Validating the cross-cultural factor structure and invariance property of the Insomnia Severity Index: evidence based on ordinal EFA and CFA

• The factorial validity and invariance of the Insomnia Severity Index (ISI) were examined.
• The validity of the three-factor model of ISI is supported.
• The strong invariance of ISI between cross-cultural data sets is also supported.

ObjectiveThe purpose of this study is to examine the factor structure of the Insomnia Severity Index (ISI) across samples recruited from different countries. We tried to identify the most appropriate factor model for the ISI and further examined the measurement invariance property of the ISI across samples from different countries.MethodsOur analyses included one data set collected from a Taiwanese sample and two data sets obtained from samples in Hong Kong and Canada. The data set collected in Taiwan was analyzed with ordinal exploratory factor analysis (EFA) to obtain the appropriate factor model for the ISI. After that, we conducted a series of confirmatory factor analyses (CFAs), which is a special case of the structural equation model (SEM) that concerns the parameters in the measurement model, to the statistics collected in Canada and Hong Kong. The purposes of these CFA were to cross-validate the result obtained from EFA and further examine the cross-cultural measurement invariance of the ISI.ResultsThe three-factor model outperforms other models in terms of global fit indices in Taiwan's population. Its external validity is also supported by confirmatory factor analyses. Furthermore, the measurement invariance analyses show that the strong invariance property between the samples from different cultures holds, providing evidence that the ISI results obtained in different cultures are comparable.ConclusionsThe factorial validity of the ISI is stable in different populations. More importantly, its invariance property across cultures suggests that the ISI is a valid measure of the insomnia severity construct across countries.