Journal of Vocational Behavior 74 (2009) 190–198

Contents lists available at ScienceDirect

Journal of Vocational Behavior

journal homepage: www.elsevier .com/locate / jvb

Social cognitive predictors of academic adjustment and life satisfactionin Portuguese college students: A longitudinal analysis q

Robert W. Lent a,*, Maria do Céu Taveira b, Hung-Bin Sheu c, Daniel Singley d

a Department of Counseling and Personnel Services, University of Maryland, 3214 Benjamin Building, College Park, MD 20742, USAb Department of Psychology, University of Minho, Braga, Portugalc Division of Psychology in Education, Arizona State University, USAd Counseling Center, University of California, San Diego, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 25 September 2008Available online 25 December 2008

Keywords:Social cognitive career theorySelf-efficacyAcademic satisfactionAcademic adjustmentLife satisfactionSubjective well-being

0001-8791/$ - see front matter � 2008 Elsevier Incdoi:10.1016/j.jvb.2008.12.006

q An earlier version of this study was presented atItaly, September, 2007.

* Corresponding author. Fax: +1 301 405 9995.E-mail address: boblent@umd.edu (R.W. Lent).

A social cognitive model of well-being [Lent, R. W. (2004). Toward a unifying theoreticaland practical perspective on well-being and psychosocial adjustment. Journal of CounselingPsychology, 51, 482–509.] was adapted to the context of academic adjustment and testedusing a longitudinal design. Participants were 252 students at a university in northernPortugal. They completed measures of academic self-efficacy, environmental support, goalprogress, and adjustment, along with global measures of positive affect and life satisfac-tion. Path analyses indicated that the model fit the data well overall. As expected, self-effi-cacy and environmental support were predictive of goal progress and academicadjustment, and the latter was predictive of students’ global life satisfaction. Self-efficacyand positive affect were found to be reciprocally related to one another. Contrary to expec-tations, goal progress did not contribute uniquely to the prediction of academic adjustmentor life satisfaction. We consider directions for future research applying the social cognitivemodel to satisfaction in, and adjustment to, educational and work settings.

� 2008 Elsevier Inc. All rights reserved.

1. Introduction

In an effort to extend the study of subjective well-being (SWB) to vocational and counseling psychology, Lent (2004) re-cently proposed a unifying theoretical approach to domain and life satisfaction. This approach draws upon inquiry on bothhedonic and eudaimonic perspectives on well-being and includes cognitive (e.g., goals), behavioral (e.g., participation in val-ued life tasks), social (e.g., support), and trait-affective elements (Brunstein, 1993; Cantor & Sanderson, 1999; Diener & Fujita,1995; Diener, Suh, Lucas, & Smith, 1999; McCrae & Costa, 1991; Ryan & Deci, 2000; Ryff & Singer, 1998). Lent and Brown(2006, 2008) later extended Lent’s general unifying approach to the specific domains of educational and work satisfaction.This theoretical effort incorporates key elements of social cognitive theory, which has proven to be a versatile framework inthe study of adaptive processes and positive adjustment (Bandura, 1997, 2001). The educational and vocational extensions ofthe well-being model are also designed to complement the previously developed interest, choice, and performance models ofsocial cognitive career theory (SCCT; Lent, Brown, & Hackett, 1994).

According to the unifying model of well-being, domain-specific adjustment (e.g., satisfaction and functioning in educa-tional and work settings) and overall SWB (e.g., general satisfaction with life) are jointly determined by cognitive, behavioral,

. All rights reserved.

the annual meeting of the International Association for Educational and Vocational Guidance, Padova,

R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198 191

social, and personality variables. In keeping with social cognitive assumptions, this approach highlights aspects of positivefunctioning over which people can exercise agency (e.g., the setting and pursuit of personal goals, involvement in valued lifetasks, building social supports). However, it also acknowledges factors (e.g., personality dispositions) that may be more resis-tant to personal agency and traditional approaches to counseling.

Lent’s (2004) theoretical framework contains two interrelated models, one aimed at the experience of well-being undernormative life conditions and the other focusing on the recovery of well-being subsequent to stressful or traumatic lifeevents. The purpose of the current study was to test a version of the former (normative well-being) model within a sampleof Portuguese college students. The basic normative well-being model is displayed in Fig. 1. According to this model, overalllife satisfaction is influenced by certain personality variables (e.g., trait positive and negative affect), satisfaction in one’s cen-tral life domains (e.g., work, family), participation in valued life tasks, and progress at fulfilling salient personal goals.

Domain satisfaction, one of the precursors of overall life satisfaction, is seen as partly determined by personality factors,but is also posited to be affected by agenetic, social cognitive mechanisms, in particular, goal-directed activity, self-efficacy,outcome expectations, and environmental supports and resources. In other words, in addition to benefiting from particulartraits (Diener et al., 1999), people are more likely to be satisfied within a given life domain when they actively pursue andmake progress at personally valued goals (Brunstein, 1993); feel competent at the tasks required for successful performanceand goal pursuit (Bandura, 1997); anticipate the receipt of favorable outcomes (Carver & Scheier, 2002); and perceive theirenvironment as supportive and as offering resources to enable their goal pursuit (Cantor & Sanderson, 1999). Both for the-oretical reasons and because of their assumed relevance to preventive and therapeutic interventions, the model is also con-cerned with the nature of the relations among the social cognitive precursors of domain satisfaction. For instance, self-efficacy and outcome expectations are each assumed to be influenced by the availability of goal-relevant environmental sup-ports and resources.

Given the newness of the generic framework of domain and life satisfaction (Lent, 2004) and its extension to the morespecific adjustment contexts of academic and work satisfaction (Lent & Brown, 2006), this approach has thus far receivedlimited empirical study. A few recent studies have, however, tested the normative well-being model. Lent, Singley et al.(2005) reported two studies in which the model fit the data well within general samples of college students, helping to ex-plain satisfaction in particular life domains (academics, social life) and in general. In both studies, progress at personal goalswas a reliable predictor of domain satisfaction which, in turn, predicted overall life satisfaction. Lent, Singley, Sheu, Schmidt,and Schmidt (2007) also found good support for the model in accounting for academic satisfaction in a sample of engineeringstudents. In most of the model tests, domain self-efficacy and/or environmental supports (but not outcome expectations)also contributed to domain satisfaction.

The present study was designed to extend the nascent line of research on the normative well-being model in severalways. First, each of the published studies on this model has used a cross-sectional design. Such a design is helpful in explor-ing concurrent relations among the theoretical variables but is not able to test the temporal predominance of the relationsposited by the model (e.g., in order to support the assumption that goal progress leads to domain satisfaction, the formermust precede the latter in time). In the present study we used a longitudinal design to examine the nature of the temporalrelations among the variables, as specified by the theory.

Second, prior studies in this line of research have focused only on the prediction of domain and life satisfaction. Lent(2004) had advocated a multicomponent view of positive adjustment which includes other indicators of optimal functioning,such as satisfactory role performance and low levels of stress in central life domains, in addition to one’s experience of do-main and life satisfaction. We, therefore, operationalized positive adjustment in this study as a combination of satisfaction,stress, and academic functioning outcomes. Third, existing model tests have all involved US college students. The presentstudy examined the degree to which the model might help explain educational adjustment outcomes among college

Fig. 1. Integrative model depicting personality, affective, and social-cognitive contributors to well-being under normative life conditions. From ‘‘Toward aUnifying Theoretical and Practical Perspective on Well-Being and Psychosocial Adjustment,” by R.W. Lent, 2004, Journal of Counseling Psychology, 51, p. 500.Reprinted with permission. PA, positive affect; NA, negative affect; GSE, generalized self-efficacy.

192 R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198

students in Portugal. Elements of the model, such as self-efficacy, have been shown to offer explanatory utility across a num-ber of cross-cultural and cross-national applications (Lent & Sheu, in press), and previous research has specifically supportedapplications of SCCT in samples of Portuguese students (e.g., Paixão, da Silva, & Leitão, 2007; Vieira & Coimbra, 2008).

In sum, we tested the model depicted in Fig. 1 with a sample of Portuguese college students. However, we adapted themodel to a longitudinal context, examining the plausibility of the hypothesized temporal flow among the variables. In par-ticular, as shown in Fig. 2, we examined the degree to which the predictors could account for change (rather than concurrentstatus) in the criterion variables. This was accomplished by examining hypothesized lagged paths among the variables (e.g.,the path from self-efficacy at time 1 to goal progress at time 2), while controlling for the autoregressive paths (e.g., the rela-tion of goal progress at time 1 with goal progress at time 2). The inclusion of the autoregressive paths provides a rigorous testof the presumed antecedents of a given construct, particularly if the construct is fairly stable between times 1 and 2 and ifthe predictors are highly related to one another and the criterion at time 1.

Controlling for autoregressive paths, we predicted that: (a) life satisfaction at time 2 (T2) would be predicted by time 1(T1) academic domain adjustment, goal progress, and trait positive affect; (b) domain adjustment at T2 (indexed by aca-demic domain satisfaction, stress, and self-reported functioning) would be predicted by T1 progress at one’s academic goals,self-efficacy, environmental supports, and positive affect; (c) T2 goal progress would be predicted by T1 self-efficacy andenvironmental supports; (d) T2 self-efficacy would be predicted by T1 environmental supports and positive affect; and(e) T2 environmental supports would be predicted by T1 positive affect. Parenthetically, the outcome expectations variablewas not included in the present study because it had not been found to be uniquely predictive of goal progress or domainsatisfaction in previous studies (Lent, Singley et al., 2005, 2007).

In addition to the above predictions, we also tested four sets of reciprocal paths, which are shown as dashed paths inFig. 2: (f) T2 adjustment would be predicted by T1 life satisfaction, (g) T2 self-efficacy would be predicted by T1 goal pro-gress; and T2 positive affect would be predicted by T1 (h) self-efficacy and (i) environmental support. The first two setsof reciprocal relations reflect theoretical assumptions that self-efficacy is both a source and an effect of goal progress andthat positive adjustment is both a cause and an effect of life satisfaction (Lent, 2004). The third and fourth sets of reciprocalpaths, though not formally a part of Lent’s original model, test the possibility that self-efficacy and perceived environmentalsupport can promote change in positive affect (i.e., that trait-social cognitive links can be bidirectional). The latter links arebased on findings that positive affect is not static over time and that it is responsive to social factors (Watson, 2002). More-over, social cognitive theory assumes that self-efficacy beliefs both influence and are influenced by affect (Bandura, 1997).

2. Methods

2.1. Participants

Participants were 252 students (217 women, 35 men) enrolled in undergraduate studies in either psychology (n = 129),pre-medicine (n = 24), or education (n = 99) at a university in northern Portugal. The sample included 48 freshmen, 76 soph-omores, 46 juniors, 81 seniors, and one special status student. Their mean age was 21.48, SD = 4.75. Ninety percent of theparticipants were Portuguese, with the remainder from a variety of other countries (e.g., in Africa and North and SouthAmerica).

Fig. 2. Longitudinal version of the integrative model of well-being tested in this study. Note. Dotted lines = autoregressive paths; dashed paths = hypoth-esized reciprocal paths. Covariances among variables at time 1 and time 2 are not shown. Adjust = perceived adjustment.

R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198 193

2.2. Procedure and instruments

Students were recruited for participation by the second author within intact classes. They did not receive incentives toparticipate in the study. Data were gathered at each of two assessments, during the 1st and 16th weeks of the same (Spring)academic semester. We selected Spring semester so that students would have some prior college experience as a basis fortheir responses to the first assessment. The 15-week interval was chosen because domain and life satisfaction are often fairlystable over brief time periods. We reasoned that 15 weeks might be sufficient to capture some variability in these variablesbut not so long as to extend beyond the semester and make it difficult to acquire time 2 responses. There were 360 and 252participants, respectively, in the first and second assessments (70% of the initial participants provided data at both assess-ments). Given missing data, the final sample size for the main analyses was 248. Results of a multivariate analysis of varianceindicated that students who participated in both assessments did not differ from those who completed the measures only atT1 across the set of theoretical variables, F(9,345) = .52, Wilks’ k = .99, p = .86.

Participants were administered the same battery of measures in groups at both assessments. Measures included academicself-efficacy, goal progress, environmental support, satisfaction, stress, and self-reported adjustment related to students’ aca-demic behavior; life satisfaction and trait positive affect; and demographic and academic status information. The measureswere originally developed in English, then translated into Portuguese by the second author and back-translated into Englishby a Portuguese/English translator. The first author reviewed the back-translation to ensure that the translated measuresretained their original meaning. For each multi-item measure, total scores were obtained by summing item responsesand dividing by the number of items on the measure. Higher scores on all measures reflected more positive expectationsor experiences (e.g., stronger self-efficacy, greater satisfaction, lower stress).

2.2.1. Self-efficacySelf-efficacy was assessed with a Portuguese version of the academic self-efficacy measure used by Lent, Singley et al.

(2005). It contained 11 items asking students to indicate their confidence in their ability to perform well academicallyand to cope with barriers to academic success (e.g., ‘‘do your best . . . during the next semester;” deal with lack of supportby the teachers or supervisors). Responses were obtained along a 10-point scale, ranging from no confidence (0) to completeconfidence (9). Prior versions of this measure have yielded adequate internal consistency reliability estimates (e.g., a = .91)and shown theory-consistent relations with measures of academic outcomes (e.g., Lent, Brown et al., 2005). In the currentsample, we found coefficient alphas for the academic self-efficacy scale at T1 and T2, respectively, of .87 and .90.

2.2.2. Goal progressGoal progress was assessed with an 8-item measure that asked participants to indicate how much progress they were

making at a variety of academic goals (e.g., ‘‘actively participate in class”). Participants indicated their progress relative toeach goal statement, from 1 (no progress) to 5 (excellent progress). Lent, Singley et al. (2005) found that this measure producedinternal consistency reliability estimates of .84 to .86 and correlated with measures of academic self-efficacy, outcomeexpectations, environmental supports, and domain satisfaction. In the current sample, the goal progress scale yielded coef-ficient alpha values of .80 and .85, respectively, at T1 and T2.

2.2.3. Environmental supportsEnvironmental supports in the academic domain were assessed with a 9-item measure describing a set of conditions

‘‘that may support your course progress.” A sample item was ‘‘am encouraged by my friends to go on with my studies.” Par-ticipants indicated how much they agreed with each statement, from 1 (strongly disagree) to 5 (strongly agree). The versionused by Lent, Singley et al. (2005) yielded adequate reliability estimates (.81–.84) and correlated with measures of self-effi-cacy, outcome expectations, goal progress, and domain satisfaction. The internal consistency estimate of this scale in the cur-rent study was .76 at T1 and .81 at T2.

2.2.4. Domain adjustmentDomain adjustment was indexed with measures of academic domain satisfaction, stress, and perceived functioning. A 7-

item measure was used to assess academic satisfaction. It asked participants to indicate their level of satisfaction with var-ious aspects of their academic experience (e.g., ‘‘In general, I am satisfied with my academic life”) along a 1 (strongly disagree)to 5 (strongly agree) scale. Lent, Singley et al. (2005) found that the measure yielded adequate reliability estimates(a = .86�.87) and correlated as expected with measures of positive affect, social domain satisfaction, and overall life satis-faction. We found alpha coefficients of .85 and .89 for this measure, respectively, at T1 and T2 in our sample.

The Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983) was used to measure stress related to the aca-demic domain. The PSS contains 4-items that reflect general stress; we modified the items slightly to link the stress expe-rience to academics (e.g., ‘‘How often did you feel that academic difficulties were piling up in such a way that you could notovercome them?”). Ratings were made on a 1–5 scale, from 1 = never to 5 = frequently. Item responses were reversed so thathigher scores would reflect less stress. Cohen et al. (1983) found that the PSS yielded adequate reliability estimates (a = .84and above in three samples) and correlated with indicators of distress (e.g., depression) and physical problems. It has alsobeen used to index college adjustment in prior research (Aspinwall & Taylor, 1992). In our sample, T1 and T2 coefficient al-pha values were, respectively, .75 and .76.

194 R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198

As a third indicator of adjustment, we asked participants to indicate how well they felt they were adjusting to the overallacademic demands at their university, using a single item (‘‘generally speaking, I would describe my academic adjustmentas. . .”) developed for this study. Participants responded along a 5-point scale, from 1 = relatively poor to 5 = excellent.

2.2.5. Positive affectWe used the Positive Affect (PA) scale of the Positive and Negative Affect Schedule (Watson, Clark, & Tellegen, 1988) to

assess the tendency to experience positive emotions. Participants were asked to indicate the extent to which they generallyfeel 10 positive emotions (e.g., ‘‘proud”) along a 5-point continuum (1 = very little or not at all; 5 = extremely). PA has yieldedadequate internal consistency and stability coefficients and correlated as expected with measures of emotional distress andextraversion in prior research (Watson & Clark, 1992; Watson et al., 1988). Lent, Singley et al. (2005) found that the PA scalewas related to life satisfaction as well as academic self-efficacy and environmental supports. Internal consistency estimatesfor PA in this sample were .86 at both T1 and T2.

2.2.6. Life satisfactionThe Satisfaction With Life Scale (SWLS; Diener, Emmons, Larsen, & Griffen, 1985) was used to assess global life satisfac-

tion. Participants indicated their level of agreement with each of the five items (e.g., ‘‘I am satisfied with my life”) along a 7-point scale (1 = strongly disagree; 7 = strongly agree). The SWLS has produced adequate internal consistency, temporal stabil-ity, and validity estimates in past research (e.g., Compton, Smith, Cornish, & Qualls, 1996; Diener et al., 1985). Coefficientalpha values in our sample were .88 at T1 and .91 at T2.

3. Results

The means, standard deviations, and correlations among the measures at T1 and T2 are presented in Table 1. To test theeducational adjustment model longitudinally, we compared the model-data fit of three model variations. First, consistentwith prior longitudinal analyses (e.g., Nauta, Kahn, Angell, & Cantarelli, 2002; Sher, Wood, Wood, & Raskin, 1996), we createda base model containing covariances among all of the T1 variables, covariances among the errors of the T2 variables, andautoregressive paths, indexing the relation of each T1 variable to the same variable at T2. The base model estimates the sta-bility of each variable over a 15-week interval.

Second, the bidirectional model, shown in Fig. 2, builds on the base model by adding the hypothesized cross-lagged pathsfrom the T1 to T2 variables, including the reciprocal paths from (a) T1 goal progress to T2 self-efficacy, (b) T1 life satisfactionto T2 adjustment, (c) T1 self-efficacy to T2 positive affect, and (d) T1 environmental supports to T2 positive affect. The thirdmodel, termed the unidirectional model, was identical to the second model, only without the reciprocal paths. The second andthird models were used to test whether particular sets of relationships are either bidirectional (e.g., self-efficacy and positiveaffect relate to one another reciprocally) or unidirectional (e.g., positive affect is an antecedent of self-efficacy but not viceversa).

All models were tested using the path analysis procedures of EQS 6.1 (Bentler & Wu, 2005) and the covariance matrices.Because the value of Mardia’s normalized estimate reflected multivariate kurtosis, we used robust maximum likelihood esti-mation. To reduce the number of parameter estimates in relation to sample size, observed variables were employed to rep-resent most constructs. Educational adjustment was, however, modeled as a latent variable, with academic satisfaction,stress, and perceived adjustment as its observed indicators; the factor loading for satisfaction was fixed to 1, and the other

Table 1Means, Standard Deviations, and Correlations Among Predictor and Criterion Variables.

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 M SD

1. T1 Self-efficacy – 6.27 1.152. T1 Goal progress .61 – 3.45 .483. T1 Support .23 .4 – 3.89 .474. T1 Acad. satisfac. .3 .46 .49 – 3.82 .595. T1 Stress .41 .48 .25 .32 – 3.36 .676. T1 Acad. adjust. .38 .49 .3 .42 .46 – 3.48 .767. T1 Positive affect .48 .56 .42 .47 .49 .41 – 3.39 .558. T1 Life satisfac. .23 .36 .49 .44 .41 .47 .45 – 4.76 1.29. T2 Self-efficacy .71 .48 .3 .24 .42 .32 .47 .28 – 6.03 1.2210. T2 Goal progress .45 .54 .37 .33 .35 .28 .45 .2 .56 – 3.4 .511. T2 Support .17 .25 .66 .4 .26 .21 .29 .41 .34 .35 – 3.91 .5112. T2 Acad. satisfac. .21 .33 .46 .74 .27 .25 .34 .33 .28 .42 .54 – 3.86 .6213. T2 Stress .35 .33 .2 .24 .7 .38 .38 .37 .47 .37 .33 .27 – 3.34 .6614. T2 Acad. adjust. .39 .38 .28 .3 .34 .55 .32 .3 .44 .44 .26 .3 .35 – 3.53 .6815. T2 Positive affect .44 .41 .34 .33 .36 .22 .61 .27 .61 .62 .41 .47 .43 .31 – 3.41 .5316. T2 Life satisfac. .24 .35 .46 .4 .47 .39 .4 .82 .38 .31 .46 .45 .46 .36 .43 – 4.88 1.19

Note. N = 248. T1, time 1; T2, time 2; effic., efficacy; acad., academic; adjust., adjustment; satisfac., satisfaction. Scale ranges were 0-9 for self-efficacy, 1–7for life satisfaction, and 1–5 for all other scales. Correlations P .17 are significant, p < .01; r P .21, p < .001.

R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198 195

loadings were freely estimated. (Covariances among the errors of like adjustment indicators across time were also modeled,e.g., the correlated errors of academic satisfaction at T1 and T2.) Two primary fit indices were employed: the comparative fitindex (CFI) and the root mean square error of approximation (RMSEA). According to Hu and Bentler (1999), CFI values closeto .95 and RMSEA values close to .06 imply good model-data fit (Hu & Bentler, 1999).

The base model yielded good fit to the data on the CFI (.95) and an acceptable RMSEA value (.07); Satorra–Bentler scaledv2(75) = 157.39, p < .001. All of the autogressive paths between the T1 and T2 variables were significant and moderate tolarge in magnitude, suggesting that scores on these variables were moderately to highly stable over the 15-week interval(the autoregressive path coefficients for positive affect, environmental supports, self-efficacy, goal progress, adjustment,and life satisfaction were, respectively, .52, .62, .65, .47, .79, and .81).

The unidirectional model fit the data well across both primary fit indices, CFI = .96, RMSEA = .07, Satorra–Bentler scaledv2(63) = 130.88, p < .001. Contrasting the unidirectional and base models with the corrected difference in Satorra–Bentler v2

(Satorra & Bentler, 2001), we found that the unidirectional model yielded improved fit over the base model, Dv2(12) = 26.53,p < .05. The bidirectional model also produced good fit on both the CFI (.97) and RMSEA (.06) indices; Satorra–Bentler scaledv2(59) = 110.17, p = .001. Like the unidirectional model, it was found to yield improved fit compared to the base model,Dv2(16) = 48.17, p < .05. However, the bidirectional model was also found to improve upon the fit of the unidirectional mod-el, Dv2(4) = 23.43, p < .05. Thus, the results of the model tests suggest that the bidirectional model provides the best fit to thedata.

Fig. 3 displays a trimmed version of the bidirectional model, deleting the non-significant lagged paths. As can be seen,consistent with expectations (and with the autoregressive paths controlled), T2 self-efficacy was predicted by positive affect(.11) and environmental supports (.10); T2 goal progress was predicted by environmental supports (.18) and self-efficacy(.20); T2 positive adjustment was predicted by environmental supports (.17) and self-efficacy (.17); and T2 life satisfactionwas predicted only by positive adjustment (.37). Although significant (p < .05, 1-tailed), each of these path coefficients wasfairly modest in magnitude (1-tailed tests were used in light of our directional hypotheses).

Contrary to expectations, there were non-significant paths from T1 positive affect to T2 environmental supports (.00),adjustment (�.02), and life satisfaction (�.11); and from T1 goal progress to T2 adjustment (.02) and life satisfaction(�.09). Thus, neither positive affect nor goal progress added directly to the prediction of the positive functioning outcomes,above and beyond the other variables. In addition, although the bidirectional model produced the best fit to the data, weobserved that only two of the four sets of posited reciprocal paths were confirmed by the data. In particular, T2 positive affectwas predicted by T1 environmental supports (.10) and self-efficacy (.21). However, the reciprocal linkages of T1 goal pro-gress to T2 self-efficacy (.00) and T1 life satisfaction to T2 adjustment (.10) were each non-significant.

4. Discussion

These findings extend earlier, cross-sectional tests of the unifying well-being model (Lent, Brown et al., 2005, 2007; Lent,Singley et al., 2005) by showing that the model fits the data well (a) within a longitudinal context, (b) when expanding thedefinition of domain well-being to include perceived academic stress and adjustment as well as satisfaction, and (c) whenapplied to Portuguese college students. The findings are also relevant to the broader literatures on domain and life satisfac-

Fig. 3. Significant coefficients in the path analysis of the bidirectional model. Note. Dotted lines = autoregressive paths; dashed paths = hypothesizedreciprocal paths. Covariances among variables at time 1 and time 2 are not shown. Adjust = perceived adjustment.

196 R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198

tion in that the model we tested attempts to integrate predictors and causal paths suggested by diverse theoretical ap-proaches (e.g., Diener et al., 1999; Ryan & Deci, 2000).

Examining the specific lagged paths in the model, we observed that positive academic adjustment predicted overall lifesatisfaction. Given the centrality of the academic domain for college students, it is not surprising that they would feel moresatisfied with their lives overall when they were functioning well in the academic realm. Academic adjustment was itselfpredicted by academic self-efficacy and by access to support for one’s academic goals. That is, students reported gains intheir academic functioning when they possessed stronger self-efficacy and environmental support. These findings are con-sistent with prior results linking self-efficacy and/or support to indicators of positive domain adjustment in college students(e.g., Aspinwall & Taylor, 1992; Brown et al., 2008; Lent, Singley et al., 2005).

In contrast to earlier findings (Lent, Singley et al., 2005, 2007), progress at academic goals did not contribute directly toacademic adjustment. This discrepancy could have been due to the fact that we operationalized domain adjustment some-what differently than in the prior studies in this line of research or that goal progress was simply less useful a predictor ofadjustment in the present context once role-related self-efficacy, environmental support, and the autoregressive paths werecontrolled. It is also possible that the predetermined, relatively general goals contained in the goal progress scale were not asmeaningful to participants as their own, more specific, self-stated goals might have been. Changes in goal progress were,however, predicted by self-efficacy and environmental supports, which is consistent with the linkages among these variablesfound in cross-sectional studies (Lent, Singley et al., 2005, 2007). In particular, stronger self-efficacy and greater access tosupport were associated with better goal progress.

As in past research, self-efficacy was predicted by environmental supports (Lent, Singley et al., 2005, 2007) and positiveaffect (Lent, Singley et al. (2005)). Thus, self-efficacy may have been bolstered by access to supports and by the cognitivefeatures of positive affect, which presumably nurture more favorable self-percepts (Lent, 2004). However, contrary to pre-dictions and prior findings, positive affect did not yield a direct path either to academic adjustment (Aspinwall & Taylor,1992) or life satisfaction (Lent, Singley et al. (2005)). Its primary role in the current study appeared to be indirect—that is,as a precursor of self-efficacy—rather than as a direct contributor to adjustment.

A particularly interesting finding was that self-efficacy and positive affect were reciprocally related to one another; thatis, self-efficacy both predicted and was predicted by positive affect. Moreover, environmental supports also predicted posi-tive affect (but not vice versa). Given that positive affect is typically considered to be a personality trait (Watson et al., 1988),one might expect it to be a precursor of other, presumably state-like variables, like self-efficacy and perceptions of socialsupport, rather than the other way around. However, stability coefficients indicate that positive affect is not an immutableperson attribute (Veenhoven, 1994), and Watson (2002) has suggested that it can be modified by particular self-directed andsocial activities, such as social support. It is also noteworthy that positive affect is often viewed as a facet of subjective well-being (Diener et al., 1999). Thus, our findings that positive affect may be responsive to cognitive and social variables holdpotential theoretical and practical import.

From a practice perspective, the current findings suggest that academic adjustment (defined in terms of domain satisfac-tion, overall role functioning, and absence of stress) and positive affect may be promoted by interventions that enhance self-efficacy and provide academic social support (e.g., access to role models, encouragement from peers). According to our find-ings, support may be helpful both as a source of self-efficacy and as a direct facilitator of adjustment. Social cognitive theory(Bandura, 1997) and its application to the career development context (Brown & Lent, 1996) suggest additional theoreticalelements that can be used to foster favorable self-efficacy and, thereby, promote enhanced role functioning. Examples ofsuch elements include providing new (or reviewing prior) personal performance accomplishments, challenging non-optimalperformance attributions, and reducing elevated performance anxiety.

Interpretation and application of these findings should be viewed in the context of the study’s limitations. First, althoughthe educational adjustment model was found to be tenable within our Portuguese sample, and steps were taken to ensurethe linguistic and cultural relevance of the measures we used, more research is needed to establish the range of the model’scross-cultural applicability and the psychometric properties of the measures with international samples. It would be partic-ularly valuable to examine the measurement and conceptual equivalence of the model’s variables across samples from dif-ferent cultures or nations (cf. Miller & Sheu, 2008). It would also be useful to employ multiple-group path analysis (e.g.,involving two or more cultural groups in the same study) so that possible differences in model fit or path coefficients canbe examined (Lent & Sheu, in press).

Second, efforts to generalize the study’s findings should consider the gender composition of our sample (86% of the par-ticipants were women). Further study is needed to examine the extent to which the findings replicate in male samples. (Priorcomparisons of other SCCT models have found comparable model-data fit across gender, e.g., Lent, Brown et al., 2005; Lent,Lopez, Lopez, & Sheu, 2008; Lent, Singley et al., 2005). Third, all of the variables in this study were assessed via self-report.Although it can be argued that most of the theoretical variables (e.g., self-efficacy, positive affect, satisfaction) are properlyand validly measured from the perspective of the individual, adjustment could be indexed in behavioral terms (e.g., gradeperformance, persistence) and from an external perspective (Lent, 2004). It would, therefore, be useful to include such broad-ened indictors of adjustment in future research on the unifying well-being model.

A fourth limitation is that there were only two assessment points in this study, which precluded an adequate test of medi-ating paths. For example, the theory suggests that positive affect leads to more favorable self-efficacy beliefs which, in turn,foster goal progress and domain adjustment. Although our findings did indicate that positive affect was associated with sub-sequent self-efficacy, and that self-efficacy predicted change in goal progress and domain adjustment, we could not directly

R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198 197

test the mediational sequence hypothesized by the theory. In order to do so, it would be valuable for future research to in-clude at least three measurement points, which can explicitly test, for example, whether T1 positive affect predicts T3 adjust-ment through its relation to T2 self-efficacy (cf. Cole & Maxwell, 2003). A fifth limitation is inherent in longitudinal designs:while such designs can be used to test the plausibility of causal relations, they cannot conclusively demonstrate causality.

Sixth, although we found that the educational adjustment model, especially the bidirectional version, provided goodoverall fit to the data, several of the individual paths yielded non-significant coefficients, and the reciprocal links fromself-efficacy and environmental supports to positive affect represented additions to Lent’s (2004) original hypotheses. Fur-ther research testing the model would help to clarify its explanatory adequacy. Such research might, for example, examinealternative, theoretically plausible, variations of the model or contrast the basic model with other theoretical models of well-being or college student adjustment. Finally, our operationalization of several of the variables could have been improved. Inparticular, the environmental support measure focused on access to general types of academic support (e.g., models, tutors,peers, social encouragement) rather than resources specifically linked to students’ personal goals. Likewise, as noted above,the goal progress measure may not have optimally represented goals that were of central relevance to individualparticipants.

Despite these limitations, the present study did provide a fairly rigorous test of the temporal relations of the predictors tothe criterion variables because, unlike prior cross-sectional research on the well-being model, the predictors and criteriawere assessed at separate times (rather than concurrently) and the autoregressive paths (i.e., change or stability in the cri-terion variables over time) were controlled. Although the lagged path coefficients were relatively small to moderate in size,they were comparable to those found in other longitudinal studies of social cognitive variables (e.g., Lent, Sheu et al., 2008;Nauta et al., 2002). In essence, it is difficult to account for large portions of variance in a dependent variable within a lon-gitudinal design when there is a great deal of stability in that variable over the predictive interval (Sher et al., 1996). Buteven modest lagged paths can suggest practically useful routes to promote adjustment and well-being outcomes.

Given the early stage of research on Lent’s (2004) well-being model and its career-specific elaboration (Lent & Brown,2006), it would be valuable for future longitudinal research to examine the temporal relations among the variables in bothstudent and worker samples. Such research might vary the interval between assessments and examine the relations amongthe variables in different developmental contexts. It is possible, for example, that stronger T1 to T2 relations will be revealedat earlier stages of adjusting to a new environment or during periods of relative flux in life roles (e.g., the first semester ofcollege or the period immediately after work entry). It would also be useful to assess goal progress in relation to self-statedgoals. Finally, it would be valuable to study the normative and restorative versions of the well-being model using experimen-tal designs, including controlled interventions. Such research could offer a more stringent test of causal hypotheses, togetherwith an exploration of the model’s clinical utility, for example, in facilitating the adjustment of college students who areexperiencing untoward academic difficulties in their new environment.

References

Aspinwall, L. G., & Taylor, S. E. (1992). Modeling cognitive adaptation: A longitudinal investigation of the impact of individual differences and coping oncollege adjustment and performance. Journal of Personality and Social Psychology, 63, 989–1003.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W.H. Freeman.Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 1–26.Bentler, P. M., & Wu, E. J. C. (2005). EQS 6.1 for Windows user’s guide. Encino, CA: Multivariate Software.Brown, S. D., & Lent, R. W. (1996). A social cognitive framework for career choice counseling. The Career Development Quarterly, 44, 354–366.Brown, S. D., Tramayne, S., Hoxha, D., Telander, K., Fan, X., & Lent, R. W. (2008). Social cognitive predictors of college students’ academic performance and

persistence. A meta-analytic path analysis. Journal of Vocational Behavior, 72, 298–308.Brunstein, J. C. (1993). Personal goals and subjective well-being: A longitudinal study. Journal of Personality and Social Psychology, 65, 1061–1070.Cantor, N., & Sanderson, C. A. (1999). Life task participation and well-being: The importance of taking part in daily life. In D. Kahneman, E. Diener, & N.

Schwarz (Eds.), Well-being: The foundations of hedonic psychology (pp. 230–243). New York: Russell Sage Foundation.Carver, C. S., & Scheier, M. F. (2002). Optimism. In C. R. Snyder & S. J. Lopez (Eds.), Handbook of positive psychology (pp. 231–243). New York: Oxford

University Press.Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396.Cole, D. A., & Maxwell, S. E. (2003). Testing mediational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal

of Abnormal Psychology, 112, 558–577.Compton, W. C., Smith, M. L., Cornish, K. A., & Qualls, D. L. (1996). Factor structure of mental health measures. Journal of Personality and Social Psychology, 71,

406–413.Diener, E., Emmons, R. A., Larsen, R. J., & Griffen, S. (1985). The Satisfaction With Life Scale. Journal of Personality Assessment, 49, 71–75.Diener, E., & Fujita, F. (1995). Resources, personal strivings, and subjective well-being: A nomothetic and idiographic approach. Journal of Personality and

Social Psychology, 68, 926–935.Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276–302.Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural

Equation Modeling, 6, 1–55.Lent, R. W. (2004). Toward a unifying theoretical and practical perspective on well-being and psychosocial adjustment. Journal of Counseling Psychology, 51,

482–509.Lent, R. W., & Brown, S. D. (2006). Integrating person and situation perspectives on work satisfaction: A social-cognitive view. Journal of Vocational Behavior,

69, 236–247.Lent, R. W., & Brown, S. D. (2008). Social cognitive career theory and subjective well-being in the context of work. Journal of Career Assessment, 16, 6–21.Lent, R. W., Brown, S. D., & Hackett, G. (1994). Toward a unifying social cognitive theory of career and academic interest, choice, and performance

[Monograph]. Journal of Vocational Behavior, 45, 79–122.Lent, R. W., Brown, S. D., Sheu, H., Schmidt, J., Brenner, B. R., Gloster, C. S., et al (2005). Social cognitive predictors of academic interests and goals in

engineering: Utility for women and students at historically Black universities. Journal of Counseling Psychology, 52, 84–92.

198 R.W. Lent et al. / Journal of Vocational Behavior 74 (2009) 190–198

Lent, R. W., Lopez, A. M., Lopez, F. G., & Sheu, H. (2008). Social cognitive career theory and the prediction of interests and choice goals in the computingdisciplines. Journal of Vocational Behavior, 73, 52–62.

Lent, R.W., & Sheu, H. (in press). Applying social cognitive career theory across cultures: Empirical status. In J.G. Ponterotto, J.M. Casas, L.A., Suzuki, & C.MAlexander, (Eds.), Handbook of multicultural counseling (3rd ed.). Thousand Oaks, CA: Sage.

Lent, R. W., Sheu, H., Singley, D., Schmidt, J. A., Schmidt, L. C., & Gloster, C. S. (2008). Longitudinal relations of self-efficacy to outcome expectations, interests,and major choice goals in engineering students. Journal of Vocational Behavior, 73, 328–335.

Lent, R. W., Singley, D., Sheu, H., Gainor, K., Brenner, B. R., Treistman, D., et al (2005). Social cognitive predictors of domain and life satisfaction: Exploring thetheoretical precursors of subjective well-being. Journal of Counseling Psychology, 52, 429–442.

Lent, R. W., Singley, D., Sheu, H., Schmidt, J. A., & Schmidt, L. C. (2007). Relation of social-cognitive factors to academic satisfaction in engineering students.Journal of Career Assessment, 15, 87–97.

McCrae, R. R., & Costa, P. T. (1991). Adding liebe und arbeit: The full five-factor model and well-being. Personality and Social Psychology Bulletin, 17, 227–232.Miller, M. J., & Sheu, H. (2008). Conceptual and measurement issues in multicultural psychology research. In S. D. Brown & R. W. Lent (Eds.), Handbook of

Counseling Psychology (4th ed., pp. 103–120). New York: Wiley.Nauta, M. M., Kahn, J. H., Angell, J. W., & Cantarelli, E. A. (2002). Identifying the antecedent in the relation between career interests and self-efficacy: Is it one,

the other, or both? Journal of Counseling Psychology, 49, 290–301.Paixão, M. P., da Silva, J. T., & Leitão, L. M. (2007). Occupational considerations in a sample of Portuguese adolescents: Replicating the Lent et al. (2003) study.

Paper presented at the meeting of the International Association for Educational and Vocational Guidance, University of Padova, Padova, Italy.Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American

Psychologist, 55, 68–78.Ryff, C. D., & Singer, B. (1998). The contours of positive human health. Psychological Inquiry, 9, 1–28.Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66, 507–514.Sher, K. J., Wood, M. D., Wood, P. K., & Raskin, G. (1996). Alcohol outcome expectancies and alcohol use: A latent variable cross-lagged panel study. Journal of

Abnormal Psychology, 105, 561–574.Veenhoven, R. (1994). Is happiness a trait? Tests of the theory that a better society does not make people any happier. Social Indicators Research, 32,

101–160.Vieira, D., & Coimbra, J. L. (2008). La transition entre l’enseignement supérieur et l’emploi: Auto-efficacité, attentes de résultats et objectifs professionnels

[School-to-work transition: Self-efficacy, outcome expectations, and goals]. L’orientation Scolaire et Professionnelle, 37, 91–112.Watson, D. (2002). Positive affectivity: The disposition to experience pleasurable emotional states. In C. R. Snyder & S. J. Lopez (Eds.), Handbook of positive

psychology (pp. 106–119). New York: Oxford University Press.Watson, D., & Clark, L. A. (1992). On traits and temperament: General and specific factors of emotional experience and their relation to the five-factor model.

Journal of Personality, 60, 441–476.Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of

Personality and Social Psychology, 54, 1063–1070.