STRUCTURAL STABILITY OF ENGLISH LEARNING MOTIVATION IN THAI OLDER ADULTS : A ROBUSTNESS INVESTIGATION USING MONTE CARLO SIMULATION

Abstract

Research involving older adult learners often encounters statistical power limitations inherent in small sample sizes, which can compromise the precision of statistical inferences. This study investigates the methodological robustness of the motivational relationship structure in learning English among Thai older adults. Adopting a simulation-based robustness approach, the research integrates Monte Carlo generation with bootstrap resampling. Secondary data from a sample of 77 Thai older adults, recruited via online snowball sampling, served as parameters to generate a pseudo-population, from which 10,000 simulation iterations were performed.

Findings reveal differential stability among motivational variables. Under the normality assumption, the relationship involving "Attitude toward Learning" demonstrated high robustness, maintaining stable correlation coefficients across the majority of simulated scenarios. Conversely, "Anxiety" exhibited sampling fragility. These results are synthesized into a conceptual proposition regarding the stability of motivation. The study advocates for the application of Monte Carlo simulation as a methodological cross-validation strategy, suggesting that classical inference alone may be insufficient for small-sample studies in social science research.