

• 

A 

Motivation-Reward 
Uithin Groups 

130 

1.717 








. 




























































APPENDIX 6.5 


SUMMARY ANCOVA TABLE FOR POST-RATING 
SCALE DATA : 2 ^GROUPS) x 5 (LEVELS) 


Source 

df 

MS 

F 


P 

Motivation Level (A) 

1 

1.116 

0.489 

P 

>.25 

Reward Condition (B) 

4 

3.399 

1.490 

P 

<. 05 

Motivation x Reward 
(AxB) 

4 

4.694 

2.059 

P 

>.05 

Motivation-Reward 
Within Groups 

130 

2.280 


























































APPENDIX 6.6 


SUMMARY ANCOVA TABLE FOR POST-EFFORT 
RATING SCALE DATA: 2 (GROUPS) x 5 (LEVELS) 


Source df 

Motivation Level (A) 1 

Reward Condition (B) 4 

Motivation x Reward 

(AxB) 4 

Motivation-Reward 

Within Groups 130 


MS 

F 


P 

0.178 

0.061 

P 

>.75 

2. 002 

0.682 

P 

<.50 

0.749 

0.255 

P 

>.75 

2.937 























. 





















































APPENDIX 6.7 

CORRELATION MATRIX OF INTRINSIC INTEREST MEASURES (N = l44) 


155 


O 




★ 

* 


* 

* 



* 

* 




* 

* 


* 

* 


* 

* 

* 


Ph 

vO 

CM 

VA 

'A 

O 

VA 

-rt 


Ov 

O 

CO 


-3- 

A- 

Ov 

A' 

VO 

CM 

rH 

rH 


CA 

CM 


O 

CM 

CM 

O 

-3" 

CM 

rH 

CM 


-3- 

rH 


8 

O 

O 

P-< 

Eh 

<5 

03 

Ph 

Eh 

§ 

.03 


Ph 


★ 


* 



o 

vO 

vO 


* 

* 


Q 

-3" 

A- 

rH 

VA 

Ov 

Ov 

03 

o 

»-H 

rH 

VA 

cc 

vO 




O 

rv 

rH 

O 

S 

VA 

♦ 

★ 

* 


* 


m 

O 

VO 

f, 


* 


ft. 

O 

-3" 

CM 

CM 

CA 


w 

1 

CM 

rH 

CM 

A 





A 

O 

CM 



O 

vO 


O 

o 


C\] 

ca 


vO 

Ov 

O 


(A 

CM 


O 

CM 

P 

I 


vO 


00 


* 

CO 

ov 


* 

A- 


* 

VO - 


CM 

a- 

o 


VO 

VA 


* 

* 

A- 

O' 

0 s? 


ca 

-b- 


va 

A- 

O 


* 

* 

00 

7$ 


ca 

CA 


* 

* 

o 

o 

va 


* 

* 

o 

VO 


* 

* 

CA 

VA 

CM 


* 

VA 

O 

CM 


VA 

VA 


* 

cc 

Ov 


* 

* 


* 

* 

VA 

Ov 

VA 


VA 

'A 


O 

t —1 

rH 


CA 


* 

A- 

co 


A- 

VA 


* 

* 

Ss 

-3* 



* 

* 

★ 

* 



* 

* 

VA 

O 

* 

* 

* 



* 

* 

O 

rH 

O 

5 

0v 

A 

Ov 

vO 

VA 

CM 


Ov 

A 

rH 

Ov 

CM 

Ov 



CM 

CM 

VA 

rH 

O 

CA 

VA 


* 

* 

co 

CA 

sj- 


Ov 

CA 


Ov 

vO 


O 


2a 

§ 


a- 

00 


* 

* 

A- 

A- 


* 

* 

A- 

N 

CA 


CO 

o 


P4 



* 









O 

rH 


CA 

vO 








Q 

A 


Ov 

rH 








W 

o 


P 

O 








CL, 



$ 









g 

00 

Ov 


& 









*03 

o 


CA 









PH 

* 











< 

Q 

fc 











03 

CM 











P-« 












Eh 












CO 












P 












W 



/»—S 









<D 



p 

P 



/-S 


/—N 


-^N 

fH 



0 

1/3 


/—s 

P 


P 


P 




0 

0 


p 

0 


w 


w 

CO 


H 


P 

w 

P 0 

P 

0 

p 

0 

p 



l 

cc) 

0 

cc) Eh 

CC) 

Eh 

cc) 

Eh 

CO 


P 

0 

p 0 

p 

Eh 

P 1 

p 

1 

p 

1 

<D 


q 

in 

p h 

p 

1 

P P 

p 

P 

p 

P 

H 

0 

0 

(C 

q Ph 

q 

0 

C W 

c 

CO 

q 

0 

<D 

o 

q 

q 


0 « 

0 

q 

0 O 

0 

o 

P 

O 

P 

0 

• 

H P 

H 

PH 

H Ph 

Eh 

Ph 

Eh 

Ph 

c 

c3 

C+-, 

0 

0 CC) 

0 

•> 

0 - 

0 

* 

0 

« 

M 

p 


> 

«H P 

p 

0 

P 0 

P 

P 

P 

0 


(0 


P 

P Eh 

P 

> 

P > 

P 


P 

> 

, v; 

•rH 

Q 

P 

p o 

p 

P 

P P 

P 

H 

P 

P 

W 

CO 


O 

Q p 

Q 

P 

n p 

Q 

O 

Q 

P 

nj 

q 

O 

0 

> 


P 

o 


P 


P 

Eh 

<D 

P 


O d 
p i: 

O 

q 

O 0 

o 

> 

O 

q 


PH 

P 


P 

bD 

p p 

p 


P 

ho 

o 

c 

p 0 

P 

O 


p 

j3 

P 

o 

P 


cc) 


q pq 

q 

O 

q 

0 

q 

o 

CO 

CO 

£ 


eg ^ 


N-' 


to 

pp 

CD 

v— 

c 

cc) 

0 


0 

0 


£; —^ 

0 

6 


e 


P 

£-* 

03 



0 


0 





03 

03 


03 

03 


03 


P 












C 

• 

• 










M 

rH 

CM 


CA 

-3* 


VA 

VO* 


A- 



0 

0 

0^ 


q 



Tj 

i—i 

P 

P 


P r*^ 

Ja 


0 
. \ 

cc) 

cc) 

CC) rH 

p 

P P 

p 


•+-’ 

o 

o 

o q 

q 

!SS 

q q 


P 

ei 

03 

03 

03 O 

o 

o o 


C 

A 






p 


u 

hD 

bD 

fckD bD 

bD 

bD 

P bD 


■ ^ 

q 

q ^ 

q q 

q 

q 

w q 


H 

p ' n 

p p 

p p 

p 

0 P 

0 P 


kH 

p p 

P w 

p p 

0 p 

H P 

q i 

0 p 


P 

0 

q w 

CC 0 

Eh 

T) » 

1 

X) P 

i 

T) P 

P q) 

S'? 

03 P 

w (S 

d i 

0 P 

VA 

O 

P, 

P 

O P 
• d 

0 0 

0 W 

0 W 

W 

p co 

0 w 


p q 

P o 

q o 

P o 

Ph O 

P o 

V 

V 

d Ph 

CC) Ph 

P Ph 

Eh Ph 

• Ph 

q Ph 

P- 

v tH 

PH O 

J £ 

1 

| W 

P ^ 

e 

o 

O ^ 

P 

<H 

t i w 

o 

P 

o 

03 

03 

o 

tl 


> 

.. 








0 



. 

. 

# 

# 

# 

P 


00 

Ov 

o 

P 

CM 

CA 

O 




p 

P 

P 

P 

2T 







































































































































































































































































156 






rH 

VO 

Ov 

rH 






it 

€0 

a- 

a~ 






rH 

CV 


VO 

CO 




<1) 









> 









P 

O 

CM 

rH 

rH 1 






CO 

if 

CM 

ol 

rH 




1-1 f?n 

o 

rH 

O 

col 

O 




viJ 

O 









a 







P 









II 









z 









g 


e 

o 

P P 

it 

CM 

VO 

O 


3 

rH 

O 


o 


M_g 

cm 

CM 

001 

o 

1 

CA 


o 


<D 







CO 


m 







6-* 











P 







H 


c 









0) 







H 


£ 







co 

szs 

p 







s 

o 

M 

cd 

CM 

■it 

ca 

-t 

CD 

00 

Cxi 

Eh 


VO 

A- 

ca 

o 

vO 


Eh 

< 

£ 

CO 

rH 

o 

rH 


VO 

S5 

O 








P 

PC 

P 









w 






X 

X 


o 






M 

Q 

sa 

1 

fin 






S 

Eh 

« 








o 

c 

-P 






On 

1—1 

> 

0 






c 

co 


CD 







‘Z 

< 

(H 







M 


<1) 







cn 


*P 







H 


G 







P 


M 

U-\ 

VO 

it 

vO 

o-| 




o 

_H- 

rH 

CM 

VO 




rH 

CO 

CM 

rH 

A-l 


o 


cd 









Eh 







CO 









w 


nd 







CO 


$ 







>H 


-p 

<2 
















<c 









Cd 

g 









o 




__ 

-P 

,-- 

-P 


< 




-p 

10 

P 

CO 


Cx, 




rH CO 

rH CD 

rH CO 

rH 0 






Cd CD 

rt Eh 

d <l) 

ct) EH 






P Eh 

P 1 


•H 1 






-P 1 

-p a> 

-P 1 

-P -P 






G CD 

c P 

c <d 

c co 





CD 

cd q 

CD 

0 H 

0 O 





O 

H it 

(H 

H it 

ft PC 





q 

CD 

0 . 

CD 

0 





to 

«H - 

«H rH 

«m - 

Cm - 





•p 

<P CD 

<n d 

Cm CD 

Cm 0 





CO 

P ^ 

P H 

P > 






•rH 


Q O 

Q -rH 

Q -H 




-p 

CO 

-P 


-P 

-P 




CO 

h 

O O 

o > 

O P 

o o 




0) 

<D 

P <D 

d p. 

P G 

*H 0 




Eh 

Ph 

-P Cm 

■P hD 

-P <M 






C Ch 
cd < 

§<8 

ScS 






CO 

B w- 

E ^ 

e ^ 






,cd 

CD 

CD 

CD 

0 





Eh 

CO 

CO 

CO 

CO 





rH 

CNfM 

(A 

-it 

VA 


o 

ON 

va 

A- 


CA 

O 



1 

1 


ca 

CO 


VO 


VO 

it 

VA 

VA 

§ 




ca 

n 

ca 

CM 

o 


Ov 

rH 

vO 

1 


co 

cn 

VA 

O 

vO 

O 

O 

CO 

A- 

00 


o 

W 

rH 

O 

O 

O 

rH 

rH 

CA 

V?S 










rH 

o 











p 


-=H 


CA 

(A 

VO 

Ov 


CM 

Ov 



O'. 

A- 

O 

VA 

CM 

A- 

VO 

it 

O 

VA 


col 

o 

O 

i — 1 

CM 

rH 

it 

rH 

O 

-t 







1 


1 

CM 

VA 











P 


o 

ca 

-it 

a- 

ON| 

i — 11 

M 

o| 

2^1 

O- 



rH 

Ov 

Ov 

-4 

vO 

Ov 



i — 1 

00 


1 

o 

O 


vO| 

uS| 

VA, 

'-Ol 

P 

CM 










CM 

vO* 











P 


CO 

CM 

Ol 

Ol 

C- 

(A 

CA 

CM 

CM 

"es. 


CO 

o 

00 

o 

Si 

11 


?M 

VA 

rH 

-if 

rH 

Ov 

A- 

00 

A- 









1 

• 

it 










CM 












P 











CM 













-p 

/•—N 










CO 

-p 





G 




Td 

rH 0 

rH CO 

0 

0 

0 


O 




0 


d a> 

n 



/• - s 


x - s 



P 

P i 

P Eh 

a3 

a) >> 


P >> 



0 

P 

P -P 

-P 1 

o 

O 

O rH 


0 P 

G rH 


O 

e 

C co 

G -P 

co 

CO 

CO C 

G 

0 G 

O G 


q 

O 

0 O 

0 0 



o 

O 

3° 

•H O 


TO 


ft PC 

u o 

to 

hD 

hD 




t 

p 

0 - 

0 P-1 

c 

c 

C hD 

hD 

hD 

W hD 


G 

Cm rH 

<H 

•H 

•rH 

•H C 

G 

0 G 

0 G 


cd 

p 

cm cd 

tp ® 

-P ■-'' 

-P -P 


0 P 


G P 


> 

o 

P H 


cO -P 
a? co 

0 

Cd 0 
Eh 

0 P 

rH P 

G P 

CJ 1 -P 



PL 

o o 

P 

O -H 


0 <d 
cd Cd 

h c$ 


p 

cd 

P 

o > 

O -rH 

tJ 9h 

Td 1 

TJ 1 

CO i 

0 i 

0 1 

0 

P 

cd 

Si 

d & 

0 1 

0 ^ 

a> -p 

P 

£ 

0 P 

O 

o 

E 

-P 0 

■P CO 

G 0 

P 0 

p 0 

c 

Eh 

P 

C 0 
ct5 PQ 

q o 

co Jh 

cd o 

•H O 

P O 

1 O 

c o 

d 


O 

co o 

e ft. 

e Pn 

P Ph 

O PL, 


3 P-, 

■H 

Cm 

fS 

e w 

e ^ 

P v_' 

E —' 

B ^ 

ch ^ 


H 

O 

0 

0 

3 

p 

O 

<H 

o 

o 

TO 



CO 

CO 

co 

CO 

O 

H 

s 

> 













01 











p 

vO 

cv 

co 

Ov 

O 


CM 

CA 



o 





rH 

rH 

rH 

i—1 



Sd 






























































APPENDIX 7 


POST HOC ANALYSIS OF MONEY CHOICE VERSUS 


NO-MONEY CHOICE GROUPS DATA 











158 


APPENDIX 7.1 

Another two-factor analysis of covariance was conducted 
on the task persistence data and the self-report measures of 
intrinsic task interest across two levels of motivation. 

This analysis was conducted to test the effects of subject’s 
taking money or not taking money (A). Appendix 7.2 presents 
the F values of main and interaction effects. Both 
motivation level main effects (B) and the money, no-money, 
and control main effects (A) were significant. The 
interaction effect was not significant. Scheffe a posteriori 
test was conducted to test differences between individual 
group means within both motivation levels (B) . A number of 
interesting significant results were revealed: (1) Both the 
money and no-money groups were significantly lower in task 
persistance than the high motivation controls, (2) within 
the low motivation level, however, only the money group mean 
was significantly lower than the control group mean, (3) the 
money group means were significantly lower in task 
persistance than the no-money group means in both low and 
high task motivation conditions, and finally, (4) there was 
a strong and significant Money x Motivation level 
interaction effect; F(df=2,84)=49.60, p <.001. Appendix 7.3 
and 7.4 provides the F values for all Scheffe tests. 

Appendic 7.5 provides the means and standard deviations for 
the groups. 




■ 


- 




159 


APPENDIX 7.2 

SUMMARY ANCOVA TABLE FOR TASK PERSISTENCE 
DATA OF MONEY CHOICE VS-NO-MONEY 
CHOICE GROUPS: 2 (GROUPS) x 2 (LEVELS) 


Source 

df 

MS 

F 


P 

Choice Level (A) 

1 

252381.875 

7.004 

P 

>.05 

Motivation Level (B) 

2 

336098.005 

9.327 

P 

<.001 

Choice x Motivation 
(AxB y 

2 

12567.746 

0. 349 

P 

>.50 

Choice-Motivation 
Within Groups 

62 

360.016 




















































160 


APPENDIX 7.3 

SUMMARY TABLE OF SCHEFFE A POSTERIORI 
COMPARISONS OF INDIVIDUAL TASK PERSISTENCE MEANS 


Comparisons 

High Task Motivation 

F 

Low Task Motivation 

F 

B1B2 

13.63*(df=2,42) 

7 .80* idf = 2,36) 

B1B3 

5. 19*idf=2,69) 

0.13 idf=2,60) 

B2B3 

42.28*idf=2,69) 

14.38* (df=2,60) 


Note: *p <.05 

B1 = No-$ Choice 
B2 = $ Choice 
B3 = Control 












































— 




















161 


APPENDIX 7.4 

SUMMARY TABLE OF SCHEFFE A 

POSTERIORI COMPARISONS OF INDIVIDUAL TASK PERSISTENCE MEANS 
, ' ACROSS MOTIVATION LEVEL. 


Comparison 

No-Money Choice 

S-Choice 

A1 A2 

7.13*(df= 1,44) 

2. 92 (df = 1*42) 

Note: *p <.05 

A1 ='High Task 

Motivation Groups 


A2 = Low Task 

Motivation Groups 



































































162 


MEANS AND STANDARD DEVIATIONS OF REWARD 
CHOICE AND CONTROL GROUPS FOR TASK PERSISTENCE IN SECONDS 


Reward 


High Task 

Motivation 

Low Task 

Motivation 

Mean 

S. D 

Mean 

S.D 

Choice 

($) 

178.59 

137.36 

86. u 7 

121.25 

Choice 

(N o- $) 

421.41 

239.91 

277.56 

244.59 

Control 


543.70 

103.51 

299.38 

242.83