Aglet

Investigate Ambiguous Product Survey Responses

A response distribution cannot show whether people understood the item the same way. Investigate ambiguity by asking respondents to explain their interpretation, comparing groups, and testing one wording change at a time. Treat clean data as unhelpful when its construct is unstable.

Build a useful investigation brief

  1. Probe how people read it

    Use a small cognitive interview or open follow-up to ask what the question means, what they considered, and how they chose an answer. Preserve the respondent's words and context. Do not correct their interpretation before recording it.

  2. Compare response pathways

    Review answers by language, familiarity, task experience, and missing or not-applicable choices. Look for respondents answering different halves of a double-barreled item or using a scale endpoint differently. Keep subgroup differences separate from random noise.

  3. Test one revision

    Change the ambiguous term, timeframe, scale, or response option and compare the interpretation with the original. Record what improved and what new limitation appeared. State whether the item measures the intended construct or needs another method.

What to carry forward

The investigation brief should link original wording, respondent interpretations, response patterns, revision, and remaining uncertainty. End with a clearer item or an evidence request. Do not infer product preference from answers that represent different questions.

Technical background: GOV.UK user research guidance.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow