Re-assessing the Psychometric Properties of Stress Appraisal Measure in Ghana Using Multidimensional Graded Response Model

The development and validation of the Chinese safety climate scale using the item-response theory approach

Introduction

To develop a valid and practical tool to measure the safety climate in China, and further raise awareness of it in Chinese industries, we developed the Chinese safety climate scale (including two subscales at the levels of organization and group separately) based on the work of Huang et al. in 2017.

Methods

A descriptive survey with the convenience sampling method was conducted in Shanghai Disney Resort, China. A sample of 1,570 employees was involved in the final data analyses. The item response theory (IRT) analyses with graded response models were conducted using.

Results

The unidimensionality and local independence assumption were held. The Cronbach's α of organization- and group- level safety climate subscale was 0.912 and 0.937, respectively. The confirmatory factor analysis (CFA) showed good model fit for a one-factor model of the organization-level safety climate subscale, χ² (df = 20) = 129.158, p < 0.001, CFI = 0.994, TLI = 0.992, NFI = 0.993, IFI = 0.994, RMSEA = 0.059, 90% CI = (0.050, 0.069), and SRMR = 0.048. A one-factor model also fits well for the group-level safety climate subscale, χ² (df = 44) = 219.727, p < 0.001, CFI = 0.996, TLI = 0.9925, NFI = 0.995, IFI = 0.996, RMSEA = 0.050, 90% CI = (0.044, 0.057), and SRMR = 0.046. Discrimination and difficulty parameters showed that all items effectively spanned the range of the latent trait and could successfully separate participants at different safety climate levels. Items in the organization-level and group-level Chinese safety climate subscales had significantly different discrimination parameters, fitted well with the models, and had a substantive relationship with the latent traits.

Discussion

The Chinese safety climate scale was reliable and valid overall. They can facilitate the research and survey regarding the safety climate in China.

Examining the underlying latent structure of the sports emotion questionnaire: Insights from the bifactor multidimensional item response theory

Background

Despite the widespread use of the sports emotion questionnaire (SEQ) in several studies, it is surprising that only a few have explicitly tested the validity and utility of the instrument in non-western populations. Besides, the issue of dimensionality and the latent structure of the instrument remain inconclusive given that several authors have revealed different factor structures across diverse populations. The central concern is whether the items on the various dimensions, proposed for the original SEQ, offer adequate information to their respective expected subscale or otherwise. This study assessed the underlying latent structure of the SEQ using confirmatory and bifactor multidimensional item response (MIRT) models.

Methods

Through a well-designed validation study 300 athletes from three West African countries, participating in the 2018 West African University Games were surveyed to respond to the SEQ. The data were analyzed using first, a 5-factor confirmatory factor analysis (CFA) via the MIRT model and second, a bifactor MIRT analysis.

Results

The results revealed that items on the SEQ were fairly good in measuring the construct under the respective domains of the instrument. However, the outcome of the bifactor model showed that the majority of the items on the SEQ explained common variance in relation to the general factor other than the specific domains (5-dimensions).

Conclusion

Findings of the bifactor model question whether the sub-dimensions of the SEQ are needed since most of the items on the SEQ explained larger variances in the general factor than any of the five domains. It is concluded that instruments like SEQ should be scored for a general factor and not as sub-dimensions. Further investigations are encouraged by scholars within the area to probe the dimensionality of the SEQ.

Validation of the WHO-5 Well-Being Scale among Adolescents in Ghana: Evidence-Based Assessment of the Internal and External Structure of the Measure

The WHO-5 well-being measure happens to be one of the most renowned measures of subjective well-being across the globe. Although the instrument has been calibrated in different countries, its psychometric properties and applicability in Africa, especially in Ghana, are not known. In this study, the WHO-5 well-being scale was validated among adolescents in Ghana by assessing the validity evidence of the measure based on the internal and external structure. In particular, the study examined the (1) dimensionality of the WHO-5 well-being scale, (2) quality of the items (including the scale functioning) for the measure, and (3) criterion validity of the well-being measure. Using a survey approach, 997 adolescents were recruited in secondary schools across the northern belt of Ghana. The study found a one-factor structure of the scale, which supports the factor solution of the original measure. The items were found to be of high quality, except for one item. The WHO-5 well-being measure was found to have sufficient evidence regarding convergent and divergent validity. The outcome of this validation study provides support for the validity and reliability of the WHO-5 well-being scale’s utility and use among adolescents in Ghana. The study encourages further validation studies to be conducted in Ghana to widen the reproducibility of the WHO-5 well-being measure.