Academic Statistics · Module 03 of 11

t-tests: independent, paired and one-sample

This module covers the one-sample, independent-samples and paired-samples t-tests, their assumptions, the Welch correction for unequal variances and the appropriate effect sizes. The t-test is the most common comparison in doctoral work, and correct choice and reporting of it are basic expectations of examiners.

Watch the first lesson of every module free. The remaining lessons, exercises and live sessions open when you enrol.

What you will be able to do

  • Choose between one-sample, independent and paired t-tests from the design
  • Check independence, approximate normality and homogeneity of variance, and apply Welch's t-test where needed
  • Calculate and interpret Cohen's d or Hedges' g with a confidence interval
  • Report a t-test in APA format, for example t(df) = value, p = value, d = value, 95% CI

Key terms

  • One-sample t-test
  • Independent-samples t-test
  • Paired-samples t-test
  • Welch's t-test
  • Levene's test
  • Cohen's d
  • Hedges' g
  • Degrees of freedom

Hands-on practice

Run and report an independent-samples t-test with Welch's correction and Cohen's d on a provided dataset in SPSS or R.

Common pitfalls this module fixes

  • Using an independent t-test on paired data such as pre-test and post-test scores; use the paired t-test, which accounts for the correlation between measures.
  • Running many t-tests to compare three or more groups; this inflates the Type I error rate, so use ANOVA with planned contrasts or post-hoc tests.
  • Reporting only the p-value; always add the mean difference with its confidence interval and an effect size.