Search Results

You are looking at 1 - 10 of 13 items for :

  • "structural invariance" x
  • Refine by Access: All Content x
Clear All
Restricted access

John B. Bartholomew, Alexandra Loukas, Esbelle M. Jowers, and Shane Allua

Background:

Design and evaluation of physical activity interventions depends upon valid instruments to assess mediating processes. The Physical Activity Self-Efficacy Scale (PASES) has been used in a variety of forms within samples of African American and Caucasian children.

Method:

This study was designed to extend the statistical validity of the scores from the PASES by comparing 1 and 3-factor models and testing measurement invariance between Hispanic and Caucasian children. 883 fourth and fifth graders were recruited (mean age, 9.71 y; 48% female, 52% male; 67% Hispanic, 33% Caucasian). The factor structure was tested with confirmatory factor analysis, using two-group analyses to model ethnic differences.

Results:

The 17-item, 3-factor version of the PASES evidenced poor fit with the data. In contrast, an 8-item, 1-factor solution provided adequate fit for both samples.

Conclusions:

The 8-item, 1-factor version of the PASES provides statistically valid scores for Hispanic and Caucasian children.

Restricted access

James W. Adie, Joan L. Duda, and Nikos Ntoumanis

Grounded in the 2 × 2 achievement goal framework (Elliot & McGregor, 2001), a model was tested examining the hypothesized relationships between approach and avoidance (mastery and performance) goals, challenge and threat appraisals of sport competition, and positive and negative indices of well-being (i.e., self-esteem, positive, and negative affect). A further aim was to determine the degree to which the cognitive appraisals mediated the relationship between the four achievement goals and the indicators of athletes’ welfare. Finally, measurement and structural invariance was tested with respect to gender in the hypothesized model. An alternative model was also estimated specifying self-esteem as an antecedent of the four goals and cognitive appraisals. Four hundred and twenty-four team sport participants (M age = 24.25) responded to a multisection questionnaire. Structural equation modeling analyses provided support for the hypothesized model only. Challenge and threat appraisals partially mediated the relationships observed between mastery-based goals and the well-being indicators. Lastly, the hypothesized model was found to be invariant across gender.

Restricted access

David E. Conroy, Andrew J. Elliot, and Scott M. Hofer

Achievement goals in sport have traditionally been defined according to the definition of competence alone (i.e., mastery/task, performance/ego). Emerging research and theory from the academic domain indicates that the utility of the achievement goal construct can be enhanced when the valence (i.e., approach, avoidance) of goals is also considered in conjunction with the definition of competence. The present study was designed to evaluate the psychometric properties of scores for mastery-approach, mastery-avoidance, performance-approach, and performance-avoidance goals from a measure of achievement goals in sport. The a priori 2 × 2 model fit the data better than any of the plausible alternative models. In a series of longitudinal factorial invariance and latent growth curve analyses, scores for the four subscales exhibited structural invariance, and both differential and latent mean stability over a 21-day period. Achievement goal scores conformed to theoretical predictions regarding their relationship with fear-of-failure scores. The AGQ-S would be an appropriate instrument for future research using the 2 × 2 model of achievement goals in sport, particularly for experimental/intervention research on change processes associated with achievement goals.

Restricted access

Fotini Venetsanou, Irene Kossyva, Nadia Valentini, Anastasia-Evangelia Afthentopoulou, and Lisa Barnett

order to choose the most parsimonious model, the Akaike Information Criterion (AIC) was used ( Vrieze, 2012 ). Fornell and Larcker’s ( 1981 ) recommendations were employed to estimate the composite reliability for each factor. To analyze the invariance of the model and the structural invariance of PMSC

Restricted access

Christine E. Pacewicz and Alan L. Smith

interval (CI) for an indirect effect did not span zero. Multigroup latent analysis was then used to test for measurement and structural invariance of the final model by sex. Base models were established for boys and girls, followed by testing configural (i.e., same latent variables present in each group

Restricted access

Nadia C. Valentini, Lisa M. Barnett, Paulo Felipe Ribeiro Bandeira, Glauber Carvalho Nobre, Larissa Wagner Zanella, and Rodrigo Flores Sartori

Information Criterion (AIC) was used to choose the most parsimonious model ( Vrieze, 2012 ). The composite reliability for each factor was estimated following the Fornell and Larcker ( 1981 ) recommendations. A multigroup analysis to assess the invariance of the model and the structural invariance of PMSC

Restricted access

Andre Koka and Heino Sildala

assessing the structural invariance by constraining factor covariances, in addition to factor loadings and item intercepts, to be equivalent across groups. Multi-group model comparison was based on the χ 2 difference test. However, because the χ 2 statistic is oversensitive to sample size (e.g., Hu

Restricted access

Jorge Zamarripa, René Rodríguez-Medellín, and Fernándo Otero-Saborido

] .919 .923 M1 Structural invariance (baseline model) 2986 15,154.75** .075 [.073, .076] .913 .915 .002 .000 .001 M2 Total invariance 3058 15,102.01** .074 [.073, .075] .915 .915 .001 .002 .000 Note. All the Δ index comparisons are made with respect to the baseline model (M1). df  = degrees of freedom

Restricted access

Yonghwan Chang, Yong Jae Ko, and Brad D. Carlson

attitudes for consumers with high involvement ( H1 ). Test of hypothesis 2 The structural invariance test, comparing the unconstrained model with the three-path constrained model, supported the moderating effects of involvement in the causal links of implicit attitudes and explicit affective attitudes, Δχ 2

Restricted access

Yonghwan Chang

.e., testing H1, H2, and H3). To test H4, the multiple-group structural equation modeling (MGSEM) including measurement and structural invariance tests were performed. Test of H1 The results of the three-way ANOVA and follow-up univariate analyses revealed the significant main effects of fit on explicit