# Comparing scores

I am trying to compare different survey questions against each other to show greatest improvement of scores. Here is a description of the case:

I have customer service questions that I am trying to compare month to month improvement to. My issue is that certain questions have scored relatively high and others very low. Theoretically, the low scoring questions have a greater opportunity for improvement than the higher scoring ones. (Example, one question has scored a 65% where the best score of another question is 85%. So it is easier for the 65% question to improve than the 85% question.) Is there a way to normalize/standardize the scores where I can show which improvement is the greater achievement?

If I understand you correctly, all you need to do is compute the relative improvement month-to-month of each of the questions. Say question $q_i$ has a score at month $j$ of $q_{i j}$, then the improvement from month $j$ to month $j+1$ is $\frac{q_{i j+1}}{q_{i j}}$. Multiply this quantity by 100 to have a percentage if you wish. Being relative to each other, these improvements are already normalized and can be compared across questions.