I am performing a meta-analysis comparing two methods for performing a test; the traditional (TRAD) method and a new experimental (EXP) method. As a secondary analysis I want to look at the maximum measure obtained from a 0–10 likert scale, let's call it "Rating A". In TRAD, rating A is obtained from the participant by having them provide the rating at the end of the test. In EXP, the rating scale is used to guide the test which terminates when the scale reaches 10, hence the maximum value for EXP will ALWAYS be 10 with no variance.
Ordinarily, within a primary study, this would be assessed via a one-tail, one-sample t-test for whether the sample mean from TRAD is less than 10.
Is there an equivalent method for a meta-analysis where I can take the mean and variance in Rating A for TRAD and compare it to a defined value (i.e., 10)? Essentially, I want to test if the mean value across studies for Rating A from TRAD is less than 10, the maximum of the scale.
That said, I'm not convinced this is an appropriate analysis given the scale has a ceiling. Are there any reasons this would not be appropriate, or is there an alternative way to approach this analysis?
The primary analyses were performed using R and metafor.
