Test whether a satisfaction trend survives comparable segment mixes and survey response patterns. Use before calling an overall survey increase a like-for-like improvement.
Installs into .claude/skills of the current project.
Are you the author of Compare Satisfaction Waves By Respondent Mix?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/alisonjieli-png-compare-satisfaction-waves-by-respondent-mix)
---
name: compare-satisfaction-waves-by-respondent-mix
description: Test whether a satisfaction trend survives comparable segment mixes and survey response patterns. Use before calling an overall survey increase a like-for-like improvement.
---
# Compare satisfaction waves by respondent mix
## When to use
Use for two or more waves of a defined satisfaction measure. This method tests comparability and composition; it cannot recover the unknown views of people who did not respond.
## Inputs
- Wave dates, question and response-scale versions, target population, invitation frames, segment definitions, and response dispositions.
- Response records or segment-level outcome counts with eligible and responding denominators.
- A supplied common weighting rule or reference mix if a standardized comparison is requested.
## Procedure
1. Check that the question, scale, eligible population, collection mode, and observation periods permit comparison. Name each change that could alter answers independently of satisfaction.
2. For each wave and segment, report invited, responded, missing or unknown disposition, and the satisfaction numerator and denominator. Keep one person's repeated responses under the supplied identity rule.
3. Compare the overall result with results inside comparable segments. Show how respondent shares changed; do not infer that nonrespondents had the same outcome as respondents.
4. If a valid common weighting rule is supplied, calculate a fixed-mix comparison and show its weights. Keep unweighted and weighted results separate and flag cells too sparse or undefined to support a stable comparison.
5. State whether the reported overall change survives the comparison, could be explained by mix, or remains undecidable because frame or nonresponse coverage is insufficient.
## Completion check
Return the comparable question and frame conditions, wave-by-segment counts and rates, respondent mix, overall trend, optional fixed-mix trend, and caveats. Every trend number must show its denominator and weighting basis.
## Stop conditions
Hold a like-for-like improvement claim if question meaning, frame coverage, segment mapping, or response denominator is unknown. Do not contact respondents, infer individual attitudes, or invent survey weights.
