PSYCHOMETRIC PROPERTIES OF ANXIETY ABOUT AGING SCALE AMONG MALAYSIAN YOUTHS: A CONFIRMATORY FACTOR ANALYSIS

Hazwan Mat Dina,Halimatus Sakdiah Minhata

doi:10.47263/jasem.5(1)03

Abstract

The Anxiety about Aging Scale (AAS) is a questionnaire developed based on multidimensional aging anxiety to measure anxiety towards aging. However, the AAS constructs and items vary depending on the study population. This study aimed to explore the validity and reliability of AAS through evaluation of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) among youths in Malaysia. A cross-sectional was conducted among 1988 university and college students in Klang Valley, Malaysia, selected using stratified proportionate random sampling. EFA results suggested four factors solution based on the interpretation of Eigenvalues and scree plot with 59.11% variance extracted. Results of CFA supported the four-factor model of revised 17-item AAS with acceptable model fit indices and high factor loading. The revised 17-item AAS has good reliability through the assessment of the internal consistency of the items. In conclusion, the revised 17-item AAS measures four distinct factors is a valid and reliable questionnaire to measure anxiety towards aging among Malaysian youths

Highlights

Partial least squares structural equation modeling (PLS-SEM), known as PLS Path Modeling, is one of the most widely used methods of multivariate data analysis among business and social science scholars
We discuss three different commercial “stand alone” PLS-SEM applications with graphical user interface, namely SmartPLS (Ringle et al, 2015), WarpPLS (Kock, 2017) and ADANCO (Henseler & Dijkstra, 2015), which are currently available on the market
The software provides additional algorithms that are useful for understanding and modelling composite-based models, such as advanced bootstrapping (Aguirre-Urreta & Rönkkö, 2018; Hair et al, 2022), confirmatory tetrad analysis (Gudergan et al, 2008), importance-performance map analysis (Ringle & Sarstedt, 2016), predictive power assessment using PLSpredict (Shmueli et al, 2016; Shmueli et al, 2019), predictive model comparison based on information criteria such as BIC (Chin et al, 2020; Liengaard et al, 2021; Sharma et al, 2019a, 2019b), multi-group analysis based on bootstrapping and permutation (Cheah et al, 2020; Chin & Dibbern, 2010; Hair et al, 2018b), latent class segmentation using finite mixture PLS (Hahn et al, 2002; Sarstedt et al, 2011), and prediction-oriented segmentation (Becker et al, 2013)

Summary

Introduction

Partial least squares structural equation modeling (PLS-SEM), known as PLS Path Modeling, is one of the most widely used methods of multivariate data analysis among business and social science scholars. We discuss three different commercial “stand alone” PLS-SEM applications with graphical user interface, namely SmartPLS (Ringle et al, 2015), WarpPLS (Kock, 2017) and ADANCO (Henseler & Dijkstra, 2015), which are currently available on the market (see Table 1).

Results

Conclusion