I am conducting a qualitative-driven approach to mixed-method research. The role of my quantitative data is to corroborate the findings of the qualitative data. Qualitative data has been collected through interviews at the end of the project, and quantitative with a pre-survey before the project and a post-survey at the end of the project. One of the questions that arises when analyzing the quantitative data from my first project is whether I should use a paired t-test or an unpaired t-test analysis. In the first project, I had a small group of participants: the same 18 participants responded to both the pre- and post-questionnaires, but in total, I have 28 participants; the remaining 10 only responded to either the pre- or the post-questionnaire. I am unsure if 28 participants are enough for an unpaired t-test; however, 18 participants might be enough for a paired t-test. I would appreciate your help regarding this question and any resources or videos I can use to analyse pre- and post-questionnaires with SPSS, as I am unsure if I need to upload all the data from the pre-and post-questionnaire in the same file in SPSS or separately. Thanks very much.
Given your data and the small sample size, a paired t-test is more appropriate for the 18 participants who completed both surveys. This approach will better control for the individual differences and allow for a more accurate comparison of the pre- and post-survey data.
To supplement Michael Hindelang response, I have included two YouTube videos and two books that may prove beneficial.
https://www.youtube.com/watch?v=MJGk2sg4EZU
https://www.youtube.com/watch?v=8alv3kZt8Ug
"SPSS for Dummies" by Arthur Griffith.
"Discovering Statistics Using IBM SPSS Statistics" by Andy Field.
I align with Michael Hindelang perspective and would like to reiterate the following for enhanced comprehension:
Sample Size Considerations
Paired T-Test: With 18 participants, the sample size is generally sufficient for conducting a paired t-test. However, larger sample sizes are always preferred to ensure greater statistical power.
Unpaired T-Test: While 28 total participants can be used for an unpaired t-test, each group would be smaller (pre- and post-test groups with fewer than 28 participants each), which might limit the test's power.
You clearly have a pre post design, so that you should incorporate this information into your analyses. A repeated measures t-test might be suitable, if all assumptions are reasonably met.
But since you have additional information, i.e. the other participants who delivered data either for the pre or the post measure, you might think about a multilevel model. This would enable you to use all participants and so the 10 participants could contribute to have a better estimate for the pre and post mean value. And with the random effect for participant, you also account for the repeated measure.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA