Aglet

Triage Bias in User Interview Evidence

Interview bias can enter through wording, order, facilitator reactions, recruiting, or the setting itself. Triage the guide and notes together. Identify where a participant may have been shown the expected answer, then separate spontaneous observations from answers that followed a cue.

Establish what is happening

  1. Mark loaded questions

    Read each prompt literally and underline words that name a problem, feature, emotion, or preferred outcome. Flag yes-or-no questions that imply approval and follow-ups that repeat the interviewer's interpretation instead of the participant's words.

  2. Review the interaction

    Compare the written guide with pauses, probes, facilitator explanations, and visible reactions in the session record. Note when the moderator introduced a label or rescued a participant. Preserve exact wording; paraphrased notes can hide the moment bias entered.

  3. Separate evidence types

    Classify each finding as observed behavior, recalled event, spontaneous opinion, prompted opinion, or facilitator inference. Record participant context and question order. Do not discard a biased session automatically; mark which parts need corroboration from another method.

What to carry forward

The triage output is a bias register linking question wording and facilitator behavior to affected findings. Stop when reviewers can distinguish direct observation from prompted opinion. Keep conclusions provisional when session recordings, question order, or participant context are missing.

Technical background: Nielsen Norman Group 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