Hi everyone! So I need some help with this problem. Anything would help at this point, thanks
A researcher is interested in examining effect of familiarity on favorability of ratings. Specifically, she hypothesizes that raters who are more familiar with a subject tend to give more favorable ratings. To test this hypothesis, she had 100 subjects rated by 200 raters. For each subject, there are two raters, one is familiar with the subject and one is not. The researcher then compared the ratings given by the “familiar” raters to those from the “not-familiar” raters using independent sample t-test (n=100 for each group). Result was not significant: t = 1.50 (p=.068, one-tailed test).
Later, the researcher realized that she could match the raters based on the subjects they rated in order to conduct repeated measure t-test, which has more power. Unfortunately, she has lost all the data due to a computer problem. The only record she still has is the correlation between the pairs of raters (familiar vs. not-familiar), which is .70.
Can you help the researcher conduct analysis using repeated-measure t-test based on the information here?
a) Present your calculation and result
b) Apparently, using repeated-measure t-test results in increased power. Calculate the sample size you would need for independent sample t-test to have the same result obtained by using repeated-measure t-test as in part a above.