Simplified Edition · Module 02

Research Methods

Sampling, estimation, error and power: the reasoning behind a result rather than the arithmetic of one.

  • 21 activities
  • 5 to 7 minutes each
  • 128 minutes end to end
Sample means vary far less than single observations A wide, even scatter of individual observations across the full width. Below, the means of samples drawn from the same population form a narrow hump near the centre. Individual observations Sample means
Single observations scatter widely. The means of samples drawn from them do not.

The activities

In teaching order

  1. Research Question to Method Mapper

    Reading a research question for what would count as an answer

    6 min

  2. Operationalisation Laboratory

    Operationalisation, construct coverage and the trade-off between under-representation and irrelevant variance

    6 min

  3. Confound Detective

    Confounding, and why bias and noise are different problems with different fixes

    7 min

  4. Sampling Bias Simulator

    Sampling bias, sampling variability, and which one a larger sample fixes

    6 min

  5. Thematic Analysis Coding Laboratory

    Coding as an analytic act, and why competent analysts differ

    5 min

  6. Theme or Topic?

    Topics, staging posts, themes and overreach in qualitative analysis

    5 min

  7. Reflexivity and Alternative Themes

    Reflexivity, and why the account follows from the question and the lens

    6 min

  8. Sampling Distribution and p-Value Simulator

    P-values as tail areas, and what they are not

    7 min

  9. Confidence Interval Laboratory

    What the ninety-five per cent in a confidence interval refers to

    7 min

  10. ANOVA F-Ratio Visualiser

    What the F ratio is actually a ratio of, and the three separate things that move it

    6 min

  11. Factorial Interaction Detective

    What an interaction is, and how much of the one you can see is the axis rather than the data

    6 min

  12. What a Covariate Can and Cannot Buy

    What adjusting for a covariate does to an estimate, and why only the design decides what the estimate means

    6 min

  13. The Same Mark, Two Different Distributions

    Why a raw score means nothing on its own, and what standardising actually does to the axis

    5 min

  14. Central Limit Theorem Simulator

    Sampling distributions, the central limit theorem and what it does not claim about data

    6 min

  15. How Much Do Two Groups Overlap?

    How much two distributions still overlap at a large effect, and why an effect size says nothing about whether an effect is real

    6 min

  16. The Same Finding, Twice

    What a t-test actually asks, and why the sample size changes the answer without changing the finding

    6 min

  17. What Gets Published From a Small Study

    Why the significant results from small studies overstate effects even when everyone is honest

    7 min

  18. What r Cannot See

    What a correlation coefficient measures, what it cannot see, and how much one observation can move it

    6 min

  19. Why It Is Called Least Squares

    What least squares actually minimises, drawn as the squares it is named after

    6 min

  20. Does the Spread Stay the Same?

    Whether the spread around a regression line stays the same across the range, and why the residual plot is the only way to see it

    7 min

  21. The Garden of Forking Paths

    Researcher degrees of freedom, and what choosing an analysis after seeing the data does to a p-value

    6 min

Elsewhere

Every activity here has a full-length twin in the main collection, linked from the foot of the activity itself.

All five modules Research Methods in the full collection