I would like to measure how similar the responses to the same survey are from two separate institutions. Would a correlation coefficient between two columns of mean scores for each question provide this information? Most of the questions are 5-point Likert scaled; the rest are rankings out of 4.
If a correlation coefficient could be meaningful, I infer from online reading that a Spearman's Rho might be appropriate due to the data being ordinal and discrete. Which t-Test would suit this - if any - for a measure of statistical significance?