Tools

Tools for clearer research thinking.

Interactive tools I designed to turn research judgment into practical workflows, from defining testable hypotheses to selecting methods and planning samples. They help make research decisions more structured and transparent without replacing the researcher’s expertise.

Generate hypothesis

Write a clear, provisional hypothesis.

Turn an initial assumption into a clear, testable hypothesis by defining the audience, context, expected behavior, comparison and measure.

PeopleWho is this about?

SituationWhat are they doing at that moment?

Change or conditionWhat will be different in the experience?

Expected behaviorWhat observable action do you expect from them?

ComparisonWhat will the new experience be compared with?

MeasureWhat observable outcome will you count or calculate?A measure is what you count. A comparison is what you compare it with. For example, purchases is a measure; purchases in the previous month is a possible comparison period.

Success criterionWhat result would support the hypothesis?

Not set yet

The target can remain undecided while you plan the research. Selecting this will keep any draft text in place.

Time periodWhen will the result be evaluated?

Found something that looks wrong? Let me know.

Which research method should I use?

Start with the question, access, and limits.

Explore suitable research approaches based on the decision you need to support, the evidence required and the practical constraints of your study.

Sample-size and invitation planner

Estimate a defensible starting point.

Estimate sample sizes and recruitment needs while making assumptions about precision, response rates and representativeness explicit.

Confidence levelHow confident the estimate should be under the sampling assumptions.

Margin of errorThe expected sampling error around a proportion estimate.

Target population size (optional)Finite-population correction matters most when the eligible population is small.

Illustrative response-rate scenarioThis scenario changes the invitation estimate only, not the number of completed responses.

385complete responses

Estimated invitations: ~1,284
Formula: n = Z² × p(1−p) / e²

About this estimate
This estimate uses the simple random sample formula for a population proportion. When the expected proportion is unknown, it assumes p = 0.5, which gives a conservative sample size. At 95% confidence, the stated margin of error describes sampling uncertainty under this design: across repeated probability samples, about 95% of intervals calculated this way would contain the true population proportion. It does not account for coverage, nonresponse, or measurement error. Weighted, clustered, or other complex designs may require adjustments. Response rates are illustrative planning assumptions, not benchmarks. Actual rates vary by audience, sampling frame, contact channel, survey length, incentives, and follow-up. Use results from a comparable survey when available, and update the invitation estimate as responses come in. Sample validity depends on who responds, not only on how many respond. A large sample from the wrong population is still the wrong sample. For reliable comparisons between segments, each segment needs enough responses. Dividing the total sample into smaller groups does not preserve the overall estimate’s precision.