• Aucun résultat trouvé

3. E MPIRICAL PART

3.4. Exploring the impact of locus of control, self-efficacy, and self-esteem on

3.4.4. Results and Discussion

3.4.4.1. Dimensionality of Individual difference Scales

Mean scores of the self-esteem scale, the general self-efficacy scale, and the three subscales (internal scale, powerful others scale, and chance scale) of the locus of control scale are shown in Table 1. The internal consistencies of these scales were between .642 and .810, indicating good reliability.

Conclusions on the mean scores are limited given that these scales do not provide cut-off criteria according to which someone would score low, medium, or high on a particular scale. However, across different countries and populations the mean score of the general self-efficacy scale has been around 2.9 (Luszczynska, Scholz, & Schwarzer, 2005). Thus, a mean score of 3.2 in our study suggests that the participants had slightly increased self-efficacy Table 1

Mean Values of the Individual Difference Scales with the Corresponding Standard Deviations (SD), and Internal Consistencies (Cronbach’s alpha)

Scale Mean SD Cronbach's alpha Range

Self-esteem 19.97 3.41 .760 11-25

General self-efficacy 3.17 4.45 .810 2.2-3.9

Locus of control

Internal Scale 32.38 6.89 .783 14-43

Powerful Others Scale 16.79 6.55 .681 6-31

Chance Scale 19.24 6.41 .642 11-37

Note. N = 34.

CHAPTER 3.4:APPRAISAL BIASES 164 beliefs. For the self-esteem scale, mean scores between 15 and 25 are within normal range (cf. Pietromonaco, & Barrett, 2009), which indicates that the mean score of our sample is in normal range. With respect to the mean scores of the subscales of the locus of control scales, our sample reported relatively high internal, high powerful others and high chance beliefs (cf.

Rossier et al., 2002, p. 148).

The theoretical concept of each scale has originated independently. However, individual difference variables of self-efficacy and self-esteem were predicted to be related, and to be distinct from locus of control (Scherer & Brosch, 2009; Wranik & Scherer, 2010).

The scales’ z-scores were calculated to make the scores of the scales comparable. First, the relationship between these scales was explored by calculating bivariate Person’s correlations.

The results are presented in Table 2. Supporting the prediction, the self-esteem and the general self-efficacy scales correlated positively, r = .59, p < .001, which suggests that both scales overlap: the higher the self-efficacy belief, the higher is also the self-esteem and vice versa. Self-esteem was also correlated with the internal scale (of the locus of control scale), r = .61, p < .001. The general self-efficacy scale was also positively related to the internal scale, r = .69, p < .001. Furthermore, the powerful others and the chance scales (each a subscale of the locus of control scale) were positively related: The more participants believed that their lives are controlled by powerful others, the more they believed that events are happening by chance, r = .47, p = .005. These results suggest that the esteem, the self-efficacy, and the internal scales relate to one dimension of internal beliefs, and that the powerful others and chance scales relate to another dimension of external beliefs. The results of the correlations indirectly support the prediction that the locus of control is related to the processing of the control check (associated with external beliefs), and self-esteem as well as self-efficacy is associated with the processing of the power check (associated with internal

CHAPTER 3.4:APPRAISAL BIASES 165 beliefs). Therefore, they indirectly suggest that the control and the power checks assess

distinct information about events.

Table 2

Pearson's Correlation Coefficients of the Individual Difference Measures

Scale GSE IS PS Chance Scale

Self-esteem (SE) .585*** .608*** -.226 .106

General self-efficacy (GSE) 1.000 .690*** -.317 -.003

Locus of control

Internal Scale (IS) .690*** 1.000 -.053 .103 Powerful Others Scale (PS) -.317 -.053 1.000 .469**

Note. N = 34. p** < .01, p*** < .001; two-tailed.

The prediction of the Component Process Model (Wranik & Scherer, 2010) that the conceptualizations of self-esteem and general self-efficacy relate to one variable (internal belief), whereas the locus of control scales is associated with another variable (external belief) was further investigated by submitting the scale scores to a principal component analysis. The analysis extracted two principal components. Table 3 provides the rotated component loadings, eigenvalues, and % of explained variance. The scales that loaded highly on Component 1 (as predicted: self-esteem, self-efficacy, internal scale) may represents a dimension of “internal belief”, whereas the scales that loaded with similar high magnitude on Component 2 (viz., powerful others and chance scales) may constitutes an “external belief”

about what causes the occurrence of events and whether one can act on them. This result further indirectly supports the prediction of the model claiming that there are separate

appraisals about the (intrinsic) controllability (external evaluation of causality) and the ability to act on the event (internal power belief).

CHAPTER 3.4:APPRAISAL BIASES 166

Table 3

Results of the Principal Component Analysis on the Individual difference Scales

Component 1 Component 2

Scale "Internal belief" "External belief"

Self-esteem 0.84 -0.03

General self-efficacy 0.87 -0.17

Locus of control scales

Internal Scale 0.88 0.09

Powerful Others Scale -0.23 0.85

Chance Scale 0.15 0.86

Eigenvalues 2.32 1.50

% of variance 46.32 29.97

Note. N = 34, Varimax rotation. Highlighted in bold text is the highest loading of one measure on a component.

3.4.4.2. Covariation Between Appraisal and the Individual Difference Scales The results of the principal component analysis on the individual difference scales suggested two uncorrelated dimensions of an internal and an external belief. The z-scores of the scales that loaded with similar high magnitude on the same component were therefore averaged. These two new variables were named accordingly: one was labeled “internal belief” and the other one “external belief”. For the analysis of covariance, one ANCOVA was calculated separately for each of the two belief variables. The results of the ANCOVAs are presented below and summarized in Tables 4 and 5, respectively.

For the sake of simplicity, in the presentation of the results of the ANCOVA,

significant interaction effects that involved the covariate (i.e., internal or external belief) as a factor were considered as “biased” effects because scale scores that deviated from the mean (i.e., increased or decreased) drove the effects. In contrast, significant ANCOVA results that

CHAPTER 3.4:APPRAISAL BIASES 167 did not have the covariate as a factor were considered as “unbiased” because for these effects, the covariate scores were around the mean.

3.4.4.2.1. Covariations with FRN amplitudes.

The four-factorial ANCOVA with internal belief as covariate revealed two significant interaction effects (see Table 4): A five-fold interaction effect of Goal Conduciveness × Control × Power × Channel × Internal Belief and a two-fold interaction effect of Control × Internal Belief (see Figure S1 of supplementary material section). These interaction effects suggest that higher or lower (“biased”) degrees of internal belief differentially affected FRN amplitudes in response to the information about goal conduciveness, control, and power.

Further, they indicate that FRN amplitudes of the control check were influenced by a “biased internal belief”. Moreover, the main effects of power and control were significant, implying that FRN amplitudes of the control and the power checks were differentiated when the internal belief was unbiased. No other effects were significant (F-values < 3.17, p-values >

.089).

The four-factorial ANCOVA on FRN amplitudes with external belief as covariate revealed two significant main effects of control and power (see Table 5). No other significant effects were found (F-values < 2.85, p-values >.106).This result indicates that when the external belief was around the mean (“ unbiased”), FRN amplitudes were sensitive to the information about the degree of control and power, conveyed in the feedback stimuli.

CHAPTER 3.4:APPRAISAL BIASES 168 Table 4

Main and interaction effects F ŋ² F ŋ² F ŋ² F ŋ² F ŋ²

GC 2.91 .12 5.16* .19 3.90 .13 0.06 .00 0.10 .00

GC × Channel 1.76 .07 5.61* .20

Control 11.03** .33 26.20*** .54 1.40 .05 3.08 .11 0.16 .01

Control × Channel 0.00 .00 7.21** .25

Power 7.75* .26 14.61** .40 2.28 .08 0.73 .03 0.08 .00

Control × Power × Channel 0.42 .02 3.83* .15

GC × Control × Power 0.42 .02 2.25 .09 0.00 .00 0.77 .03 0.74 .03

GC × Control × Power × Channel 0.65 .03 1.43 .06

GC × Internal Belief 0.64 .03 0.77 .03 1.74 .06 2.75 .10 0.04 .00

GC × Channel × Internal Belief 3.83 .15 0.19 .01

Control × Internal Belief 10.18** .32 0.06 .00 2.27 .08 3.98 .13 0.13 .01

Control × Channel × Internal Belief 0.93 .04 2.10 .09

Power × Internal Belief 0.00 .00 1.42 .06 4.31* .14 0.75 .03 0.16 .01

Power × Channel × Internal Belief 0.01 .00 1.42 .06

GC × Control × Internal Belief 0.10 .00 0.20 .01 0.07 .00 1.99 .07 0.06 .00

GC × Control × Channel × Internal Belief 1.79 .08 0.56 .03

GC × Power × Internal Belief 0.81 .04 0.76 .03 1.79 .06 0.23 .01 1.72 .06

GC × Power × Channel × Internal Belief 2.42 .10 0.82 .04

Control × Power × Internal Belief 3.17 .13 6.8* .24 0.40 .02 0.02 .00 1.41 .05

Control × Power × Channel × Internal Belief 3.04 .12 7.85** .26

GC × Control × Power × Internal Belief 0.05 .00 0.00 .00 0.70 .03 1.45 .05 0.07 .00 GC × Control × Power × Channel × Internal

Belief

13.45** .38 0.36 .02

Results of Covariate Analyses on the ERPs (FRN and P300) and the Facial EMG (frontalis, corrugator, and cheek region) with "Internal Belief" as Covariate

"Unbiased internal belief""Biased internal belief"

*p < .05, **p < .01, ***p < .001.

Note. ERPs: N = 24; facial EMG: N = 28. ERP = event-related potential, FRN = Feedback-related negativity, GC = goal conduciveness. "Unbiased internal belief" refers to ANCOVA results with internal belief z-scores around the mean (M = 0). "Biased internal belief" refers to ANCOVA results with belief z-scores that deviate from the mean (more than 1 SD).

Psychophysiological Measures

ERPs Facial EMG

FRN P300 Frontalis region Corrugator region Cheek region

CHAPTER 3.4:APPRAISAL BIASES 169 Table 5

Main and interaction effects F ŋ² F ŋ² F ŋ² F ŋ² F ŋ²

GC 2.47 .10 4.47* .17 2.55 .09 0.00 .00 0.13 .01

GC × Channel 0.86 .04 6.02** .22

Control 11.93** .35 26.54*** .55 0.50 .02 1.33 .05 0.07 .00

Control × Channel 0.04 .00 5.69** .21

Power 7.92* .27 13.47** .38 1.03 .04 1.01 .04 0.17 .01

Control × Power × Channel 0.09 .00 1.80 .08

GC × Control × Power 0.40 .02 2.28 .09 0.03 .00 0.31 .01 1.14 .04

GC × Control × Power × Channel 0.04 .00 1.22 .05

GC × External Belief 0.00 .00 0.94 .04 0.93 .04 2.29 .08 0.84 .03

GC × Channel × External Belief 0.12 .01 2.52 .10

Control × External Belief 2.26 .09 0.06 .00 3.26 .11 1.47 .05 0.48 .02

Control × Channel × External Belief 0.01 .00 0.26 .01

Power × External Belief 0.13 .01 2.34 .10 0.21 .01 0.89 .03 0.18 .01

Power × Channel × External Belief 0.13 .01 0.82 .04

GC × Control × External Belief 1.13 .05 0.23 .01 3.75 .13 1.76 .06 2.12 .08

GC × Control × Channel × External Belief 0.02 .00 4.89** .18

GC × Power × External Belief 1.69 .07 0.81 .04 0.00 .00 5.28* .17 0.01 .00

GC × Power × Channel × External Belief 0.63 .03 0.51 .02

Control × Power × External Belief 0.01 .00 5.88* .21 2.33 .08 0.26 .01 0.43 .02

Control × Power × Channel × External Belief 1.45 .06 0.56 .03

GC × Control × Power × External Belief 0.07 .00 0.00 .00 1.02 .04 0.16 .01 1.26 .05

GC × Control × Power × Channel × External Belief

0.71 .03 1.56 .07

Results of Covariate Analyses on the ERPs (FRN and P300) and the Facial EMG (frontalis, corrugator, and cheek regions) with "External Belief" as Covariate

"Unbiased external belief""Biased external belief"

*p < .05, **p < .01, ***p < .001.

Note. ERPs: N = 24; facial EMG: N = 28. ERP = event-related potential, FRN = Feedback-related negativity, GC = goal conduciveness. "Unbiased external belief" refers to ANCOVA results with external belief z-scores around the mean (M = 0). "Biased external belief" refers to ANCOVA results with belief z-scores that deviate from the mean (more than 1 SD).

Psychophysiological Measures

ERPs Facial EMG

FRN P300 Frontalis region Corrugator region Cheek region

CHAPTER 3.4:APPRAISAL BIASES 170 3.4.4.2.2. Covariations with P300 amplitudes.

With internal belief as covariate, the ANCOVA on P300 amplitudes yielded two significant interaction effects (see Table 4): A three-fold interaction effect of Control × Power × Internal Belief (Figure S2 of supplementary material section) and a four-fold interaction effect of Control × Power × Channel × Internal Belief. These interaction effects indicate that a “biased internal belief” affected the cumulative effects of the control and the power checks. In contrast, an “unbiased internal belief” resulted in significant main effects of the three appraisal checks. Moreover, the following two-fold interaction effects were

significant: Goal Conduciveness × Power, and Control × Power. Further, significant

“unbiased” interactive (cumulative) appraisal check effects involving power were: Power × Channel, Control × Power × Channel, Goal Conduciveness × Power, and Goal

Conduciveness × Power × Channel. The other “unbiased” interaction effects were Goal Conduciveness × Channel, and Control × Channel. The interaction effects with channel as a factor suggest that P300 amplitudes differed across the four channel locations. No other effects were significant (F-values < 2.10, p-values > .141).

The four-factorial ANCOVA showed two significant interaction effects with external belief as covariate (see Table 5), which were the interaction effects of Control × Power × External Belief (Figure S4 of supplementary material section), and Goal Conduciveness × Control × Channel × External Belief. When participant had “biased external beliefs”,

significant main effects of the three appraisal checks emerged. Further, cumulative appraisal check effects were found of Power × Channel, Goal Conduciveness × Channel, Goal

Conduciveness × Power, and Goal Conduciveness × Power × Channel. The ANCOVA yielded no other significant effects (F-values < 2.34, p-values > .141). These results suggest that when participants’ external belief was “biased”, P300 amplitudes were modulated by information about goal conduciveness, control, and power; whereas, when the external belief

CHAPTER 3.4:APPRAISAL BIASES 171 was “unbiased”, P300 amplitudes showed a different pattern of main and interaction effects.

These results imply that the strength of an external belief had an impact on the effects of the goal conduciveness, the control, and the power checks on P300 amplitudes.

To sum up, the results of the ANCOVAs on FRN and P300 amplitudes with internal or external belief as covariate showed differential effects on the processing and integration of the appraisal checks. In general, the impact of the belief variables was larger on the

processing of the appraisal checks on the P300 compared with the effects on the FRN. On the FRN, it seems that moderate (“unbiased”) degrees of internal and external belief were

associated with differentiated appraisal check main effects of goal conduciveness, control, and power compared with “biased” beliefs. In particular, a “biased internal belief” affected the processing of the control check on FRN amplitudes and P300; they seem to have influenced the integration of appraisal results of the control and the power checks. These results encourage specifying the importance of biased internal and external beliefs on the processing and the integration of the control and the power checks.

3.4.4.2.3. Covariations with frontalis activity.

The analysis of the frontalis activity with internal belief as covariate yielded a significant interaction effect of Power × Internal Belief (see Table 4 and Figure S3 of

supplementary material section). No other significant effects were found (F-values < 3.90, p-values > .059). This result implies that the efferent effects of the power check on frontalis activity were affected by a “biased internal belief”. However, with external belief as

covariate, the ANCOVA yielded no significant effects (F-values < 3.26, p-values > .083, see Table 5). These results suggest that frontalis activity was affected by a “biased internal belief”, whereas it was unaffected by a “biased external belief”.

CHAPTER 3.4:APPRAISAL BIASES 172 3.4.4.2.4. Covariations with corrugator activity.

With internal belief as covariate, over the corrugator region, the ANCOVA showed a significant interaction effect of Power × Goal Conduciveness (see Table 4). No other

significant effects were obtained (F-values < 3.98, p-values > .057). This result suggests that the response pattern of the goal conduciveness and the power checks yielded significant effects in corrugator activity only when participants had an “unbiased internal belief”. With external belief as covariate, the ANCOVA on corrugator activity measures revealed (see Table 5) significant interaction effects of Goal Conduciveness × Power and of Goal

Conduciveness × Power × External Belief (Figure S5 of supplementary material section). The results propose that a “biased external belief” had an impact on the efferent effect of the goal conduciveness and the power checks in corrugator activity, whereas a “biased internal belief”

seems to have not affected the efferent appraisal check effects in corrugator activity.

3.4.4.2.5. Covariations with cheek region activity.

Over the cheek region, no significant effects were found with internal or external belief as covariate (F-values < 4.14, p-values > .051, F-values < 2.12, p-values > .157, respectively). This result implies that cheek region activity might have been unaffected by internal and external beliefs.

To sum up, the results of the repeated measures ANCOVAs with internal and external belief each used as a covariate in the analysis, indicate that the upper face was differentially affected by “biased internal and external beliefs” in comparison to the lower face. A “biased internal belief” seems to have affected the power check effects on frontalis activity, whereas a “biased external belief” influenced the cumulative effects of the goal conduciveness and the power checks in corrugator activity.

CHAPTER 3.4:APPRAISAL BIASES 173