Question: I'm summarizing survey results and noticed a difference in the percentage change between the overall scores and the delta values. Can you explain how these percentages are calculated?

It's not uncommon to encounter variations between overall scores and delta percentages when comparing different survey datasets. The key to understanding these differences lies in the way the percentages are calculated.

The percentage change is calculated at the question level, not the overall scores level. This means that it reflects the changes in responses for specific questions that are common between the two surveys. Let's illustrate this with an example:

Suppose we have two surveys, Survey A and Survey B. Survey A contains questions Q1, Q2, Q3, Q4, and Q5, while Survey B includes only Q1 and Q4. When we calculate the delta percentages, we focus on the changes between the responses to Q1 and Q4 in both surveys, as these are the common questions.

For instance, if for Q1 Survey A scores 80 and Survey B scores 85, resulting in a delta of +5, and for Q4, Survey A scores 60 and Survey B scores 63, resulting in a delta of +3, the average delta between these two surveys would be +4. However, the gap in the overall scores may differ because they take into account all the questions in each survey. In Survey A, Q2, Q3, and Q5 may have different responses that impact the overall score, while Survey B doesn't include those questions. This is why the overall score in Survey A might be 75 and 74 in Survey B, even when the delta is +4.

So, to understand percentage changes accurately, focus on the delta percentages for the specific questions that are common between the surveys. The overall scores may not provide an accurate representation of changes when different sets of questions are used.