دانلود مقاله ISI انگلیسی شماره 38286
عنوان فارسی مقاله

مقیاس کارگر:توسعه یک اقدام برای توضیح تفاوت های جنسیتی در خود ناتوان سازی رفتاری

کد مقاله سال انتشار مقاله انگلیسی ترجمه فارسی تعداد کلمات
38286 2008 22 صفحه PDF سفارش دهید محاسبه نشده
خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.
عنوان انگلیسی
The worker scale: Developing a measure to explain gender differences in behavioral self-handicapping
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Journal of Research in Personality, Volume 42, Issue 4, August 2008, Pages 949–970

کلمات کلیدی
خود ناتوان سازی - تفاوت های جنسیتی - دست آورد
پیش نمایش مقاله
پیش نمایش مقاله مقیاس کارگر:توسعه یک اقدام برای توضیح تفاوت های جنسیتی در خود ناتوان سازی رفتاری

چکیده انگلیسی

Abstract Research has consistently found that men engage in more behavioral self-handicapping than do women. We first review evidence suggesting that these gender differences result from women placing more importance on displaying effort than do men. We then present the results of two studies seeking to develop measures of beliefs about effort that might explain these gender differences in behavioral self-handicapping. Women, across a wide range of measures, placed more importance on effort than did men. However, only a new measure of more personalized effort beliefs, dubbed the Worker scale, uniquely explained gender differences in dispositional tendency to behaviorally self-handicap. The Worker scale also predicted academic performance, consistent with the notion that these effort beliefs would predict engagement in actual behavioral self-handicaps that undermine performance.

مقدمه انگلیسی

Introduction Self-handicapping involves the creation or claiming of obstacles to successful performance, in the hopes that any subsequent failure can be blamed on the handicap (Jones & Berglas, 1978). For example, individuals have been shown to take drugs or alcohol (Berglas & Jones, 1978), withhold practice effort (Pysczczynski & Greenberg, 1983), or claim handicaps such as stress (Hirt, Deppe, & Gordon, 1991) or bad mood (Rosenfarb & Aron, 1992) prior to important tests. The more active forms of self-handicapping such as effort withdrawal or drug use have been referred to as behavioral or acquired handicaps, whereas claims of the presence of an obstacle such as stress or bad mood have been referred to as claimed or self-reported handicaps (Arkin and Baumgardner, 1985 and Leary and Shepperd, 1986). This theoretical distinction is important, in that behavioral self-handicaps are generally more observable, controllable, and likely to be directly related to performance than are claimed handicaps (Hirt et al., 1991 and Leary and Shepperd, 1986). With regard to performance, recent evidence increasingly has shown that behavioral self-handicaps, in particular reduced study effort, impair academic performance in the long term (McCrea and Hirt, 2001, Murray and Warden, 1992, Urdan, 2004 and Zuckerman et al., 1998). This theoretical distinction is also critical with regard to several individual differences that exist in self-handicapping behavior. Jones and Rhodewalt (1982) developed the self-handicapping scale (SHS) to measure the tendency to engage in self-handicapping and excuse-making more generally. A number of studies have demonstrated the predictive validity of the scale (see Rhodewalt, 1990 for a review). Recently, McCrea, Hirt, and Hendrix (2006) factor analyzed the SHS, showing that the scale is comprised of two subscales corresponding to the distinction between claimed and behavioral self-handicapping (see Appendix). These subscales were specifically predictive of corresponding forms of self-handicapping. For example, use of claims of stress as a handicap were predicted only by the claimed subscale of the SHS, whereas use of reduced practice effort as a handicap was predicted only by the behavioral subscale of the SHS. The prediction of these self-handicapping behaviors was also improved when using the corresponding SHS subscale rather than the complete 25-item scale. Furthermore, the behavioral subscale was more predictive of lower academic performance than was the claimed subscale (McCrea et al., 2006). One of the most consistent individual differences found in the self-handicapping literature is that men are more likely than women to engage in behavioral forms of self-handicapping, whereas both men and women are equally likely to engage in claimed self-handicapping (see Harris and Snyder, 1986, Hirt et al., 1991, McCrea et al., in press and Rhodewalt, 1990). Similarly, men tend to score higher on the behavioral subscale of the SHS, whereas women tend to score higher on the claimed subscale of the SHS (McCrea et al., 2006). To date, an explanation for these findings has not been forthcoming. Previous research has largely examined possible reasons why men would be more motivated to self-handicap than women, either because of differences in the importance of the task domain (Dietrich, 1995, Hirt, 1993 and Kimble et al., 1990), differences in the attributions made after self-handicapping by the self or others (Berglas and Jones, 1978 and Hirt et al., 2003) or differences in level of concern about managing public impressions of their ability (see Hirt et al., 2000, Rhodewalt, 1990, Rhodewalt and Davison, 1986 and Snyder et al., 1985). The present work takes a different approach to explaining this finding. Specifically, we suggest that women view certain types of self-handicapping behavior more negatively than do men, and that they therefore choose to self-handicap in other ways (see also Harris and Snyder, 1986 and Hirt et al., 1991). We first review prior research which supports this view. 1.1. Evidence supporting the role of effort beliefs in gender differences in self-handicapping Earlier work on the gender difference in behavioral self-handicapping largely focused on the idea that men experience more evaluative threat and thus have greater motivation to self-handicap. For example, based on the finding that self-handicapping is at least partly motivated by impression management concerns (Kolditz & Arkin, 1982), Hirt et al. (2000) examined whether increasing the evaluative threat of a performance via public self-focus would increase behavioral self-handicapping among men and women. Although men responded to public self-focus with increased behavioral self-handicapping, women did not. In contrast, women as well as men demonstrate increased claimed self-handicapping when placed under conditions of public self-focus (Koch, Hirt, & McCrea, 2003). Furthermore, studies examining behavioral self-handicapping in less stereotypically masculine domains have also observed the same gender difference (Dietrich, 1995 and Hirt, 1993). Thus, the gender difference in self-handicapping appears to be related to the characteristics of the handicap rather than differences in the motivation to protect the self. Similar conclusions are drawn from a recent study on observer reactions to behavioral self-handicapping (Hirt et al., 2003). One could argue that women do not expect the same attributional benefits of self-handicapping in terms of observer reactions, compared to men. Indeed, the failure of women tends to be attributed to lack of ability, whereas the failure of men tends to be attributed to lack of effort (Dweck et al., 1978 and Swim and Sanna, 1996). Women may therefore have less to gain from a self-handicapping strategy. To test this possibility, Hirt et al. (2003) examined whether observers would be more critical of a woman who behaviorally self-handicaps than a man who does so. Surprisingly, there were none of the expected effects of target gender, indicating that the self-handicap was equally unacceptable when used by a man or by a woman. However, female participants were much more critical of a self-handicapper of either gender. Although women accepted that the self-handicapper had the potential to do well in the future, they felt the target person lacked ability and was not particularly likeable. Women were more likely than were men to explain the self-handicapper’s behavior in terms of more enduring character flaws such as a lack of self-control or laziness, and to suspect that the individual was self-handicapping. These findings suggest that women, more than men, place considerable value on putting forth effort, and expect others to do so as well. As a result, women may choose to claim self-handicaps rather than self-handicap behaviorally because they view the latter behaviors more negatively (Hirt et al., 1991 and Hirt et al., 2003). Thus, a review of the self-handicapping literature suggests that women may place greater importance on putting forth effort in performance contexts, and this difference might explain the corresponding gender difference in behavioral self-handicapping, at least those handicaps involving reduced effort. Certainly, other researchers have also proposed that women appear to value effort more than do men. For example, female students report studying harder, procrastinate important tasks less, and adopt more effortful learning goals and strategies than do male students (Ablard and Lipschultz, 1998, Cooper et al., 1991, Stricker et al., 1993 and Zimmerman and Martinez-Pons, 1990). The present studies sought to examine these beliefs in more detail and thereby identify which specific effort beliefs might relate to behavioral self-handicapping and gender differences in their use. Although we focus on handicaps involving reduced effort, we suggest beliefs about effort more broadly explain gender differences in other types of behavioral self-handicapping, a point which we also address in the studies presented here. To examine this potential explanation for gender differences in behavioral self-handicapping, we first set out to develop and identify relevant measures of effort beliefs. There are of course a myriad of effort beliefs that might be relevant to self-handicapping and academic performance. In the following section, we detail each of the measures we considered in the present studies. 1.2. Types of effort beliefs First, individuals may differ in the extent to which they view effort as an end in itself. This belief could be more normative in nature, reflecting the impression that effort is respected by others. For example, there are pressures from parents and teachers to “always try your best.” These exhortations could presumably become internalized through socialization, such that the individual views effort as an important aspect of their self-concept and value system. Those placing more importance on effort would thus be less likely to self-handicap and more likely to negatively evaluate those who do. To our knowledge, there has not been any previous research attempting to examine gender differences in these types of beliefs. A related possibility is that self-handicapping behavior is influenced by beliefs concerning the effectiveness of effort in improving performance. Thus, effort might be seen as a tool for achievement. For example, the Protestant work ethic suggests that personal success is a result of hard work (Mirels & Garrett, 1971). On the other hand, no such gender differences have been found in Protestant work ethic beliefs (Mirels & Garrett, 1971). Similarly, individuals differ in the extent to which they believe abilities are malleable. Dweck and colleagues (e.g., Dweck & Leggett, 1988) have examined naı¨ve theories of intelligence and abilities, specifically whether individuals believe that they can improve over time as a result of learning (so-called incremental theorists) or that abilities are fixed (so-called entity-theorists). Individuals holding a more incremental-theorist view (and thus believing that effort enhances ability) may take a dim view of self-handicapping. There is some evidence to suggest that women are in fact more likely to subscribe to such an incremental theory perspective (Rhodewalt, 1994). Individuals may also view self-handicapping negatively simply because they generally find putting forth effort enjoyable. It is not clear whether there might be gender differences in this type of belief. Finally, women could value effort more than men because they are more concerned with their academic performance. Academics, at least in areas such as math, is considered stereotypically masculine (Swim & Sanna, 1996), and so women might be more concerned that they will not perform well based on their ability alone. Furthermore, female high school seniors report enjoying and placing more importance on school than do male students (U.S. Department of Education, 2004). Thus, it could be that women value effort more as a result of placing more emphasis on academic performance. In the present studies, we examined whether any of these measures exhibit gender differences. Next, we tested which, if any, of these measures were predictive of individual differences in the dispositional tendency to behaviorally self-handicap as measured by the behavioral subscale of the SHS (Jones and Rhodewalt, 1982 and McCrea et al., 2006). As discussed earlier, the behavioral subscale of the SHS is more predictive of actual behavioral self-handicapping and academic performance than is the claimed subscale of the SHS or the complete 25-item scale. Furthermore, prior studies have shown that men tend to score higher on the behavioral subscale of the SHS (McCrea et al., 2006). Thus, a measure seeking to explain gender differences in behavioral self-handicapping would ideally mediate gender differences on this subscale. Finally, given past findings that the tendency to self-handicap, particularly via behavioral self-handicapping, is linked to poor academic performance (e.g., McCrea and Hirt, 2001, Murray and Warden, 1992, Urdan, 2004 and Zuckerman et al., 1998), effort beliefs should also mediate the relationship between behavioral subscale scores and academic performance.

نتیجه گیری انگلیسی

Results 8.1. Confirmatory factor analysis To further test the factor structure of the three effort scales developed in Study 1, we conducted a confirmatory factor analysis using LISREL 8 (Jöreskog & Sörbom, 2005) on the items from the Worker, Prescriptive effort norm scales, and Ability vs. effort tradeoff scales. We fit a three-factor solution to the data using a maximum likelihood estimation method. We used a covariation matrix to run the analysis. For ease of interpretation, a correlation matrix is presented in Table 7. The three-factor model was specified with all factors correlated. Initial runs suggested that the model would be considerably improved if measurement errors for one pair of items were allowed to correlate. Namely, Worker scale item 1 (“I tend only to work as hard as I have to in my classes”, reversed) and Worker scale item 4 (“I push myself a lot to perform well academically”) shared a focus on putting forth effort in order to do well academically. This modified model had a significant chi-square, χ2=χ2=957.85, df = 148, p < .001. This significant test was not surprising given the large sample size. However, other indices revealed a good fit for the three-factor model (GFI = .94, CFI = .94, RMSEA = .06). Table 7. Correlation matrix, confirmatory factor analysis (Study 2) W1 W2 W3 W4 W5 W6 W7 W8 A1 A2 A3 A4 A5 A6 N1 N2 N3 N4 N5 W1 1.22 W2 .17 0.95 W3 .30 .54 1.04 W4 .39 .29 .39 1.32 W5 .28 .43 .54 .38 1.10 W6 .22 .53 .52 .31 .44 1.00 W7 .21 .39 .49 .30 .43 .45 1.06 W8 .08 .30 .31 .15 .31 .40 .27 1.21 A1 −.13 −.15 −.12 −.11 −.04 −.14 −.09 .00 1.33 A2 −.15 −.21 −.13 −.18 −.04 −.15 −.12 .00 .41 1.39 A3 −.17 −.16 −.13 −.17 −.08 −.20 −.13 −.05 .38 .34 1.21 A4 −.08 −.14 −.07 −.10 −.02 −.08 −.07 .05 .36 .33 .32 1.26 A5 −.13 −.06 −.02 −.12 −.01 −.07 −.06 −.05 .24 .15 .24 .19 1.25 A6 −.10 −.06 −.02 −.07 .00 −.07 −.05 −.02 .31 .24 .23 .23 .25 1.29 N1 −.01 .23 .18 .03 .16 .20 .16 .13 −.23 −.19 −.18 −.13 −.04 −.09 1.05 N2 .05 .45 .29 .09 .23 .31 .22 .17 −.14 −.23 −.15 −.19 −.07 −.05 .20 0.85 N3 .07 .18 .34 .08 .27 .22 .16 .17 −.08 −.05 −.11 −.05 −.01 .03 .19 .16 1.21 N4 .05 .30 .25 .00 .18 .24 .20 .12 −.22 −.22 −.21 −.25 −.08 −.08 .24 .37 .25 0.87 N5 .05 .35 .24 .05 .20 .41 .22 .15 −.21 −.21 −.18 −.22 −.05 −.06 .26 .38 .17 .44 0.84 Note: Standard deviations for each of the items are presented along the main diagonal. Table options As a comparison, we ran an identical analysis with a single-factor model (χ2=χ2=3549.62, df = 151, p < .001, GFI = .80, CFI = .83, RMSEA = .12). The single-factor model performed more poorly, with a higher chi-square (View the MathML sourceχchange2=2591.77, df = 3, p < .001) and lower fit indices than the three-factor solution. Thus, separating the Worker scale, Ability vs. effort tradeoff scale, and the Prescriptive effort norm scale appears to be well justified. 6 8.2. Relationships between variables Correlations between the measures are presented separately for women (Table 8) and men (Table 9). As in Study 1, the effort measures were only weakly to moderately correlated with each other and the other individual difference measures, with the exception that the Worker scale correlated strongly and positively with both Need for achievement and Conscientiousness. Correlations did not differ greatly between men and women. Table 8. Correlations between measures for women (Study 2) 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Worker — 2. Prescriptive effort norm .38‡ — 3. Ability vs. effort tradeoff −.23‡ −.27‡ — 4. Protestant work ethic .21‡ .32‡ −.01 — 5. Need for achievement .55‡ .25‡ −.23‡ .24‡ — 6. Concern .29‡ .26‡ −.13‡ .16‡ .27‡ — 7. Doubt −.26‡ −.01 .29‡ .08† −.26‡ −.16‡ — 8. Agreeableness .31‡ .27‡ −.24‡ .09† .18‡ .13‡ −.24‡ — 9. Conscientiousness .64‡ .29‡ −.26‡ .17‡ .49‡ .24‡ −.37‡ .40‡ — 10. Claimed subscale −.27‡ −.07∗ .43‡ .06 −.21‡ −.15‡ .54‡ −.33‡ −.39‡ — 11. Behavioral subscale −.71‡ −.27‡ .15‡ −.18‡ −.43‡ −.24‡ .22‡ −.23‡ −.56‡ .21‡ — 12. College entrance exam −.04 −.06∗ .02 −.07∗ .06∗ −.01 −.09† −.12‡ −.05 .00 .01 — 13. GPA −.32‡ −.05 .05 −.02 −.20‡ −.08† .12‡ −.04 −.21‡ .07∗ .26‡ −.26‡ — ∗ p < .05. † p < .01. ‡ p < .001. Table options Table 9. Correlations between measures for men (Study 2) 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Worker — 2. Prescriptive effort norm .42‡ — 3. Ability vs. effort tradeoff −.23‡ −.39‡ — 4. Protestant work ethic .28‡ .45‡ −.23‡ — 5. Need for achievement .59‡ .28‡ −.26‡ .35‡ — 6. Concern .28‡ .27‡ −.20‡ .24‡ .26‡ — 7. Doubt −.18‡ −.11∗ .32‡ .03 −.23‡ −.19‡ — 8. Agreeableness .30‡ .35‡ −.26‡ .15† .22‡ .13† −.23‡ — 9. Conscientiousness .61‡ .33‡ −.31‡ .31‡ .55‡ .37‡ −.35‡ .35‡ — 10. Claimed subscale −.17‡ −.09 .44‡ −.12∗ −.20‡ −.27‡ .53‡ −.30‡ −.35‡ — 11. Behavioral subscale −.69‡ −.26‡ .10∗ −.20‡ −.48‡ −.19‡ .15† −.18‡ −.53‡ .06 — 12. College entrance exam −.18‡ −.12∗ .09 −.11∗ −.12∗ −.04 −.03 −.06 −.11∗ .08 .09 — 13. GPA −.28‡ .00 .00 −.02 −.16† −.15† .11∗ −.01 −.17 .07 .22‡ −.25‡ — † p < .01. ∗ p < .05. ‡ p < .001. Table options 8.3. Gender differences Reliability scores as well as gender differences for the measures are reported in Table 10. Cases in which respondents had failed to provide their gender were treated as missing for these analyses. Men scored higher than women on the behavioral subscale of the SHS. In contrast, women scored higher than men on the claimed subscale of the SHS. Consistent with the findings of Study 1, women reported placing more importance on effort across all three of the effort measures that we developed. Women reported placing more personal value on effort as measured by the Worker scale, and agreed that effort was more normative as measured by the Prescriptive effort norm and Ability vs. effort tradeoff scales, than did men. In contrast, gender differences were not observed on the Protestant work ethic scale or on the Subjective overachievement subscales of Concern and Doubt. Gender differences were also found on several other measures. Women were higher than men in Need for achievement, Agreeableness, and Conscientiousness. Women also reported higher GPA and lower college entrance exam scores than did men. Table 10. Gender differences on measures (Study 2) Measure Cronbach’s α Men Women t η2η2 M SD M SD Worker .81 19.99 5.16 22.26 5.17 7.69‡ .04 Prescriptive effort norm .63 19.17 3.41 20.56 2.85 8.04‡ .04 Ability vs. effort tradeoff .71 14.56 5.06 13.58 4.84 3.49‡ .01 Protestant work ethic .68 10.16 12.34 10.69 11.64 <1 .00 Need for achievement .69 9.25 3.31 10.05 3.13 4.32‡ .01 Concern .80 38.04 7.77 38.65 7.19 1.46 .00 Doubt .78 27.03 6.12 27.26 6.29 <1 .00 Agreeableness .81 32.44 5.52 34.92 5.25 8.04‡ .04 Conscientiousness .77 30.46 5.15 31.69 5.09 4.18‡ .01 Claimed subscale .72 19.46 6.85 21.51 6.73 5.26‡ .02 Behavioral subscale .65 19.60 5.28 17.82 5.42 5.81‡ .02 College entrance exam — 6.63 1.50 6.09 1.51 6.39‡ .03 GPA — 4.29 1.62 3.93 1.56 4.03‡ .01 ‡ p < .001. Table options 8.4. Predicting self-handicapping subscale scores The next set of analyses was directed at identifying whether effort beliefs would be linked to behavioral and claimed self-handicapping, and specifically whether effort beliefs would mediate the gender difference on the behavioral subscale of the SHS. We therefore entered gender as an initial step into the regression model, and the measures that had demonstrated gender differences in a second step. These analyses are presented in Table 11. With regard to the behavioral subscale of the SHS, controlling for the individual differences reduced the gender effect to non-significance. The Worker, Ability vs. effort tradeoff, Need for achievement, and Conscientiousness scales were all significant predictors of behavioral subscale scores. Individuals placing more personalized value on effort had lower scores on the behavioral subscale, as did individuals high in Need for achievement and Conscientiousness. Finally, those believing ability (relative to effort) is more valued by others had lower behavioral subscale scores, indicative of a suppression effect of controlling for the other individual differences. Table 11. Prediction of SHS items (Study 2) Term Behavioral subscale Drug use Inadequate sleep Claimed subscale ββ t p ββ t p ββ t p ββ t P Initial model Gender −.142 5.00 <.001 −.171 6.10 <.001 −.072 2.53 <.05 .135 4.74 <.001 R2=R2=.02, FChangeFChange(1, 1215) = 25.01, p < .001 R2=R2=.03, FChangeFChange(1, 1237) = 37.16, p < .001 R2=R2=.01, FChangeFChange(1, 1242) = 6.41, p < .05 R2=R2=.02, FChangeFChange(1, 1202) = 22.43, p < .001 Final model Gender −.014 <1 ns −.114 4.07 <.001 −.026 <1 ns .200 8.11 <.001 Worker −.591 20.37 <.001 −.082 2.07 <.05 −.259 6.43 <.001 −.004 <1 ns Prescriptive effort Norm .004 <1 ns −.089 2.82 <.01 .036 1.12 .26 .184 6.59 <.001 Ability vs. effort tradeoff −.052 2.41 <.05 .012 <1 ns −.002 <1 ns .334 12.72 <.001 Need for achievement −.054 2.16 <.05 −.069 2.02 <.05 .048 1.38 .17 .008 <1 ns Agreeableness .038 1.72 .09 −.014 <1 ns −.017 <1 ns −.196 7.17 <.001 Conscientiousness −.175 6.26 <.001 −.097 2.55 <.05 −.044 1.13 .26 −.277 8.13 <.001 R2=R2=.53, ΔR2=ΔR2=.51, FChangeFChange(6, 1209) = 217.09, p < .001 R2=R2=.10, ΔR2=ΔR2=.07, FChangeFChange(6, 1231) = 15.99, p < .001 R2=R2=.07, ΔR2=ΔR2=.06, FChangeFChange(6, 1236) = 14.25, p < .001 R2=R2=.31, ΔR2=ΔR2=.29, FChangeFChange(6, 1196) = 83.91, p < .001 Table options Given that the Worker scale was a particularly strong predictor of scores on the behavioral subscale, we conducted a mediational analysis examining whether the Worker scale alone would explain the gender differences observed in tendency to behaviorally self-handicap. In this analysis, the gender difference on the behavioral subscale was significant (β = −.154, t = 5.85, p < .001). Adding the Worker scale alone in a second step in the regression model revealed a significant Worker scale effect (β = −.708, t = 37.08, p < .001). Furthermore, including the Worker scale in the model reduced the gender effect to nonsignificance (β = −.013, t < 1, ns), reliably mediating this effect, Sobel test z = 7.53, p < .001. 7 Furthermore, additional analyses revealed that the gender effect on the behavioral subscale remained significant when controlling for Need for achievement or Conscientiousness alone. We also investigated the specificity of this effect, examining whether effort beliefs would mediate gender differences on two individual items of the behavioral subscale related to drug use (item 20) and inadequate sleep (item22), see Appendix. In the regression model predicting scores on the drug use item, gender was entered in a first step. This analysis revealed a significant gender difference, such that men reported being more willing to let drugs interfere with their performance than did women. The individual difference measures that had shown gender differences were then entered into the regression model in a second step. Individuals scoring higher on the Worker scale, Prescriptive effort norm, Agreeableness, and Conscientiousness scales indicated being less willing to let drugs interfere with their performance. Furthermore, controlling for these variables reduced the gender effect, although it remained significant. We then conducted a mediational analysis examining whether the Worker scale alone would explain the gender differences observed on this item. In this analysis, the gender difference on the behavioral subscale was significant (β = −.168, t = 6.49, p < .001). Adding the Worker scale alone as a second step in the regression model revealed a significant Worker scale effect (β = −.233, t = 9.04, p < .001). Furthermore, including the Worker scale in the model reduced the gender effect (β = −.122, t = 4.76, p < .001), partially mediating this effect, Sobel test z = 5.84, p < .001. In the regression model predicting scores on the inadequate sleep item, gender was entered in a first step. This analysis revealed a significant gender difference, such that men reported being more willing to let inadequate sleep interfere with their performance than did women. The individual difference measures that had shown gender differences were then entered into the regression model in a second step. In this analysis, only the Worker scale was significant, and inclusion of the individual difference measures reduced the gender effect to nonsignificance. We again conducted a mediational analysis examining whether the Worker scale alone would explain the gender differences observed on this item. In this analysis, the gender difference on the behavioral subscale was significant (β = −.087, t = 3.32, p < .01). Adding the Worker scale alone as a second step in the regression model revealed a significant Worker scale effect (β = −.268, t = 10.40, p < .001). Furthermore, including the Worker scale in the model reduced the gender effect to nonsignificance (β = −.034, t = 1.31, p > .19), reliably mediating this effect, Sobel test z = 6.17, p < .001. Gender was again entered in an initial step into the regression model predicting claimed subscale scores. Women had higher scores on the claimed subscale than did men. The individual differences that had shown gender differences were then entered in a second step into the model. Controlling for the individual difference measures did not eliminate the gender difference. Rather the effect became stronger, replicating the results of Study 1. In contrast to the effects on the behavioral subscale, there was no significant effect of the Worker scale. The Ability vs. effort tradeoff and Prescriptive effort norm scales were positively related to the claimed subscale of the SHS, indicating that those believing ability (relative to effort) is more valued by others as well as those believing that others expect effort had higher scores on the claimed subscale. Those lower in Agreeableness and Conscientiousness also scored higher on the claimed subscale. 8.5. Predicting academic performance We next examined which of these scales might best explain the finding (see also Table 8 and Table 9) that those scoring higher on the behavioral subscale of the SHS have lower academic performance. Therefore, we regressed GPA onto the behavioral subscale of the SHS while controlling for college entrance exam score (see Table 12). Higher scores on the behavioral subscale were related to worse academic performance, over and above the effects of entrance exam score. We then added all of the individual difference measures to the model in a second step. The Worker scale significantly predicted better GPA whereas higher Prescriptive effort norm scores predicted lower academic performance. None of the other individual difference measures were significant. Furthermore, including these terms reduced the behavioral subscale effect, although not completely eliminating it. To further examine this finding, we added the Worker scale alone in a second step to the model predicting GPA. In the initial model, scores on the behavioral subscale were again related to worse academic performance (β = .274, t = 10.97, p < .001), over and above the effects of entrance exam score (β = −.253, t = 10.12, p < .001). Including the Worker scale in the model in a second step revealed a significant Worker scale effect (β = −.280, t = 8.04, p < .001), and reduced the behavioral subscale effect (β = .076, t = 2.19, p < .05), demonstrating reliable partial mediation of this effect, Sobel test z = 7.86, p < .001. The effect of entrance exam score (β = −.272, t = 11.08, p < .001) remained significant. Furthermore, additional analyses revealed that Need for achievement and Conscientiousness did not significantly add to the prediction of GPA beyond the effects of the Worker scale. Table 12. Regression models predicting GPA (Study 2) Term GPA ββ t p Initial model Behavioral subscale .307 10.98 <.001 College entrance exam −.245 8.77 <.001 R2=R2=.14, F(2,1102) = 91.68, p < .001 Final model Behavioral subscale .118 2.91 <.01 College entrance exam −.247 8.84 <.001 Worker −.324 6.85 <.001 Prescriptive effort norm .088 2.61 <.05 Ability vs. effort tradeoff −.021 <1 ns Protestant work ethic .022 <1 ns Need for achievement .009 <1 ns Concern −.001 <1 ns Doubt −.006 <1 ns Agreeableness .017 <1 ns Conscientiousness .006 <1 ns R2=R2=.19, ΔR2=ΔR2=.04, FChangeFChange(9, 1093) = 6.29, p < .001

خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.