KembaraXtra – Psychology: Common Language Effect Size Statistic

Definition

The common language effect size statistic, abbreviated CL or CLES, expresses an effect size as the probability that a randomly selected score from one distribution will be greater than a randomly selected score from another. It translates a group difference into an intuitive probability rather than presenting it only in standardized units.

Statistical Mechanism

When two distributions are assumed to be normal, the statistic can be calculated by dividing the difference between their means by the square root of the sum of their variances. This produces a z score, whose cumulative probability gives the common language effect size. For example, a value of .92 means that a randomly chosen person from one group has a 92 per cent chance of scoring higher than a randomly chosen person from the comparison group.

Role in Psychological Research

The measure was developed to make effect sizes more understandable to non-specialists. Unlike Cohen’s d, which describes a difference in standard-deviation units, CLES converts the same general comparison into an everyday probability statement. It is useful in psychology, education, medicine, and communication of research findings to the public.

Interpretation and Limitations

The statistic does not describe every feature of the distributions and may depend on assumptions about their shape and variability. A high probability does not necessarily imply practical importance, causal influence, or overlap-free groups. It should be reported alongside descriptive statistics, confidence intervals, study design, and contextual interpretation.

Summary

The common language effect size statistic expresses a group difference as the probability that a randomly selected score from one group exceeds one from another. It offers an intuitive alternative to standardized effect sizes. The measure is especially useful for communicating findings clearly. Its interpretation still requires attention to assumptions and practical context.