cohensd_rm: Cohen's d standardized mean change effect size for repeated...

In gitrman/itns: Datasets from the book Introduction to the New Statistics

Description

Computes Cohen's d effect size for repeated measures designs (paired samples), using the standardizer recommended by Cumming (2012). In other words, the standardizer is the average of the pre and post-treatment standard deviations, rather than the standard deviation of the change scores. An approximate noncentral-t confidence interval is computed using the method proposed by Algina & Keselman (2003), Equations 7 to 9. The effect size estimate can be corrected for sample sample bias (see Cumming, 2012, p.294) by setting the Unbiased argument to TRUE.

Usage

cohensd_rm(x, y, ci = 95, unbiased = FALSE)

Arguments

x	Numeric vector of observations at time 1 (e.g., pre-test)
y	Numeric vector of observations at time 2 (e.g., post-test)
ci	Confidence level. Default is 95 (for a 95 percent CI).
unbiased	Logical. If TRUE, the estimated effect size is corrected for small-sample bias. Default is FALSE.

Value

A numeric vector of length three comprising the estimated effect size (est), lower limit of the confidence interval (ll), and upper limit of the confidence interval (ul).

References

Algina, J. A., & Keselman, H. J. (2003). Approximate Confidence Intervals for Effect Sizes. Educational and Psychological Measurement, 63, 537-553.

Cumming, G. (2012). Understanding The New Statistics: Effect Sizes, Confidence Intervals, and Meta-Analysis. Routledge; New York.

Examples

## Not run: 
   cohensd_rm(x = thomason1$pre, y = thomason1$post)

## End(Not run)

gitrman/itns documentation built on May 17, 2019, 5:29 a.m.