Resources / research-methods-selector

URI for your assistant: designers://reference/research-methods-selector Free

Choosing a research method

Pick the method from the question you must answer, not from habit. Two axes help: attitudinal (what people say) versus behavioural (what they do), and qualitative (why, how) versus quantitative (how many, how much).

By question

• "What problem do these people have, and how do they cope now?" Generative interviews or field visits. 5 to 8 participants per segment; you stop when answers repeat. Analytics cannot answer this.
• "Can people complete this task, and where do they get stuck?" Moderated usability tests on a prototype or build. 5 per round per distinct user group, several small rounds. Use lint_research_questions on the guide; use usability_stats (mode discovery) to plan.
• "How usable do people find this overall, and is it better than last time?" SUS after a task session, 12 or more participants for a stable number; usability_stats (mode sus) scores and gives the interval.
• "Which of two designs performs better?" A/B test with a pre-set sample size and run length; usability_stats (mode ab).
• "How many people want or do X?" A survey with a representative sample; usability_stats (mode survey) for sample size. Surveys measure prevalence, not reasons.
• "Where do people look for things in the structure?" Tree testing (findability) and card sorting (grouping). 15 to 30 participants for stable results.
• "What do people do in the product at scale?" Analytics plus session recordings for the why behind the numbers. Needs a clean tracking plan; lint_tracking_plan.
• "Do people understand this message or concept?" A five-second test, a comprehension test (ask them to say it in their own words), or a first-click test.
• "What happens over weeks?" A diary study or longitudinal interviews.

Recruiting

Recruit for behaviour, not demographics ("bought a flight in the last 3 months", not "age 25 to 34"). Include people who struggle: older users, assistive technology users, low bandwidth. Screen out people who work in design or research and your competitors. Pay fairly. Record with consent and say how data is stored.

Plan template

Objective (the decision this will inform), questions, method and why, participants (who, how many, how recruited), tasks or script, materials, timeline, how findings will be used and who decides, and what would change the plan.

Interview technique

Ask for stories of the last time, not opinions or predictions. Start broad, then narrow. Follow "what happened next". Stay silent after an answer. Do not explain the design. Do not ask "would you use this". Ask "tell me about the last time you did X" and "what did you do about it". Close by asking what you did not ask about.

Synthesis

Write observations on separate notes (one idea each, in participant words), cluster by theme bottom-up, name each theme as a finding with evidence count, then turn findings into insights ("people do X because Y, which means Z") and opportunities. Keep quotes and clips with findings. Separate what you observed from what you infer.

Common failures

Leading questions, hypothetical questions, testing with colleagues, treating five interviews as statistics, treating a survey as a source of reasons, reporting findings no decision depends on.