Measurement of Attributional Style for Positive and Negative Events Using the Children’s Attributional Style Interview-II

Abstract

The current study provides psychometric support for the use of the Children’s Attributional Style Interview-II (CASI-II) to assess attributional style for positive and negative events with young adolescents. The paper-and-pencil version of the CASI-II was administered to a sample of 546 seventh- and eighth-grade students attending a Midwest suburban middle school (50% women, 54% white, 22% Hispanic, 10% Asian American, and 14% other). A multilevel confirmatory factor analysis demonstrated a factor structure consistent with attributional style theory and evidence for internal consistency reliability and temporal stability. Construct and convergent validity were strong. Dimensional subscale scores as well as composite scores for both positive and negative events showed adequate reliability and strong validity evidence. These results offer confidence for testing theory-driven predictions related to attributional style for both positive and negative events, as well as specific attributional dimensional patterns, using the CASI-II.

Keywords

attributional style attributional dimensions positive events psychometrics measurement

The construct of attributional style (AS) comes from the reformulated theory of learned helplessness (RTLH; Abramson, Seligman, & Teasdale, 1978), which proposes that people who consistently attribute negative life events to internal, stable, and global causes can potentially be at risk for developing depression. A related construct referred to as attributional generality, focused only on stable and global attributional dimensions, is central to the hopelessness theory of depression (Abramson, Metalsky, & Alloy, 1989). Most of the past research on AS testing both of these theories has been focused on the attributional dimensions related to negative events due to the assumption that attributional processes are more often produced by negative events rather than by positive events (Wong & Weiner, 1981). However, AS for positive events has also demonstrated importance.

Specifically, research has demonstrated support for the recovery model, which proposes that an enhancing AS (internal, stable, and global attributions for positive events) contributes to a quicker recovery from depression (Needles & Abramson, 1990). In addition, a recent study demonstrated support for the role of AS for positive events as a protective factor with a sample of early adolescents. An enhancing AS led to decreases in depressive symptoms, mediated through higher self-esteem (Rueger & George, 2017). Furthermore, there is increasing interest in the role of AS for positive events as related to other important outcomes, such as academic achievement (Gordeeva, Sheldon, & Sychev, 2019). Thus, instruments with strong psychometric properties that assess AS for both positive and negative events are important to investigate the full range of attributional patterns.

The Children’s Attributional Style Interview-II (CASI-II) has demonstrated strong psychometric support for the attributional dimensions of internality, stability, and globality for both positive and negative events (Rueger & Malecki, 2007; Rueger, Haines, & Malecki, 2010), which allows for fully testing theories related to AS. The internality dimension refers to the extent to which people believe they can influence or change the cause of an event. The stability dimension refers to the perceived consistency and persistence of causal factors. Finally, the globality dimension is characterized by the extent to which people believe causes of specific events will generalize to other situations (Abramson et al., 1978). A notable benefit to using the CASI-II instrument is the parallel structure to assessment tools for older adolescents (Adolescent Cognitive Style Questionnaire-ACSQ; Hankin & Abramson, 2002) and adults (Attributional Style Questionnaire-ASQ; Peterson et al., 1982), which allows for consistency of measurement in longitudinal studies across developmental periods of childhood, adolescence, and adulthood. It was originally posited that measures with an open-ended format consistent with measures of AS for older adolescents and adults would be too complex for children to understand (Reivich, 1995). Most of the AS literature focused on youth has used the Children’s Attributional Style Questionnaire (CASQ; Kaslow, Tanenbaum, & Seligman et al., 1978) and the shortened revised version (CASQ-R; Kaslow & Nolen-Hoeksema, 1991). However, initial psychometric evidence for the CASI measures support the utility of using this open-ended format from childhood and into early adolescence (Rueger et al., 2010; Rueger & Malecki, 2007).

The CASI-II has been used in theory-testing with AS to demonstrate the importance of AS in understanding depressive symptoms in the vulnerable developmental period of early adolescence. In a study focused on AS for negative events, results demonstrated findings consistent with RTLH with some interesting gender differences (Rueger & Malecki, 2011). Specifically, the predicted protective nature of an adaptive AS for negative events (i.e., causes of negative events are external, temporary, and circumscribed) in times of heightened stress was demonstrated for boys who reported higher levels of parental support; this three-way interaction was associated with decreases in depression over a 4-month time frame. For girls with higher levels of parental support, an adaptive AS was associated with decreases in depression, regardless of stress. In another study focused on AS for positive events, results demonstrated findings consistent with the recovery model (Rueger & George, 2017). Specifically, there was evidence for an indirect effect of positive AS on lower depressive symptoms as mediated by self-esteem.

Regarding the role of attributional dimensions, the CASI-II was used in one study of 5th- and 6th-grade students and results showed that stable attributions for negative events were related to learned helplessness, as predicted by RTLH, and stable and global attributions for negative events were related to depressive symptoms, as predicted by hopelessness theory. In addition, internal and stable attributions for positive events were related to lower depressive symptoms, and stable attributions for both positive and negative events were related to higher optimism (Rueger & Malecki, 2007). These findings were replicated in a study of 6th and 7th graders using the CASI-II (Rueger et al., 2010). Stable and global attributions for negative events were related to depressive symptoms, whereas only stable attributions for negative events were related to learned helplessness. These researchers also found that internal and stable attributions for positive events were negatively related to depressive symptoms, and that stable attributions for positive events were related to learned helplessness. Notably, optimism was significantly related to stable attributions for both positive and negative events in the expected directions.

These results testing the attributional dimensions demonstrated convergent validity of the CASI-II in two independent samples (Rueger et al., 2010; Rueger & Malecki, 2007). In addition, these studies demonstrated a factor structure consistent with attributional theories (Abramson et al., 1989, 1978) and strong internal consistency reliability of scale scores. However, although there is evidence to suggest the psychometric utility of the CASI-II, further psychometric support for this measure would increase confidence to use this relatively new measure to assess AS in this developmental period. Specifically, a confirmatory factor analysis with a consideration for both the individual and classroom levels would add significantly to the literature. Educational research often involves the use of nested or clustered data (e.g., students within classrooms within schools), and multilevel confirmatory factor analysis (MCFA) should be used to account for nestedness in data, especially when intraclass correlations among variables are high (Pornprasertmanit, Lee, & Preacher, 2014). In addition, there is evidence to suggest that Cronbach’s alpha is not the most accurate estimation of reliability. Thus inclusion of McDonald’s omega, which provides a more accurate estimation of internal consistency reliability (Dunn, Baguley, & Brunsden, 2014; Revelle & Zinbarg, 2009), as well as a test of temporal stability would add to reliability evidence. Finally, additional evidence of convergent and discriminant validity of AS dimensions using both positive and negative outcomes would add to the psychometric support for the use of dimensional subscale scores in theory-testing.

Thus, the goal of the current study is to test these psychometric properties of the CASI-II in a large independent sample of middle school students in order to replicate support for the measurement of AS for both positive and negative events. It is predicted that the CASI-II will demonstrate (1) construct validity by a factor structure that supports the distinctness of AS for positive and negative events, and attributional dimensions; (2) adequate internal consistency reliability of AS composite scores for both positive and negative events, and stable and global attributional dimensions for negative and positive events; (3) adequate temporal stability; (4) and evidence of convergent validity of composite and dimensional scores by a pattern of stronger intercorrelations between AS for positive events with positive outcomes, compared to negative outcomes, and between AS for negative events with negative outcomes compared to positive outcomes.

Method

Participants and Procedures

Data were collected at two timepoints in January and May of the same school year at a large suburban middle school in the Midwest using convenience sampling (Rueger & Malecki, 2011). Active consent procedures were used to invite all students in the school to participate in this longitudinal study, and 69% of the students returned signed parent permission slips. At both timepoints, students completed written assent before completing survey packets, and teachers completed behavior ratings of learned helplessness and mastery orientation. Student and teacher data from the first timepoint are used in the current study for the majority of analyses; student CASI-II data from the second timepoint are used to test for temporal stability.

The current sample includes 546 (50% female) students in seventh and eighth (49%) grade, with ages that ranged from 12 to 15 (M = 13.21 and SD = .71). The sample was 54% white (n = 297), 22% Hispanic (n = 121), 3% African American (n = 16), 10% Asian American (n = 53), 8% biracial (n = 43), and 3% other or unreported (n = 16). Although data were not collected on socioeconomic status of the sample, a report on this middle school showed that 25% of the students qualified for free or reduced lunch at the time of data collection.

Student survey

Data were collected in large groups in the school cafeteria (ns ranged from 76 to 106). The CASI-II was administered first, followed by other measures, including those for the current study. For all measures, the researcher read the directions and all items to control for the possible effects of varying reading abilities and to help keep the students focused. Research assistants were available to assist with the group administration and answer any individual questions.

Teacher-report survey of student learned helplessness and mastery orientation

In the participating middle school, students were assigned to one of six teaching teams with five teachers who teach the core subjects. 30 participating students from each of the six teaching teams were randomly selected for teacher-report ratings of learned helplessness and mastery orientation using the Student Behavior Checklist (SBC; Fincham, Hokoda, & Sanders, 1989). Five of the six teaching teams completed and returned the SBC ratings in sealed envelopes within 2 weeks of the student data collection for an 83% return rate. A total of 144 students had SBC teachers’ ratings and complete self-report survey data.

Measures

The Children's Attributional Style Interview-II (CASI-II; Haines et al., 2005)

The CASI-II is a 16-item self-report inventory used to assess AS with children in grades 5 through 8. It measures three attributional dimensions (internality, stability, and globality) in two domains (achievement and interpersonal) using eight positive and eight negative events with an open-ended rating format. Assessing attributions from the achievement and interpersonal domains allows researchers to examine the extent to which children attribute causality to social interactions, effort, and/or ability. Each item presents a hypothetical situation, followed by a prompt for a self-generated attribution for the event. Three follow-up questions based on a 7-point Likert scale focused on rating the internality, stability, and globality of the attribution. Procedures for group administration described in the CASI manual (Haines et al., 2005) were followed. Full Positive (FP) and Full Negative (FN) composite scores, which can range from 24 to 168, are computed by summing the 24 ratings (8 for internality, stability, and globality each) for positive or negative events. The Generality Negative (GN) and Generality Positive (GP) composite scores, which can range from 16 to 112, were computed by summing the 8 stability and 8 globality items for positive and negative events respectively. The three-dimensional subscale scores for positive and negative events were computed by summing the 8 items for internality, stability, and globality respectively. Dimensional scores can range from 8 to 56.

The Center for Epidemiological Studies Depression Scale for Children (CES-DC; Weissman, Orvaschel, & Padian, 1980)

The CES-DC is a 20-item self-report inventory of depressive symptoms that has demonstrated strong psychometric properties with adolescents, including internal consistency reliability, test-retest reliability, and concurrent validity (Faulstich, Carey, Ruggiero, Enyart, & Gresham, 1986). Participants are instructed to indicate the response that most closely matches their experience over the course of the previous week. Items are rated on a 4-point rating style scale ranging from 0 (not at all) to 3 (a lot). Scores can range from 0 to 60. Four positively phrased items are reverse scored with higher scores indicating higher levels of depression.

The Children’s Hope Scale (CHS; Snyder et al., 1997)

The CHS is a 6-item measure of dispositional hope for children ages 8 through 16 that has demonstrated strong psychometric properties, including internal consistency reliability, test-retest reliability, convergent, and discriminant validity (Valle, Huebner, & Suldo, 2004). The six items consist of two components (agency and pathways) with three items for each component. Each item is rated on a 6-point scale ranging from 1 (none of the time) to 6 (all of the time). Scores can range from 6 to 36 with a higher score indicating a higher hope.

The Youth Life Orientation Test (YLOT; Ey et al., 2005)

The YLOT is a 16-item measure of optimism and pessimism designed for children in 3rd through 12th grade containing six optimism items, six pessimism items, and four filler items. The measure has demonstrated strong psychometric properties, including internal consistency reliability, stability, test-retest reliability, convergent and discriminant validity, and predictive validity (Ey et al., 2005). Each item is rated on a 4-point scale ranging from 0 (not true for me) to 3 (true for me). The optimism and pessimism subscales are calculated by taking the sum of the six optimism and six pessimism items respectively. Scores on each subscale can range from 0 to 18, with a higher score indicating higher levels of either optimism or pessimism.

The Student Behavior Checklist (SBC; Fincham et al., 1989)

The SBC is a 24-item teacher-report measure of mastery orientation and learned helplessness in the classroom that has demonstrated strong psychometric properties, including internal consistency reliability, convergent validity, and incremental validity (Nolen-Hoeksema, Girgus, & Seligman, 1992). Teachers rate students on a 5-point scale from 1 (not true) to 5 (very true). Learned helplessness and mastery orientation subscales are computed by taking the sum of each subscale’s set of 12 items. For each subscale, scores can range from 12 to 60, with higher scores indicating higher learned helplessness or mastery orientation.

Data Analytic Plan

A MCFA was conducted using R software (R Core Team, 2016) with the lavaan package (Rosseel, 2012). In total, five models were tested: (a) a 6-factor model representing an overarching AS for positive and negative events (Model 1: Full Composite-FC); (b) a three-factor model including all items related to each negative AS dimension (Model 2: Full Negative-FN); (c) a three-factor model including all items related to each positive AS dimension (Model 3: Full Positive-FP); (d) a two-factor model including the stability and globality dimensions for negative events (Model 4: Generality Negative-GN); and (e) a two-factor model including the stability and globality dimensions for negative events (Model 5: Generality Positive-GP). The three-factor models are consistent with the RTLH, and the two-factor models are consistent with the hopelessness theory of depression. Because no past research has examined the extent to which models are different at the individual or classroom levels, identical models were fit at the between and within levels of each analysis. Variability of scores between classroom clusters was assessed with intraclass correlation coefficient (ICC) statistics.

Model fit was evaluated using Satorra–Bentler chi-square, the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and standardized root mean residual (SRMR). The chi-square test was included as a measure of exact model fit, with non-significant p-values indicating no significant differences between the models and observed data (Barrett, 2007). However, it has been documented that chi-square tests are sensitive to sample size (Vandenberg, 2006). CFI values above .90, RMSEA values below .08, and SRMR values below .10 were criteria for an adequate fit, whereas values above .95, below .05, and below .08, respectively, were criteria for an excellent fit (Vandenberg & Lance, 2000). For all models, residuals for items that shared a common stem were allowed to correlate to account for covariation associated with items sets that are context dependent (Higgins, Zumbo, & Hay, 1999). Internal consistency reliability was assessed using McDonald’s omega and Cronbach’s alpha, and convergent validity was tested using Pearson’s r correlations. Magnitude of effect sizes were interpreted using Cohen’s (1988) rules of thumb.

Results

MCFA

The underlying structure of the CASI-II was tested using MCFA (see Table 1 for fit indices for all models) to address Hypothesis 1. Model 1 tested the six-factor model representing the attributional dimensions (Internality, Stability, and Globality) for both positive and negative events together. At both the first (individual) and second (classroom) levels, the RMSEA and SRMR scores indicated an excellent fit (Vandenberg & Lance, 2000), but the CFI scores indicated a poor fit. Thus, no further analyses were conducted with this composite score.

Table 1.

Fit Indices Supporting the CASI-II Composite Scores at Level 1 (Within/Individual) and Level 2 (Between/Classroom).

	Model 1: Full Composite	Model 2: Full Negative	Model 3: Full Positive	Model 4: Generality Negative	Model 5: Generality Positive
Level 1
SB X² (df)	1752.06*** (1026)	374.50*** (225)	532.26*** (225)	181.01*** (95)	288.61*** (95)
CFI	.85	.92	.90	.96	.91
RMSEA	.04	.04	.05	.04	.06
SRMR	.05	.05	.06	.04	.06
Level 2
SB X² (df)	1758.335*** (1017)	333.725*** (224)	366.166*** (219)	146.651** (104)	185.906*** (88)
CFI	.86	.93	.91	.97	.91
RMSEA	.03	.03	.04	.03	.05
SRMR	.05	.05	.06	.04	.06

Note. SB X² = Satorra–Bentler chi-square; df = degrees of freedom; CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual; * p < .05, ** p < .01, and *** p < .001.

Models 2 and 3 tested the three-factor model representing the attributional dimensions for positive and negative events, respectively. At both levels and for both models, the CFI scores indicated an adequate fit, and the RMSEA and SRMR scores indicated an excellent fit (Vandenberg & Lance, 2000). Models 4 and 5 tested the two-factor model representing stability and globality for positive and negative events, respectively. At both levels, the CFI, RMSEA, and SRMR scores for Model 4 indicated an excellent fit to the data, and the fit indices for Model 5 indicated adequate to excellent fit. All path coefficients (presented in Supplemental Tables 1 and 2) were statistically significant across all four models. Factor loadings for all models were generally adequate (≥.30), though eight factor loadings were below .30 (see Supplemental Tables 1 and 2). Specifically, five items in the internality dimension and three items in the stability dimension had factor loadings below .30. There were no items in the Globality dimension measuring positive or negative events with factor loadings below .30.

Model variability across classroom clusters was very low, as evidenced by ICC statistics (ICCs ranged from .0109 to .0111). There were a total of six classroom clusters, with an average cluster size of ∼82 children per classroom. Model fit at the second level was slightly better as indicated by .01 increases in CFI scores and .01 decreases in RMSEA scores as compared to Level 1 fit indices scores, with the exception of the GP model.

Reliability of Composite and Dimensional Scores

Internal consistency reliabilities of CASI-II composite and dimensional scores were tested using McDonald’s omega and Cronbach’s alpha (Table 2) to address Hypothesis 2. Only the Negative Internality dimensional subscale score had alpha and omega values below .60, which is considered unacceptable. Internal consistency reliability of other measures were minimally acceptable to very good (DeVellis, 2016; Taber, 2018). Temporal stability of scale scores was tested with bivariate correlations between CASI-II scores from the two timepoints to address Hypothesis 3. All correlations were statistically significant (ps < .01) and rs ranged from .53 to .56 for the four composite scores, .43 to .53 for the three negative dimensional subscale scores, and .37 to .54 for the three positive dimensional subscale scores (Table 2).

Table 2.

Descriptive Statistics and Correlations among Constructs.

Scale	N	M	SD	ω	α	FN	FP	GN	GP	Neg-Int	Neg-Sta	Neg-Glo	Pos-Int	Pos-Sta	Pos-Glo
FN	526	101.38	17.78	.79	.79	.53**
FP	527	130.55	15.61	.81	.80	.27**	.54**
GN	526	60.34	15.34	.83	.83	.92**	.18**	.55**
GP	527	82.21	12.65	.79	.78	.28**	.94**	.25**	.56**
Neg-Int	527	41.01	7.04	.57	.54	.52**	.27**	.15**	.16**	.43**
Neg-Sta	527	30.24	7.99	.71	.71	.79**	.09*	.86**	.15**	.12**	.51**
Neg-Glo	526	30.08	9.41	.79	.78	.83**	.22**	.90**	.28**	.15**	.55**	.53**
Pos-Int	528	48.32	5.66	.63	.60	.11*	.66**	−.05	.36**	.39**	−.07	−.01	.37**
Pos-Sta	528	42.91	6.30	.64	.63	.09	.78**	.03	.78**	.15**	.06	.00	.40**	.52**
Pos-Glo	527	39.29	8.69	.78	.77	.35**	.81**	.34**	.89**	.13**	.17**	.41**	.23**	.41**	.54**
Depression	527	18.52	11.08	.89	.89	.16**	−.14**	.22**	−.10**	−.09	.16**	.21**	−.17**	−.21**	.00
Hope	527	26.89	5.40	.83	.82	−.14**	.28**	−.21**	.26**	.10*	−.19*	−.17**	.20**	.32**	.14**
Pessimism	526	5.70	3.80	.79	.79	.16**	−.18**	.22**	−.15**	−.07	.17**	.21**	−.17**	−.24**	−.04
Optimism	527	12.84	3.08	.71	.71	−.04	.32**	−.10**	.30**	.14**	−.12**	−.07	.21**	.29**	.24**
Helplessness	144	24.78	10.65	.90	.89	.09	−.09	.14	−.04	−.08	.18*	.08	−.13	−.16	.05
Mastery	144	45.54	11.81	.96	.96	−.13	.08	−.16	.03	.03	−.19*	−.11	.15	.15	−.07

Note. Full Composite (FC) model omitted due to poor model fit; FN = Full Negative, FP = Full Positive, GN = General Negative, GP = General Positive, Neg-Int = Negative Internal, Neg-Sta = Negative Stable, Neg-Glo = Negative Global, Pos-Int = Positive Internal, Pos-Sta = Positive Stable, Pos-Glo = Positive Global, Depression = Center for Epidemiological Services Depression Scale for Children, Hope = Children’s Hope Scale, Pessimism = Youth Life Orientation Test Pessimism Subscale, Optimism = Youth Life Orientation Test Optimism Subscale, Helplessness = Student Behavior Checklist-Learned Helplessness subscale, Mastery = Student Behavior Checklist-Mastery Orientation Subscale; ω = McDonald’s omega, α = Cronbach’s alpha; Correlations between T1 and T2 CASI-II scores are reported along the diagonal; p < .05 = *, p < .01 = **, p < .001 = ***.

The theoretical relationships between the attributional dimensions were supported by significant correlations among the dimensional subscale scores (Table 2). Negative Globality was significantly related to Negative Stability, Positive and Negative Internality were significantly related, and all three dimensions for positive events were significantly related in the expected directions.

Convergent Validity of Composite and Dimensional Scores

Patterns of intercorrelations between scores of theoretically related constructs and CASI-II tested for convergent validity to address Hypothesis 4 (Table 2). FN Composite demonstrated correlations with negative outcomes (i.e., depression and pessimism) that were statistically significant and small in magnitude. The correlation with hope was also small but statistically significant; however, the correlation with optimism was not significant. Compared to FN, GN Composite demonstrated stronger correlations, but the pattern was similar: Correlations with negative outcomes (i.e., depression and pessimism) were moderately small in magnitude, as was the correlation with hope, but the correlation with optimism was only small in magnitude. Similarly, FP and GP composites demonstrated correlations with positive outcomes (i.e., hope and optimism) that were statistically significant and moderately small in magnitude, but correlations with negative outcomes (i.e., depression and pessimism) were only small in magnitude.

Similar patterns were noted with some of the dimensional subscale scores. Negative Globality was significantly associated with negative outcomes (i.e., depression and pessimism) as well as hope, but not optimism, whereas Positive Globality was significantly associated with positive outcomes but not negative outcomes. However, Negative Stability, Positive Internality, and Positive Stability were associated with both negative and positive outcomes with correlations that were small to moderately small. Interestingly, Negative Stability was the only dimension to be significantly associated with Learned Helplessness and Mastery Orientation. Correlations between Learned Helplessness and Mastery Orientation with Positive Internality and Positive Stability were also small in magnitude but only marginally significant due to the smaller sample size for this teacher-report assessment. Finally, Negative Internality was significantly related to hope and optimism but in the unexpected direction.

Discussion

The purpose of this investigation was to further the psychometric support for the CASI-II by conducting a MCFA using individual items on an independent adolescent sample, and testing for temporal stability of scale scores and internal consistency reliability, as well as convergent and divergent validity with both positive and negative outcome variables theoretically linked to AS. For positive events, analyses testing the underlying factor structure, and reliability and validity of scores, demonstrated support for the use of the FP and GP composite score and all three dimensional scores. For negative events, the CFA supported the use of the FN and GN composite scores. However, there was lack of both reliability and validity support for the Negative Internality score. Thus, the current results support the use of the GN composite score, and the Negative Stable and Negative Global subscale scores.

In addition, the results of the item-level CFA in the current study are consistent with previous studies using individual items with older children and young adolescents (Rueger et al., 2010) and adults (Higgins et al., 1999), and item parcels with children (Cole et al., 2008; Rueger & Malecki, 2007). Additionally, the current study added to this psychometric support by conducting a multilevel analysis and found that model fit was similar at both the individual and classroom levels. This is potentially attributable to low variability in responses across classroom clusters. Overall, these results indicate that AS can be measured consistently with the CASI-II at both the individual and classroom levels.

The patterns of intercorrelations between CASI-II scores and the scores of theoretically related constructs confirmed previous research (Rueger et al., 2010), such as significant correlations with depressive symptoms and optimism. Extending previous research findings, comparisons of associations with positive and negative outcomes highlighted some noteworthy findings. First, this is the first study to investigate the relationship between AS and optimism and pessimism as distinct constructs, and these results suggest that AS for positive events, compared to AS for negative events, may be more related to optimism. Second, correlations between attributional dimensions and theoretically related variables highlighted nuances in the ways in which attributions for positive and negative events may differentially influence positive and negative outcomes. For example, the pattern of findings demonstrated with global attributions suggests that cognitive generalizations about negative events affect negative outcomes and cognitive generalizations about positive events affect positive outcomes. For example, globality attributions for positive events was significantly related to optimism but not pessimism, and globality attributions for negative events was significantly related to pessimism but not optimism. On the other hand, stability attributions for negative events was related to all outcomes, both positive and negative. Internality and stability for positive events were also related to almost all outcomes: depression, hope, pessimism, and optimism, but not learned helplessness or mastery orientation.

Interestingly, teacher-reported learned helplessness and mastery orientation were significantly related only to stable attributions for negative events. This suggests that stable attributions for negative events play a larger role in predicting learned helplessness than global or internal attributions for negative events, or attributions for positive events. Notably, the stable dimension for both positive and negative events generally had the strongest correlations with almost all outcomes. These results suggest that the stability dimension plays a significant role in comprising adolescents’ attributions for both positive and negative events. The belief that either a positive or negative event will continually occur throughout time appears to have a more salient effect than whether adolescents attribute agency and/or generalizability to events.

Taken together, the different pattern of correlations between AS for positive versus negative events and theoretically related constructs emphasizes the potential problem with combining composite scores for positive and negative events in order to determine an overall AS. Poor fit of the six-factor model that combined attributions for negative and positive events in one model also suggests that creating an overall composite score of AS for both positive and negative events may lead to misleading results.

Implications for Research and Clinical Practice

This study has several important implications. First, adding to the psychometric support for the CASI-II provides confidence in measurement to both researchers and clinicians who are interested in considering the role of AS in the well-being of older children and young adolescents. For researchers, support for the use of the CASI-II to assess AS for positive as well as negative events allows for continued examination of how AS for positive events relates to recovery from depression and other important outcomes, including academic achievement. Furthermore, support for the use of the CASI-II to assess attributional dimensions allows for specific theory-testing about stable and global attributions for negative events (i.e., hopelessness theory), and theory development about internal, stable, and global attributions for positive events.

Additionally, this study offers support for the consideration and measurement of AS in cognitive therapy in clinical settings. For example, because the CASI-II has been demonstrated to have strong psychometric properties, it could be used in depression treatment programs focused on changing maladaptive attributions (e.g., the Penn Resiliency Project; Gillham, Reivich, Jaycox, & Seligman, 1995). In addition, the current study highlights the potential value of intervening in youth’s attributions for positive events. The significant correlations between AS for positive events and depression, hope, pessimism, and optimism suggest that encouraging youth to identify positive events in their lives and attribute those events as internal, stable, and global causes could lead to a greater sense of well-being in these areas of functioning. More research is needed to support this line of clinical research, but this type of treatment strategy, which focuses on identifying positive events, is consistent with behavioral activation (Ritschel, Ramirez, Cooley, & Edward Craighead, 2016).

Strengths, Limitations, and Future Directions

There were several unique strengths of this study as well as limitations that should be addressed in future investigations. A strength of the study is the use of a large and ethnically/racially diverse sample, which allows for greater generalizability of findings. Another strength is the use of multiple reporters to address the problem of shared method variance with the use of a teacher-report measure of learned helplessness. This multi-method approach provides the opportunity to understand children’s cognitive and behavioral tendencies from the perspective of the child and outside observer, which offers confidence in interpretations of findings (Holmbeck, Li, Schurman, Friedman, & Coakley, 2002). A third strength was the inclusion of both adaptive and maladaptive outcome variables to tease apart the potential differential impact that AS for positive and negative events may have on theoretically related positive and negative outcomes. Additionally, this was the first study to test the extent to which the CASI-II’s factor structure is consistent with theory at both the individual and school levels using a MCFA. Furthermore, this study was the first to explicitly test the fit of GP model examining stable and global attributions.

Some limitations of this study include a limited age range and a sample size not large enough to test for ethnic/racial differences. Future studies could examine the utility of the CASI-II with a wider age range of children and explore whether the measurement of children’s AS differs by ethnicity/race. Furthermore, the CASI-II could be compared to other theoretically unrelated constructs to test for discriminant validity to supplement the convergent validity demonstrated in the current study. In addition, the current study only investigated cross-sectional associations to test convergent validity. Future research should test predictive validity of the CASI-II using longitudinal data. Finally, additional research could further examine the measurement and functioning of internality in attributional processes, especially as it relates to negative events. The lack of reliability and validity evidence suggest that more work is needed to assess this dimension with confidence, and the unexpected pattern of correlations highlight a need for more theoretical clarity.

Conclusion

The majority of the body of literature on AS has been focused on the role that attributions for negative events play in the development of depression. The CASI-II can be useful to expand this literature from an exclusive focus on negative events, to include the role of AS for positive events. In addition, the CASI-II, with its adequate psychometric properties, can be useful to expand this literature from a focus on general AS to a focus on attributional dimensions of internality, stability, and globality. Overall, the results from the current study offer confidence for the use of the CASI-II in important theory-testing studies focused on risk and protective factors related to depression and in the emerging field of positive psychology (Rigby & Huebner, 2005).

Supplemental Material

SupplementalTable – Supplemental Material for Measurement of Attributional Style for Positive and Negative Events Using the Children’s Attributional Style Interview-II

Supplemental Material, SupplementalTable for Measurement of Attributional Style for Positive and Negative Events Using the Children’s Attributional Style Interview-II by Jake C. Steggerda and Sandra Y. Rueger in Journal of Psychoeducational Assessment

Footnotes

Acknowledgments

The authors would like to thank Kenyon Woods Middle School for their participation in the data collection phase of this project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Wheaton College Alumni Association and the Wheaton College Summer Fellowship Program.

ORCID iD

Jake C. Steggerda

Supplemental Material

Supplemental material for this article is available online.

References

Abramson

L. Y.

Metalsky

G. I.

Alloy

L. B.

(1989). Hopelessness depression: A theory-based subtype of depression. Psychological Review, 96(2), 358-372. doi:10.1037/0033-295X.96.2.358

Abramson

L. Y.

Seligman

M. E.

Teasdale

J. D.

(1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87(1), 49-74. doi:10.1037/0021-843X.87.1.49

Barrett

(2007). Structural equation modelling: Adjudging model fit. Personality and Individual differences, 42(5), 815-824. doi:10.1016/j.paid.2006.09.018

Cohen

(1988). Statistical power analysis for the behavioural sciences (2nd ed.). Hillsdale, NJ: L. Erlbaum Associates.

Cole

D. A.

Ciesla

J. A.

Dallaire

D. H.

Jacquez

F. M.

Pineda

A. Q.

LaGrange

, ... Felton

J. W.

(2008). Emergence of attributional style and its relation to depressive symptoms. Journal of Abnormal Psychology, 117(1), 16-31. doi:10.1037/0021-843X.117.1.16

DeVellis

R. F.

(2016). Scale development: Theory and applications (Vol. 26). Newbury Park, CA: Sage publications.

Dunn

T. J.

Baguley

Brunsden

(2014). From alpha to omega: A practical solution to the pervasive problem of internal consistency estimation. British Journal of Psychology, 105(3), 399-412. doi:10.1111/bjop.12046

Hadley

Allen

D. N.

Palmer

Klosky

Deptula

, … Cohen

(2005). A new measure of children’s optimism and pessimism: The youth life orientation test. Journal of Child Psychology and Psychiatry, 46(5), 548-558. doi:10.1111/j.1469-7610.2004.00372.x

Faulstich

M. E.

Carey

M. P.

Ruggiero

Enyart

Gresham

(1986). Assessment of depression in children and adolescence: An evaluation of the center for epidemiological studies depression scale for children (CES-DC). The American Journal of Psychiatry, 143, 1024-1027. doi:10.1176/ajp.143.8.1024

10.

Fincham

F. D.

Hokoda

Sanders

Jr. (1989). Learned helplessness, test anxiety, and academic achievement: A longitudinal analysis. Child Development, 60, 138-145. doi:10.2307/1131079

11.

Gillham

J. E.

Reivich

K. J.

Jaycox

L. H.

Seligman

M. E. P.

(1995). Prevention of depressive symptoms in schoolchildren: Two-year follow-up. Psychological Science, 6(6), 343-351. doi:10.1111/j.1467-9280.1995.tb00524.x

12.

Gordeeva

Sheldon

Sychev

(2019). Linking academic performance to optimistic attributional style: Attributions following positive events matter most. European Journal of Psychology of Education, 35, 1-28. doi:10.1007/s10212-019-00414-y

13.

Haines

B. A.

Wells

R. M.

Rueger

S. Y.

Conley

C. S.

Louie

B. K.

Lukk

A. A.

(2005). Manual for the children’s attributional style Interview: Paper-and-pencil versions. (Unpublished manuscript). Department of Psychology, Lawrence University, Appleton, WI.

14.

Hankin

B. L.

Abramson

L. Y.

(2002). Measuring cognitive vulnerability to depression in adolescence: Reliability, validity, and gender differences. Journal of Clinical Child & Adolescent Psychology, 31(4), 491-504. doi:10.1207/S15374424JCCP3104_8

15.

Higgins

N. C.

Zumbo

B. D.

Hay

J. L.

(1999). Construct validity of attributional style: Modeling context-dependent item sets in the attributional style questionnaire. Educational and Psychological Measurement, 59(5), 804-820. doi:10.1177/00131649921970152

16.

Holmbeck

G. N.

S. T.

Schurman

J. V.

Friedman

Coakley

R. M.

(2002). Collecting and managing multisource and multimethod data in studies of pediatric populations. Journal of Pediatric Psychology, 27(1), 5-18. doi:10.1093/jpepsy/27.1.5

17.

Kaslow

N. J.

Nolen-Hoeksema

(1991). Children’s attributional style questionnaire revised. (Unpublished manuscript). Emory University, Atlanta, GA.

18.

Kaslow

N. J.

Tanenbaum

R. L.

Seligman

M. E. P.

(1978). The children’s attributional style Questionnaire. Unpublished manuscript.

19.

Needles

D. J.

Abramson

L. Y.

(1990). Positive life events, attributional style, and hopefulness: Testing a model of recovery from depression. Journal of Abnormal Psychology, 99(2), 156-165. doi:10.1037/0021-843X.99.2.156

20.

Nolen-Hoeksema

Girgus

J. S.

Seligman

M. E.

(1992). Predictors and consequences of childhood depressive symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology, 101(3), 405-422. doi:10.1037/0021-843X.101.3.405

21.

Peterson

Semmel

Von Baeyer

Abramson

L. Y.

Metalsky

G. I.

Seligman

M. E.

(1982). The attributional style questionnaire. Cognitive Therapy and Research, 6(3), 287-299. doi:10.1007/BF01173577

22.

Pornprasertmanit

Lee

Preacher

K. J.

(2014). Ignoring clustering in confirmatory factor analysis: Some consequences for model fit and standardized parameter estimates. Multivariate behavioral research, 49(6), 518-543.

23.

R Core Team . (2016). R: A language and environment for statistical computing [computer software manual]. Vienna, Austria.

24.

Reivich

(1995). The measurement of explanatory style. In Buchanan

G. M.

Seligman

M. E. P.

(Eds.), Explanatory style (pp. 57-70). Mahwah, NJ: Lawrence Erlbaum Associates.

25.

Revelle

Zinbarg

R. E.

(2009). Coefficients alpha, beta, omega, and the glb: Comments on Sijtsma. Psychometrika, 74(1), 145-154. doi:10.1007/s11336-008-9102-z

26.

Rigby

B. T.

Huebner

E. S.

(2005). Do causal attributions mediate the relationship between personality characteristics and life satisfaction in adolescence? Psychology in the Schools, 42(1), 91-99. doi:10.1002/pits.20026

27.

Ritschel

L. A.

Ramirez

C. L.

Cooley

J. L.

Edward Craighead

(2016). Behavioral activation for major depression in adolescents: Results from a pilot study. Clinical Psychology: Science and Practice, 23(1), 39-57. doi:10.1111/cpsp.12140

28.

Rosseel

(2012). “Lavaan: An R package for structural equation modeling.” Journal of Statistical Software, 42(2), 1-36. http://www.jstatsoft.org/v48/i02/

29.

Rueger

S. Y.

George

(2017). Indirect effects of attributional style for positive events on depressive symptoms through self-esteem during early adolescence. Journal of Youth and Adolescence, 46, 701-708. doi:10.1007/s10964-016-0530-2

30.

Rueger

S. Y.

Haines

B. A.

Malecki

C. K.

(2010). Children’s attributional style from middle childhood to early adolescence: Further validation of the paper-and-pencil versions of the children’s attributional style interview. Assessment, 17(3), 294-307. doi:10.1177/1073191109356528

31.

Rueger

S. Y.

Malecki

C. K.

(2007). Group administration of the children’s attributional style interview. Journal of Clinical Child & Adolescent Psychology, 36(3), 476-482. doi:10.1080/15374410701448711

32.

Rueger

S. Y.

Malecki

C. K.

(2011). Effects of stress, attributional style and perceived parental support on depressive symptoms in early adolescence: A prospective analysis. Journal of Clinical Child & Adolescent Psychology, 40(3), 347-359. doi:10.1080/15374416.2011.563461

33.

Snyder

C. R.

Hoza

Pelham

W. E.

Rapoff

Ware

Danovsky

, … Stahl

K. J.

(1997). The development and validation of the children’s hope scale. Journal of Pediatric Psychology, 22(3), 399-421. doi:10.1093/jpepsy/22.3.399

34.

Taber

K. S.

(2018). The use of Cronbach’s alpha when developing and reporting research instruments in science education. Research in Science Education, 48(6), 1273-1296. doi:10.1007/s11165-016-9602-2

35.

Valle

M. F.

Huebner

E. S.

Suldo

S. M.

(2004). Further evaluation of the children’s hope scale. Journal of Psychoeducational Assessment, 22(4), 320-337. doi:10.1177/073428290402200403

36.

Vandenberg

R. J.

(2006). Introduction: Statistical and methodological myths and urban legends: where, pray tell, did they get this idea? Organizational Research Methods, 9(2), 194-201. doi:10.1177/1094428105285506

37.

Vandenberg

R. J.

Lance

C. E.

(2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 4-70. doi:10.1177/109442810031002

38.

Weissman

M. M.

Orvaschel

Padian

(1980). Children’s symptom and social functioning self-report scales: Comparison of mothers’ and children’s reports. The Journal of Nervous and Mental Disease, 168(12), 736-740. doi:10.1097/00005053-198012000-00005

39.

Wong

P. T.

Weiner

(1981). When people ask “why” questions, and the heuristics of attributional search. Journal of Personality and Social Psychology, 40(4), 650-663. doi:10.1037/0022-3514.40.4.650

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB

0.03 MB

0.04 MB