Research method selector
Which research method fits your decision?
Describe what people need to compare and how you will use the answer. Get an approach to investigate, an alternative, and the questions to resolve before designing the study.
Start with the question you need to answer
Choose your own answers or load an illustrative example. The recommendation updates as you go. Unknown or conflicting answers remain visible in the result.
Your method recommendation
Start with your research decision
Fit describes the match to your answers. It is not a confidence percentage or evidence that a study will succeed.
What to clarify
- Choose the business decision.
- Clarify what respondents will evaluate.
- Clarify the role of price.
- Choose whose preferences you need.
Answer the decision, comparison, price, and output questions to see a recommendation.
From a method to a study
A method match is the first decision.
The exercise, sample, and validation plan follow from what you need to learn. This selector records the assumptions behind a recommendation; it does not evaluate an experimental design or certify a study.
Discuss your research questionWhen should we do discovery first?
When needs, concepts, or attributes are unclear, use qualitative work and comprehension testing to define them. Asking people to compare poorly understood options cannot resolve that uncertainty. Explore qualitative research; research-design guidance.
Can MaxDiff tell us what people will pay?
Standard MaxDiff measures relative item preferences. An absolute anchor adds a threshold such as acceptance; it does not identify feature-price trade-offs. Monetary estimates need an appropriate price-sensitive model and checks on its stability. Anchoring; willingness-to-pay methods.
Does a larger sample solve a short questionnaire?
More people can improve estimates for a population without adding observations for any one person. If you need individual predictions or segment assignments, check the evidence collected within each respondent and the uncertainty of those outputs. Individual-level parameters.
What does “conditional” method fit mean?
The approach is a candidate with an unresolved issue, such as conflicting objectives, uncertain comprehension, or specialist allocation. “High” means the stated question fits the method family. Neither label is a statistical confidence level, a sample-size approval, or evidence that predictions will be accurate.
Do all recommended methods have a planner?
The working conjoint planners cover conventional choice-based conjoint, and the MaxDiff planner covers stated exposure and assignment assumptions. Adaptive conjoint, menu-based choice, Bandit, and relevant-items designs need separate review. The selector links to a relevant solution when the available calculators do not support the recommended method.
Can a generic power calculator assess these scores?
Conjoint utilities, MaxDiff scores, modeled choice shares, and willingness-to-pay estimates need uncertainty methods appropriate to their design and model. A generic percentage or mean comparison does not automatically apply. For an ordinary independent-group survey outcome, first confirm the assumptions in the detectable-difference planner.
Research behind the selector
The reasoning is open to review.
These sources inform the method distinctions. The routing rules are Russell's planning policy, not a published or empirically calibrated prediction model. Study evidence remains unassessed. Sources reviewed September 29, 2026.
- Bridges et al. (2011). Conjoint Analysis Applications in Health: A Checklist.
Research question, attributes, tasks, design, and analysis are separate stages. Applying that sequence to this selector is our routing policy.
- Sawtooth Software. Designing the Study: MaxDiff.
Item exposure, balance, and connected comparisons. Practitioner guidance does not establish an optimal study for every decision.
- Orme (2019). Making Sense of All Those MaxDiffs!
Applied descriptions of sparse, Express, Bandit, and relevant-items designs; allocation changes the evidence collected.
- Sawtooth Software. Anchored MaxDiff.
Adding an absolute threshold changes the response task. It does not add a monetary trade-off.
- Scarpa, Thiene & Train (2008). Utility in Willingness to Pay Space.
Monetary trade-offs require an appropriate price model; coefficient ratios can produce unstable willingness-to-pay estimates.
- Train (2009). Discrete Choice Methods with Simulation, Chapter 11.
More respondents and more observations within each respondent address different sources of uncertainty.
- Sawtooth Software. Adaptive Choice-Based Conjoint.
Adaptive interviewing is a specialist option with additional respondent work; it needs its own design and timing review.
- Sawtooth Software. Menu-Based Choice.
Multiple selections from a menu differ from choosing one preconfigured alternative. Design and analysis are more specialized.
- Sawtooth Software. Adaptive Conjoint Analysis.
A legacy adaptive method, with limitations for pricing. It is not the selector's default for a new study.
- Sawtooth Software. Pricing Research Techniques.
Distinguishes fixed-offer price research from conjoint. The selector routes to a pricing discussion; it does not calculate price curves.