Development of a Short Version of the Perceived Social Support Scale: Based on Classical Test Theory and Item Response Theory
Yunchou Wu,Jingqi Tang,Zhuoran Du,ChenKe,Xinyi Zhang,Fei Wang,Xiaonan Sun,Xinying Sun,Yibo Wu
DOI: https://doi.org/10.31234/osf.io/d95zg
2022-01-01
Abstract:Perceived social support is defined as the cognitive assessment of an individual’s ability to establish reliable relationships with others in society. The most commonly used measurement tool for this construct is the Multidimensional Scale of Perceived Social Support (MSPSS) created by Zimet et al. (1988). This scale has been widely used in foreign countries, applied and validated in different populations, and gradually extended to different age groups and populations in China. Although the MSPSS scale consists of only 12 items, large cross-sectional surveys often include multiple scales of different types. Using shorter scales can make it quicker and more convenient to apply them in research and clinical work. Therefore, this study aims to simplify the MSPSS scale for Chinese residents and examine its psychometric properties. A multi-stage stratified cluster sampling method was employed to collect data from 10,914 residents in 120 cities across the country. After randomly dividing the data into two parts, exploratory analysis and confirmatory analysis were conducted separately. During the exploratory analysis of Dataset 1, the scale was simplified using both Classical Test Theory (CTT) and the Mokken model in Item Response Theory (IRT), resulting in one short form for each approach. Additionally, three short forms were obtained by introducing other short forms of the MSPSS scale derived from multidimensional item response theory in previous studies. The empirical validity and criterion-related validity of the three short forms were tested. Subsequently, the reliability, dimensionality, validity, and assumption of the Mokken model for the two short forms of the MSPSS scale were examined using Dataset 2. Based on the data analysis results, the 7-item short form developed based on the Mokken model assumption exhibited the best psychometric properties among all the short forms. Moreover, the 3-item short form composed of items with the highest factor loadings in each dimension also demonstrated good psychometric properties. Therefore, this study suggests that in future research, if a short form of the MSPSS scale is needed for measurement purposes, the 3-item short form is suitable for clinical settings, large-scale cross-sectional studies, or longitudinal studies, while the 7-item short form is more appropriate for more precise measurements.