Paired samples t-test

Aim

Calculate The aim of the analysis: N for required N, Power for estimated power and Effect Size for minimal detectable effect size

Input

Effect information

Expected Cohen’a d (dz) Expected Cohe’s d_z, defined as the mean difference divided by the standard deviation of the difference score
N Expected sample size (number of paires)

Parameters

Minimal desired power Minimal desired power (1-Beta)
Alpha (Type I rate) Critical alpha (significance level to be used in the analysis)
Tails Wheather the test is two-tailed (two.sided) or one-tailed (greater).

Equivalence Testing

Perform equivalence testing Perform an equivalence test to show that a true effect is small enough to be considered practically negligible
Equivalence limit

Options

Explanatory text Produces some text to help interpreting the results

Sensitivity Analysis

Sensitivity analysis, exploring different plots of possible combinations of parameters, can be carried out like for any other PAMLj sub-modules. Please visit Sensitivity analysis page for more details.

Effect Size Conversion

With this panel is possible to estimate the expected effect size based on observed t-test, taken from empirical studies or other empirical sources.

Convert from t-test

Observed t-test The observed t-test
N The observed t-test sample size (number of pairs)
Cohen’s dz The resulting d_z

Convert d to dz

Cohen’s d (between) Cohen’s d uncorrected for pairs correlation
Correlation Correlation between repeated measures
Cohen’s dz Resulting Cohen’s d_z

It should be noted that PAMLj uses Cohen’s \(d_z\) as the effect size for power analysis. This is defined as the mean of the difference scores divided by their standard deviation (Cohen 1988). Some authors instead suggest starting from Cohen’s \(d\) for independent samples, defined as the mean difference divided by the square root of the sum of the variances of the two measures. If you want to convert this independent-samples \(d\) into a paired-samples \(d_z\), accounting for the expected correlation \(r\) between measures, you can use the Convert to dz , which applies the following formula:

\[ d_z = \frac{d}{\sqrt{2(1 - r)}} \]

0.7.0

Comments?

Got comments, issues or spotted a bug? Please open an issue on PAMLj at github or send me an email

Return to main help pages

Main page T-test
Cohen, J. 1988. Statistical Power Analysis for the Behavioral Sciences. Lawrence Erlbaum Associates.