after the pink slip: applying dynamic ......after the pink slip: applying dynamic motivation...

AFTER THE PINK SLIP: APPLYING DYNAMIC MOTIVATIONFRAMEWORKS TO THE JOB SEARCH EXPERIENCE

CONNIE R. WANBERGUniversity of Minnesota

JING ZHUHong Kong University of Science and Technology

RUTH KANFERGeorgia Institute of Technology

ZHEN ZHANGArizona State University

We propose and examine a self-regulatory framework focused on understanding thedynamics of job search intensity and mental health over the first several months ofunemployment. We use a repeated-measures design, surveying newly unemployedindividuals weekly for 20 weeks. Through the lens of our framework, we test relation-ships pertaining to the role of motivational “traits” (i.e., temporally stable approachand avoidance motivations) and self-regulatory “states” (i.e., more transient motiva-tion control and self-defeating cognition) in predicting job seekers’ search intensity andmental health over the duration of our study. The findings provide evidence on thedynamics of the job search journey.

In late 2008, the U.S. and much of the industri-alized world entered a recession, resulting in sig-nificant job loss around the globe. For most indi-viduals, job loss is a highly negative event,signaling the loss of a steady income and an end tothe routine, identity, and daily social connectionsassociated with employment. Because of the impor-tance of this issue to individuals and society, theexperience of job loss and subsequent search foremployment has attracted steady research attentiondating back to the Great Depression (Hanisch, 1999;Saks, 2005).

From a micro or individual-level perspective,two of the most extensively studied constructs ex-amined in this literature have been job search in-tensity and job seeker mental health. Job searchintensity refers to the amount of time or effort in-dividuals spend on their job search. Search inten-sity is an essential behavior for most individualswho wish to find work (Prussia, Fugate, & Kinicki,2001). Meta-analytic findings show that individu-als who allocate more time to job search find workmore quickly (Kanfer, Wanberg, & Kantrowitz,2001). Mental health refers to individuals’ psycho-

logical distress and well-being (Ware, Manning,Duan, Wells, & Newhouse, 1984). Meta-analyticfindings suggest that the experience of unemploy-ment is associated with poor mental health andstress-related physical symptoms such as head-aches and stomachaches (McKee-Ryan, Song, Wan-berg, & Kinicki, 2005; Paul & Moser, 2009).

Significant research attention has been devotedto understanding individual differences in effortdevoted to job search during unemployment, aswell as why some individuals have a more negativeexperience with job loss than others (e.g., Creed &Bartrum, 2006; Kanfer et al., 2001; McKee-Ryan etal., 2005). As a recent part of this research stream,self-regulation theories, which address the extentto which individuals are able to successfully mod-ulate their emotions, attention, effort, and perfor-mance during goal-directed activity, have emergedas having promising potential to help deepen un-derstanding of the experience of unemployment.

Like many other common issues individualsstruggle with, such as overeating, procrastination,or smoking cessation, the unemployment experi-ence is inherently a self-regulatory process. In ad-dition to requiring goal setting and attention to thedirection of one’s efforts, issues that are outside thefocus of our study, being unemployed demandsself-regulation to sustain job search effort and man-age negative emotions over time. As delineated by

Editor’s note: The manuscript for this article was ac-cepted for publication during the term of AMJ’s previouseditor-in-chief, R. Duane Ireland.

� Academy of Management Journal2012, Vol. 55, No. 2, 261–284.http://dx.doi.org/10.5465/amj.2010.0157

261

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

Kanfer and colleagues (2001), looking for a job is anunfolding task that is highly autonomous, self-or-ganized, loosely structured, and ill-defined. Indi-viduals must decide on their own how and howoften to search, and they rarely receive feedbackabout the effectiveness of the job search activitiesand strategies they are using. A myriad of issueschallenge job seeker psychological well-being, suchas low opportunities for control and skill use, re-duced daily variety and opportunity for social con-tact, shortage of money, and loss of a valued socialposition (Jahoda, 1987; Warr, 1987). In this context,emotion regulation is required to manage feelingsof discouragement and frustration, fears about fi-nancial challenges, and embarrassment about beingunemployed. Motivation to persist is needed tosustain search efforts despite possible distractionsat home, multiple rejections, or disliking the searchprocess (i.e., identifying job openings, revisingone’s résumé, networking, applying for jobs, pre-paring for interviews, and more).

Empirical work applying self-regulation perspec-tives to the unemployment experience, althoughvaluable, has tended to focus on the study of jobsearch behavior and has almost exclusively used aset of measures assessed at one time point to pre-dict between-individual differences in search be-havior at a later time point (e.g., Creed, King, Hood,& McKenzie, 2009; van Hooft, Born, Taris, Flier, &Blonk, 2005; Van Hoye & Saks, 2008; Zikic & Saks,2009). Self-regulatory frameworks are also usefulfor examining how individuals psychologically ad-just and respond to their unemployment experi-ence (Feather & Davenport, 1981; Latack, Kinicki, &Prussia, 1995; Niessen, Heinrichs, & Dorr, 2009). Todate, however, applications of self-regulatory per-spectives to mental health during unemploymenthave been sparse. No current work, furthermore,has applied self-regulatory theory to the examina-tion of job search behavior or mental health overseveral weeks, limiting the types of self-regulationquestions that can be asked.

Our study investigates self-regulatory processesinvolved in unemployment in a dynamic, temporalframework. The incorporation of time opens thedoor to new questions and insight. Steel (2002), forexample, developed an extensive argument for theneed to study the process of job search over time.He distinguished between longitudinal research(i.e., constructs measured at time 1 are used topredict outcomes at time 2) and repeated-measuresresearch (i.e., the same constructs are measured atmultiple time points). He argued that repeated-measures research is essential for understandingphenomena such as job search that involve con-structs that change and evolve over time. Research

measuring focal outcomes such as job search inten-sity or mental health at only one time point isinsufficient for understanding self-regulatory pro-cesses that unfold over time, such as persistence insearch behavior, maintenance of mental health, de-cline (or increase) in job search or mental health,and vacillation in self-regulated job search activi-ties. A few recent studies have shown that time putinto job search and levels of distress during unem-ployment vary from day to day and week to week(Song, Uy, Zhang, & Shi, 2009; Wanberg, Glomb,Song, & Sorenson, 2005; Wanberg, Zhu, & vanHooft, 2010), but researchers know very little aboutthe factors that contribute to changes in job searchand distress levels over the course of unemploy-ment. Static research is also not well suited toexamining process questions. For example, throughwhat dynamic, more transitory, “state-related”mechanisms do stable “trait variables” exert theirimpact on study outcomes? It is furthermore impor-tant (although to date unusual) for studies employ-ing self-regulation theories to incorporate time,since self-regulation theories are by nature dy-namic (Dalal & Hulin, 2008). For many real-worldproblems, studying a person’s motivated behaviorat one point in time is not compelling. The largerinterest is “the manifestation of motivation in somesustained way” (Ployhart, 2008: 54).

In this article, we extend theory and empiricalfindings by proposing and testing a comprehensiveself-regulatory framework focused on understand-ing the dynamics of job search intensity and mentalhealth over the first several months of the unem-ployment experience. We used a repeated-mea-sures design, surveying newly unemployed indi-viduals once per week for either 20 weeks or untilan individual was reemployed. Contributing to the-ory, we use our proposed framework to delineateexpectations about how self-regulatory constructsare relevant in a time-based context. The dynamicnature of the model allows us to propose bothbetween and within-individual relationships. Forexample, we develop arguments about how self-regulatory constructs can help explain (1) interin-dividual differences in levels of job search intensityand mental health over continued weeks of unem-ployment, (2) interindividual differences in de-clines in job search intensity and mental healthover time, and (3) intraindividual differences inself-regulation over time and relationships to jobsearch level and mental health. Empirically, we putforth the most comprehensive repeated-measuresinvestigation within the job loss domain that weknow of to date. Beyond the aforementionedpoints, our data provide practical new informationabout within-person changes in job search and

262 AprilAcademy of Management Journal

mental health over the duration of an unemploy-ment experience and advance understanding ofhow (i.e., through what mechanisms) stable indi-vidual differences exert their influence on the jobsearch experience.

A DYNAMIC SELF-REGULATORYPERSPECTIVE

Our conceptual model, based in extant self-regu-latory theorizing and prior research in motivationalscience and personality psychology, is shown inFigure 1. Our expectations regarding the impor-tance of time in the relationships shown are elabo-rated in the presentation of our hypotheses. Wefirst posit that individual differences in broad andstable motivational traits (specifically, approachand avoidance motivations) exert direct effects onobserved levels and changes in job search intensityand mental health throughout a spell of unemploy-ment. We then propose that these motivation traitsinfluence job search intensity and mental health inpart through two key self-regulatory state variables:

motivational control and self-defeating cognition.We conceptualize these processes as operating dy-namically, in such a way that self-regulatory statesvary as a function of time and exert correspondingtime-based changes and vacillation in search inten-sity and mental health observed during a period ofunemployment. Finally, although the central focusof our study is on understanding the manifestationand dynamics of job search intensity and mentalhealth over the unemployment experience, we alsoexamine two outcomes relevant to reemploymentsuccess: speed of reemployment and number ofinterviews each week.

The Effect of Approach and AvoidanceMotivations on Job Search Intensity

An “achievement goal” refers to a desired, com-petence-related, future end state that guides indi-vidual behavior (see, for example, Elliot &McGregor, 2001; Hulleman, Schrager, Bodmann, &Harackiewicz, 2010). Although such goals havebeen considered from a variety of perspectives that

FIGURE 1Theoretical Framework

EmploymentOutcomes

Speed ofReemployment

Number ofInterviewsEach Week

Job SearchIntensity

Mental Health

Motivational TraitsApproach MotivationAvoidance Motivation

Self-Regulatory States

Search Effort/Personal Well-being

Self-DefeatingCognition

MotivationControl

Repeating Cycle

Baseline (Week t) End of Study

Each Week of Job Search(Week t + 1, week t + 2, week t + 3,

week t + 4, week t + 5, week t + 6 ...)

2012 263Wanberg, Zhu, Kanfer, and Zhang

are still evolving into a cohesive framework (Hul-leman et al., 2010; Payne, Youngcourt, & Beaubien,2007), deeply steeped in the motivational literatureis the recognition that individuals may differ intheir tendency to conceptualize the goals they wishto achieve from either an approach or an avoidanceorientation (Elliot, 2008). Individuals high in ap-proach orientation (related concepts include “pro-motion focus,” “learning orientation,” and “per-sonal mastery”) are posited to engage in goalstriving for the purpose of personal growth anddeveloping competencies. In contrast, individualshigh in avoidance orientation (related concepts in-clude “prevention focus,” “performance orienta-tion,” and “anxiety-related goal orientation”) areposited to strive to avoid failure, preserve re-sources, protect self-concept, prevent emotionaldisruption of action, and fulfill obligations (Elliot &Harackiewicz, 1996; Higgins, 1998; Kanfer &Heggestad, 1997; VandeWalle, 1997). Evidence sug-gests that distinct and independent systems of ap-proach and avoidance exist and that the approachand avoidance motivations capture different as-pects of behavioral, neural, and affect-related pro-cesses (Robinson, Meier, Tamir, Wilkowski, & Ode,2009; Watson, Wiese, Vaidya, & Tellegen, 1999).

In goal situations, approach-oriented individualshave been characterized as focused on possibility,opportunity, challenge, and what they want toachieve. In keeping with this description, researchsuggests that individuals with higher levels of ap-proach-oriented motivational traits demonstrateenhanced performance outcomes in a variety ofgoal-oriented contexts (e.g., Galinsky, Leonardelli,Okhuysen, & Mussweiler, 2005; Hinsz & Jundt,2005; Liberman, Molden, Idson, & Higgins, 2001;Payne et al., 2007; Porath & Bateman, 2006). Forexample, approach motivation was positively re-lated to goal commitment and performance on anidea generation task (Hinsz & Jundt, 2005) and salesperformance in a field context (Porath & Bateman,2006). Specific to the job search context, Creed etal. (2009) found that higher levels of approach ori-entation on the part of job seekers was related tohigher levels of job search intensity measuredfour months later.

In this study, we examined approach orientationas an individual-difference predictor of time spentin job search over several weeks. Because individ-uals with high approach orientation tend to ac-tively pursue their goals and demonstrate higherlevels of goal-related performance, we expect suchindividuals will report higher levels of search in-tensity at the start of their unemployment experi-ence than individuals with lower levels of ap-proach orientation. With respect to what happens

over time, Dweck and Leggett (1988) suggested thatindividuals with higher approach motivation arewell suited to persist in their goal pursuit efforts.Their research suggests that children with high ap-proach orientation who do not immediatelyachieve their goals tend not to feel they are failing.Instead, they view unsolved problems as chal-lenges to be overcome. We expect that, owing tothis constructive interpretation of failure and per-sistence, individuals with higher levels of ap-proach orientation will report higher average levelsof search intensity over the duration of their unem-ployment experience. We propose:

Hypothesis 1a. Approach-oriented trait moti-vation is positively related to higher levels ofjob search intensity at the start of an unem-ployment experience.

Hypothesis 1b. Approach-oriented trait moti-vation is positively related to higher averagelevels of job search intensity over the durationof an unemployment experience.

In a goal situation, individuals high in avoidancemotivation tend to focus heavily on risks of failure,threats to themselves, and what they ought to do orprevent from happening. Research has been some-what inconsistent but has tended to show thatavoidance orientation tends to hinder goal-relatedperformance outcomes. In a meta-analysis, Payneand colleagues (2007) found avoidance perfor-mance goal orientation had estimated true meancorrelations of –.13 with task performance on lab-oratory tasks and –.06 with academic performance.In the context of job search, Creed et al.. (2009),however, found that higher levels of avoidance ori-entation on the part of job seekers were not relatedto levels of job search intensity measuredfour months later. Dweck and Leggett (1988) arguedthat the passage of time during goal pursuit mayhave a special relevance for individuals with highavoidance motivation. Specifically, these authorssuggested that individuals with high avoidancemotivation may be able to pursue goals with a highintensity of effort early in the goal pursuit process.However, when faced with failure, these individu-als show a helpless, maladaptive response involv-ing avoiding challenge and deteriorating perfor-mance. In a study with children, for example,individuals with high avoidance motivationshowed a clear decline in their problem-solvingstrategies following failure (Diener & Dweck, 1978).

We propose that over time, job seekers withhigher avoidance motivation will show a decliningtrend (i.e., a negative slope of change) in their jobsearch intensity. We further expect that individuals


with high avoidance motivation will show loweraverage levels of search intensity when consideredover multiple weeks of unemployment than indi-viduals low in avoidance motivation.

Hypothesis 2a. Avoidance-oriented trait moti-vation is negatively associated with the slopeof change in job search intensity over the du-ration of an unemployment experience.

Hypothesis 2b. Avoidance-oriented trait moti-vation is related to lower average levels of jobsearch intensity over the duration of an unem-ployment experience.

The Effects of Approach and AvoidanceMotivations on Mental Health

To date, the impact of motivational traits on af-fect during goal striving has been largely limited tothe study of task-related variables such as perfor-mance satisfaction and self-efficacy. Nonetheless,emerging research suggests that individual differ-ences in motivational traits may exert influence onboth immediate and longer-term affective states,such as mental health (e.g., Carver & White, 1994;Tamir & Diener, 2008). For example, Tamir andDiener (2008) suggested the inherent tendency ofindividuals high in approach motivation to meetgoal-oriented situations with an appetitive ratherthan a defensive posture is conducive to well-be-ing. In contrast, the tendency for individuals highin avoidance motivation to view goals with fear andanxiety and to be sensitive to criticism and riskpropagates low well-being. Strauman (2002) simi-larly proposed that individuals who have difficultypursuing promotion goals (i.e., making good thingshappen, fulfilling personal aspirations) and thosewho emphasize prevention goals (i.e., keeping badthings from happening, emphasizing obligations,and fulfilling duties) are at risk for lower well-being. Although the study outcome was frustration,rather than mental health, Whinghter, Cunning-ham, Wang, and Burnfield (2008) found that highworkload was more associated with frustrationfor individuals with higher avoidance orientationthan it was for individuals with higher approachorientation.

In the emotionally demanding job loss–jobsearch context, it is reasonable to expect that ap-proach and avoidance tendencies may exert influ-ences that are reflected in levels of well-being.First, we expect that individuals with higher levelsof approach orientation, because of their higherlikelihood of positively framing job search as anopportunity for personal growth and career en-

hancement, will report higher levels of well-beingat the start of their unemployment experience aswell as over its duration.

Hypothesis 3a. Approach-oriented trait motiva-tion is related to higher levels of mental health atthe start of an unemployment experience.

Hypothesis 3b. Approach-oriented trait moti-vation is related to higher average levels ofmental health over the duration of an unem-ployment experience.

Our expectation with regard to avoidance moti-vation highlights the temporal nature of our dataset. Specifically, our expectation is that individualswith higher levels of avoidance orientation willshow lower levels of mental health over the courseof unemployment as a consequence of their higherlevels of apprehension, worry, and anxiety relatedto employment loss and reemployment success. Inaddition, these individuals are also expected toshow greater sensitivity to the negative conse-quences of unemployment that accrue over time.With the passage of time, the threat of staying un-employed is likely to be more manifest (McFadyen& Thomas, 1997). Although speculative, given thelack of field research on trait avoidance motivationand mental health over time, our proposition is thatcontinued unemployment among individuals highin avoidance motivation will exacerbate the possi-bility of failure, a sensitivity that tends to bring outincreased anxiety in these individuals (Carver &White, 1994). We thus propose that individualshigh in avoidance motivation will show a declinein mental health over the course of job search.

Hypothesis 4a. Avoidance-oriented trait moti-vation is negatively associated with the slopeof change in mental health over the duration ofan unemployment experience.

Hypothesis 4b. Avoidance-oriented trait moti-vation is related to lower average levels of men-tal health over the duration of an unemploy-ment experience.

Self-Regulatory States as Mediating Processes

Although individual differences in motivationaltraits are purported to directly affect both jobsearch intensity and mental health, we posit thatself-regulatory processes partially mediate this re-lationship (see Figure 1). Specifically, drawing onKanfer and Ackerman (1996) and Kanfer andHeggestad (1997), we suggest the relationship be-tween motivational traits and mental health andsearch intensity can be explained in part through


experienced levels of two fundamental self-regula-tory states: motivation control and self-defeatingcognition. Motivation control refers to the inten-tional cognitive redirection of attention, use of goalsetting, and/or use of environmental managementstrategies to stay on course and sustain effort withrespect to a job search (Kanfer & Heggestad, 1997;Kuhl, 1985). Self-defeating cognition refers to neg-ative, rigid, and dysfunctional “self-talk” (Alloy,Abramson, Grant, & Liu, 2009; Glass & Arnkoff,1997; Kanfer & Ackerman, 1989). In our study, weassess self-talk specific to feelings of hopelessness,giving up, and negative expectations about beingsuccessful in a job search. Motivational control ischaracterized by cognition directed toward increas-ing task effort. In contrast, self-defeating cognitionis a failure in emotional control, characterized bycognition directed toward off-task negative affectand task withdrawal.

Effects of Motivational Traits on Self-RegulatoryStates

Kanfer and Heggestad (1997) argued that motiva-tional traits influence the use of self-regulatorystrategies in specific goal-oriented situations. Thiswork suggests that, when confronted with the needto continue with a difficult problem or task, ap-proach-oriented individuals will tend to engage inactivities that exemplify motivational control, suchas planning or instructing themselves to exert extraeffort. Even when faced with failure, they will tendto direct their attention toward goal pursuit andovercoming the obstacles at hand. For example,when faced with a difficult problem, approach-oriented individuals tend to make remarks such as“The harder it gets, the harder I need to try” (Diener& Dweck (1978: 459). In contrast, when faced witha difficult situation, avoidance-oriented individu-als will tend to display self-defeating cognition anda conviction that they cannot overcome the situa-tion and make remarks such as “I give up” (Diener& Dweck, 1978; Kanfer & Heggestad, 1997). Ratherthan engaging in extra effort when the going getsrough, individuals high in avoidance motivationinstead tend to view challenge as a threat. Consis-tently with these findings, in a recent meta-analysisof related constructs (Payne et al., 2007), a learninggoal orientation was negatively associated withstate anxiety, and an avoidance performance goalorientation was positively associated with stateanxiety (k’s � 16, 7; estimated true mean correla-tions � –.09, –31, respectively). Given this work,we expected to observe the following relation-ships between the trait-based predictors shown

in Figure 1 (approach motivation and avoidancemotivation) and motivational control and self-defeating cognition:

Hypothesis 5a. Higher levels of approach-ori-ented trait motivation are associated with higheraverage levels of motivation control and loweraverage levels of self-defeating cognition over theduration of an unemployment experience.

Hypothesis 5b. Higher levels of avoidance-ori-ented trait motivation are related to lower av-erage levels of motivation control and higheraverage levels of self-defeating cognition overthe duration of an unemployment experience.

Effects of Self-Regulatory States on Job SearchIntensity and Mental Health

Kanfer and Heggestad (1997) further argued thatself-regulatory strategies impact goal-relevant out-comes. In the dynamic unemployment context, weposit that within-individual changes in the use ofself-regulation from week to week will predictchanges in demonstrated search intensity and men-tal health.

First, when an individual exerts motivationalcontrol, it involves purposeful management of theenvironment to stay on task, generate next steps,and direct attention to the task at hand (Kanfer &Heggestad, 1999). Findings in the goal orientationand goal setting literatures (e.g., Kanfer & Acker-man, 1996; Kozlowski & Bell, 2006; Lee, Sheldon, &Turban, 2003) suggest that higher levels of motiva-tion control are associated with higher levels ofperformance on goal-related activities. In the jobsearch context, Wanberg, Kanfer, and Rotundo(1999) showed that motivation control was posi-tively associated with job search intensity both atthe same point in time and three months later.Creed and colleagues (2009) obtained similar find-ings. Although the relationship between motiva-tional control and well-being has not been directlystudied, Kanfer and Ackerman (1996) suggestedthat individuals have limited attentional resourcesand that motivational control strategies will divertattention from feelings of unhappiness or dissatis-faction about performance. In this study, becausemotivational control and our outcomes were mea-sured each week as repeated measures, we wereable to use within-person analyses to examine theextent to which changes in the extent to whichindividuals display motivational control help ex-plain vacillations in their job search intensity andmental health over time. We propose:


Hypothesis 6a. During weeks when individualsexert more motivation control, we expect theirjob search intensity to be higher, compared toweeks when they exert less motivation control.

Hypothesis 6b. During weeks when individualsexert more motivation control, we expect theirmental health to be higher, compared to weekswhen they exert less motivation control.

The effects of self-defeating cognition on effortdevoted to goal attainment have yet to receive sub-stantial research attention. Available work suggeststhat the coactivation of negative cognition or affectwith goal-directed activity will reduce or even stopgoal-directed effort (Aarts, Custers, & Holland,2007). Self-destructive self-talk can be expected toresult in reduced effort when it reflects a lack ofhope (Lopez, Snyder, & Pedrotti, 2003). It further-more reduces cognitive resources needed for per-forming well on the task at hand (Kanfer &Heggestad, 1999). A more substantial literature ex-ists with respect to the relationship between self-defeating cognition and well-being. Dysfunctionalthought processes, including a broad array of neg-ative cognitive styles, negative self-attributions,and self-defeating cognition, have been shown toundermine self-worth (Kuiper, Olinger, & Swallow,1987) and to diminish individual well-being (Judge& Locke, 1993; Petrocelli, Glaser, Calhoun, & Camp-bell, 2001). High levels of self-defeating cognitionspecifically related to lack of hope and low expec-tancy of success are problematic because they lockindividuals into viewing their problems as unsolv-able, a core issue that triggers depression (Beck,1972, 1987). Despite some consistency in individ-ual tendencies to engage in dysfunctional thinking,environmental stressors and events that vary fromweek to week intensify the extent to which indi-viduals engage in self-defeating cognition (Zuroff,Blatt, Sanislow, Bondi, & Pilkonis, 1999). In thisstudy, we propose that variability in job seekerself-defeating cognition from week to week willhelp explain changes in their job search intensityand mental health.

Hypothesis 7a. During weeks when individualsreport higher levels of self-defeating cognition,job search intensity is lower than in weekswhen they report lower levels of self-defeatingcognition.

Hypothesis 7b. During weeks when individualsreport higher levels of self-defeating cognition,mental health is lower than in weeks when theyreport lower levels of self-defeating cognition.

Effects of Motivational Traits on Self-RegulatoryStates Affecting Job Search Intensity and MentalHealth

Although motivational traits may exert some di-rect effects on goal outcomes, theorizing by Kanferand colleagues suggests that the impact of traits onbehavior occurs in part through the influence thattraits have on the initiation and execution of self-regulatory processes that govern the direction andintensity of action. We propose that the relation-ship between the broad motivational traits in-cluded in our model (approach and avoidance mo-tivation) and job search intensity and mental healthare partially mediated by context-specific self-reg-ulatory states (specifically, motivation control andself-defeating cognition).

Hypothesis 8. The relationship between ap-proach and avoidance motivation and searchintensity and mental health are partially me-diated by motivation control and self-defeatingcognition.

Effects of Job Search Intensity and Mental Healthon Job Search Success

Although our hypotheses for the most part focuson job search intensity and mental health, we alsoassess two important distal job search outcomes ofour participants, reemployment speed and numberof interviews. Search intensity has been studied asa predictor of job search success more frequentlythan has mental health. A meta-analysis by Kanferand colleagues (2001) suggests job search intensityis a positive predictor of later employment status(rc � .21, k � 21) and number of offers (rc � .28,k � 11), and a negative predictor of unemploymentduration (rc � –.14, k � 9). The more intensive aperson’s search behaviors are and the more time isspent on search, the more employers potentiallyhave the job seeker’s information brought to theirattention. On the basis of this literature, we expectthat search intensity is positively related to successin getting interviews as well as to reemploy-ment speed.

Hypothesis 9. Job search intensity at the startand over the duration of an unemploymentexperience is positively related to success ingetting interviews and reemployment speed.

Although a wealth of studies have examined howunemployment affects mental health, surprisinglyfew have addressed mental health as a predictor ofjob search success. Authors have argued that lowmental health slows reemployment because de-pressed mood deprives job seekers of the mental


and physical energy needed to engage in an effec-tive job search (Viinamäki, Koskela, & Nishanen,1996). At least one study, however, has shown thatunemployed individuals with higher depressionwere reemployed more quickly (Kessler, Turner, &House, 1988). These authors argued that stress maymotivate individuals to find work faster. One of themost extensive studies of the relationship betweenwell-being during unemployment and the probabil-ity of reemployment is by Ginexi, Howe, and Cap-lan (2000). The authors assessed the mental healthof 254 unemployed individuals three times, firstwithin 49 days of job loss, then 5 and 11 monthslater. Their times 1 and 2 measures of depressionwere not predictive of time to reemployment. Arecent meta-analysis compared the mental healthlevels of continuously unemployed individualswith the mental health levels of individuals whofound jobs during the study reporting periods (Paul& Moser, 2009). Although effect sizes were small,continuously unemployed individuals had lowerlevels of mental health than individuals who foundjobs (k � 49, n � 13,259, D � .15).

Overall, the literature regarding mental health asa predictor of reemployment speed is somewhatunclear. Because few previous studies (Ginexi et al.[2000] is an exception) have assessed mental healthearly in an unemployment experience and also re-peated those assessments over its duration, we sug-gest our study has the potential to contributeunique insight into mental health as a predictor ofjob search success. Specifically, our study allowsusing both baseline and aggregate levels of mentalhealth measured over multiple weeks (thus remov-ing some of the issues related to self-report biasfrom one-time assessments and also providing acollective assessment over time). Given the resultsof the Paul and Moser (2009) meta-analysis, wepropose a positive relationship between mentalhealth and search success:

Hypothesis 10. Mental health at the start andover the duration of an unemployment experi-ence is positively related to success in gettinginterviews and reemployment speed.

METHODS

Participants

Participants were drawn from a pool of unem-ployment insurance recipients identified by theU.S. Department of Employment and Economic De-velopment as having been unemployed for 3 weeksor less and eligible for a full-duration claim (i.e., 25or 26 weeks) of unemployment insurance. Wesought out individuals unemployed for 3 weeks or

less because our aim was to get individuals into thestudy as quickly after job loss as possible (to reduceproblems of “left-censoring” and allow a logicalstarting point for examination of changes in the jobsearch experience over time). In the United States,unemployment insurance is available to individu-als who lose their jobs through no fault of their own(i.e., they did not quit and were not fired for mis-conduct); individuals also have to be available forwork and seeking full-time employment. Individu-als eligible for fewer than 25 weeks of unemploy-ment insurance tend to have more intermittent con-nections with the workplace.

To preclude the possibility of potential influ-ences associated with recent unemployment expe-rience or age-related differences in job search andreemployment likelihood, we required that poten-tial participants be between the ages of 25 and 50and have had no unemployment insurance claim inthe last four years. In addition, to reduce the impactof occupational level, we required potential partic-ipants to have at least a bachelor’s degree. Accord-ing to U.S. Department of Labor statistics, the sea-sonally adjusted unemployment rate in the state inwhich we conducted the study ranged from 4.7 to5.4 percent over its duration (from the end of Jan-uary 2008 to the beginning of July 2008).

Our study design required individuals to com-plete a weekly online survey for 20 weeks or untilthey found reemployment, whichever came first.This time period was chosen to exceed the mediantime reported by the Bureau of Labor Statistics forweeks unemployed (8.9 weeks for the study period[Bureau of Labor Statistics, 2009]). Our chosen timeperiod also allowed us to follow most of the indi-viduals in our sample throughout the duration oftheir unemployment experience.

Study invitations were sent by mail, and individ-uals were asked to enroll in the study by visitingthe study’s website. We recognized the seriouschallenge posed by sending individuals a “cold”invitation in the mail for a study that asked them tocomplete a survey once a week for 20 weeks. Toenhance our response rate, we enclosed a profes-sionally printed brochure with clear informationabout the study requirements in our invitation. Weoffered individuals a $20 incentive to enroll in thestudy and complete the baseline survey withinone week. Individuals were offered an additional$20 to complete the second weekly survey and anadditional $75 again if they completed at least 16 ofthe 20 surveys involved in the project.

A total of 508 individuals were invited to be inthe study. Of these, 182 enrolled and completedbaseline surveys (36%). Participants were com-pared with invited nonparticipants using database


elements available from the state. Participantsdid not differ from nonparticipants in terms of theirmaximum allowed unemployment insuranceamount (t � –0.68, p � .50), their unemploymentinsurance account balance at the end of the study(t � –1.40, p � .17), whether they had exhaustedtheir unemployment insurance at the end of thestudy (t � 0.47, p � .64), proportion of females(t � 1.15, p � .26), age (t � –0.26, p � .80), educa-tion level (t � –0.41, p � .72), or length of unem-ployment up to the start of the study (t � –0.98,p � .33). However, the proportion of white individ-uals among respondents was higher than that in thepool of potential participants (t � 2.73, p � .01).

Individuals enrolled in the study were sent on-line surveys every week for the 20 weeks followingthe baseline survey. An e-mail with a survey linkwas sent at noon each Friday. Individuals wereasked to complete the survey by Sunday evening ofeach week to ensure current reflection on the lastweek’s job search as well as equal intervals be-tween surveys. Response rates to the weekly sur-veys ranged from 73 to 95 percent.

A total of 177 individuals completed at least thebaseline survey and one weekly survey; these indi-viduals constituted the sample for this study. Ofthe 177 individuals, the average age was 37 years(s.d. � 7.52). Forty percent of the sample memberswere female, and 93.8 percent were white. On av-erage, individuals in the final sample had beenunemployed for 28 days (s.d. � 10.5) at the timethey enrolled in the study. Approximately 60 per-cent of the sample reported their last job had beenin a professional occupation; 19 percent reportedclerical or sales-related fields, and 21 percent re-ported other job types.

A total of 128 individuals became reemployedduring the duration of the study (n � 35 in week 1;9 in week 2; 2 in week 3; 2 in week 4; 15 in week 5;2 in week 6; 4 in week 7; 6 in week 8; 5 in week 9;6 in week 10; 5 in week 11; 6 in week 12; 3 in week13; 4 in week 14; 5 in week 15; 5 in week 16; 2 inweek 17; 4 in week 18; 5 in week 19; and 3 in week20). These individuals were included in the studyup to the point of reemployment (i.e., their datacontribute to the analyses up to the week theyfound a job).

Measures

Demographic characteristics, control variables,and motivational traits were assessed in the base-line survey. Self-regulatory states, job search inten-sity, and mental health were assessed in everyweekly survey during the period when respondentswere looking for a job.

Baseline survey measures. Approach andavoidance motivation were measured using thepersonal mastery (16 items) and motivation relatedto anxiety (19 items) subscales from the Motiva-tional Trait Questionnaire (MTQ; Heggestad & Kan-fer, 2000; Kanfer & Ackerman, 2000). Respondentswere asked to indicate the extent to which eachitem describes them (1 � “very untrue of me,” and6 � “very true of me”). Example items for personalmastery include “When I become interested insomething, I try to learn as much about it as I can”and “If I already do something well, I don’t see theneed to challenge myself to do better” (reverse-coded). The coefficient alpha for this scale was .90.Example items for motivation related to anxietyinclude “When working on important projects, Iam constantly fearful that I will make a mistake”and “I do not get nervous in achievement set-tings” (reverse-coded). The coefficient alpha was.93. Heggestad and Kanfer (2000) and Kanfer andAckerman (2000) provide evidence supportingthe discriminant validity of these two motiva-tional traits from each other and other personal-ity measures.

Age in years, education level (0 � “bachelor’sdegree,” 1 � “master’s degree or above”), gender (0� “male,” 1 � “female”), and ethnicity (0 � “non-white” and 1 � “white”) were included as controlvariables. A methodological contribution of ourstudy is that we followed individuals from the startof their unemployment; the significant majority ofexisting research includes individuals at many dif-ferent stages of their job search (Steel, 2002). Owingto some variability in time enrolled in the study,however, number of days unemployed at the timeof baseline was controlled. To account at least at acoarse level for differences in labor market demandas well as possible differences in the nature of jobsearch, occupation (three categories: professional,technical, and managerial; clerical and sales; andothers) was also controlled. Finally, one item wasused to control for individual differences in em-ployment commitment (Rowley & Feather, 1987),although all individuals in our sample indicatedthey were engaged in active work search. The item,“Having a job is very important to me,” was ratedon a scale ranging from 1 (“strongly disagree”) to 5(“strongly agree”).

Dynamic survey measures. Motivation controlwas assessed with four items developed to examineindividuals’ intentional, cognitive redirection of at-tention toward their job search, use of goal setting,and/or environmental management of their jobsearch over the previous week. Individuals wereasked to what extent four statements were descrip-tive of their past week with regard to their job


search (“Despite difficulties that passed my waythis week, I was able to stay focused on my jobsearch”; “If I got interrupted, I worked hard to getback on track”; “I mentally pushed myself to workharder on my job search”; “I boosted my motivationto look for a job”). Responses to the items wereprovided on a scale ranging from 1 (“not at all trueof me”) to 5 (“very true of me”). We wrote newitems rather than use an existing measure for twoprimary reasons. First, previous assessment (e.g.,Creed et al., 2009; Wanberg et al., 1999) has been ata trait level (e.g., “In general, do you have motiva-tion control skills?”) rather than at a state level; weassess the extent to which motivation control wasused each week. Second, the trait-based itemswere not easily transferrable to a state level becausethey referred to specific event-based behavioral ex-amples that might not occur every week in a jobsearch (e.g., “I practice my conversations with po-tential employers ahead of time”). The coefficientalphas for the motivation control scale ranged from.84 to .92 over the 20-week study span; we inter-wove items with the self-defeating cognition itemsto reduce the possibility of high internal consis-tency simply due to response bias.

Self-defeating cognition was assessed with sixitems that asked individuals about the occurrenceof negative cognition over the past week specific tofeelings of hopelessness, giving up, and negativeexpectations related to their job search. We againdeveloped items specific to the job search context.We patterned the item content from the negativeexpectations and giving up/helpless dimensions ofthe Automatic Thoughts Questionnaire (Hollon &Kendall, 1980) and the Hopelessness Scale (Beck,Weissman, Lester, & Trexler, 1974), ensuring ouritems were broad enough to capture thoughts thatparticipants might actually have weekly (Glass &Arnkoff, 1997). Specific items included “I thoughtabout simply giving up on job search”; “I felt like Icouldn’t continue to do this anymore”; “I thoughtabout how much more job search I could tolerate”;“I thought about how hopeless it was to look for anew job”; “It crossed my mind that I would neverfind a new job”; and “It occurred to me that all myefforts to get a job weren’t worth the trouble” [1,“not at all true of me,” to 5, “very true of me”]). Weaveraged the scores of the items to measure self-defeating cognition for each week (� � .88–.97 overthe span of the study). Higher scores indicate moreself-defeating cognition.

Job search intensity was assessed by asking indi-viduals the following question in each weekly sur-vey: “How many hours did you spend on your jobsearch each day this week?” Participants wereasked to write down the number of hours for each

day of the week. We summed these reported num-bers to get the measure of hours spent in search forthe particular week. Hours spent in job search hasbeen shown to be highly correlated (r � .56–.68)with alternative measures (e.g., multiple-item as-sessments) of job search intensity (Wanberg etal., 2005).

Mental health was measured each week over theduration of the study using the five-item MentalHealth Inventory (MHI; Berwick, Murphy, Gold-man, Ware, Barsky, & Weinstein, 1991), a shortversion of the 38-item MHI (Veit & Ware, 1983).The inventory taps psychological distress (anxietyand depression) and psychological well-being (gen-eral affect). Respondents were asked to respond toeach item with respect to the past week using thescale 1, “none of the time,” to 6, “all of the time.”Items include “Have you felt downhearted andblue?” and “Have you been a very nervous person?”The items were reverse-coded and averaged so thathigher scores indicate better mental health (� �.85–.93 over the span of the study). The MHI-5 ishighly correlated with other well-known measuresof mental health such as the GHQ, and research hassupported its ability to distinguish between groupswith health differences (e.g., McCabe, Thomas, Bra-zier, & Coleman, 1996).

Indicators of search success. Length of unem-ployment was used in survival analysis (with Coxregression) to examine reemployment speed. Fromstate records, we obtained each job seeker’s last dayon the previous job. Those who were reemployedreported their starting dates for the new jobs. Wecalculated the difference between these two days toget the length of unemployment for those reem-ployed. For those who remained unemployed, thelength of unemployment was “right-censored” atthe end of the study. At that time, 27.7 percent ofour sample was still unemployed (n � 49), and 72.3percent had found jobs (n � 128).

Number of interviews was calculated as the totalnumber of interviews for an average week during aperson’s unemployment experience. Individualsgave their answers to the question “How manyinterviews did you have this past week?” in theweekly survey.

Confirmatory Factor Analysis

We conducted several sets of confirmatory factoranalyses (CFAs) to establish the distinctiveness ofour measures. These CFAs supported the use of thescales as distinct measures. First, self-defeatingcognition refers to negative self-talk and is distinctconceptually from mental health. However, to es-tablish empirically that these variables are distinct,


we entered the items for the MHI and self-defeatingcognition into a CFA with separate but related fac-tors. The fit indexes indicated a good fit for thetwo-factor model (�2 � 75.57, df � 41, p � .001,CFI � .96, RMSEA � .08). Combining the two intoone construct produced a significantly inferior fit(�2 � 261.77, df � 42, p � .001, CFI � .73, RM-SEA � .20). These CFA results were based on datafrom the first week; similar results were obtainedfrom subsequent weeks.

More broadly, we also completed a CFA of all ofthe multiple-item self-report measures in our study(approach motivation, avoidance motivation, moti-vation control, self-defeating cognition, and mentalhealth). To achieve a sufficient ratio between sam-ple size and the number of estimated parameters(Bentler & Chou, 1987), we generated item parcelsinstead of using individual items for the approachand avoidance scales; we used individual items forthe other scales. The fit indexes indicated a good fitfor the five-factor model (�2 � 343.69, df � 220,p � .001, CFI � .94, RMSEA � .06). Combining theitems into competing models of two, three, orfour factors produced a significantly inferior fit(p � .001).

Analyses

The survey questions were set to require a re-sponse to all questions. However, participants wereable to skip a week’s survey, which would in turnlead to missing data for that week. We examinedthe electronic time stamps associated with all sur-veys to ensure they were submitted within the ap-propriate time frame (i.e., between Friday noon andSunday midnight). Surveys containing electronictime stamps outside of the acceptable time framewere not used.

The data set has a hierarchical structure, inwhich repeated measures (level 1) were nestedwithin individuals (level 2). For hypotheses inwhich the dependent variable is not job searchintensity (i.e., Hypotheses 3a–5b, 6b, 7b, and part ofHypothesis 8), we used linear mixed models, asimplemented by SAS “proc mixed” (Fitzmaurice,Laird, & Ware, 2004), to test the within-individualrelationships among approach and avoidance mo-tivation, motivation control and self-defeating cog-nition, and mental health. For hypotheses in whichjob search intensity is the dependent variable (i.e.,Hypotheses 1a, 1b, 2a, 2b, 6a, 7a, and part of Hy-pothesis 8), because job search intensity is a countvariable (e.g., number of hours) and its distributionwas positively skewed, generalized linear models,as implemented by SAS “proc genmod” with a

Poisson distribution and a repeated function,were used.

In the analyses of dynamic variables, we in-cluded a linear term (i.e., week, coded 0 for the firstweek, 1 for the second week, and so forth) and aquadratic term for time (i.e., the squared term ofweek). The inclusion of week and week squaredserved three purposes. First, the coefficients of thelinear and quadratic terms allowed us to estimatethe growth trends of the dynamic variables (e.g., jobsearch intensity and mental health) over time. Sec-ond, for testing Hypotheses 1a and 3a, the inclusionof week and week squared in their current codingtheme allow us to predict the relationships be-tween motivational traits and the dynamic depen-dent variables at the start of the job search duration.In contrast, when week and week squared wereexcluded from the model, motivational traits werepredicting the average or overall levels of the dy-namic dependent variables over the unemploymentexperience, as in the testing of Hypotheses 1b, 2b,3b, 4b, 5a, and 5b. The third purpose of includingweek and week squared, pertinent to the testing ofHypotheses 2a and 4a, lay in the fact that the coef-ficients of the linear and quadratic term can bepredicted by individual-level (level 2) variables.Thus, we could show that a motivational trait wasrelated to the decline or increase of a dynamicdependent variable over time.

Hypotheses 6a through 8 use dynamic predictors,that is, motivation control and self-defeating cogni-tion each week, to predict that week’s job searchintensity and mental health. In testing these level 1relationships, we included week and week squaredas control variables to partial out any associationbetween the independent and dependent variablesdue to the growth trends of the dependent vari-ables. For Hypotheses 9 and 10, we used Cox re-gression (i.e., proportional hazard model [Klein &Moeschberger, 1997]) and ordinary least squares(OLS) regression to predict the speed of reemploy-ment and number of interviews. OLS regression wasused because the number of interviews (per week)follows an approximately normal distribution.

RESULTS

Descriptives

Table 1 portrays descriptive statistics and corre-lations for baseline variables (variables 1–11), re-peated measures (variables 12–15), and distal out-come measures (variables 16–17). We have 20 timewaves of data for variables 12–15, but as this is anextensive amount of data, we only present correla-tions for these variables based on the measures


TA

BL

E1

Des

crip

tive

Sta

tist

ics

and

Cor

rela

tion

sof

Sel

ecte

dS

tud

yV

aria

bles

a

Var

iabl

ebn

Mea

ns.

d.

12

34

56

78

910

1112

1314

1516

1.A

ge17

737

.02

7.52

2.E

du

cati

on17

70.

110.

32.1

7*3.

Gen

der

177

0.40

0.49

�.1

2.0

04.

Eth

nic

ity

177

0.94

0.24

.02

.02

.02

5.D

ays

un

emp

loye

du

pto

stu

dy

star

t17

728

.02

10.5

0.0

5.1

6*�

.13

.05

6.O

ccu

pat

ion

–Pro

fess

ion

al17

70.

600.

49.1

0.1

8*.1

3�

.02

�.1

87.

Occ

up

atio

n–C

leri

cal

177

0.19

0.39

.00

�.0

8.1

1.0

0.0

3�

.59*

*8.

Occ

up

atio

n–O

ther

177

0.21

0.41

�.1

2�

.14

�.2

6**

.02

.18*

�.6

4**

�.2

5**

9.E

mp

loym

ent

com

mit

men

t17

74.

481.

17.0

1.0

8.0

6�

.09

.05

.02

.06

�.0

910

.A

pp

roac

h-o

rien

ted

mot

ivat

ion

177

4.69

0.06

�.0

5�

.09

.11

.14

.01

.00

.03

�.0

2.0

2(.

90)

11.

Avo

idan

ce-o

rien

ted

mot

ivat

ion

177

3.30

0.84

.04

.01

.20*

*.0

2.0

3�

.01

.08

�.0

6.0

8–.

29**

(.93

)12

.W

eek

1m

otiv

atio

nco

ntr

ol12

93.

330.

85�

.06

.00

�.1

2�

.09

.03

.03

�.0

7.0

3.1

6.4

0**

–.26

**(.

84)

13.

Wee

k1

self

-def

eati

ng

cogn

itio

n12

91.

580.

72�

.05

�.0

4.1

6.0

2.1

0�

.03

�.0

5.1

0�

.02

–.20

*.3

7**

–.36

**(.

88)

14.

Wee

k1

job

sear

chin

ten

sity

129

17.8

012

.85

.12

.04

�.1

7�

.13

�.0

1.2

4**

�.1

0�

.20*

.20*

.09

–.06

.48*

*–.

0815

.W

eek

1m

enta

lh

ealt

h12

94.

390.

98�

.04

.12

�.0

2�

.04

.00

.06

�.0

4�

.04

.01

.21*

–.48

**.3

0**

–.43

**–.

11(.

89)

16.

Nu

mbe

rof

inte

rvie

ws

132

1.93

0.93

.10

.15

�.0

7.0

0.0

6.0

6.0

2�

.10

.10

.10

–.22

*.3

6**

–.01

7.4

5**

.13

17.

Ree

mp

loye

dd

uri

ng

the

stu

dy

177

0.72

0.45

�.2

1**

�.0

6�

.04

.00

.00

.06

�.1

6*.0

8.0

8.2

3**

–.17

*.1

8*.0

7.1

1–.

03.2

5**

aV

aria

bles

1–11

wer

eas

sess

edin

the

base

lin

esu

rvey

(n�

177)

.V

aria

bles

12–1

5w

ere

asse

ssed

each

wee

kfo

ra

tota

lof

20m

easu

rem

ents

;on

lyth

ew

eek

1m

easu

res

are

incl

ud

edh

ere.

En

trie

son

the

dia

gon

alin

par

enth

eses

are

alp

has

for

the

list

edva

riab

les.

bE

du

cati

onw

asco

ded

0�

“bac

hel

or’s

orbe

low

,”1

�“m

aste

r’s

orab

ove”

;ge

nd

er,

0�

“mal

e,”

1�

“fem

ale”

;et

hn

icit

y,0

�“n

onw

hit

e,”

1�

“wh

ite.

”Jo

bse

arch

inte

nsi

tyw

asm

easu

red

asn

um

ber

ofh

ours

spen

tin

job

sear

chp

erw

eek.

Nu

mbe

rof

inte

rvie

ws

isth

eav

erag

en

um

ber

ofin

terv

iew

sin

div

idu

als

had

per

wee

kd

uri

ng

un

emp

loym

ent.

Ree

mp

loye

dd

uri

ng

the

stu

dy,

0�

“no,

”1

�“y

es.”

*p

�.0

5**

p�

.01

assessed in the week 1 survey for illustration pur-poses. Before testing the hypotheses, we examinedwhether systematic within- and between-individ-ual variance existed in the repeated-measures vari-ables by running a series of null (intercept-only)models. The analyses supported using hierarchicallinear modeling (HLM) on these data, as there wassufficient within-individual and between-individ-uals variance in the measures over time. For exam-ple, within-individual variance was 53 percent formotivation control, 41 percent for self-defeatingcognition, 22 percent for job search hours, and 26percent for mental health. Although not reported inTable 1, the average number of hours spent in jobsearch per week over the 20 weeks of our study was14.5 hours, ranging from an average of 11.2–17.8.

We examined the trends of our repeated mea-sures with unconditional HLM (only including theintercept, linear term and quadratic term of time) toportray the growth patterns of these variables overtime. Table 2 shows the results of these models.The intercept, linear (week), and quadratic (weeksquared) coefficients are significant for all vari-ables, indicating that there was an average within-person change on these variables over time. For jobsearch intensity, these data show that individualsin our study reported spending less time in jobsearch as their unemployment spell continued,with a slight uptick in later weeks. For mentalhealth, there was an improvement over time, with aslight downturn in later weeks (see Figure 2). Thetrends shown in Figure 2 must be interpreted withthe caveat that, over time, some individuals in the

sample are becoming reemployed. All participantsare represented in the data up to the point at whichthey became reemployed. This means that the datapoints for early weeks are based on a larger sample.

The trends observed across individuals variedconsiderably, in particular with regard to the linearslope coefficients, meaning that changes in jobsearch and mental health are not the same for ev-eryone over time. We model these changes depen-dent on our focal study variables in the next sec-tions. As the coefficients of the quadratic term ofjob search intensity and mental health were ex-tremely small (see Table 2), to obtain parsimoniousmodels we did not use individual-level (level 2)variables to predict these quadratic term coeffi-cients. Thus, in testing Hypotheses 2a and 4a, weused motivational traits to predict the coefficient ofthe linear term only (i.e., the interaction terms be-tween motivational traits and week).

Motivational Traits Affecting Search Intensity

Hypothesis 1a, suggesting that approach-orientedtrait motivation is related to more time spent in jobsearch at the start of an unemployment experience,was supported (� � .21, model 1a, Table 3). Sup-porting Hypothesis 1b, individuals with higher lev-els of approach-oriented trait motivation also re-ported more job search hours over the duration oftheir unemployment spell. As shown in model 1b,Table 3, a one-point increase in approach-orientedmotivation was associated with a .25 times (� e(.22)

–1) increase in average job search hours per week

TABLE 2Generalized Linear Model and Hierarchical Linear Modeling Results for Linear and Quadratic Coefficients of

Repeated Measuresa

Variables

Intercept �00 Linear Term of Time �10 Quadratic Term of Time �20

CoefficientBetween-Individual

Variances CoefficientBetween-Individual

Variances CoefficientBetween-Individual

Variances

Job search intensity 0.95** n.a.b –0.04** n.a. 1.70E-3** n.a.Mental health 4.45** 0.69** 0.03* 5.25E-3** –1.40E-3* 1.50E-5**Motivation control 3.29** 0.28** –0.04** 4.80E-5 2.27E-3** 4.76E-6Self-defeating cognition 1.60** 0.26** 0.06** 8.13E-3** –1.88E-3* 2.50E-5**

a The linear term of time was coded 0 for week 1, 1 for week 2, and so forth. Generalized linear model (SAS “proc genmod”) with aPoisson distribution and AR (1) variance-covariance structure (i.e., first-order autoregressive) correlation matrix structure was used to testthe unconditional model for job search intensity. Linear mixed model (SAS “proc mixed”) with the RE-AR (1) variance–covariancestructure (i.e., random effects plus auto-regression [1]) structure was used to test the unconditional model for all other variables.

b Between-individual variances were not available for job search intensity with the generalized linear model (SAS “proc genmod”)because of its “population-average” feature. Instead, we tested the between-individual variances (i.e., random effects) for job searchintensity with a linear mixed model (SAS “proc mixed”). The variance of the intercept was significant (variance � 134.50, p � .01); thevariance of the linear term of time was significant (variance � 0.6040, p � .05); and the variance of the quadratic term of time wasnonsignificant.

* p � .05** p � .01


over the unemployment duration. Stated anotherway, given a one-point increase in approach-ori-ented motivation, average job search hours wouldbe increased from 1 to 1.25 hours, 2 to 2.5 hours, 10to 12.5 hours, and so on, each week. Figure 3 isprovided to illustrate the job search hours trendover time of people who were high (in the top third)versus people who were low (in the bottom third)on approach motivation. The figure helps portrayhow approach motivation makes a differenceover time.

Hypothesis 2a indicates that avoidance-orientedtrait motivation will be negatively associated withthe slope of job search intensity over time. Model2a of Table 3 shows that the interaction term ofavoidance-oriented trait motivation and week (i.e.,the linear term of time) was not significant. There-

fore, Hypothesis 2a was not supported. Hypothesis2b suggests that avoidance-oriented trait motiva-tion will be negatively related to average levels ofjob search intensity over an unemployment experi-ence, which was not supported in model 2b ofTable 3.

Motivational Traits Affecting Mental Health

Supporting Hypothesis 3a, approach-orientedtrait motivation was positively related to mentalhealth at the start of unemployment (see model 3a,Table 3). Hypothesis 3b was also supported. Higherlevels of approach-oriented trait motivation wereassociated with higher average levels of mentalhealth over the duration of unemployment (model3b, Table 3). Results of the tests of the two hypoth-

FIGURE 2Mean Plots of Job Search Hours and Mental Health during Unemployment Experience

(2A) Trend of Job Search Hours during Unemployment Experience 20

19

18

17

16

15

14

13

12

11

101 2

Job SearchHours

3 4 5

1 2 3 4 5

6 7

Week8 9

Week

10 1112

6 7 8 9 10 1112

13 14 1516 1718 19 20

1314 1516 171819 204

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5

MentalHealth

(2B) Trend of Mental Health During Unemployment Experience


TA

BL

E3

Gen

eral

ized

Lin

ear

Mod

elan

dH

iera

rch

ical

Lin

ear

Mod

elin

gR

esu

lts

for

Intr

ain

div

idu

alan

dC

ross

-Lev

elE

ffec

tsa

Var

iabl

esb

Mod

el1a

:Jo

bS

earc

hIn

ten

sity

Mod

el1b

:Jo

bS

earc

hIn

ten

sity

Mod

el2a

:Jo

bS

earc

hIn

ten

sity

Mod

el2b

:Jo

bS

earc

hIn

ten

sity

Mod

el3a

:M

enta

lH

ealt

h

Mod

el3b

:M

enta

lH

ealt

h

Mod

el4a

:M

enta

lH

ealt

h

Mod

el4b

:M

enta

lH

ealt

h

Mod

el5:

Mot

ivat

ion

Con

trol

Mod

el6:

Sel

f-D

efea

tin

gC

ogn

itio

n

Mod

el7:

Job

Sea

rch

Inte

nsi

ty

Mod

el8:

Men

tal

Hea

lth

Mod

el9:

Job

Sea

rch

Inte

nsi

ty

Mod

el10

:M

enta

lH

ealt

h

Mod

el11

:M

enta

lH

ealt

h

Inte

rcep

t1.

010.

831.

87**

1.76

**3.

38**

3.87

**5.

89**

6.13

**1.

75*

2.27

**1.

28**

5.27

**0.

904.

55**

6.10

**C

ontr

ols

and

stat

icp

red

icto

rsA

ge0.

010.

010.

010.

01�

0.01

�0.

010.

000.

000.

01�

0.01

0.00

�0.

010.

00�

0.01

0.00

Ed

uca

tion

�0.

06�

0.08

�0.

08�

0.10

0.31

0.15

0.07

0.07

�0.

08�

0.08

�0.

180.

18�

0.17

0.21

0.15

Gen

der

�0.

27*

�0.

27*

�0.

21�

0.21

�0.

07�

0.08

0.16

0.15

�0.

110.

19�

0.30

**0.

09�

0.32

**0.

050.

21E

thn

icit

y�

0.05

�0.

030.

090.

11�

0.29

�0.

34�

0.08

�0.

06�

0.16

�0.

03�

0.01

�0.

21�

0.07

�0.

30�

0.11

Day

su

nem

plo

yed

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Occ

up

atio

n—

Pro

fess

ion

al0.

44*

0.43

*0.

44*

0.43

*0.

320.

310.

300.

290.

08�

0.30

0.49

**0.

190.

49**

0.20

0.19

Occ

up

atio

n—

Cle

rica

l0.

210.

190.

230.

20�

0.02

�0.

10�

0.07

�0.

070.

11�

0.29

0.24

�0.

130.

24�

0.12

�0.

10E

mp

loym

ent

com

mit

men

t0.

080.

080.

100.

11�

0.04

�0.

07�

0.01

�0.

010.

06�

0.06

0.11

�0.

050.

09�

0.08

�0.

04A

pp

roac

hm

otiv

atio

n0.

21*

0.22

*0.

34**

0.31

*0.

32**

�0.

170.

110.

21A

void

ance

mot

ivat

ion

�0.

06�

0.07

�0.

46**

�0.

51**

�0.

16*

0.27

�0.

37S

tati

cm

oder

ator

sA

pp

roac

hm

otiv

atio

n�

wee

kA

void

ance

mot

ivat

ion

�w

eek

�0.

00�

0.01

**

Wee

kly

pre

dic

tors

Wee

k�

0.04

**�

0.03

0.02

*0.

06**

�0.

03**

0.05

**�

0.03

**0.

05**

0.05

**W

eek

squ

ared

0.00

1*0.

001

�0.

001*

�0.

001*

*0.

00�

0.00

2**

0.00

�0.

002*

*�

0.00

2**

Mot

ivat

ion

con

trol

0.23

**0.

09**

0.23

**0.

08**

0.08

**S

elf-

def

eati

ng

cogn

itio

n0.

03�

0.34

**0.

04�

0.34

**�

0.34

**M

odel

fitc

�8,

576.

12�

8,74

0.87

�8,

315.

31�

8,20

1.69

2,42

5.30

2,46

4.10

2,58

2.80

2,43

5.80

3,17

5.50

2,69

0.90

�10

,395

.93

2,13

5.50

�10

,512

.30

2,13

4.60

2,11

3.40

an

�1,

455

obse

rvat

ion

s.S

AS

“pro

cge

nm

od”

was

use

dw

ith

Poi

sson

dis

trib

uti

onan

dan

AR

(1)

(i.e

.fir

st-o

rder

auto

regr

essi

ve)

corr

elat

ion

mat

rix

stru

ctu

refo

rm

odel

sin

wh

ich

the

dep

end

entv

aria

ble

isjo

bse

arch

inte

nsi

ty.C

oeff

icie

nts

nee

dto

bean

ti-l

og-t

ran

sfor

med

and

trea

ted

asm

ult

ipli

ers

wh

enca

lcu

lati

ng

the

pre

dic

ted

hou

rsof

job

sear

ch.S

AS

“pro

cm

ixed

”an

alys

isw

asu

sed

wit

hth

eR

E-A

R(1

)va

rian

ce-c

ovar

ian

cest

ruct

ure

(i.e

.,ra

nd

omef

fect

sp

lus

auto

regr

essi

on[1

]st

ruct

ure

)fo

ral

lot

her

mod

els.

En

trie

sco

rres

pon

din

gto

the

pre

dic

tors

inth

efi

rst

colu

mn

are

esti

mat

esof

fixe

def

fect

s,�’s

.b

For

occu

pat

ion

,th

eco

mp

aris

ongr

oup

is“o

ther

occu

pat

ion

s.”

cT

he

mod

elfi

tin

dex

isQ

ICfo

r“p

roc

gen

mod

”m

odel

san

d–2

log-

like

lih

ood

for

“pro

cm

ixed

”m

odel

s.T

he

smal

ler

the

valu

e,th

ebe

tter

the

fit.

*p

�.0

5**

p�

.01

eses together showed that individuals one pointhigher in approach-oriented trait motivation were.34 points higher on their mental health level at thestart of their unemployment and .31 points higheron average over its course.

Hypothesis 4a suggests that avoidance-orientedtrait motivation will be negatively associated withthe slope of mental health over time. This wastested in model 4a of Table 3. As predicted, avoid-ance-oriented trait motivation was negatively asso-ciated with the slope of mental health (� � –.01,p � .01). Therefore, Hypothesis 4a was supported.Hypothesis 4b suggests that avoidance-orientedtrait motivation will be negatively related to aver-age levels of mental health over an unemploymentexperience. This was supported in model 4b ofTable 3 (� � –.51, p � .01).

Motivational Traits Affecting Self-RegulatoryStates Affecting Search Intensity and MentalHealth

Hypothesis 5a suggests that approach motivationwill be positively related to motivation control andnegatively related to self-defeating cognition dur-ing unemployment, and Hypothesis 5b suggeststhat avoidance motivation will be negatively re-lated to motivation control and positively related toself-defeating cognition during unemployment.Models 5 and 6 of Table 3 are the tests of thesehypotheses. Hypothesis 5a was partially sup-ported, and Hypothesis 5b was fully supported.Specifically, a one-point increase in approachmotivation was associated with an average .32point increase in motivation control but was not

FIGURE 3Job Search Hours Trend Breakdown on High versus Low Levels of Approach Motivation

25

20

15

10

5

Job SearchHours

Week

Low approach motivation (bottom third of sample)High approach motivation (top third of sample)

1 2 20191817161514131211109876543


significantly related to self-defeating cognition.In contrast, a one-point increase in avoidancemotivation was associated with an average .16point decrease in motivation control and a .27point increase in self-defeating cognition over thecourse of unemployment.

The goal of the next set of analyses was to inves-tigate whether motivation control and self-defeat-ing cognition reported in each week account fordifferences in weekly job search intensity and men-tal health levels.1 Hypotheses 6a and 6b suggestthat in weeks in which motivation control washigher, job search intensity and mental health willalso be higher. Results in models 7 and 8 supportedthese two hypotheses. Specifically, a one-point in-crease in motivation control in a given week wasassociated with a .26 times (� e(.23) –1) increase injob search hours and a .086 point increase in men-tal health in that same week.

Hypothesis 7a was not supported. Specifically,levels of self-defeating cognition in a given weekdid not predict levels of job search intensity in thesame week (� � .034, n.s.; see model 7 in Table 3).Hypothesis 7b, however, was supported, withhigher levels of self-defeating cognition in a givenweek associated with lower levels of mental healthin the same week (� � –.34, p � .01; see model 8 inTable 3).

Hypothesis 8 proposes that motivational statesmediate the relationship between approach andavoidance motivation and mental health and jobsearch intensity. We examined the indirect effectsvia motivational states using the bootstrappingmethod recommended by Bauer, Preacher, and Gil(2006). As suggested by prior research, the conven-tional method (i.e., Baron & Kenny, 1986) may in-

clude unnecessary steps and is inferior in terms ofstatistical power and type I error rates (MacKinnon,Lockwood, Hoffman, West, & Sheets, 2002). Basedon the direct effects of approach and avoidancemotivation on the motivational states mentionedabove and the direct effects of the motivationalstates on job search intensity and mental healthshown in models 9, 10, and 11 of Table 3, ourbootstrapping results indicated the following.

For approach motivation, the 95% confidenceinterval of the indirect effect via motivation controlwas .15, .65 for job search intensity and .0085, .05for mental health. This suggests that motivationcontrol mediated the relationships between ap-proach motivation and job search intensity andmental health. Self-defeating cognition was not amediator of the relationships between approach mo-tivation and job search intensity or mental health.

For avoidance motivation, the 95% confidenceinterval of the indirect effect via motivation controlwas –0.38, –0.021 for job search intensity and–.028, –.001 for mental health. This suggests thatmotivation control mediated the relationships be-tween avoidance motivation and job search inten-sity and mental health. In addition, the 95% confi-dence interval of the indirect effect of avoidancemotivation via self-defeating cognition on mentalhealth was –.14, –.046. Self-defeating cognitionwas not a mediator of the relationships betweenavoidance motivation and job search intensity.

In sum, out of the eight possible indirect effectsof approach and avoidance motivation on mentalhealth and job search intensity via motivationalstates, we found five indirect effects that were sta-tistically significant. Thus, Hypothesis 8 wassupported.

Search Intensity and Mental Health Affecting JobSearch Success

Hypothesis 9 suggests that job search intensity atthe start and over the course of unemployment willbe positively related to number of interviews andreemployment speed.2 We conducted Cox regres-

1 We report the unlagged instead of lagged model re-sults. It is the motivational states displayed in a givenweek that are most central, theoretically, to behavior andmood in that same week. For example, it is possible for aperson to engage in self-defeating cognition in one weekbut recover by the following week, making current levelsof self-defeating cognition most relevant. However, forcomprehensiveness and to assess lingering effects, weexamined whether motivational states in each week twere related to job search intensity and mental health inthe following week (t � 1). Our lagged results were sim-ilar to the nonlagged results. Specifically, our findingsshowed that motivation control in each week was posi-tively related to job search intensity in weeks t � 1(� � .11, p � .001), but self-defeating cognition was notrelated to job search intensity in weeks t � 1. Motivationcontrol was positively related to mental health in weekst � 1 (� � .04, p � .05), and self-defeating cognition wasnegatively related to mental health in weeks t � 1 (� �–.21, p � .001).

2 Although, given the many relationships we focusedon in our literature review, approach and avoidance mo-tivation and the two motivational states were not hypoth-esized as predictors, they were examined as such inanalyses reported in Table 4. Approach motivationwas not associated with either outcome. Avoidance mo-tivation was significantly and negatively associated withnumber of interviews but not with reemployment speed.Motivational states were unrelated to the distal out-comes, with the exception of motivation control, whichwas a positive predictor of number of interviews.


sions and OLS regression for the dependent vari-ables reemployment speed and number of inter-views, respectively (see Table 4). Using averagesearch intensity over the duration of unemploy-ment as the predictor (see columns 1 and 3), wefound that it was positively related to both reem-ployment speed (hazard ratio � 1.03, p � .05) andnumber of interviews (� � .05, p � .001). Using jobsearch intensity at the beginning of unemploymentas the predictor (see columns 2 and 4), we foundthat it was not related to reemployment speed butwas positively related to average number of inter-views per week (� � .04, p � .001).

Hypothesis 10 suggests that job seekers’ mentalhealth at the start and over the course of unemploy-ment will be positively related to reemploymentspeed and number of interviews. The results fromTable 4 indicate that mental health at the start waspositively related to number of interviews (� � .17,p � .05; column 4), but whether measured at thestart or over the unemployment experience, itwas not related to reemployment speed.

DISCUSSION

Over the course of job search, individuals expe-rience a panoply of events and emotions that mayimpact their mental health or derail their searchprocess. However, very little research data address

what happens over time during a job search expe-rience. Our findings address this gap by providingevidence on temporal dynamics in the job searchprocess. Building upon recent theorizing and stud-ies of job search from the self-regulation perspec-tive, we proposed a framework to examine the roleof motivational traits and states and their relevanceto the job search journey. Our findings offer bothempirical and theoretical contributions to the liter-ature and suggest implications for practice as wellas several directions for future research.

The design of our study permits evaluation of thefindings at several levels of analysis. At the aggre-gate level, focusing on the temporal dimensionalone, we examined general trends in how jobsearch and mental health changed during unem-ployment. Among all the individuals in our sam-ple, significant within-person decline in the timespent in job search occurred over time, with a slightuptick in later months. This finding is similar tothat reported by Wanberg et al. (2005), the onlyother repeated-measures (over several weeks),within-person assessment of job search that weknow of. For mental health, we found that partici-pants showed gradual improvement over the first10–12 weeks. This was followed by a slight down-turn in mental health in later weeks. This finding iscontrary to the negative relationship betweenlength of unemployment and mental health re-

TABLE 4Relationship between Mental Health and Job Search Intensity and Indicators of Search Successa

Variables

Reemployment SpeedAverage Number of

Interviews per Week

Model 1 Model 2 Model 3 Model 4

Intercept 0.49 0.03Age 0.98 0.98 0.00 0.00Education 0.63 0.60 0.36 0.31Gender 1.06 1.02 –0.05 0.04Ethnicity 1.40 1.75 0.10 0.30Days unemployed before the start of the study 1.00 1.00 0.00 0.00Occupation—Professional 0.71 0.69 –0.08 –0.13Occupation—Clerical 0.45* 0.37* 0.15 0.03Employment commitment 1.19 1.23 –0.01 –0.01Average mental health over unemployment duration 0.97 0.12Average job search hours over unemployment duration 1.03* 0.05***Week 1 mental health 0.92 0.17*Week 1 search hours 1.01 0.04***nb 125 123 132 129Adjusted R2 .28 .20

a Hazard ratios are reported for the reemployment rate-speed analysis. A coefficient greater than 1.0 means that the variable is associatedwith faster reemployment speed.

b Sample size varies owing to missing data on the two dependent variables and the week 1 and average levels of job search hours andmental health.

* p � .05*** p � .001


ported in meta-analyses by both McKee-Ryan et al.(2005) and Paul and Moser (2009). However, thework summarized in these studies was largelycross-sectional, based on observations of the men-tal health levels of different individuals unem-ployed for various lengths of time rather than onfollowing the same individuals over time. In con-trast, the design of our study permitted us to assesswithin-person changes in mental health by follow-ing individuals from the start of their unemploy-ment experience over time. It is possible that theimprovement in mental health we observed stemsfrom an inherently low level of mental health asso-ciated with the start of an involuntary unemploy-ment spell. For example, Amundson and Borgen(1982), drawing on their counseling experience, ob-served that it is typical for newly unemployed in-dividuals to feel angry and powerless. As timepasses, they accept their unemployment situation.There is an increase in energy, high hopes, andsometimes unrealistic expectations that can liftmood. As time passes further, individuals begin tofeel burned out and frustrated as they encounterrepeated rejections. The Amundson and Borgen(1982) portrait of the dynamics of unemploymentfrom an experiential standpoint fit our data reason-ably well. Further repeated-measures research, be-ginning at the onset of job loss, is needed to extendour results as well as to ascertain whether the av-erage search intensity and mental health trajecto-ries we observed are robust.

At a second, more person-oriented level of anal-ysis, our findings indicate the importance of deter-mining not only what the average experience lookslike for individuals over time, but also the impactof individual differences on experienced trajecto-ries of job search intensity and mental health overtime. Specifically, our investigation into the role ofindividual differences in motivational traits andstates showed that higher approach-oriented traitmotivation was related to higher levels of jobsearch intensity and mental health, both at the startof and over the course of unemployment. Higheravoidance-oriented trait motivation, on the otherhand, predicted lower levels of mental health atboth the start of unemployment and over time. Inaddition, we further found that avoidance-orientedtrait motivation was associated with a negativeslope for mental health over time. Taken together,the findings contribute new information about theuse of self-regulatory variables to predict goal-based effort as well as mental health over time in afield context. The time-based effects of motiva-tional traits on job search intensity, for example,are illustrated in Figure 3.

Finally, analyses conducted on motivationalstates by week showed that changes in motivationcontrol in any given week were significantly asso-ciated with within-person changes in search inten-sity and mental health, and changes in self-defeat-ing cognition from week to week were significantlyassociated with changes in mental health. On thebasis of our proposed conceptual model, we furtherexamined these motivational states as mediatingvariables between motivational traits, job searchintensity, and mental health. These findings ad-vance understanding of how (i.e., through whatmechanisms) stable individual differences exerttheir influence and the effects of these mechanismson the job search experience. Our results especiallyhighlight motivation control as an explanatorymechanism. Motivation control mediated the rela-tionship between approach motivation and bothjob search intensity and mental health, mediatedthe relationship between avoidance motivation andjob search intensity, and mediated the relationshipbetween avoidance motivation and mental health.In contrast, self-defeating cognitions only mediatedthe relationship between avoidance motivation andmental health and had no relationship with jobsearch intensity.

Although the mediating role of self-regulatoryprocesses in trait-performance relations has beenshown previously in nondynamic contexts (e.g.,Creed et al., 2009; see Kanfer & Kantrowitz, 2003),the pattern of findings obtained in this study sug-gests a new perspective on the nature of this medi-ation. For example, previous findings by Creed andcolleagues (2009) show that individual differencesin avoidance motivational traits were not related toemotion control states. In our study, however, in-dividuals higher in avoidance motivation experi-enced higher levels of self-defeating cognition, sug-gesting that it may be valuable to study emotion-related states in terms of both the experience ofnegative cognitions and attempts to control nega-tive affect.

One particularly interesting pattern of findingspertains to the differential relationship of self-de-feating cognition to job search intensity and mentalhealth. Contrary to expectations, no significant re-lationship was observed between self-defeatingcognition and job search intensity. However, a neg-ative relationship was observed between self-de-feating cognition and mental health. Although itmight be tempting to conclude that self-defeatingcognitions do not play a role in job search intensity,it is important to consider whether these cognitionsmay have other undesirable effects that were notmeasured in this study. For example, it may be thathigh levels of self-defeating cognition exert their


influence on other dimensions of search behaviorthat we did not measure, such as breadth or qualityof search activities. It is also possible that the be-havioral impact of thinking negatively about one’ssearch may be more immediate and thus morelikely to be captured in a repeated-measures studythat assesses behavior daily.

Alternatively, our findings may reflect importantcontextual differences in behavioral sensitivity tonegative affective states. For example, behavioralsensitivity to negative self-defeating cognition maybe lower in the job search context than in time-bounded, routine achievement contexts. Or, from acontrol theory perspective (Carver, 2003, 2006), itis possible that self-defeating cognition stems fromlow levels of perceived progress, creating a discrep-ancy to be resolved. Some individuals may usetheir negative emotions to highlight the specificthreat or issue that needs to be resolved (Wanberget al., 2010). Finally, it is possible a broader oralternative measure of self-defeating cognition mayproduce the expected findings. Our measure fo-cused on cognition specific to feelings of hopeless-ness, giving up, and negative expectancies, funda-mental triggers of depression and reduced activity(Beck et al., 1974), but self-defeating cognition canmore broadly include destructive thoughts aboutoneself, such as “What is wrong with me?” (Hollon& Kendall, 1990). Our measure was, furthermore,developed for this study and did not undergo ex-aminations of concurrent or predictive validity,something that may be considered a limitation toour study. A similar limitation applies to our mea-sure of motivational control; it was developed for thisstudy, and further validation and development of thismeasure is desirable before it is used more widely.

With regard to helping individuals find jobsmore quickly, our results showed that individualswho reported higher initial levels of search inten-sity and mental health at the beginning of job lossexperienced greater success in obtaining job inter-views and that individuals who maintained highlevels of job search intensity over the duration ofjob search had more interviews and found jobsmore quickly. These findings are consistent withprevious research results that show a positive rela-tionship between job search intensity and reem-ployment (Kanfer et al., 2001). In addition, ourfindings indicate that although the importance ofjob search intensity continues throughout unem-ployment, the mental health of a person at the startof job search also plays an important role in searchoutcomes. Future research is needed to determinethe full range of mechanisms by which mentalhealth at the onset of a job search affects job searchoutcomes.

Implications for Practice and Future Research

The average tendency for individuals to decreasetheir search behavior as their unemployment expe-rience continues suggests that it may be construc-tive for job seekers, as well as organizations thatwork with job seekers, to monitor job search levelsover time to keep persistence in the search going.Although other factors, such as a job seeker’s “hu-man capital” and quality of search, are critical toreemployment, research has established that timespent in search distinguishes between successfuljob seekers and unsuccessful job seekers (Kanfer etal., 2001). We additionally recognize that in somesituations, such as that faced by a job seeker in asmall town without the choice of relocating, a timemay come when there are few new job search leadsto pursue.

The process orientation of our study yields otherfindings that have implications for job search inter-ventions. In keeping with models of action thatconceptualize traits as inputs to proximal determi-nants of action, we found that motivational statesmediated the relationship between motivationaltraits and the focal variables. In particular, strate-gies to enhance motivation control not only facili-tated job search intensity but also had a salutaryeffect on mental health. These findings suggest thattraining individuals to employ self-regulation strat-egies that “pump up” attentional effort for jobsearch activities may be effective in enhancing jobsearch intensity and may in fact also confer moremental health protection (by attenuating the nega-tive influence of avoidance-oriented traits on men-tal health) than training directed toward reducingself-defeating cognitions. Consistently with this no-tion, van Hooft and Noordzij (2009) showed theutility of helping unemployed individuals withself-regulatory states (i.e., situational state on learn-ing goal orientation). To date, many interventionsfor job seekers place equal emphasis on training inboth motivation control and emotion control. Re-search to identify a more complete understandingof the consequences of both motivational and var-ious forms of emotional control, including self-de-feating cognition, is needed.

Overall, substantially more can be learned abouthow the two focal variables included in our study(search intensity and mental health) change overtime during unemployment, why they change, forwhom they change, the implications of suchchange, and how new knowledge about the dynam-ics of job search might be used to help job seekers.With regard to how these focal variables changeover time, repeated-measures studies of bothshorter durations (see, e.g., Song et al., 2009; Wan-


berg et al., 2010) and longer durations would bevaluable. It will also be useful to examine the gen-eralizability of the average trajectories we report forjob search and mental health over the duration ofunemployment. Specifically, to what extent do thetrajectories we show generalize to other economicor cultural contexts, to very low or very high in-come groups, to more ethnically diverse samples,or to individuals in very narrow fields with limitedjob openings versus fields with many opportuni-ties? For example, regarding economic context andtype of profession, individuals may be more likelyto decline in their job search intensity when jobsare scarce (Micklewright & Nagy, 1999). Regardingcultural context, the mental health of unemployedindividuals tends to be higher in countries withgenerous safety nets (Bambra & Eikemo, 2009) andin countries that place a lower importance on work(Marsh & Alvaro, 1990). Scholars are in the veryearly stages of understanding what happens withinthe job search experience over time. There is exten-sive opportunity to build upon our self-regulatorymodel as well as to use other theoretical perspec-tives to understand such dynamics. It would bevaluable to incorporate additional distal outcomesinto future studies, beyond the two we used, suchas reemployment quality.

REFERENCES

Aarts, H., Custers, R., & Holland, R. W. 2007. The non-conscious cessation of goal pursuit: When goals andnegative affect are coactivated. Journal of Personal-ity and Social Psychology, 92: 165–178

Alloy, L. B., Abramson, L. Y., Grant, D., & Liu, R. 2009.Vulnerability to unipolar depression: Cognitive-be-havioral mechanisms. In K. Salzinger & M. R. Serper(Eds.), Behavioral mechanisms and psychopathol-ogy: Advancing the explanation of its nature,cause, and treatment: 107–140. Washington, DC:American Psychological Association.

Amundson, N. E., & Borgen, W. A. 1982. The dynamics ofunemployment: Job loss and job search. Personneland Guidance Journal, 60: 562–564.

Bambra, C., & Eikemo, T. A. 2009. Welfare state regimes,unemployment, and health: A comparative study ofthe relationship between unemployment and self-reported health in 23 European countries. Journal ofEpidemiology and Community Health, 63: 92–98.

Baron, R. M., & Kenny, D. A. 1986. The moderator-medi-ator variable distinction in social psychological re-search: Conceptual, strategic, and statistical consid-erations. Journal of Personality and SocialPsychology, 51: 1173–1182.

Bauer, D. J., Preacher, K. J., & Gil, K. M. 2006. Conceptu-

alizing and testing random indirect effects and mod-erated mediation in multilevel models: New proce-dures and recommendations. PsychologicalMethods, 11: 142–163.

Beck, A. T. 1972. Depression: Causes and treatment.Philadelphia: University of Pennsylvania Press.

Beck, A. T. 1987. Cognitive models of depression. Jour-nal of Cognitive Psychotherapy, 1: 2–27.

Beck, A. T., Weissman, A., Lester, D., & Trexler, L. 1974.The measurement of pessimism: The HopelessnessScale. Journal of Consulting and Clinical Psychol-ogy, 42: 861–865.

Bentler, P. M., & Chou, C. P. 1987. Practical issues instructural equation modeling. Sociological Methodsand Research, 16: 78–117.

Berwick, D. M., Murphy, J. M., Goldman, P. A., Ware,J. E., Barsky, A. J., & Weinstein, M. D. 1991. Perfor-mance of a five-item mental health screening test.Medical Care, 29: 169–176.

Bureau of Labor Statistics. 2009. Labor force statisticsfrom the Current Population Survey. http://www.bls.gov/webapps/legacy/cpsatab9.htm, accessed No-vember 23, 2009.

Carver, C. S. 2003. Pleasure as a sign you can attend tosomething else. Cognition and Emotion, 17: 241–261.

Carver, C. S. 2006. Approach, avoidance, and the self-regulation of affect and action. Motivation and Emo-tions, 30: 105–110.

Carver, C. S., & White, T. L. 1994. Behavioral inhibition,behavioral activation, and affective responses to im-pending reward and punishment: The BIS/BASscales. Journal of Personality and Social Psychol-ogy, 67: 319–333.

Creed, P. A., & Bartrum, D. 2006. Explanations for dete-riorating wellbeing in unemployed people: Specificunemployment theories and beyond. In T. Kie-selbach, A. H. Winefield, & S. Anderson (Eds.), Un-employment and health: International and inter-disciplinary perspectives: 1–20. Bowen Hills,Queensland: Australian Academic Press.

Creed, P. A., King, V., Hood, M., & McKenzie, R. 2009.Goal orientation, self-regulation strategies, and job-seeking intensity in unemployed adults. Journal ofApplied Psychology, 94: 806–813.

Dalal, R. S., & Hulin, C. L. 2008. Motivation for what? Amultivariate dynamic perspective of the criterion. InR. Kanfer, G. Chen, & R. D. Pritchard (Eds.), Workmotivation: 64–100. New York: Routledge.

Diener, C. I., & Dweck, C. S. 1978. An analysis of learnedhelplessness: Continuous changes in performance,strategy, and achievement cognitions following fail-ure. Journal of Personality and Social Psychology,36: 451–462.


Dweck, C. S., & Leggett, E. L. 1988. A social-cognitiveapproach to motivation and personality. Psycholog-ical Review, 95: 256–273.

Elliot, A. J. 2008. Approach and avoidance motivation. InA. J. Elliot (Ed.), Handbook of approach and avoid-ance motivation: 415–428. New York: PsychologyPress.

Elliot, A. J., & Harackiewicz, J. M. 1996. Approach andavoidance achievement goals and intrinsic motiva-tion: A mediational analysis. Journal of Personalityand Social Psychology, 70: 3–14.

Elliot, A. J., & McGregor, H. A. 2001. A 2 � 2 achievementgoal framework. Journal of Personality and SocialPsychology, 80: 501–519.

Feather, N. T., & Davenport, P. R. 1981. Unemploymentand depressive affect: A motivational and attribu-tional analysis. Journal of Personality and SocialPsychology, 41: 422–436.

Fitzmaurice, G. M., Laird, N. M., & Ware, J. H. 2004.Applied longitudinal analysis. Hoboken, NJ: Wiley.

Galinsky, A. D., Leonardelli, G. J., Okhuysen, G. A., &Mussweiler, T. 2005. Regulatory focus at the bargain-ing table: Promoting distributive and integrative suc-cess. Personality and Social Psychology Bulletin,31: 1087–1098.

Ginexi, E. M., Howe, G. W., & Caplan, R. D. 2000. De-pression and control beliefs in relation to reemploy-ment: What are the directions of effect? Journal ofOccupational Health Psychology, 5: 323–336.

Glass, C. R., & Arnkoff, D. B. 1997. Questionnaire meth-ods of cognitive self-statement assessment. Journalof Consulting and Clinical Psychology, 65: 911–927.

Hanisch, K. A. 1999. Job loss and unemployment re-search from 1994 to 1998: A review and recommen-dation for research and intervention. Journal of Vo-cational Behavior, 55: 188–220.

Heggestad, E. D., & Kanfer, R. 2000. Individual differ-ences in trait motivation: Development of the Moti-vational Trait Questionnaire. International Journalof Educational Research, 33: 751–776.

Higgins, E. T. 1998. Promotion and prevention: Regula-tory focus as a motivational principle. In M. P. Zanna(Ed.), Advances in experimental social psychology,vol. 30: 1–46. San Diego: Academic Press.

Hinsz, V. B., & Jundt, D. K. 2005. Exploring individualdifferences in a goal-setting situation using the Mo-tivational Trait Questionnaire. Journal of AppliedSocial Psychology, 35: 551–571.

Hollon, S. D., & Kendall, P. C. 1980. Cognitive self-state-ments in depression: Development of an automaticthoughts questionnaire. Cognitive Therapy and Re-search, 14: 383–395.

Hulleman, C. S., Schrager, S. M., Bodmann, S. M., &Harackiewicz, J. M. 2010. A meta-analytic review ofachievement goal measures: Different labels for thesame constructs or different constructs with similarlabels? Psychological Bulletin, 136: 422–449.

Jahoda, M. 1987. Unemployed men at work. In D. Dryer &P. Ullah (Eds.), Unemployed people: Social andpsychological perspectives: 1–73. Milton Keynes,U.K.: Open University Press.

Judge, T. A., & Locke, E. A. 1993. Effect of dysfunctionalthought processes on subjective well-being and jobsatisfaction. Journal of Applied Psychology, 78:475–490.

Kanfer, R., & Ackerman, P. L. 1989. Motivation and cog-nitive abilities: An integrative/aptitude-treatment in-teraction approach to skill acquisition. Journal ofApplied Psychology–Monograph, 74: 657–690.

Kanfer, R., & Ackerman, P. L. 1996. A self-regulatoryskills perspective to reducing cognitive interference.In I. G. Sarason, B. R. Sarason, & G. R. Pierce (Eds.),Cognitive interference: Theories, methods, andfindings: 153–171. Hillsdale, NJ: Erlbaum.

Kanfer, R., & Ackerman, P. L. 2000. Individual differ-ences in work motivation: Further explorations of atrait framework. Applied Psychology: An Interna-tional Review, 49: 470–482.

Kanfer, R., & Heggestad, E. D. 1997. Motivational traitsand skills: A person-centered approach to work mo-tivation. In R. Sutton & B. Staw (Eds.), Research inorganizational behavior, vol. 19: 1–56. Greenwich,CT: JAI.

Kanfer, R., & Heggestad, E. D. 1999. Individual differ-ences in motivation: Traits and self-regulatory skills.In P. L. Ackerman & P. C. Kyllonen (Eds.), Learningand individual differences: Process, trait, and con-tent determinants: 293–313. Washington, DC:American Psychological Association.

Kanfer, R., & Kantrowitz, T. M. 2003. Individual differ-ences in affectivity and emotion regulation duringjob search. Paper presented at the Society for Indus-trial and Organizational Psychology meetings, Or-lando.

Kanfer, R., Wanberg, C. R., & Kantrowitz, T. M. 2001. Apersonality-motivational analysis and meta-analyticreview. Journal of Applied Psychology, 86: 837–855.

Kessler, R., Turner, J. B., & House, J. S. 1988. Effects ofunemployment on health in a community survey:Main, modifying, and mediating effects. Journal ofSocial Issues, 44: 69–85.

Klein, J. P., & Moeschberger, M. L. 1997. Survival anal-ysis: Techniques for censored and truncated data.New York: Springer.

Kozlowski, S. W. J., & Bell, B. S. 2006. Disentangling


achievement orientation and goal setting: Effects onself-regulatory processes source. Journal of AppliedPsychology, 91: 900–916.

Kuhl, J. 1985. Volitional mediators of cognitive-behav-ioral consistency: Self-regulatory processes and ac-tion vs. state orientation. In J. Kuhl & J. Beckmann(Eds.), Action control: From cognition to behavior:101–128. New York: Springer Verlag.

Kuiper, N. A., Olinger, L. J., & Swallow, S. R. 1987.Dysfunctional attitudes, mild depression, views ofself, self-consciousness, and social perceptions. Mo-tivation and Emotion, 11: 379–401.

Latack, J. C., Kinicki, A. J., & Prussia, G. E. 1995. Anintegrative process model of coping with job loss.Academy of Management Review, 20: 311–342.

Lee, F. K., Sheldon, K. M., & Turban, D. B. 2003. Person-ality and the goal-striving process: The influence ofachievement goal patterns, goal level, and mentalfocus on performance and enjoyment. Journal ofApplied Psychology, 88: 256–265.

Liberman, N., Molden, D. C., Idson, L. C., & Higgins, E. T.2001. Promotion and prevention focus on alternativehypotheses: Implications for attributional functions.Journal of Personality and Social Psychology, 80:5–18.

Lopez, S. J., Snyder, C. R., & Pedrotti, J. T. 2003. Hope:Many definitions, many measures. In S. J. Lopez &C. R. Snyder (Eds.), Positive psychological assess-ment: A handbook of models and measures: 91–106. Washington, DC: American Psychological Asso-ciation.

MacFadyen, R. G., & Thomas, J. P. 1997. Economic andpsychological models of job search behavior of theunemployed. Human Relations, 50: 1461–1484.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M.,West, S. G., & Sheets, V. 2002. A comparison ofmethods to test mediation and other interveningvariable effects. Psychological Methods, 7: 83–104.

Marsh, C., & Alvaro, J. L. 1990. A cross-cultural perspec-tive on the social and psychological distress causedby unemployment: A comparison of Spain and theUnited Kingdom. European Sociological Review, 6:234–255.

McCabe, C. J., Thomas, K. J., Brazier, J. E., & Coleman, P.1996. Measuring the mental health status of a popu-lation: A comparison of the GHQ-12 and the SF-36(MHI-5). British Journal of Psychiatry: The Journalof Mental Science, 169: 516–521.

McKee-Ryan, F., Song, Z., Wanberg, C. R., & Kinicki, A. J.2005. Psychological and physical well-being duringunemployment: A meta-analytic study. Journal ofApplied Psychology, 90: 53–76.

Micklewright, J., & Nagy, G. 1999. The informationalvalue of job search data and the dynamics of

search behaviour: Evidence from Hungary. Work-ing paper no. 1999/1), Labour Research Department,Institute of Economics, Hungarian Academy of Sci-ences; and Department of Human Resources, Buda-pest University of Economics.

Niessen, C., Heinrichs, N., & Dorr, S. 2009. Pursuit andadjustment of goals during unemployment: The roleof age. International Journal of Stress Manage-ment, 16: 102–123.

Paul, K. I., & Moser, K. 2009. Unemployment impairsmental health: Meta-analyses. Journal of VocationalBehavior, 74: 264–282.

Payne, S. C., Youngcourt, S. S., & Beaubien, J. M. 2007. Ameta-analytic examination of the goal orientationnomological net. Journal of Applied Psychology,92: 128–150.

Petrocelli, J. P., Glaser, B. A., Calhoun, G. B., & Campbell,L. F. 2001. Cognitive schemas as mediating variablesof the relationship between the self-defeating per-sonality and depression. Journal of Psychopathol-ogy and Behavioral Assessment, 23: 183–191.

Ployhart, R. E. 2008. The measurement and analysis ofmotivation. In R. Kanfer, G. Chen, & R. D. Pritchard(Eds.), Work motivation: 18–61. New York: Rout-ledge.

Porath, C. L., & Bateman, T. S. 2006. Self-regulation:From goal orientation to job performance. Journal ofApplied Psychology, 91: 185–192.

Prussia, G. E., Fugate, M., & Kinicki, A. J. 2001. Explica-tion of the coping goal construct: Implications forcoping and reemployment. Journal of Applied Psy-chology, 86: 1179–1190.

Robinson, M. D., Meier, B. P., Tamir, M., Wilkowski,B. M., & Ode, S. 2009. Behavior facilitation: A cog-nitive model of individual differences in approachmotivation. Emotion, 9: 70–82.

Rowley, K. M., & Feather, N. T. 1987. The impact ofunemployment in relation to age and length of un-employment. Journal of Occupational Psychology,60: 323–332.

Saks, A. M. 2005. Job search success: A review and inte-gration of the predictors, behaviors, and outcomes.In S. D. Brown & R. W. Lent (Eds.), Career develop-ment and counseling: Putting theory and researchto work: 155–179. Hoboken, NJ: Wiley.

Song, Z., Uy, M. A., Zhang, S., & Shi, K. 2009. Daily jobsearch and psychological distress: Evidence fromChina. Human Relations, 62: 1171–1197.

Steel, R. P. 2002. Turnover theory at the empirical inter-face: Problems of fit and function. Academy of Man-agement Review, 27: 346–360.

Strauman, T. J. 2002. Self-regulation and depression.Self and Identity, 1: 151–157.


Tamir, M., & Diener, E. 2008. Approach-avoidance goalsand well-being: One size does not fit all. In A. J.Elliot (Ed.), Handbook of approach and avoidancemotivation: 415–428. New York: Psychology Press.

van Hooft, E. A. J., Born, M., Taris, T. W., Flier, H., &Blonk, R. W. B. 2005. Bridging the gap betweenintentions and behavior: Implementation intentions,action control, and procrastination. Journal of Vo-cational Behavior, 55: 238–256.

van Hooft, E. A. J., & Noordzij, G. 2009. The effects of goalorientation on job search and reemployment: A fieldexperiment among unemployed job seekers. Journalof Applied Psychology, 94: 1581–1590.

Van Hoye, G., & Saks, A. M. 2008. Job search as goal-directed behavior: Objectives and methods. Journalof Vocational Behavior, 73: 358–367.

VandeWalle, D. 1997. Development and validation of awork domain goal orientation instrument. Educa-tional and Psychological Measurement, 57: 995–1015.

Veit, C. T., & Ware, J. E. 1983. The structure of psycho-logical distress and well-being in general popula-tions. Journal of Consulting and Clinical Psychol-ogy, 51: 730–742.

Viinamäki, H., Koskela, K., & Nishanen, L. 1996. Rapidlydeclining mental well-being during unemployment.European Journal of Psychiatry, 10: 215–221.

Vohs, K. D., & Baumeister, R. F. 2004. Understandingself-regulation. In R. F. Baumeister & K. D. Vohs(Eds.), Handbook of self-regulation: 1–9. New York:Guilford.

Wanberg, C. R., Glomb, T. M., Song, Z., & Sorenson, S.2005. Job-search persistence during unemployment:A ten-wave longitudinal study. Journal of AppliedPsychology, 90: 411–430.

Wanberg, C. R., Kanfer, R., & Rotundo, M. 1999. Unem-ployed individuals: Motives, job-search competen-cies, and job-search constraints as predictors of jobseeking and reemployment. Journal of Applied Psy-chology, 6: 897–910.

Wanberg, C. R., Zhu, J., & van Hooft, E. 2010. The job-search grind: Perceived progress, self-reactions, andself-regulation of search effort. Academy of Man-agement Journal, 53: 788–807.

Ware, J. E., Jr., Manning, W. G., Jr., Duan, N., Wells, K. B.,& Newhouse, J. P. 1984. Health status and the use ofoutpatient mental health services. American Psy-chologist, 39: 1090–1100.

Warr, P. 1987. Work, unemployment, and mentalhealth. Oxford, U.K.: Clarendon Press.

Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. 1999.

The two general activation systems of affect: Struc-tural findings, evolutionary considerations, and psy-chobiological evidence. Journal of Personality andSocial Psychology, 76: 820–838.

Whinghter, L. J., Cunningham, C. J. L., Wang, M., &Burnfield, J. L. 2008. The moderating role of goalorientation in the workload-frustration relationship.Journal of Occupational Health Psychology, 13:283–291.

Zikic, J., & Saks, A. 2009. Job search and social cognitivetheory: The role of career-relevant activities. Journalof Vocational Behavior, 74: 117–127.

Zuroff, D. C., Blatt, S. J., Sanislow, C. A., III, Bondi, C. M.,& Pilkonis, P. A. 1999. Vulnerability to depression:Reexamining state dependence and relative stability.Journal of Abnormal Psychology, 108: 76–89.

Connie Wanberg ([email protected]) is the IndustrialRelations Faculty Excellence Chair in the Carlson Schoolof Management at the University of Minnesota. She re-ceived her Ph.D. in industrial/organizational psychologyfrom Iowa State University. Her research has been fo-cused on the individual experience of unemployment,job-search behavior, organizational change, employee so-cialization, and employee development.

Jing Zhu ([email protected]) is an assistant professor in theDepartment of Management at the School of Businessand Management, Hong Kong University of Science andTechnology. She received her Ph.D. from the CarlsonSchool of Management, University of Minnesota. Herresearch interests include team process and effectiveness(with a focus on expertise utilization, conflict, composi-tion, and performance), expatriate adjustment, dynamicjob search behaviors, and motivation.

Ruth Kanfer ([email protected]) is a professor ofindustrial/organizational psychology in the School ofPsychology at Georgia Institute of Technology. She re-ceived her Ph.D. from Arizona State University. Her re-search focuses on the impact of personality, motivation,and affective processes in workforce aging, job searchand work transitions, teams, cognitive fatigue, and self-regulated learning.

Zhen Zhang ([email protected]) is an assistant profes-sor of management in the W. P. Carey School of Businessat Arizona State University. He received his Ph.D. inhuman resources and industrial relations from the Uni-versity of Minnesota. His research interests includeleadership process and development, biological basesof organizational behavior, intersections between or-ganizational behavior and entrepreneurship, and re-search methods.


Copyright of Academy of Management Journal is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

after the pink slip: applying dynamic ......after the pink slip: applying dynamic motivation...

Documents