Build the comparison record
Put the two periods side by side with exact wording, response choices, eligible group, invitation method, timing and valid-answer count. If any of these changed, label the comparison accordingly. A tidy chart cannot repair incompatible inputs. Preserve an explicit break rather than presenting a continuous trend that was never measured consistently.
A fictional denominator trap
Period A has eight positive ratings among ten answers. Period B has eighteen among thirty. The positive count rose, but the observed share fell from 80% to 60%. Neither fact alone explains service quality. Perhaps the invitation group expanded, or the experience changed, or different people chose to respond. Those remain questions to investigate.
Keep the action separate
If you changed an instruction between periods, record when and where it was used. A before-and-after difference does not isolate that change from every other influence. Do not claim causation or a statistically meaningful improvement from this basic comparison. Qualified research design may be needed for higher-stakes conclusions.
A useful review outcome
The review can conclude that the periods are not comparable and still be valuable. It may identify the wording and collection conditions to keep stable next time. Record that decision and the remaining uncertainty instead of searching for a graph that makes the intervention look successful.
A visible comparability decision
Use three possible labels: reasonably comparable on the documented conditions, partially comparable with a stated change, or not comparable for the proposed conclusion. These are editorial descriptions, not a statistical test. A fictional team might compare the wording and period but acknowledge that invitations expanded from collection customers to all service customers. It can still read the comments, while refusing to interpret the percentage difference as a like-for-like movement in satisfaction.
Sources used for this page
These records support the facts and comparisons above. Merchant-controlled records are labelled so you can separate product claims from independent evidence.
- AAPOR survey-research best practices — Standards and certification reference · aapor.org · Publisher independence not verified · checked 2026-10-01