Blog

Worked analyses with the R script and the output shown.

• ANOVA post hoc tests: Tukey, Games-Howell, Dunn

  Which post hoc test follows which ANOVA: Tukey after the classic F-test, Games-Howell after Welch, Dunn after Kruskal-Wallis, all three run on one dataset.

  October 22, 2026, Comparing groups, 8 min read

• Pretest posttest design: change scores or ANCOVA?

  How to analyse a two-group pretest posttest design: change scores, ANCOVA or a mixed model, when each one fits, and a worked ANCOVA example in R.

  October 22, 2026, Statistics for a thesis or dissertation, 10 min read

• Multiple linear regression, and when to switch

  Choosing a regression model from your outcome type (linear, logistic, Poisson, negative binomial or ordinal), with a multiple linear regression worked in R.

  October 22, 2026, Regression models, 14 min read

• APA style statistics reporting: every common test

  The APA 7 rules for reporting statistics (italics, leading zeros, p values, degrees of freedom, intervals) and a fill-in template for each common test.

  October 22, 2026, Reporting results in APA style, 19 min read

• Survival analysis: Kaplan-Meier, Cox and AFT in R

  Survival analysis in R on public lung cancer data: censoring, Kaplan-Meier, log-rank, Cox regression with its proportional hazards check, and AFT.

  October 22, 2026, Survival analysis, 15 min read

• T-tests explained: one-sample, paired and Welch

  What a t-test does, how to choose the one-sample, paired, Student or Welch version, and a worked Welch example in R with Cohen's d and its interval.

  October 22, 2026, Comparing groups, 25 min read

• Do you need to test for normality before a t-test?

  Should you test normality before a t-test? Usually not. Why a Shapiro-Wilk pre-test misleads at small and large n, what to check instead, and an R example.

  October 22, 2026, Assumptions and diagnostics, 7 min read

• Two-sample t-test: formula, assumptions, example

  The two-sample t-test formula worked by hand and confirmed in R, its assumptions, and Student's versus Welch's version on the same unequal-sized groups.

  October 22, 2026, Comparing groups, 9 min read

• Welch's t-test: when to use it instead of Student's

  Why Welch's t-test is the safer default for two groups, how its degrees of freedom work, and a worked R example where group sizes and spreads both differ.

  October 22, 2026, Comparing groups, 8 min read