an opponent process theory of job satisfaction

15
Journal of Applied Psychology 1978, Vol. 63, No. 5, 533-547 An Opponent Process Theory of Job Satisfaction Frank J. Landy Pennsylvania State University The role of job satisfaction in research and theory in the area of industrial and organizational psychology is considered. Job satisfaction seems to occupy a position as the hedonic or affective component in theories of motivation. In spite of its importance, little theory is available for understanding the affective state represented by the concept of job satisfaction. Opponent process theory is suggested as a reasonable deductive statement for a consideration of the phenomenon of satisfaction. The theory proposes that every excursion from hedonic neutrality is accompanied by an attempt to bring the excursion back within "normal" limits. This return to normal levels is accomplished via an opponent process. The theory further suggests that the opponent process grows in strength with use. The theory is applied to some current questions regarding the relationship between job satisfaction and work motivation. The major parameters of the theory are represented by a series of research hypotheses and corollaries. Job satisfaction is a popular concept in industrial and organizational psychology. The reasons for its popularity have been well documented in Locke's (1976) recent chapter in the Handbook of Industrial and Organiza- tional Psychology (Dunnette, 1976). At various times, the concept has been a dependent variable, an independent variable, a covariate, and a moderator variable. It has been linked to productivity, motivation, absenteeism and tardiness, accidents, mental health, physical health, and general life satisfaction. Fifty years of research have attempted to document the relationship between an individual's feelings about his or her job and that individ- ual's behavior. There does seem to be one common theme that pervades the research on job satisfaction. An affective state is implied. It is assumed that at any specific point in time, an individual occupies a point on a continuum that ranges from a strongly positive emotional state to a The author is grateful to Mary Dunnette, Jim Farr, John Hall, Jeanne Herman, and to the late Don Trurabo for comments on an earlier version of this manuscript. Requests for reprints should be sent to Frank J. Landy, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania 16802. strongly negative one. The position that the individual occupies depends on both internal and external variables. Job-related stimuli comprise a class of these variables. Thus, it is assumed that, at least in part, a person's emotional state is affected by interactions with the work environment. It is normally this portion of general hedonic or affective variance that is referred to as job satisfaction. Locke (1976) defines job satisfaction as "a pleasurable or positive emotional state resulting from the appraisal of one's job or job experiences" (p. 1300). This definition high- lights one of the major questions addressed in job satisfaction research: Under what conditions does a positive or negative state arise? The answer to this question has implications for theories of work motivation. At a basic level, all models of work motivation imply that individuals will expend energy in main- taining or increasing pleasurable experiences; conversely, energy will be expended in min- imizing or decreasing unpleasant experiences. Thus, the affective reaction of an individual to work-related stimuli (either positive or negative) is an indication of the potential power that those stimuli have for affecting the individual's behavior. As an illustration of the role of job satisfac- Copyright 1978 by the American Psychological Association, Inc. 0021-9010/78/6305-0533$00.75 533

Upload: conlib

Post on 18-Nov-2014

117 views

Category:

Documents


1 download

DESCRIPTION

By Landy, Frank J.Journal of Applied Psychology. Vol 63(5), Oct 1978, 533-547.AbstractConsiders the role of job satisfaction in research and theory in the area of industrial and organizational psychology. Job satisfaction seems to occupy a position as the hedonic or affective component in theories of motivation. In spite of its importance, little theory is available for understanding the affective state represented by the concept of job satisfaction. Opponent process theory is suggested as a reasonable deductive statement for a consideration of the phenomenon of satisfaction. The theory proposes that every excursion from hedonic neutrality is accompanied by an attempt to bring the excursion back within "normal" limits. This return to normal is accomplished via an opponent process. The theory further suggests that the opponent process grows in strength with use. The theory is applied to some current questions regarding the relationship between job satisfaction and work motivation. The major parameters of the theory are represented by a series of research hypotheses and corollaries.doi:10.1037/0021-9010.63.5.533

TRANSCRIPT

Page 1: An opponent process theory of job satisfaction

Journal of Applied Psychology1978, Vol. 63, No. 5, 533-547

An Opponent Process Theory of Job Satisfaction

Frank J. LandyPennsylvania State University

The role of job satisfaction in research and theory in the area of industrial andorganizational psychology is considered. Job satisfaction seems to occupy aposition as the hedonic or affective component in theories of motivation. Inspite of its importance, little theory is available for understanding the affectivestate represented by the concept of job satisfaction. Opponent process theoryis suggested as a reasonable deductive statement for a consideration of thephenomenon of satisfaction. The theory proposes that every excursion fromhedonic neutrality is accompanied by an attempt to bring the excursion backwithin "normal" limits. This return to normal levels is accomplished via anopponent process. The theory further suggests that the opponent process growsin strength with use. The theory is applied to some current questions regardingthe relationship between job satisfaction and work motivation. The majorparameters of the theory are represented by a series of research hypothesesand corollaries.

Job satisfaction is a popular concept inindustrial and organizational psychology. Thereasons for its popularity have been welldocumented in Locke's (1976) recent chapterin the Handbook of Industrial and Organiza-tional Psychology (Dunnette, 1976). At varioustimes, the concept has been a dependentvariable, an independent variable, a covariate,and a moderator variable. It has been linkedto productivity, motivation, absenteeism andtardiness, accidents, mental health, physicalhealth, and general life satisfaction. Fiftyyears of research have attempted to documentthe relationship between an individual'sfeelings about his or her job and that individ-ual's behavior.

There does seem to be one common themethat pervades the research on job satisfaction.An affective state is implied. It is assumed thatat any specific point in time, an individualoccupies a point on a continuum that rangesfrom a strongly positive emotional state to a

The author is grateful to Mary Dunnette, Jim Farr,John Hall, Jeanne Herman, and to the late DonTrurabo for comments on an earlier version of thismanuscript.

Requests for reprints should be sent to Frank J.Landy, Department of Psychology, PennsylvaniaState University, University Park, Pennsylvania 16802.

strongly negative one. The position that theindividual occupies depends on both internaland external variables. Job-related stimulicomprise a class of these variables. Thus, it isassumed that, at least in part, a person'semotional state is affected by interactions withthe work environment. It is normally thisportion of general hedonic or affective variancethat is referred to as job satisfaction.

Locke (1976) defines job satisfaction as"a pleasurable or positive emotional stateresulting from the appraisal of one's job or jobexperiences" (p. 1300). This definition high-lights one of the major questions addressedin job satisfaction research: Under whatconditions does a positive or negative state arise?The answer to this question has implicationsfor theories of work motivation. At a basiclevel, all models of work motivation implythat individuals will expend energy in main-taining or increasing pleasurable experiences;conversely, energy will be expended in min-imizing or decreasing unpleasant experiences.Thus, the affective reaction of an individualto work-related stimuli (either positive ornegative) is an indication of the potentialpower that those stimuli have for affectingthe individual's behavior.

As an illustration of the role of job satisfac-

Copyright 1978 by the American Psychological Association, Inc. 0021-9010/78/6305-0533$00.75

533

Page 2: An opponent process theory of job satisfaction

534 FRANK J. LANDY

tion in current theories of work motivation,consider Maslow's (1943) need hierarchytheory. In this framework, unfulfilled lowerneeds represent a deficiency in the individual;this deficiency is experienced as discomfort bythe individual. Presumably, if an individualwere asked a question concerning well-beingwhile "under the influence" of this deficiency,the answer would imply some dissatisfaction.The theory proposes that the individual willengage in actions that will diminish thisdiscomfort. The same basic mechanism seemsto be implied in the fulfillment of upper levelneeds. Once lower level needs are fulfilled, thedegree to which upper levels needs are fulfilledassumes major importance for determiningthe degree of satisfaction of dissatisfactionthat an individual experiences. Once again,the theory implies that if an individual wereasked questions concerning his or her well-being, the answers would reflect the degree towhich the individual perceived these upperlevel needs as being met, and concomitantly,the degree to which the individual wouldengage in activities aimed at fulfilling theseneeds. Thus, in need hierarchy theory, statesof satisfaction-dissatisfaction precede directedbehavior. Additionally, since there is no firmcontinuum underlying the ordering of Maslow'sneeds in the hierarchy, one might infer thatsatisfaction at one level functions as a releasemechanism for dissatisfaction at the nexthigher level.

In equity theory, dissatisfaction is anunpleasant aftereffect of discordant cognitions.As such, this dissatisfaction represents a sourceof tension to be reduced, and the organismexpends energy in service of this reduction.The discordant cognitions are the result ofindividuals' comparisons of their own inputsand outcomes to the inputs and outcomes ofsignificant others. In that sense, equity theoryemphasizes the role of social stimuli in generalhedonic states. Nevertheless, the implicitassumption that individuals will engage inactivities to reduce tension or discomfort issimilar to the one proposed by Maslow.Research in equity theory has attempted toshow that individuals with discordant cogni-tions report dissatisfactions; this dissatisfac-tion is thought to represent felt tension; itis proposed that individuals will engage in

activities related to the reduction of thistension. The role of satisfaction in equitytheory is less clear. The most reasonableinterpretation of available theoretical discus-sion and empirical research is that satisfactionis the absence of dissatisfaction. There hasbeen little attention devoted to the phenom-enon of the positive end of the hedonic contin-uum by researchers using the equity paradigm.

Finally, in the Porter and Lawler (1968)version of Vroom's (1964) VIE model, satisfac-tion is hypothesized to be a derivative variable,depending for its value on the match betweenexpected and obtained rewards. Satisfaction,in turn, has a general excitatory or inhibitoryeffect on future estimations made by theperson concerning the value of promisedrewards. By virtue of the capacity to rememberpleasant and unpleasant experiences andrelate these experiences to specific contextsand stimuli, the individual is able to anticipatefuture hedonic experiences and choose behav-ioral alternatives accordingly.

Even though the work motivation theoriespresented above are only a sample of a muchlarger domain, they are representative of theimplied role of job satisfaction in the area ofwork motivation. The following general prin-ciples would hold for most of these theories:(a) The hedonic state of an individual canbe located on a continuum from positive(pleasant) to negative (unpleasant), and (b)there are systematic interactions between theemotional state of an individual and work-related variables. These work-related variablesmay either be temporally antecedent to theemotional state (e.g., reward policies of theorganization) or temporally consequent to theemotional state (e.g., absenteeism or tardiness).

In the motivation theories presented above,job satisfaction had clear emotional overtones.It represented a psychophysiological state ofpleasure or displeasure. This psychophysiolog-ical state had implications for future actionchoices by the individual. Thus, it is surprisingthat there have been no well-articulatedtheories that include hedonic components inaccompanying postulates or corollaries. Mostresearch attention has been directed towardidentifying the environmental elements orconditions capable of producing positive ornegative emotional states. Taylor (1947)

Page 3: An opponent process theory of job satisfaction

OPPONENT PROCESS 535

implied that monetary rewards mediated theseemotional states in workers. The Hawthorneresearchers replaced monetary rewards withsocial rewards and incentives (Roethlisberger& Dixon, 1939). Schaffer (1953) introducedthe notion of individual differences in thesalience of rewards. Maslow (1943) suggestedfive sources of satisfaction and dissatisfaction.Herzberg, Mausner, and Snyderman (1959)proposed two general sources, and Alderfer(1969) suggested three classes of potentialrewards. Even though they were not clearly"content theorists," the Cornell researchers(Smith, Kendall, & Hulin, 1969) implied fivesources of job satisfaction. Locke (1976)proposes that feelings of well-being or satisfac-tion derive from a system of values rather thanfrom needs. While needs are innate, valuesare assumed to be learned. It is these valuesthat determine the individual's actual choicesand emotional reactions. Thus, values arethought to be the determiners of hedonic states.

In addition to content theories of satisfac-tion, emphasizing the source of the emotionalstate, there have been various process theoriesof satisfaction, suggesting how various reac-tions to environmental stimuli combine toproduce the state called satisfaction ordissatisfaction. Early theorists implied a singlevariable, linear model. Schaffer (1953) madesome radical changes in that model. Hesuggested multiple variables with weightsdetermined by need strength. This might bethought of as the discrepancy model (Lawler,1973) and is characteristic of the work ofmany satisfaction researchers (Katzell, 1964;Likert, 1961; Morse, 1953). Herzberg et al.(1959) suggested a radically different processmodel proposing that satisfaction and dissatis-faction were not on a single continuum andthat there were limits to the effects of increased"rewards," yielding asymptotic curves. Recentwork by more cognitively oriented theoristshas generally followed a discrepancy frameworkin which rewards are thought to be linearlyand additively related to the general emotionalstate on a satisfaction-dissatisfaction con-tinuum.

If one accepts the proposition that jobsatisfaction represents some affective statethat is an important component of mosttheories of work motivation, it is distressing

to recognize that attention has been paidalmost exclusively to the conditions antecedentto that affective state (Lawler, 1973). Littleor no attention has been paid to the character-istics of the state itself or to the intra-individualpast history of that state. For example, fewresearchers in the area would quarrel with theproposition that the perception of constantlevels of work-related stimuli (such as pay,co-workers, or challenge) changes systemat-ically over time. Yet no current theories ofjob satisfaction suggest any mechanisms forunderstanding this change. A valuable theoryof hedonic states and their effects on behaviorshould be able to deal with these systematicand gradual changes in hedonic states overtime. There have been some attempts to dealwith these changes in the general experimentalparadigm as adaptation (Helson, 1964) orhabituation (Groves & Thompson, 1970). Inthe industrial and organizational area, mech-anisms like arousal or activation have beeninvoked to deal with perceptions of stimuluschange (Duffy, 1962; McGrath, 1970; Scott,1966). Nevertheless, the issue of satisfactionis seldom explicitly addressed in these con-siderations. As I have indicated above, thehedonic state represented by the term satisfac-tion has some rather important implicationsfor more general theories of industrial behavior.Consequently, theories of satisfaction shouldmore clearly address the characteristics of thestate itself, not solely the antecedents of thisstate.

The present article is an attempt to adapta general theory of derived motivation(Solomon & Corbit, 1973) to a specific con-sideration of the phenomenon of job satisfac-tion. In this adaptation, specific emphasiswill be placed on affective arousal, main-tenance, and decay as separate processes. It ishoped that by decomposing the phenomenonof job satisfaction-dissatisfaction and reducingit to simpler process components, the flexibilityand utility of the concept will be broadened.

An Opponent Process Theory ofJob Satisfaction

In several recent articles, Solomon (Hoffman& Solomon, 1974; Solomon & Corbit, 1973,1974) proposes that many emotional or hedonic

Page 4: An opponent process theory of job satisfaction

536 FRANK J. LANDY

Primary Process

Hedonic _^Neutrality \.

_ ittr&shold

Opponent Process

I STIMULUS Ionset termination

Figure 1. Underlying opponent processes after fewstimulus presentations.

states that can be commonly observed inhumans seem to follow a regular pattern ofchange. This pattern of change suggests thatthere is a two-phase underlying process govern-ing manifest emotional behavior. The firststage is excitatory and commences withstimulus presentation. This is called theprimary process. The second state is inhibitory(at least initially) and is instigated by theexcitatory or primary state. The second stageis called the opponent process. Figure 1graphically depicts the interaction of the twoprocesses. Stimulus onset is shown at thebottom of the figure. The upper portion of thefigure shows the effect of the stimulus onseton the primary process and the sequentialeffect of the primary process on the opponentprocess when a hedonic threshold is exceeded.

Solomon proposes that the inhibitory oropponent process seems to function as acontrol on the level of stimulation. The modelimplies that there are limits to a departurefrom hedonic neutrality that are acceptable tothe organism. The interaction of the primaryand opponent processes is thought to begoverned by neural mechanisms that placelimits on levels of excitation. Thus, theoperation of the mechanism is straightforward:(a) the organism is excited by external stimula-tion; (b) when excitation exceeds criticallevels, an opposing inhibitory process com-mences to bring excitation within normal oracceptable levels; (c) when external stimulationdisappears, the primary process ceases, and

shortly thereafter, the opponent processdecays as well. Figure 2 presents the neteffect of stimulus onset and termination andthe joint operation of the primary and oppo-nent processes. Thus, Figure 2 representsthe behavioral manifestation of the opposingprocesses.

Solomon suggests that this type of model ismost parsimonious for explaining a number ofhedonic phenomena. In support of the model,he presents data from both infrahuman andhuman studies of "emotionality." Thesestudies include examinations of imprintingbehavior in young ducklings, responses ofdogs to varying amounts of electrical shock,reactions of parachutists to free-fall conditions,responses to the loss of a loved one, cyclicvariations in opiate usage, and cyclic variationscharacteristic of cigarette addiction. He pointsout that when individuals are stimulatedabove some level, they do not return imme-diately to some "base" level when the sourceof stimulation is removed. Instead, they tendto "overshoot" the base level in a directionopposite to that produced by the originalstimulation. Only gradually do they return tobase level. This overshoot phenomenon canbe seen both physiologically and behaviorally.Solomon stresses the capacity of the opponentprocess mechanism to explain systematicchanges in hedonic responses over time. Itis this capacity that makes the theory poten-tially useful in understanding the phenomenonof job satisfaction. There are a number oftheoretical postulates that must be presented

Hedonic

Neutrality

STIMULUS!onset termination

Figure 2. Manifest hedonic response after few stimuluspresentations.

Page 5: An opponent process theory of job satisfaction

OPPONENT PROCESS 537

in order to fully understand how the proposi-tions relate to hedonic change.

The first major postulate is that there arecentral nervous system mechanisms thatfunction in such a manner as to reduce allvariations beyond some normal excursionsfrom hedonic neutrality. Such mechanismsmust assume the role of a hypothetical con-struct at this point. Nevertheless, it is noless reasonable than other similar constructssuch as "activation" or "arousal" levels(Duffy, 1962; Scott, 1966).

The second postulate of the theory is thatthese mechanisms work identically for positiveand negative stimuli. As far as central nervoussystem responses are concerned, both positiveand negative variations from hedonic neutral-ity are equally threatening. Once again, thisassumption is similar to the inverted-Uhypothesis of the arousal models (Duffy, 1962).The overarousal end of the continuum impliesthat too much stimulation is counterproductiveto the organism regardless of whether thesource is thought to be pleasant or unpleasanton a normative basis.

The third postulate is that, at least initially,both the primary response to the stimulus(primary process) and the mechanism attempt-ing to control this primary process (theopponent process) are automatic rather thanlearned.

The fourth postulate is that the process ofstimulus reduction begins at the point atwhich some hedonic threshold is exceeded.This stimulus reduction or opponent processmight be thought of as a negative feedbackloop or dampening mechanism. Further, it isassumed that the opponent process is somewhatsluggish in decay. Thus, while the primaryaffective response system decays quickly afterexternal stimulation is terminated, the oppo-nent process disappears more slowly.

The final postulate is, by far, the mostcrucial. It is assumed that the opponentprocess is strengthened with use and weakenedwith disuse. It is this final assumption thatlays the groundwork for explaining systematicchange in hedonic response levels to constantstimuli. A corollary of this proposition isthat the primary process or response tostimulation never changes in intensity orreactivity. It is only the opponent process

that changes as a function of stimulus presenta-tion.

It is important to note that the two processesare inextricably bound when some thresholdlevel of stimulation is exceeded. In this case,the opponent process is referred to as a " slave"process. Although the opponent process as-sumptions do not deny the propositions ofarousal theorists, the theory does imply thatonce some critical level of arousal is attained,monotonicity of hedonic responsiveness dis-appears.

Although Solomon suggests that the oppo-nent process mechanism operates identicallyfor both positive and negative primarystimulation, there is one basic differencebetween the positive and negative primaryprocess with respect to the behavior ofindividuals. A negative primary process isproduced by an aversive stimulus. At leastinitially, individuals are unlikely to engagein actions voluntarily that would lead toaversive stimulus onset. This would continueto be true until the opponent process (oppositein sign, thus, positive) grew in strengthsufficient to suppress the aversive effects ofthe primary stimulus and produce a pleasantdecay period following stimulus termination.This is not true with primary positive states.Individuals will engage in activities necessaryto produce the positive primary states vol-untarily. As the opponent process grows instrength (a negative opponent process, inthis case), individuals will increasingly engagein activities that will cause the reoccurrenceof the pleasant primary state and the termina-tion of the unpleasant opponent state. Thus,there is some asymmetry to the positive andnegative primary and opponent states. Thisasymmetry will be important in applying theopponent process theory to the concept ofjob satisfaction.

Even though the focus of the present articleis not on the content aspects of job satisfaction(i.e., exactly which factors "cause" satisfactionor dissatisfaction) but rather on the processaspects (i.e., the operations which govern theemergence, maintenance, and decay of job-related emotional states), some attentionshould be given to how job-related stimulimight be categorized as positive or negative onan a priori basis. Such a categorization is

Page 6: An opponent process theory of job satisfaction

538 FRANK J. LANDY

Primary Process

HedonicNeutrality

Opponent Process

[STIMULUS!onset terminationFigure 3. Underlying opponent processes after manystimulus presentations.

important in avoiding tautologies similar tothose that hindered the behaviorists in theirattempts to define reinforcement. There areseveral alternatives that might be chosen foran initial attempt at categorization. One mightchoose the operant paradigm by definingpositive stimuli as those that increase theprobability of a class of responses when thepresentation of these stimuli is made contingentupon a response of that class. Most contenttheories of satisfaction presume such para-metric information (e.g., two-factor theory;the Cornell studies) or ignore the impliedcircularity (e.g., pleasant working conditionsare satisfying). The asymmetry of primaryhedonic states under conditions of pleasurableand aversive initial stimulation demandsrigorous and independent definitions of theproperties of these stimuli. The operant para-digm peresents one avenue for resolution.

It was indicated above that the theoryproposes that the opponent process grows instrength with use and decreases in strengthwith disuse. This implies that the manifesthedonic response of an individual afterfew stimulations will be different than thatfor the same individual after many stimula-tions. Figure 2 depicted the manifest hedonicresponse level after few stimulations. InFigure 3, hypothetical primary and opponentprocesses are shown after many stimulations.Notice the increase in strength of the opponentprocess compared to the strength of the process

in Figure 1. Figure 4 presents the manifesthedonic response after many stimulus presenta-tions. Once again, notice the dramatic dif-ference between the two manifest hedonicresponses represented by Figures 2 and 4.Notice particularly the effect of the terminationor weakening of stimulation after manystimulus presentations. If the original stimuluswould be considered "positive," the aftereffectwould be negative. If the original stimuluswere aversive, the aftereffect would be positive.

Based on the assumption that the opponentprocess grows in strength and that speed ofinitiation is directly proportional to its use,we would predict that after many stimulations,an individual has a much more modest hedonicresponse to stimulus presentation but a muchmore dramatic and opposite response tostimulus termination.

Since his original statement of the theory,Solomon (Star & Solomon, Note 1) hasinvestigated conditions surrounding the de-velopment and decay of opponent processes.Of major importance was a recent findingthat the opponent process seemed to increasein strength only when the interstimulus-interval period was less than the decay periodfor the opponent hedonic response. Thissuggests the operation of classically conditionedmechanisms in opponent process operation.

There is a good deal of work to be done onthe basic propositions of the model. Forexample, Solomon suggests that there may beprimary processes that have no correspond-ing opponent processes (Solomon & Corbit,1974). It is more likely that the speed withwhich opponent processes develop, and theirultimate strength, is related to the degree towhich stimulus termination can be identified

Hedonic LNeutrality

ISTIMULUSlonset termination

Figure 4. Manifest hedonic response after manystimulus presentations.

Page 7: An opponent process theory of job satisfaction

OPPONENT PROCESS 539

by the organism. Consequently, we are morelikely to find well-developed opponent proc-esses in situations in which stimulus presenta-tion follows cyclic parameters closely. Examplesof such situations would be invariant task-specific cycle time, fixed reward intervals,and other consistent temporal parametersgoverning individual-environment interactions.It will be necessary to conduct basic investiga-tions in the area of contexts that are moreor less conducive to the development ofopponent processes. In addition, basic researchmust be conducted to determine the propertiesof several parameters of the model, for example,initiation time, response intensity, and decaytime for primary and opponent processes.Nevertheless, the model provides some ex-planatory mechanisms that are crucial forunderstanding the role of hedonic states intheories of satisfaction. In particular, thetheory provides an excellent framework forunderstanding the affective responses ofworkers to work-related stimuli as well as adescription of how those responses change as afunction of repeated stimulus presentation.Opponent process theory describes a manifestemotional state (e.g., job satisfaction-dissatis-faction) as the resolution of two opposinghedonic processes. Further, it suggests amechanism for understanding how stimuli inthe work environment lose their value.

The theory is not particularly parsimoniouscompared to other process theories such asSchaffer's (1953) or Lawler's (1973). Never-theless, when the simpler theories attempt toexplain why two individuals respond differentlyto the same job conditions, their argumentsoften become circular (e.g., "individual dif-ferences in the salience of various rewardsmust be taken into account as well . . .")•In such cases, parsimony is a vice not a virtue.We are asked to accept individual differencesas an explanation rather than an admission ofignorance. In the area of job satisfaction, theopponent process theory implies the oppositeof parsimony, reductionism. It implies thatprevious process theories have stopped shortof efficient explanation. To be sure, therewill be individual difference parameters inthe opponent process model as well. Some ofthese will be introduced in a later section.Nevertheless, it should be clear that the

opponent process theory has both methodolog-ical and theoretical implications for the areaof job satisfaction. In the next section, someof these implications will be introduced.

Measurement Implications

If the opponent process theory is a reason-able process structure for understanding anindividual's response to work-related stimuli,there are a number of serious implications forthe ways in which that hedonic response ismeasured. The most serious of these implica-tions is the role of temporal variables in themeasurement of satisfaction. From an inspec-tion of Figure 4, it should be clear that aresponse to a satisfaction question will varydepending upon whether the question is askedprior to stimulus onset, prior to stimulustermination, or after stimulus termination.Further, the intensity of the response willdepend on whether the stimulus has beenencountered often or seldom. Finally, thedirection of the satisfaction response (posi-tive-negative) will depend on the interactionbetween the direction of the primary state(positive-negative) and the point on thehedonic continuum when the response isprovided (stimulus on-stimulus off).

A second major measurement implication isthat correlational within-subject designs pur-porting to address stability of satisfaction areinappropriate. The opponent process modelwould suggest that while correlatons betweenresponses on two separate occasions may behigh, absolute levels of satisfaction can varydramatically as a function of the number ofstimulus presentations between the first andsecond measurements. In addition, low correla-tions over time may be a result of unequalnumbers of stimulus presentations acrossindividuals rather than inherent structuralflaws in the instrument used to measuresatisfaction.

A third measurement implication has to dowith the role of reward salience, valence, orimportance. It is common in satisfactionstudies to ask a subject how important orsalient a particular reward is. A numberrepresenting the degree of that importance isthen used as a multiplier of some kind in a"need strength-need satisfaction" paradigm

Page 8: An opponent process theory of job satisfaction

540 FRANK J. LANDY

(e.g., Schaffer, 1953; Vroom, 1964). Theopponent process theory suggests that theimportance an individual attaches to apotential reward will be affected by both thedecay period of the last presentation of thatreward and the number of previous presenta-tions of that reward. An examination of Figure4 might make this point clearer. If relativeimportance of a particular reward is assessedby means of self-report measures prior tostimulus termination, it is likely to be con-sidered less important than after the individualreturns to base or prereward level; if impor-tance is measured after stimulus terminationbut prior to return to base or prereward level,the factor will likely be considered veryimportant. In fact, reported satisfactionwith the element should be identical toreported importance of the element. Thisis exactly what has been demonstrated inprevious research (Ewen, 1967; Locke, 1976).In a similar manner, individuals who haveexperienced many previous stimulus presenta-tions should assign relatively less importanceto stimuli prior to stimulus termination thanindividuals with fewer previous stimuluspresentations. Such a mechanism was suggestedoriginally by Maslow (1943) in an attemptto build in individual differences parameters tohis concept of a need hierarchy.

There are certainly many other measurementimplications. Those mentioned above rep-resent a sample of a larger domain. Neverthe-less, even this small sample suggests that(a) satisfaction reponses should be gatheredwithin individuals across occasions, (b) analysesof change over time should include both leveland shape changes, and (c) verbal report ofsatisfaction and/or importance responses mustbe examined for systematic temporal hedoniceffects. A corollary to this third suggestion isthat physiological, psychophysiological, andbehavioral indicators of hedonic level orstate be examined for estimates of satisfaction.While the physiological and psychophysiolog-ical effects of satisfaction and dissatisfactionhave been considered by researchers in therecent past (Frankenhauser, 1974; Kahn,1973), few researchers have considered anchor-ing satisfaction-dissatisfaction in physiolog-ical or psycholophysiological networks.

Theoretical Implications

In a recent examination of the currentstate of work motivation theory, Campbelland Pritchard (1976) identify the role ofreinforcement as crucial in experimentallygrounded theories of work motivation. Thiswould be particularly true for unadornedassociationist and reinforcement positions butwould also have implications for need anddrive theorists as well. The basic question is"What makes a reinforcer reinforcing?" Whileopponent process theory does not purport toanswer the content question implied, it doesdeal more effectively with a corollary to thatquestion than other theories, that is, Underwhat conditions will reinforcers increase anddecrease in strength. Campbell and Pritchardsuggest that the mechanism of reinforcementmight be found in the central nervous system,citing the work of Miller (1957) and Olds(1962), among others. They also imply that thework of the "activation" and arousal-leveltheorists leads to an examination of physio-logical changes as reinforcements in and ofthemselves. If that were to be the case, theopponent process theory would assume contentaspects as well. If one assumes (a) that jobsatisfaction (regardless of how it is operational-ized) represents a point on a hedonic con-tinuum, (b) that hedonic level covaries withreinforcement, and (c) that reinforcement hasimplications for directed energy, then theopponent process theory would seem to repre-sent a move in the direction suggested byCampbell and Prichard.

The relationship between the opponentprocess theory and the construct of arousal ishard to ignore. Both imply some optimal levelof stimulation, and both imply that beyondthat level both "positive" and "negative"stimulation are equally counterproductive.Nevertheless, opponent process theory isable to extend the construct of arousal oractivation by suggesting a mechanism forunderstanding shifting base arousal levels.Berlyne (1967) has suggested that progressin understanding the impact of arousal onmanifest behavior can be made by consideringthe interaction of arousal and reinforcement.These mechanisms and interaction processesare well represented in the primary and

Page 9: An opponent process theory of job satisfaction

OPPONENT PROCESS 541

opposing processes suggested by Solomon. Inthis light, arousal occupies the role of a hypo-thetical construct while opponent processtheory helps to define that construct operation-ally and to suggest theoretical postulates andtheir concomitant corollaries.

Campbell and Pritchard (1976) suggest thatthere is sufficient evidence available toconclude that the "overall judgment about thejob (job satisfaction) is made up of two,sub-general factors corresponding roughlyto the intrinsic vs. extrinsic breakdownoriginally indentified by Herzberg" (p. 103).Thus, independent of Herzberg's propositions,a review of the available literature suggeststhat there are at least two distinct classes ofjob-related factors that have the capacity toeffect manifest behavioral patterns of workers.

Several researchers have suggested thatthese two classes of factors have differentialmain effects as well as interactive effects onmanifest behavior (deCharms, 1968; Deci, 1972,1975; Herzberg et al., 1959). Opponent processtheory would also suggest differential maineffects, although interactive effects would notbe as clear. Let us assume that the presentationof extrinsic factors generally results in apositive primary state. The presentation ofmoney, the opportunity to interact withco-workers, a promotion, etc., are much morelikely to result in positive hedonic states thannegative ones in a random sample of subjects.That is not to say that every individual willbe equally pleased with the every extrinsicelement, but that in a normative sense, thepresentation of stimulus elements comprisingthe class "extrinsic" will most likely yieldpleasant hedonic states.

Intrinsic factors are represented by suchterms as responsibility, challenge, demand, etc.A number of theories (Herzberg et al., 1959;Locke," 1965; McClelland, 1961) suggest thatit is not the challenge itself that produces apleasurable state, but successfully meetingthat challenge. In a recent consideration of therelationship of challenge or increased respon-sibility to job satisfaction, Locke (1976)proposed that "these challenges must besuccessfully overcome for the individual toexperience pleasure" (p. 1320). If one wereto invoke the propositions of the opponentprocess theory to describe the phenomenon, one

would deduce that the challenge represents a"stimulus on" condition and meeting thechallenge or overcoming the obstacle representsstimulus termination. Further, if the "stimulusoff" condition produces a pleasurable state,then the stimulus on state must be opposite insign or unpleasant. If this were the case, oneshould be able to demonstrate that individualsinitially require some urging to accept increasedchallenge and responsibility since the primarystate is negative and the opponent state isinitially weak. This further suggests thatindividuals will initially attempt to avoidchallenge when possible rather than meetingit and overcoming obstacles to success. Afterrepeated presentations of challenge or in-creased responsibility followed by successfulstrategies for meeting that challenge, thechallenge itself should become less aversiveand the hedonic state following successfulmeeting of that challenge should become morepleasurable. These predictions would followdirectly from the proposition that the opponentprocess grows in strength with use.

If we were to return briefly to a considerationof the effect of repeated presentations ofextrinsic factors on hedonic state, we wouldobserve the converse of the effect of intrinsicfactors such as challenge. The repeatedpresentation of the extrinsic factor shouldresult in diminishing levels of pleasure. The pre-sentation of these factors would be consideredthe stimulus on condition. The absence of thesefactors would be considered the stimulus offcondition, and the theory would predict thatincreasing levels of displeasure would beassociated with these stimulus elements indirect proportion to the number of stimuluspresentations. Thus, after extended periodsof time, elements that originally had powerover behavioral choice by virtue of theirpresence now have control only by virtue oftheir actual or threatened absence.

In summary, opponent process theorypredicts the affective reactions among experi-enced workers to intrinsic and extrinsic factorson the job; intrinsic factors have the capacityto yield neutral hedonic states in their presenceand positive hedonic aftereffects; extrinsicfactors produce neutral hedonic states intheir presence and negative hedonic aftereffects.A simpler way of stating this relationship

Page 10: An opponent process theory of job satisfaction

542 FRANK J. LANDY

might be that extrinsic factors can producedissatisfaction after their termination andintrinsic factors can produce satisfaction aftertheir termination. If an experienced workerwere asked to recall pleasant work-relatedsituations, these situations would involveintrinsic factors (e.g., having overcome adifficult challenge); if the same worker wereasked to recall unpleasant work-related situa-tions, these situations would involve extrinsicfactors (e.g., the withdrawal of environmentalattributes normally seen as pleasurable). Thus,opponent process theory would lead one todeduce the principles that form the basis ofHerzberg's inductive theory of job satsifaction.

If one considers that challenge produces anegative hedonic state that is replaced witha positive hedonic state when the challenge ismet, the implications for feedback in the worksetting are rather dramatic. In most worksettings, individuals require outside confirma-tion that a challenge has been met. The mostobvious source for this feedback is the super-visor. Thus, if individuals are provided withchallenge yet never or seldom told when theyhave successfully met the challenge (or madeprogress toward meeting it), one would expectconstant levels of negative tension. Feedbackindicating that a challenge has been met isnecessary for the " termination" of the stimulusrepresented by the challenge. In this frame-work, feedback is a necessary but not sufficientcondition for producing positive hedonicstates. The sufficient condition is controlledby the number of previous presentations ofchallenge.

Mechanisms such as those described abovemight also help account for the effects of jobenrichment. If the enriched job representsincreased levels of challenge, initial reactionsof workers should be negative to varyingdegrees (depending on the degree of perceivedchange between the old levels of challenge andthe new levels). As the "new" job is learnedand the challenges successfully met, resistanceand negative responses should decrease andsatisfaction with mastery of the new jobshould increase. The degree to which the jobis enriched should have implications forresulting hedonic states. If the enriched jobrepresents dramatically increased levels ofchallenge, the potential for dissatisfaction with

the new job is high and prompt performancefeedback imperative.

An extension of the consideration of jobenrichment is an examination of the mirrorimage, that is, boring tasks. Boredom anddissatisfaction with the work itself are treatedsynonymously, and both are thought to be aresult of lack of mental challenge (Locke,1976). In spite of the importance of thephenomenon of boredom for understandingworkers' reactions to jobs, there are fewtheories that provide any insights into howtasks become boring. Activation and arousaltheory (Duffy, 1962; Scott, 1966) suggestthat the particular stimulus element fallsbelow some optimal level, which results inunderarousal, a negative hedonic state. Never-theless, it is commonplace to note that thesestimulus elements become "boring" over time.Helson (1964) has suggested that this increas-ing impotence of stimuli to arouse the orga-nism is a result of some neural mechanismrepresented by the hypothetical construct of"adaptation level." The operations of hedonicexcitation, maintenance, and decay are con-siderably clearer in the opponent processtheory than in the adaptation or arousallevel frameworks. This is due in part to thefact that both arousal and adaptation proposi-tions are more directly concerned with actionthan emotion.

Opponent process theory requires one toview the issue of boredom as a more generalphenomenon than currently implied by extantliterature. The issue becomes one of thehabituation level of an organism relative tovarious classes of stimuli. Thus, one can bejust as easily "bored" with levels of pay asone can with tasks required by a particularjob. In this framework, job enrichment mightbe thought of as an instance of stimulussensitization; boring tasks would be aninstance of habituation. In this framework, theadjective boring might be applied to any oneof a number of stimulus elements that char-acterize the job, not just the tasks that definethat job.

When the concept of boredom is extendedin this way, there are few theories of jobsatisfaction capable of explaining the phenom-enon. Groves and Thompson (1970) havesuggested that a model of habituation must

Page 11: An opponent process theory of job satisfaction

OPPONENT PROCESS 543

include at least the following parameters:time course and recovery, effect of repeatedseries, effect of stimulus frequency, effect ofstimulus intensity, and stimulus generalizationgradients. With the exception of the finalparameter, opponent process theory meets therequirements. In their consideration of habit-uation, Groves and Thompson suggest thathabituation and stimulus frequency are expo-nentially related. Thus, opponent processtheory would seem uniquely suited for adetailed decomposition of the phenomenonof boredom and an extension of the conceptto cover stimulus elements other than jobtasks.

There are some clear differences betweenthe approach that activation theorists (Scott,1966) take to decompose the hedonic phenom-enon accompanying boredom and that pro-posed by opponent process theory. Theactivation position would be that all departuresfrom some idiographic "normal" level wouldbe noxious to the organism and that theorganism would attempt to escape from thatstate and ultimately avoid that state bycontrolling levels of stimulation. Opponentprocess theory would separate manifest hedonicbehavior from latent states represented by theprimary and opponent processes. Thus, individ-uals would be capable of both positive andnegative affective reactions toward stimulithat exceed some threshold. A good deal ofrecent physiological research (Frankenhauser,1974) on stress suggests that there are "hidden"costs to the organism for overarousal. Thisresearch suggests that individuals are notalways capable of recognizing the harmfuleffects or costs of manifestly pleasant ex-periences. This line of research clearly impliesthat not all suprathreshold arousal is con-sidered noxious by the organism, contrary tothe propositions of activation theory.

Finally, the propositions of opponent processtheory suggest some modifications of Locke'sgoal-setting theory of work motivation and jobsatisfaction (Locke & Bryan, 1969; Locke,Cartledge, & Knerr, 1970). A major propositionof Locke's approach is that individuals whoaccept hard goals expend greater energy andhave the capacity to experience greatersatisfaction than individuals who accepteasier goals or no goals at all. As indicated

earlier, Locke (1976) implies that achievinggoals produces pleasure, and opponent processtheory would imply that setting goals mayhave negative hedonic effects. Thus, early ina goal-setting "career," individuals will ac-tively resist goal setting and receive minimalamounts of pleasure from achieving thosegoals. As experience with goal setting andgoal attainment increases, resistance shoulddecrease (since the primary state is reduced bythe increasingly powerful opponent state)and derived pleasure from goal attainmentshould increase (since the aftereffect rep-resented by the opponent state becomesincreasingly stronger and decays more slowly).Thus, opponent process theory providesgoal-setting propositions with a mechanismfor describing some affective history. Thisaffective history, in turn, represents an in-dividual differences parameter affecting thelikelihood that an individual will accept ahard goal and be satisfied with its achievement.

Parametric Investigations

It should be obvious that the first majorobstacle to be overcome in the application ofopponent process theory to the study ofjob statisfaction is a specification of theparameters of the model. An examination ofFigures 1-4 is sufficient to identify most ofthem.

The first major question concerns thedefinition of a "stimulus." The work environ-ment is rich with stimuli, and one must beable to decompose that environment intomain effects and interactions. Opponent proc-ess theory demands such dec&mposition toa greater degree than other approaches tosatisfaction. The content theories of job satis-faction ask us to accept a limited number ofhomogeneous stimulus elements. Neverthe-less, both the opponent process theory andalternative approaches face this obstableequally. A reasonable strategy for an examina-tion of this question might be borrowed fromthe experimentalists. The concept of stimuluscontrol (Terrace, 1966) might be a usefulstrategy for addressing this issue. In such aparadigm, the organism's ability to dis-criminate among stimuli is of central concern.

Page 12: An opponent process theory of job satisfaction

544 FRANK J. LANDY

The experimental designs resemble thoseassociated with constructing generalizationgradients for various stimulus elements. Opera-tionally, one might attempt to determinethe independence of stimulus elements in termsof the response characteristics of the organism.If the response levels to various stimuluselements is relatively flat, we might concludethat the stimulus elements are not distinctfrom the point of view of the respondent. Onthe other hand, if the generalization gradientsare steep, we would conclude that the stimuluselements are perceived as independent by theindividual. Historically, researchers in thearea of the job satisfaction have proposedstimulus groupings on the basis of the verbalresponses across individuals. These responseshave traditionally been factor or clusteranalyzed and functional generalization gra-dients assumed. This is an imprecise paradigmfor determining stimulus properties. Thestimulus generalization or stimulus controlparadigm is an alternative approach to theproblem that may be of some value.

A second question relates to the "acceptable"levels of neural stimulation. Opponent processtheory proposes that when some threshold isexceeded, the opponent process is activated.The definition of this acceptable-unacceptableexcursion from neutrality must be considerablytightened. Once again, this is not a problempeculiar to industrial and organizationalpsychologists. In some senses, this parameteris tied to the question of when a stimulus is"present." Experimental psychology has dealtwith variations of this question in the justnoticeable differences paradigm. To someextent, we are describing a process wherebythe organism "notices" that stimulus intensityhas changed. The question becomes one ofidentifying the level of stimulus changenecessary for the perception of change bythe organism. Weber's law and Steven's powerfunctions (Stevens, 1957) are operationaldefinitions of such changes for various modal-ities. The concept of just noticeable differencesis at the heart of equity theory (Adams, 1965;Jaques, 1961). In addition, there have beensporadic attempts to use classic psychophysicaloperations to plot the effect of increasedstimulus intensity against satisfaction in thework setting (Zedeck & Smith, 1968). These

strategies might be easily extended to plot the"power" of various work-related stimuli. Thiswould provide a preliminary specification ofthe threshold parameter of the model.

A third parameter of importance is thedecay functions of both the primary andopponent states. When the stimulus is ter-minated, it is unlikely that the primary statedisappears simultaneously. There is bound tobe some latency or decay. The same is true,and to a greater extent, of the opponentstate. In the simplest terms, the question is"How long does a stimulus last?" The experi-mental paradigm for studying such a questionis one of extinction. Since longitudinal studiesof satisfaction have been so infrequent, weknow little or nothing about the decay func-tions of stimulation as a molar concept, letalone the decay of two separate functions suchas those represented by the primary andopponent states in opponent process theory.Nevertheless, it is reasonable to propose thata decay parameter interacts with stimulusfrequency and intensity.

A fourth parameter is suggested by somerecent work done by Solomon (Starr &Solomon, Note 1). He suggests that theinterstimulus interval is the key to the increas-ing strength and decreasing latency of theopponent process. Specifically, in laboratorystudies it has been found that if the stimulusof interest is presented before the opponentstate of the previous stimulus presentationdisappears or decays, the opponent state willgrow in strength and decrease in latency.Conversely, if the next stimulus presentationconies after the decay or disappearance of theprevious opponent state, the opponent statewill decrease in strength. This suggests aclassical conditioning mechanism connectingthe primary and opponent states. This condi-tioning paradigm might be useful for examiningthe effect of interstimulus intervals. Intui-tively, the concept of interstimulus intervalsis an appealing one for work-related phenomenasuch as reward periods, job cycles, etc., butthere has been little or no systematic investiga-tion of this parameter.

Although there are other, less obvious,parameters that might be suggested by theory,the major ones have been presented above.These parameters should be important for

Page 13: An opponent process theory of job satisfaction

OPPONENT PROCESS 545

any theory of job satisfaction, yet they havenot been directly addressed by those consider-ing the question of job satisfaction. Theadvantage of opponent process theory is thatit provides a reason to investigate thoseparameters as well as a model for placing themin coherent relationships with each other.

These conceptual applications of opponentprocess theory to satisfaction and motivationin work settings are intended to demonstratethe value of considering a new paradigm inthe study of job satisfaction. The kinds ofquestions that opponent process theory raisesconcerning the phenomenon of job satisfactionhave not often been asked in publishedresearch. Nevertheless, they are appropriateand even crucial questions if we accept theproposition that the concept of job satisfactionis intended to represent some hedonic state.Most current approaches to job satisfaction(Lawler, 1973) imply that it is a cognitiveoutcome with simple hedonic consequences.Opponent process theory requires one toconsider the nature and consequences of thathedonic state in a more direct and complexmanner.

To be sure, the opponent process approachis problematic in any setting, but even moreso in attempting to explain the field behaviorof humans rather than the laboratory behaviorin infrahumans. As mentioned earlier, one ofthe most difficult problems in the applicationof the theory to the concept of job satisfactionis the notion of stimulus onset and stimulustermination. Since humans have the well-developed capacity to abstract, form concepts,and deal with the real world symbolically, onecannot easily control, or even measure,temporal stimulus properties. In addition,there are multiple primary and opponentprocesses operating simultaneously in the realworld. The manner in which these multiplehedonic systems interact becomes a centralquestion for the researcher. The intricacy andcomplexity of opponent process theory suggeststhat it is not terribly parsimonious whencompared to more compact approaches. Never-theless, the parsimony of past approachesto job satisfaction has not proven useful inincorporating the concept of satisfaction intothe broader question of motivation. Parsimonyin theory building is only desirable to the

degree that it improves our understanding orprediction.

Suggested Research

The immediate research questions suggestedby the opponent process theory are parametricin nature. As such, research generated by thesequestions should prove useful to all theorybuilding in the area of job satisfaction. Below,several of these questions have been formallystated as general hypotheses with specificcorollaries:

1. There is a systematic recruitment, main-tenance, and decay function that can beused to describe the hedonic response of anindividual over time to the presentation of aparticular stimulus element, (a) After thepresentation of a reward, an individual'ssatisfaction with that reward will systemat-ically change over time, eventually returningto prereward levels of satisfaction with thatreward, (b) Individuals will report bothsatisfaction and dissatisfaction with constantlevels of a particular reward at differingpoints in time.

2. Repeated stimulus presentation affectsthe manifest response of an individual to aparticular stimulus element, (a) The greaterthe number of times that a particular rewardor punishment has been presented, the longerit will take for an individual to return toprereward or punishment level, (b) Thegreater the number of times a reward has beenpresented, the less the satisfaction of theindividual with each succeeding presentation,(c) The greater the number of times a rewardor punishment has been presented, the greaterwill be the effect of terminating that reward orpunishment on job satisfaction of individuals.

3. The shorter the interval between succes-sive presentations of stimulus elements, thestronger the opposing response to the primarystate generated by those stimulus elements,(a) Rewards or punishments that are presentedfrequently will have smaller immediate effectson job satisfaction than those presentedinfrequently.

4. Goal-directed activity produces negativeprimary states and positive opponent states,(a) Individuals will report less satisfactionwith a particular task before meeting a goal

Page 14: An opponent process theory of job satisfaction

546 FRANK J. LANDY

than after meeting a goal, (b) Individualswill report less satisfaction with difficulttasks than with easy tasks prior to meetinga task goal, (c) Individuals will report greatersatisfaction with difficult tasks than with easytasks after meeting a task goal, (d) Individualswith a history of goal setting will report lessdissatisfaction during goal-seeking periodsthan individuals with no history of goal setting.

These are just a few of the hypotheses thatmight be posed on the basis of the opponentprocess theory and its application to thephenomenon of job satisfaction. There aremany other, less obvious, implications of thetheory. For example, stimulus intensity seemsless critical than stimulus frequency inunderstanding hedonic response level; it maymake little difference for hedonic states ifrewards are contingent or noncontingent uponsome particular behavior; avoidance andescape mechanisms are as crucial for under-standing in industrial behavior as are approachmechanisms.

In summary, the questions posed by aconsideration of opponent process theory andits applicability to job satisfaction are ques-tions that are crucial for any reasonabletheory of job satisfaction. Opponent processtheory represents the most well-developeddeductive research currently available. Assuch, it might be efficiently interlaced withcurrent content theories of job satisfaction toyield a more reasonable decomposition of thesatisfaction phenomenon.

Reference Note

1. Starr, M. D., & Solomon, R. L. Imprinting withbrief exposures: Factors affecting the development ofseparation distress in ducklings. Paper presentedat the meeting of the Eastern Psychological Associa-tion, New York, April 1976.

References

Adams, J. S. Injustice in social exchange. In L. Berko-witz (Ed.), Advances in experimental social psychology(Vol. 2). New York: Academic Press, 1965.

Alderfer, C. P. An empirical test of a new theory ofhuman needs. Organizational Behavior and HumanPerformance, 1969, 4, 142-175.

Berlyne, D. E. Arousal and reinforcement. In D. Levine(Ed.), Nebraska Symposium on Motivation (Vol. 15).Lincoln: University of Nebraska Press, 1967.

Campbell, J. P., & Pritchard, R. D. Motivation theoryin industrial and organizational psychology. InM. Dunnette (Ed.), Handbook of industrial andorganizational psychology. Chicago: Rand McNally,1976.

deCharms, R. Personal causation: The internal affectivedeterminants of behavior. New York: AcademicPress, 1968.

Deci, E. L. The effects of contingent and non-contingentrewards and controls on intrinsic motivation.Organizational Behavior and Human Performance,1972, 8, 217-229.

Deci, E. L. Intrinsic motivation. New York: Plenum,1975.

Duffy, E. Activation and behavior. New York: Wiley,1962.

Dunnette, M. Handbook of industrial and organizationalPsychology. Chicago: Rand McNally, 1976.

Ewen, R. B. Weighting components of job satisfaction.Journal of Applied Psychology, 1967, 51, 68-73.

Frankenhauser, M. Overstimulation: a threat to thequality of life. In, Man in the communications systemof the future. Stockholm: Swedish Cabinet Office,Secretariat for Future Studies, 1974.

Groves, P. M., & Thompson, R. F. Habituation: Adual process theory. Psychological Review, 1970,77, 419-450.

Kelson, H. Adaptation level theory: An experimental andsystematic approach to behavior. New York: Harper& Row, 1964.

Herzberg, F., Mausner, B., & Snyderman, B. B. Themotivation to work. New York: Wiley, 1959.

Hoffman, H. S., & Solomon, R. L. The opponentprocess theory of motivation: III. Some affectivedynamics in imprinting. Learning and Motivation,1974, 5, 149-164.

Jaques, E. Equitable payment. New York: Wiley, 1961.Kahn, R. L. Conflict, ambiguity, and overload: Three

elements in job stress. Occupational Mental Health,1973, 3, 2-9.

Katzell, R. Personal values, job satisfaction, and jobbehavior. In H. Borow (Ed.), Man in a world at work.Boston: Houghton-Mifflin, 1964.

Lawler, E. E. Motivation in work organizations.Monterey, Calif.: Brooks/Cole, 1973.

Likert, R. New patterns in management. New York:McGraw-Hill, 1961.

Locke, E. The relationship of task success to taskliking and satisfaction. Journal of Applied Psychology,1965, 49, 379-385.

Locke, E. A. The nature and causes of job satisfaction.In M. Dunnette (Ed.), Handbook of industrial andorganizational psychology. Chicago: Rand McNally,1976.

Locke, E. A., & Bryan, J. F. The directing function ofgoals in task performance. Organizational Behaviorand Human Performance, 1969, 4, 35-42.

Locke, E. A., Cartledge, N., & Knerr, C. S. Studies ofthe relationship between goal setting, satisfaction,and performance. Organizational Behavior and HumanPerformance, 1970, 5, 135-139.

Maslow, A. A theory of human motivation. Psycholog-' ical Review, 1943, 50, 370-396.

Page 15: An opponent process theory of job satisfaction

OPPONENT PROCESS 547

McClelland, D. C. The achieving society. Princeton,N.J.: Van Nostrand, 1961.

McGrath, J. Social and psychological factors in stress.New York: Holt, Rinehart & Winston, 1970,

Miller, N. E. Experiments on motivation: Studiescombining psychological, physiological, and phar-macological techniques. Science, 19S7, 126', 1271-1278.

Morse, N. C. Satisfactions in the white collar job. AnnArbor: University of Michigan Survey ResearchCenter, 1953.

Olds, J. Hypothalamic substrates of reward. Physiolog-ical Review, 1962, 42, 554-604.

Porter, L. W., & Lawler, E. E. Managerial attitudes andperformance. Homewood, 111.: Irwin-Dorsey, 1968.

Roethlisberger, F. J., & Dixon, W. J. Management andthe worker. Cambridge, Mass.: Harvard UniversityPress, 1939.

Schaffer, R. H. Job satisfaction as related to needsatisfaction in work. Psychological Monographs,1953, <J7(14, Whole No. 304).

Scott, W. E. Activation theory and task design.Organizational Behavior and Human Performance,1966, ;, 3-30.

Smith, P. C., Kendall, L. M., & Hulin, C. L. Themeasurement of satisfaction in work and retirement.Chicago: Rand McNally, 1969.

Solomon, R. L., & Corbit, J. D. An opponent processtheory of motivation: II. Cigarette addiction.Journal of Abnormal Psychology, 1973, 81, 158-171.

Solomon, R. L., & Corbit, J. D. An opponent processtheory of motivation: I. Temporal dynamics of affect.Psychological Review, 1974,81, 119-145.

Stevens, S. S. On the psychophysical law. PsychologicalReview, 1957, 64, 153-181.

Taylor, F. W. Principles of scientific management.New York: Harper & Row, 1947.

Terrace, H. S. Stimulus control. In W. H. Honig (Ed.),Operant behavior: areas of research and application.New York: Appleton-Century-Crofts, 1966.

Vroom, V. Work and motivation. New York: Wiley,1964.

Zedeck, S., & Smith, P. C. A psychophysical determina-tion of equitable payment. Journal of AppliedPsychology, 1968, 52, 343-347.

Received November 28, 1977 •