d.dep.t.rm: d for Repeated Measures with Average SD Denominator

In doomlab/MOTE: Effect Size and Confidence Interval Calculator

d for Repeated Measures with Average SD Denominator

Description

This function displays d and the non-central confidence interval for repeated measures data, using the average standard deviation of each level as the denominator, but controlling for r.

Usage

d.dep.t.rm(m1, m2, sd1, sd2, r, n, a = 0.05)

Arguments

m1	mean from first level
m2	mean from second level
sd1	standard deviation from first level
sd2	standard deviation from second level
r	correlation between first and second level
n	sample size
a	significance level

Details

To calculate d, mean two is subtracted from mean one, which is divided by the average standard deviation, while mathematically controlling for the correlation coefficient (r).

d_rm = ((m1 - m2) / sqrt(( sd1^2 + sd2^2 ) - (2 x r x sd1 x sd2))) x sqrt(2 x (1-r))

Learn more on our example page.

Value

Controls for correlation and provides the effect size (Cohen's d) with associated confidence intervals,m the confidence intervals associated with the means of each group,mstandard deviations and standard errors of the means for each group.

d	effect size
dlow	lower level confidence interval d value
dhigh	upper level confidence interval d value
M1	mean one
sd1	standard deviation of mean one
se1	standard error of mean one
M1low	lower level confidence interval of mean one
M1high	upper level confidence interval of mean one
M2	mean two
sd2	standard deviation of mean two
se2	standard error of mean two
M2low	lower level confidence interval of mean two
M2high	upper level confidence interval of mean two
r	correlation
n	sample size
df	degrees of freedom (sample size - 1)
estimate	the d statistic and confidence interval in APA style for markdown printing

Examples

#The following example is derived from the "dept_data" dataset included
#in the MOTE library.

#In a study to test the effects of science fiction movies on people's
#belief in the supernatural, seven people completed a measure of belief
#in the supernatural before and after watching a popular science fiction
#movie. Higher scores indicated higher levels of belief.

    t.test(dept_data$before, dept_data$after, paired = TRUE)

    scifi_cor = cor(dept_data$before, dept_data$after, method = "pearson",
                use = "pairwise.complete.obs")

#You can type in the numbers directly, or refer to the dataset,
#as shown below.

    d.dep.t.rm(m1 = 5.57, m2 = 4.43, sd1 = 1.99,
                sd2 = 2.88, r = .68, n = 7, a = .05)

    d.dep.t.rm(5.57, 4.43, 1.99, 2.88, .68, 7, .05)

    d.dep.t.rm(mean(dept_data$before), mean(dept_data$after),
                sd(dept_data$before), sd(dept_data$after),
                scifi_cor, length(dept_data$before), .05)

#The mean measure of belief on the pretest was 5.57, with a standard
#deviation of 1.99. The posttest scores appeared lower (M = 4.43, SD = 2.88)
#but the dependent t-test was not significant using alpha = .05,
#t(7) = 1.43, p = .203, d_rm = 0.43. The effect size was a medium effect suggesting
#that the movie may have influenced belief in the supernatural.

doomlab/MOTE documentation built on April 17, 2022, 2:08 a.m.