Skip to content
Tutorials Center
My learning

Topics

Lesson 43 of 44 ยท Projects and Practice

Research Methods Cheat Sheet

A one-page recap of research methods in psychology: designs, measurement, statistics, ethics, replication and reporting, with a poster, formulas and a test chooser.

Here is the whole course on one page. Print it, pin it above your desk, and check each new study (yours or someone else's) against it.

Six cards titled Design, Measure, Analyze, Ethics, Replicate and Report, each with three short rules.
Research methods on one page: six habits that cover most of what can go wrong.

What you will learn

  • A compact checklist from design to report.
  • Which common statistic fits which design.
  • The formulas and numbers worth memorizing.

Key terms

Operationalize: turn an idea into a measurable procedure.

Randomize: let chance assign participants or order.

Preregister: record your plan before collecting data.

Report fully: include all measures, exclusions and tests you planned.

From question to report in eight steps

  1. Ask a testable question and turn it into a hypothesis.
  2. Operationalize variables: define each IV and DV concretely and pick reliable, valid measures.
  3. Choose a design: experiment (manipulate and randomize) for causes; correlational or observational for relationships; between, within or factorial as needed.
  4. Plan power: decide the smallest effect that matters and compute the sample size.
  5. Get ethics approval, prepare the consent form and debrief.
  6. Preregister the hypotheses, exclusions and main analysis.
  7. Collect and analyze as planned, reporting effect sizes and confidence intervals; label exploratory work.
  8. Write up with the hourglass structure; state limits honestly; share data and materials where possible.

Which test for which design?

Situation Typical analysis
Two independent groups, numeric outcome Independent (Welch) t-test; Cohen's d
Same people, two conditions Paired t-test; d_z
Three or more groups or conditions ANOVA; eta squared
Two factors Factorial ANOVA; check the interaction first
Two numeric variables, relationship Correlation r (association, not cause)
Two categorical variables Chi-square test; odds ratio or Cramer's V
Predict a numeric outcome from others Regression

(This course explained the t-tests and the logic of effect sizes; the tracks listed in the next lesson cover ANOVA, regression and more.)

Numbers worth memorizing

Item Value
Conventional alpha .05
Typical power target 80% to 90%
Cohen's d anchors 0.2 small, 0.5 medium, 0.8 large
n per group for d = 0.5, 80% power about 64
n per group for d = 0.2, 80% power about 394
False-positive chance across 5 independent null tests about 23%
95% CI of a mean (large n) mean ยฑ about 1.96 standard errors
Complete counterbalancing orders for k conditions k!
Cells in an a x b x c design a * b * c

Try it

A test chooser: describe your design in four answers and get a suggested analysis. It is a study aid built from the table above, not a replacement for judgment.

Python Which analysis? (a study aid)

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. You compare memory scores in the same 30 people before and after sleep. Which analysis fits?

    Two measurements of the same people call for a paired test.
  2. 2. Which number is closest to the n per group needed for 80% power at d = 0.5?

    About 64 per group, from the standard two-sample t-test power calculation.
  3. 3. Which step comes BEFORE data collection?

    Planning, ethics approval and preregistration all precede data collection.
  4. 4. In a factorial design, what should you check first?

    A significant interaction can make main effects misleading.

Summary

  • Eight steps carry a study from question to report; most protections happen before data collection.
  • Match the analysis to the design: independent versus paired, number of groups, number of factors.
  • Remember the key numbers: alpha .05, power 80% to 90%, d anchors, sample sizes, 1 - 0.95^k.
  • Randomize, preregister, report fully and be honest about limits.

Next: Further study: where to go next