I would be very grateful for any advice on how to analyze the data of a small study I have conducted. The study has only one group of participants and their performance in a test was measured for 5 consecutive days. My aims would be:
- Assess that practice improved performance (number of correct responses)
- Obtain an individual score of "improvement" to see whether or not this "improvement" is related to baseline levels and/or to other collected variables.
From what I have learnt, I would approach aim #1 by using a suitable test for paired data (e.g. comparing average performance with a t-test for paired samples) and aim #2 by calculating individual difference scores (score_Day5 - score_Day1) and using these scores in posterior analyses.
However, I have read that this approach seems to be essentially flawed (e.g. https://pubmed.ncbi.nlm.nih.gov/22192231/ https://ijbnpa.biomedcentral.com/articles/10.1186/s12966-015-0162-z . I have seen some proposals of using ANCOVA/ ordinal regression (https://www.fharrell.com/post/errmed/#change). However, I did not really understood how I should implement them and whether or not these analyses would also provide individual "improvement" scores (or if they are also inappropriate).
Would appreciate any suggestion and guidance, specially if they do not assume deep stats knowledge.
