Abstract

In the context of EFL research, Exploratory Factor Analysis (EFA) is one of the first methods used to reveal a latent construct. However, a very significant challenge, one of the key issues in sample size determination for the analysis, continues to be a sustainable problem with the need to have clear-cut standards for attaining reliable factor structures. This gap, therefore, leads to the present study whose purpose is to investigate the optimal sample size for EFA in EFL contexts. The study employed a reading self-efficacy questionnaire with 25 items (Sukserm et al., 2023). For purposes, seven subsamples were systematically generated with the sample sizes of 125, 175, 250, 325, 375, 425, and 500 in the ratios of variable to samples of 5, 7, 10, 13, 15, 17, and 20 participants respectively. The results showed that for stable factor loadings, high communalities and explained variance to be possible, a sample of 250 participants, or 1 variable per 10 samples, could be optimum. It would be possible to obtain equally reliable results from larger samples such as those with 325 to 500 participants, but such numbers did not seem to add much more value above the minimum.

This work is licensed under a Creative Commons Attribution 4.0 License.