The same students provide two related scores, such as reading performance before and after tutoring.
Question: Did this group change?
Lesson 07.01 | Group Comparison Toolkit
Choose the Group Comparison Test
The test name comes last. First identify who is measured, how many grouping variables you have, how many groups or conditions are being compared, and whether a meaningful continuous covariate must be controlled.
The course-wide habit
Read the design before you read the data. A correct statistical test is a match between the research question, participant structure, and variable roles.
The five destinations
One continuous outcome, five different designs
Use the same academic-support setting to see how one change in the design changes the correct analysis.
Two different groups provide one score each, such as a tutoring class and a comparison class.
Question: Do these two groups differ?
One categorical factor creates three or more independent groups, such as three enrichment programs.
Question: Does at least one group differ?
Two categorical factors are studied together, such as program type and grade level.
Question: Do the factors matter, and do they interact?
Groups are compared after controlling for a meaningful continuous variable that existed beforehand, such as pretest score.
Question: Do groups differ after adjustment?
First safety check: these tests compare a continuous dependent variable. When the outcome is categorical, or the goal is prediction rather than group comparison, this decision map is not the correct endpoint.
Interactive decision map
Follow the design, not your favorite test
Answer one question at a time. The route will identify the best match among the five tests covered in Chapters 1–6.
Build the route
Recommended test
Transfer practice
Match each study to its design
Every scenario uses a continuous outcome. Focus on participant structure and variable roles.
| Study structure | Key clue | Best match |
|---|---|---|
| Same nurses report burnout before and after a scheduling change. | Same people, two related measurements | Paired Samples t Test |
| Day-shift and night-shift nurses are compared on burnout. | Two separate groups | Independent Samples t Test |
| Burnout is compared across day, evening, and night shifts. | One factor with three groups | One-Way ANOVA |
| Burnout is examined by shift and hospital unit. | Two categorical factors and an interaction question | Two-Way ANOVA |
| Shift groups are compared while controlling for years of experience. | One grouping factor plus a continuous pre-existing covariate | ANCOVA |
Final synthesis
Explain the test before you name it
A strong choice can be defended in plain language without referring to SPSS menus or output tables.
Course checkpoint: choosing the test is not a memory contest. It is a structured reading task: identify the design, follow the branches, and stop when the research question falls outside group comparison.