1972 deci jpsp

Upload: viorel

Post on 03-Jun-2018

225 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/12/2019 1972 Deci Jpsp

    1/8

    Journal oj Personality and Social Psychology1972, Vol. 22, No. 1, 113-120

    INTRINSIC MOTIVATION, EXTRINSIC REINFORCEMENT,AND INEQUITYE DW ARD L.DECI1Management Research Center, University oj Rochester

    If a person who is intrinsically motivated to perform an activity begins toreceive external reinf orcem ent for the activity, wha t w ill happ en to h is in-trinsic motivation? Previous studies and the present study indicate that moneydecreases intrinsic motivation, while verbal reinforcements tend to enhanceintrinsic m otivatio n. The b egin ning of a cognitive evaluation the ory is dis-cussed, and an apparently discrepant prediction between this theory and in-equity theory is pointed out . It is argued, however, that th e theories are notconceptually discrepant, and the present study gives support fo r this argu-ment .It is possible to distinguish between tw obroadclasses ofmotivation toperform an ac-tivity: intrinsic motivation andextrinsic moti-vation. A person is intrinsically motivated ifhe pe r fo r ms an activity for no apparent re -ward except the activity itself (c f . Berlyne,1966; H unt , 19 65; W hi te , 1959) . Extrinsicmotivation, on the other hand, refers to theperformance of an activity because it leads to

    external rewards (e.g., status, approval, orpassing grades) . The question of interest inthis study is whether there will be changes ina person's intrinsic motivation for an activitywhen he receives external rewards for per-forming thatactivity.Deci (1971) reported that external rein-forcements do affect intr insic motivation, andhe suggested th e initial elements of a cognitiveevaluation theory to account for the changesin intrinsic motivation following an experiencewith extrinsic rewards. The theory focusesona person's cognitive evaluation of an activityand th e reasons for his engaging in the activ-ity. It suggests that distinctions should bemade among different kinds of external re -wards, since aperson's evaluation of differentrewards may be different. In turn, this wouldThe author would l ike to thank Wayne Casciofo r serving as the first experimenter and for helping

    with the data analysis; Victor Vroom for makinghelpful comments on an earlier draf t of the manu-script; and Larry Coff for being th e second experi-menter .Requests for reprints should be sent to Edward L.Deci, Management Research Center, University ofRochester, Rochester, Ne w York 14627.

    lead to different effects on the person's in -trinsic motivation.For example, Deci (1971) stated thatmoney i s f requent ly used as a means of buyingservices which would probably not otherwise berendered. Perhaps, then, the presence of money asan external rewa rd suggests to the subjects thatthey should pro bab ly not rend er this activity with-out pay, tha t is, they should not be so intrinsicallymotivated to do the activity.This could lead the subjects to a process of cogni-tive reevaluation of the activity from one which isintrinsically motivated to one which is motivated byth e anticipation of m o n e y [p . 107].

    O n the other hand, when verbal reinforce-ments are given as external rewards, theserewards may not be phenomenologically dis-tinguishable f rom th e feelings of satisfactionthat th e subjects get from th e activity, so thetotal positive value properties (Koch, 1956)associated wi th th e activity have increased.Therefore, th e tendency to per form th e activ-ity in the absence of external rewards will bestrengthened; that is, the person's intrinsicmotivation will be increased.The important difference in the effects ofmoney and verbal reinforcements lies in aperson's perception of the locus of causalityof his behavior (deC harm s, 1968; H eider,1958) . When he receives money, he could eas-ily come to accept th e money as the reasonfor his behavior, but he is less likely to dothis if the rewards are social approval. Hence,th e theory posits that a person can come toperceive that his behavior is controlled byexternal rewards and that this will lead to a

    113

  • 8/12/2019 1972 Deci Jpsp

    2/8

    114 E D W A R D L. DECIdecrease in intr insic motivation. However,when he receives interpersonal rewards, hewill not perceive them as controls of his be-havior, so they will strengthen his intr insicmotivat ion because they s trengthen his senseof competence and se l f -de termina t ion .Inequity Theory

    Ad ams (1963, 1965) has presented atheory of inequi ty wh ich migh t at first seemto be inconsistent with the predictions of thecognitive evaluation theory. Adams proposedthat wh ena person is in an exchange relation-ship with another person, he will be con-cerned about th e outcomesa nd inputs of each.Outcomes are all of the compensat ions a per-so n gets, such as money, independence, per-sonal satisfaction, etc. Inputs are all of theth ingshe gives to the situation, such as train-ing, intellect, effort, etc. Adam s argues th atth e person will evaluate his own ratio of ou t -comes to inputs and compare it to the otherperson's ratio. If the two ratios are unequal ,th e person will feel inequi tyand will be moti-vated to reduce th is inequi ty . Inequi ty wil lexist for a person both when he is overcom-pensated ( i .e. , when his ratio of outcomes toinputs is greater than th e other 's ) and whenhe is undercompensated ( i .e. , his ratio is lessthan th e other's). This discussion is con-cerned only with overcompensat ion.When a person is overcompensated, he maytry to lower his own outcomes to restoreequity (e.g. , Wood & Lawler, 19 70 ) . How-ever, if a person's outcomes were fixed, an-other way he could restore equi ty would beto increase his own inputs, thereby also de-creasing his ratio of outcomes to inputs (e.g. ,Ad ams &Rosenbaum, 1 9 6 2 ) . A large numberof studies supportingthese assertions were re-viewed by Lawler (1968) , Pritchard ( 1 9 6 9 ) ,and Goodman and Fr iedman ( 1 9 7 1 ) .Ad ams (196S) mentioned that the personm ay also use an internal standard as his com-parison other. Pritchard (1 9 6 9 ) elaboratedthis point by defining internal standard as the amount of outcome Person perceives asbeing commensurate with his own inputs,without regard to any comparison person [p.205]. H e suggested that feelings ofinequi tyarise first and foremost f rom the correspon-

    dence between Person 's own inputs and out-comes. If his inputsaregreater than [his] ou t-comes, he will experience inequity, which willlead to feelings of dissatisfaction [p . 2 0 6 ] .Similarly, if his outcomes are substantiallygreater than h is inputs , he would exper iencei nequ i ty .Now, imagine a subject who i s intr insicallymotivated to per form an activity and hasreceived substant ial overpayment . H e migh tbe expected to exper ience ineq ui ty because ofh is internal s tandard. If a f t e r th e specifiedperiod ended he were allowed to continue towork for a short period for no pay, equitytheory would predict that he would tend tocont inue work ing, since at the end of thepayment per iod he had more outcomes thaninputs and making addi t ional inputs for nopay would help to restore equi ty .The cogni t ive evaluat ion theory predict ion,however, is that he would be less likely tocontinue to wo rk, because the exper ience withmoney made h im dependent on the moneyand decreased his intr insic motivation to per-form th e task. It wil l now be proposed thatthese tw o theories are not conceptually dis-crepant.

    In the cognitive evaluation theory, it issuggested that wh en a person per forms anintr insically motivated task for money , hisperception of the reason for p e r f o r mi n g th etask shif ts f rom it is intrinsically motivatedto it is motivated by the mo ney. Sincetheperson is then pe r forming for money (andtherefore has less intr insic motivat ion) , th eprinciple of inequity could certainly be rele-van t . In o ther words , in the example above,he would con t inueto p e r f o r m a f t e r th e moneystopped, if he we re feeling ineq uitably over-payed ; however , he would not per form if hefelt equi tably paid. In either case his in-tr insic motivation would have decreased, andhe would have lessintrinsicreason top e r f o r m,so th e high per formance predicted in the firstcase would occur because of inequi ty in hisow n mind, not because intr insic motivationincreased. This addi tional per form ance wouldlast only until equity w as restored.

    In a previous studyby Deci ( 1 9 7 1 ) , severalsubjects were paid as much as $4 forabout 20minutes of work on an interesting puzzle. I t

  • 8/12/2019 1972 Deci Jpsp

    3/8

    MOTIVATION, REINFORCEMENT, A N D INEQUITY 115was only at the time that the subjects wereactually handed the money that they ap-peared to feel uneasy and inequitablyoverpaid. During the session when the ex-perim enter said things like, You have nowearned $3, or You have now earne d $4,th e subjects said nothing and did not appearto feel discomfort . However, when they wereactually handed the money, they became veryuneasy, and many of them said things l ike, That'sa lot ofmoneyforw h a t I did, or Ifeel funny taking a l lthis money.Now, imagine again a laborato ry experi-ment such as the example givenabove of theoverpaid subject. If the subject were actuallygiven the money before he had the free timeto continue working (i .e. , the period with nop a y ) , it seems likely that he would experi-ence inequity (since he has the money as anin i t ia t ing s t imulus) , so he would be morelikely to make addi t ional inputs to restoreequi ty . However, if he knew he was going tobe paid,but paymen t was withheld unt i l theend of the free time, he would not experiencethe same inequity (sincehe doesnot yet haveth e init iating stimulus, money) , so he wouldbelesslikelyto domore work .He is no longeras intrinsically motivated, and heapparentlydoes not experience inequity f rom over-compensation, so he would be less likely tocont inue to makeinputs.Thus, the t iming ofthe payment may determine whether thechange in per fo rmance is governed by theprocesses of inequity or cognitive reevalua-tion.Hypothesis I: W h e n a person is rewardedwi th money for per fo rming an intrinsicallymotivated activity, his intrinsic motivationwill decrease. Hypothesis II: W h e n a personis rewarded with verbal reinforcements forper fo rming an intrinsically motivated activ-ity, his intrinsic motivation will increase.Hypothesis HI: However, whena personwhois p e r f o rm i n gan intrinsically motivated ac-tivity feels inequitably overpayed, he willincrease his performance (i.e., make-addi-t ional inputs) to restore equity.

    Two other questions are investigated in thepresent experiment . First, since money andverbal re inforcement appear to have oppositeeffects on intrinsic motivation (Deci, 1 9 7 1 ) ,

    a cell containing both rewards will allow fornew empirical evidence.There is , however ,n otheoreticalbasisfor m aking a predict ion aboutthe scores in this cell.

    Second, one might ask whether the effectsof external rewards on intrinsic motivationare the same for males and females. Hence,half o f the subje cts assigned to each cell w eremales,and half were femalessothat thisques-tion could be investigated.

    METHODThe subjects for this experiment were 96 under -graduates at the University of Rochester who were

    randomly assigned to one of six conditions, with 8males and 8 females in each of theconditions.Each subject participated for a 1-hoursession d u r -in g which he spent most of his time working on apuzzle called Soma. The puzzle is composed ofseven different pieces, each of which is m ad e to looklike it consists of three or four 1-inch cubes. Thepieces can be fitted together to form millions ofconfigurationsonly four of which were used for theexperiment.Dur ing a session, th e subject w as asked to repro-duce th e four configurations tha t had been d r aw n onpaper for h i m .The time to complete each conf igura-tion w as measured with a s topwatch, and if a sub-ject was unable to do a configuration within 10min-utes, he was s toppedand told how to do it . This le thim k n o w t ha t all the configurations were possible.When a subject repor ted to the waiting room, hewas met by the f irst experimenter w ho took himto the exper imentalroom where he was seated at atable. The experimenter then lef t th rough a seconddoor at the back of the exper imentalroom so tha the would be outside th e room observing throug h aone-way window. The subject knew, of course, thatthe experimenter w as observing h i m , and hec o m m u -nicated with th e experimenter through an intercom.On th e table in front of the subject were th e sevenpuzzle pieces-each with a n u m b e r on it so tha t th eexperimenter could refer to it over th e intercom. Toth e left of the subject was a stack of the f ou rconfigurations t ha t he was asked to reproduce. Tohis right were three other configurations. The topon e of the three was a sample ; th e other two arediscussed below. On another table to the subject 'sright were th e microphone, th e speak er, recent issuesof three magazines New Yorker, Time, and Play-boy , and an ashtray.When the experimenter got to his position behindthe one-way window, he read the instructions to thesubject. The subject w as told that it was an experi-ment to study certain problem-solving concepts,andt ha t he would be asked to solve f ou r puzzle prob-lems. Af ter th e instructions were read, th e experi-menter asked th e subject to look at the sample tohis right. He was told how i t could be solved and

  • 8/12/2019 1972 Deci Jpsp

    4/8

    116 E D W A R D L. DECIw as allowed 1 m i n u t e to man ipu la te th e pieces andreproduce i t . The subject then worked on the f o u rpuzzles in tu rn .During the puzzle solving, some of the subjectswere rewarded, and others were not . Rewards con-s is ted both of verbal re inf orcem ents and m oney . Theindependent manipula t ion cons is ted of reward ing (o rnot rewarding) each puzzle that was solved wi th inth e 10 minu tes allotted to it. The man ipu la t ion isexplained in detail below.T o obtain th e dependent measure o f mot iva t ion ,th e exper imenter lef t h is position for a period of 8minutes following the puzzle solving. The pretextwas as fol lows: When a s ub jec t h ad completed th ef o u r puzzles, th e experimenter told h im that he hadcompleted all of the problem solving that he had todo, but there was one more th ing tha t he would beasked to do, and that was to complete a shor t ques-tionnaire. Since it was an exper iment in problemsolving, the subject would be asked a few questionsa b o u t the way he had solved th e puzzles . However ,there were f o u r different sets of questions, only one ofwhich was the most appropriate for this subject andth a t wou ld be determined by how he had done onthe puzzles. To select the appropriate set of ques-tions, data from the session would be fed into aco m p u te r th rough a teletype. T o do this , th e experi-m e n t e r wou ld h ave to leave for a shor t time , 5-10m in u te s . T he subject w as told that he could do any-th ing that he l iked dur ing that time, but he was askedto s tay in the room. The exper imenter lef t his posi-t ion and entered the exper imenta l room through theback door and exited through th e f ron t door . H e th enclimbed (noisily, so the subject could hear ) a smallset of steps outside th e room and left the lab areath rough a door (noisily opened and closed) at thetop of the stairs.

    Just after the f irs t experimenter le f t the experi-mental room, he signaled to a second experimenterw ho w e n t to the outside of the one-way windowth rough a different door which the subject d id notknow existed.Hence, th e subject w as alone in the room and wasu n a w a r e that he was being observed, so he was freeto work on the puzzles, read magazines , or do any-th ing he liked. Interest was in the amoun t o f t imeout of the 8 minutes that he spent working on thepuzzles. It was reasoned that if he worked on thepuzzles dur ing this free choice time whe n he cou lddo other things, then he m u s t be intrinsically moti-vated to do the activity. T he a m o u n t of time outof the 8 minutes that the subject spent working onthe puzzle wa s determined by the second experi-men te r w ho observed h im th rough th e one-waywindow and used a s topwatch to record th e t ime . T hesecond experimenter w as blind to conditions andalso to the hypotheses of the exper imen t . There isno indication that th e subjects suspected that theywere being observed during this free choice period.Since any subject w h o w a s unable to do a con-f igura t ion within the 10 minu tes allowed w as shownth e solution, th e possibility that th e Zeigarnik (1927)effect would inf luence whether or not he worked on

    th e puzzle in the 8 - minu te free choice period w asminimized.The conf igura t ions that were in the pile to hisr igh t under th e sample were impossible ones to do.This precluded th e possibility that a subject wouldfinish one of these conf igura tions and there fore be acausal f a c t o r in determining whether or not he spentmore t ime working on the puzzles .Then af te r 8 minutes , the f irst experimenter re-turned to the room and asked the subjects to com-plete the questionnaire.T he exper imenta l des ign was a randomized th ree-factor design using after-only scores. The first varia-ble was verbal re in forcem ent wi th two levels , re in-f o rc e me n t or no re i n fo rc e me n t . For subjects in theverbal re inforc em ent condi tions , they were rew ardedwith sta tements f rom the exp erim ente r (e.g., Good,that's th e quickest that one has been done ) eachtime they solved a puzzle . T h e same f o u r state-ments (or a subset of them if the subject did notsolve all f o u r puzzles) were used for each subject .T he second variable w as money. Subjects eitherreceived no mo ney, received mo ney a t the end ofth e puzzle solving bu t be fore th e 8 - minu te freechoice per iod, or received money a f t e r th e freechoice period. A subject was given $1 in cash foreach puzzle that he solved. The third variable w asse x of the subject. The three variables crossed,th e reby mak ing 12 condi t ions , 6 for each sex.The t wo mo ney t rea tments (actual paym ent b e f o reth e 8-minute free choice period and actual paymentaf t e r that period) were exactly the same except forthe point a t which th e cash actually changed hands.Both groups were to ld a t the beginning of the ex-per iment that they would get $1 per puzzle solved,an d both were told ho w much money th ey h adearned a f t e r each puzzle they solved. But in theactual-payment-before groups , they were handed thecash by the first experimenter as hewalked throughthe room to leave for the 8-minute period, whereasin the actual-payment-af ter groups , the subjects weregiven th e cash a f t e r th e 8-minute free choice period.A s ment ioned in the in t roduct ion, i t was felt thatif a subject actually received h is m o n e y be fore th e8-minute free choice period he would experienceinequi table overcompensat ion, so he wou ld worklonger than nonrewarded subjects . However , if hedid not receive money unt i l a f t e r th e 8-minute freechoice per iod, th e money would initiate a process ofcognitive reevaluation, causing a decrease in intrinsicmotivat ion .

    To summarize , each subject solved puzzles in oneof six conditions (a) no t rew arded , (6) rewa rdedwi th money b e f o re th e free choice period, c) re -w a r d e d wi th money af te r the free choice period, o r d ) , e ) , a n d ( / ) verbally rewarded in combinat ionwith one of the f irs t three. The puzzles were thoughtto be intrinsically interesting (pilot testing substan ti-a ted th is ) , so some subjects were pe r form ing bothfo r external rewards and intrinsic motives. Subjectswere then given 8 minutes of free choice time duringwhich observat ions were made on the am oun t oftime they spent working on the puzzles.

  • 8/12/2019 1972 Deci Jpsp

    5/8

    MOTIVATION, R E I N F O R C E M E N T , A N D I N E Q U I T Y 117R E S U L T S

    I t was predicted: a) tha t subjects w howere rewarded with money would be less in-trinsically motivated (i.e., th e ones rewardedat the end would spend less free choice timeworking on the puzzle than th e nonmoneysubjects), b) that subjects verbally rein-forced would be more intrinsically motivatedto perform th e activity than those w how e r enot verbally reinforced (i.e., they would spendmore free choice time workingon thepuzzles) ,and c) that subjects w ho actually receivedmoney before th e free choice period wouldfeel inequitably overpaid (i.e., $3 or $4 is toomuch to get for about 30 minutes of workingon interesting puzzles), so although theywould be less intrinsically motivated thansubjects who got no money, they would workon the puzzles for more of the free choicetime as a way of expending additional efforton th e task for which they were overpaid. d) There was also interest in the experi-mental group that received both verbal andmonetary rewards, since the two would be,according to the theory, pulling in oppositedirections on the person's intrinsic motiva-tion. No prediction w as made about this con-dition, e) Finally, since there was an equalnumber of males and females in each cell,comparisons were made between males andfemales.

    Table 1 presen ts the cell means for thea moun t of t ime tha t subjects in the 12 condi-tions spent working on the puzzles during th e8-minute free choice period. The first vari-able w as money and t iming, w ith threelevels. The prediction w as tha t th e groupthat actually received cash after th e freechoice period would spend the least a mou n t o ftime on the puzzles, th e subjects who got nomoney would spend more time, and the sub-jects who were given money before th e freechoice period would spend th e most amounto f time working on the puzzles. The secondvariable w as verbal re inforcement, and thepredictionw asthatsubjects w howere verballyreinforced would spend more time than non-reinforced subjects.Of the 12 cells, all but 2 were orderedproperly to support th e predict ions perfectly.For males, th e money-after-verbal-reinforce-

    TABLE 1MEAN N U M B E R or S E C O N D S SPENT BY SUBJECTSO N P U Z Z L E S IN THE S-MINUTE FREECHOICE PERIOD

    C ond i t i on

    Money a f t e rN o moneyMoney before

    No v erb a lr e in fo r c e me n tFemales

    151.6292.4346.0

    M ales65.6124.4248.0

    Verb a lr e in fo r c e me n tFemales

    240.4142.5384.4

    Males219.9197.8392.9

    ment cell should have been less than th e no-money-verbal-reinforcement cell. How-ever, this reversal is very small. Also, for fe-males, th e no-money-verbal-reinforcementcell should have been greater than th e no-money-no-verbal-reinforcement ' ' cell.T he three hypotheses were tested using a3 X 2 X 2 (Money a n d Timing XVerbal R e -inforcements X Sex) analysis of variance onthe amount of t ime that subjects spent work-ing on the puzzles during the 8-minute freechoice period. These results are presented inTable 2.Variable A with three levels (money af ter ,no money, and money before) is highly sig-nificant p

  • 8/12/2019 1972 Deci Jpsp

    6/8

    118 E D W A K D L. DECITABLE3

    MEAN NUMBER OF S E C O N D S SPENT W O R K I N G ON THEPUZZLES D U R I N G 8-MiNUTE FREE CHOICE PERIODS(SCORES ARE COLLAPSED A C R O S S THE THREECONDITIONS ON THE MONEY V A R I A B L E )

    TABLE 5MEAN NUMBER O P S E C O N D S SPENT W O R K I N G O NTHE P U Z Z L E S D U R I N G THE S-MINUTEFREE CHOICE PERIODS

    G ro u pFemalesMalesDifference

    No ve rba lre inforce-m e n t263.33146.00117.33*

    Verba lre inforce-m e n t255.75269.41-13.67

    Diffe rence-7.58123.41

    ftThis d i f fe rence istested us inga n ana lys is of var iance, seeTable 4 .*p

  • 8/12/2019 1972 Deci Jpsp

    7/8

    MOTIVATION, REINFORCEMENT, AND INEQUITY 119rewards would not be phenomenologically dis-tinguishable f rom the internal satisfactionhence, there would be an increase in intrinsicmotivation. In other words, it was suggestedthat the feedbackaspect of the reward wouldbe more p rominen t than the controll ing as-pect . However, when one looks closely atthe no-money-verbal-reinforcement-femalescell, which was not in accord with the pre-dictions, it appears that there may also be astrong controlling aspect to verbal re inforce-ments under some circumstances. Let uslook carefully at the rela t ion between th eexperimenter and the subject in o rder totry to unders tand the sex differences in thisexperiment .T he experimenter in this experiment w asa veryattractiveand personable male gradu-a te s tudent . H is in teract ion wi th the femalesat the beginning of the experimen t and h iscommunicat ion w i t h t h e m d u r i n g th e experi -men t may have served as a reinforcementsimilar to the ve rbal re inforc em ents . Theirl iking of the si tuation, due partly to theirin teract ion with the experimenter , may havebeen a reward which, l ike the verbal rein-forcement , was cognitively indistinguishablef rom the in ternal feelings of satisfaction thatthey got from engaging in an intrinsicallyintere sting activity. Hence, th e fem ales w horeceived no specific verbal re inf orce m ent s ti llexperienced positive interpersonal reinforce-ment which served to increase their intrinsicmotivat ion as much as the other femalesw ho received verbal re infor cem ent and themales w ho received verbal re inforcement (seeTable 3).The males, on the other hand, showed asignificant difference between th e verbal-rein-forcement and no-verbal-reinforcement condi-t ions. This would be expected, since maleswould probably not f ind the mere in teract ionwith a male experimenter as rewarding.Hence, the specif ic verbal re inforcem entswould be expected to make a difference intheir intrinsic motivation, which in fact it did(/K.05).If this in terpreta t ion were correct , it wouldgive support to the cognitive evaluation idea,since it would be a case of positively affectiveinterpersonal re in forcement (of which verbal

    reinforcement is one example) s t rengtheningin t r insic motivat ion .Now, consider the unexpected decrease inin t r insic motivation in the no-money-verbal-r e in fo rcemen t cell for females . Al though th edifference between th e no-money-no-verbal-reinforcement and the no-money-verbal-reinforcement cells (149.9 seconds) is notsignificant p

  • 8/12/2019 1972 Deci Jpsp

    8/8

    120 E D W A R D L. DEC:dents , etc., should not concentrate on external-control systems such as monetary rewards,which are linked directly to pe r formance , bu t ,rather , he should concentrate on st ructur ingsi tuations that are intrinsically interest inga n dthen be interpersonally supportive and re-warding toward th e persons in the si tuat ion.Whi le large payments can lead to increasedperformance due to feelings ofinequity,thesepayments wil l , however, be making th e peopledependent on the money, thereby decreasingtheir intrinsic motivation.

    R E F E R E N C E SA D A M S , J . S. Toward an unde rs t and ing of inequi ty.Journal of Abnormal and Social Psychology, 1963,67, 422-436.A D A M S , J. S. Inequity in social exchange . In L.Berkowitz (Ed.), Advances in experimental socialpsychology. Vol . 2. New Y o r k : Academic Press,1965.A D A M S , J . S., & R O S E N B A U M , W. B. The relat ionshipof worker productivity to cognit ive dissonanceabout wage inequi t ies . Journal of Applied Psychol-ogy, 1962,4 6, 161-164.B E R L Y N E , D. E. Curios i ty and explorat ion. Science,

    1 9 6 6 , 153, 25-33.D E C H A R M S , R . Personal causation: Th e internal

    affective determinants of behavior. Ne w Y o r k :Academic Press, 1968.

    DECI, E. L. Effects of external ly mediated rewardson int r ins ic mot ivat ion. Journal of Personality an dSocial Psychology, 1971, 18, 105-115.G O O D M A N , P. S., & F R IE D M A N , A. An examinat ionof Adams ' theory of inequi t y . AdministrativeScience Quarterly, 1971, 16 , 271-288.HEIDER, F. Th e psychology of interpersonal relations.New York: Wiley, 1958.HUNT, J . McV. Int r ins ic mot ivat ion and i ts role inpsychological development . Nebraska Symposiumon Motivation, 1965, 13 , 189-282.KOCH, S. Beh avior as intrinsical ly regu lated : W orknotes towards a pretheory of phenomena cal ledmotivational. Nebraska Symposium onMotivation,1956, 4 , 42-87.L A W L E R , E. E. Equity theo ry as a predic tor ofproduct ivi ty and work qua l i t y . PsychologicalBulletin, 1968, 70, 596-610.P R I T C H A R D , R. D. Equity theory: A review andcri t ique . Organizational Behavior andHuman Per-formance, 1969, 4 , 176-211.WHITE, R. W. Mot ivat ion reconsidered: The conceptof competence. Psychological Review, 1959, 66,297-333.WOOD, L, & L A W L E R , E. E. Effects of piece-rateo v e r p a y m e n t on produc t iv i t y . Journal of AppliedPsychology, 1970, 54, 234-238.Z E I G A R N I K , B. Das behalten erledigter und u n e r -ledigter handl ungen . Psychologische F orschung,

    1 9 2 7 , 9, 1-85.(Received February 26 , 1971)