Development of Psychological Safety Scales for Observational and Experimental Groups

—Using Multi-level Model—

Abstract

The purpose of this study was to develop a psychological safety scale for science observation and experiment groups. Validity was examined by adding novel items to the four that have been ordinarily used up to now. From the results, certain validity could be confirmed from the evidence of structural and external aspects. The analysis also showed that most of the variance in psychological safety came from within-level, since the ICC₁ values for each item were not very large. Since the value of the ω coefficient was sufficient, it can be confirmed that this scale is indeed capable of accurately assessing the psychological safety status of individual students. When dealing with psychological safety in basic research, it is desirable to conduct the analysis after decomposing the between- and within- variance components via methods such as hierarchical linear models and MCFA, while considering the possibility of increasing the probability of Type I error.