# A reliability generalization meta-analysis of self-report measures of statistics anxiety

## TL;DR

This study evaluates the reliability of self-report statistics anxiety scales using meta-analysis to determine overall reliability coefficients.

## Contribution

The study provides a reliability generalization meta-analysis of statistics anxiety scales, revealing their average reliability and influencing factors.

## Key findings

- The mean Cronbach’s alpha coefficient across scales was .927 under a random effects model.
- Scale type significantly influenced the mean alpha value.
- Mean scale score and its standard deviation were significant predictors of reliability.

## Abstract

In this study, it was aimed to obtain a general reliability coefficient for each of the statistics anxiety scales. For this purpose, Web of Science, ERIC and Scopus databases and Google Scholar search engine were searched according to certain criteria and 84 Cronbach’s alpha coefficients reported for the whole scale were reached. Reliability generalization meta-analysis method was applied to the obtained alpha coefficients within the scope of meta-analysis, which is a quantitative method.

The mean value of the alpha coefficients for which the transformation formula was applied was found to be .927 under the random effects model, and the findings were statistically significant. In addition, the mean alpha value of each statistics anxiety scale was .931 for STARS, .917 for SAS, .918 for SAS-10 and .951 for WAESTA. Analog to the ANOVA and meta-regression analyses were conducted to reveal the heterogeneity of alpha coefficients in the overall analysis. Analog to ANOVA was applied for five different categorical variables, and according to the findings, it was observed that the mean alpha value differed statistically significantly depending on the scale type variable. Moreover, it was found that the mean scale score and the standard deviation of the mean scale score were statistically significant predictors of the mean alpha value.