Marcus Andrey Study
Ler em português Library

Framework · Research · 3 min read

Research method chooser

Three plain questions (say or do, why or how many, real use or a set-up) that narrow dozens of research methods down to two or three, plus a second method from another corner to cover the first one's blind spot.

WhyQualitative or quantitativeHow many
What people say, and why

Interviews, diary studies, participatory design

"How do…?" "Why do…?"
What people do, and why

Usability tests, field studies, contextual inquiry

"Can people…?"
What many people say

Surveys, card sorting, satisfaction scores

"How many say…?" "How do they group…?"
What many people do

Analytics, A/B tests, usability benchmarks

"How often…?" "Which is better?"
SayAttitudinal or behaviouralDo
Two questions place your research question in a quadrant; the third, the context of use, picks the method inside it.Simplified from Christian Rohrer, When to Use Which User-Experience Research Methods (NN/g, 2022).

Use it when

  • Someone opens with a method ("let's run a survey") before anyone has said what they need to learn.
  • You have a research question and want the first method that can answer it.
  • Your team's last studies all used the same method, and you suspect a blind spot.

How to use it

  1. Start from the research question. Write it down first. If you don't have one, go back and climb from the business question to the decision. Output: one research question.
  2. Say or do? Decide whether you need what people say (attitudes, beliefs, words) or what they do (behaviour). For usability questions, lean towards behaviour. Output: a column on the map.
  3. Why or how many? Decide whether you need reasons and fixes (qualitative) or how many and how much (quantitative). Output: a row on the map, so one quadrant.
  4. Pick the context. Choose how participants will use the product in the study: natural use, scripted tasks, a limited version such as a card sort or a sketch, or no product at all. Natural use gives realism, scripted use gives control. Output: a method from the quadrant.
  5. Add a second method from another corner. Pair it with data or a method that covers its blind spot, often data you already have. Output: two methods that triangulate.
  6. Check the real limits. Check time, access to the right people, budget and risk to participants. Then decide moderated or unmoderated, remote or in person. Output: the method line of your research plan.

Worked example

A team on a field-service scheduling app has four research questions. The first words of each point to a method:

Research question Say/do · why/how many First method Second method
How do dispatchers rearrange a day when a technician calls in sick? Do · why Field visit during a real morning Analytics on how often schedules change
Can new dispatchers assign an urgent job without help? Do · why Moderated usability test with tasks Support tickets about assigning
How many technicians check the app before their first job? Do · how many Analytics Short interviews to learn why the others don't
How do dispatchers group job types? Say · how many Card sort Tree test of the new menu

No question ended up with a focus group: none of them was about what a group thinks of a brand or concept.

Watch out for

  • The method first. Choosing the method you know and bending the question to fit it. Start with the question every time.
  • A survey for "why". A survey counts answers, and it can't follow up. Use interviews or observation to find reasons.
  • Interviews for "how many". Five people can't tell you how common something is. Measure it with analytics or a survey.
  • A focus group for usability. People in a group influence each other, and nobody does a real task. Run individual sessions.

Template

Answer the three questions for each research question, then fill in the last two columns.

Research question:

Say or do?            [ ] what people say      [ ] what people do
Why or how many?      [ ] why / how to fix     [ ] how many / how much
Context?              [ ] natural  [ ] scripted  [ ] limited  [ ] no product

First method:
Second method (another corner, or data we already have):

Limits:  time ........  access ........  budget ........  risk ........
Moderated or unmoderated:          Remote or in person:
Source
Christian Rohrer, "When to Use Which User-Experience Research Methods" (NN/g)
Where to find it
UX Research: Choosing a research method

"See on Amazon" links are affiliate links. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close