How skewed are psychological data? A standardized index of effect size.

In this article the author discusses the lack of attention given to the concept of skewness in research reported in the general psychological literature. Although the report of descriptive statistics, such as means and standard deviations, is standard practice, skewness is rarely reported. Significance tests are available for determining departure from symmetry and normality, but do not indicate the magnitude of skew. Indexes for standardizing skew have been offered in specialized areas of research, such as structural equation modeling but no standardized index of effect size exists on a unit (0-1) scale for the pragmatic interpretation of the meaning of a given skewness value. Therefore, in this article the author proposes an index that standardizes the raw skewness value relative to its maximum possible value, enabling an interpretation of the mystical raw score value in 0-1 terms. The author discusses utility of the index for general psychological research, and suggestions for future research to provide an empirical basis for establishing effect-size benchmarks of low, medium, and high skewness.