Hello,
I'm looking at calculating the TAU-U effect size for my single case research data set. I'm struggling to find any literature on the minimum number of baseline data points to compute TAU-U, does anybody know the answer to this? I'm aware of the recommended baseline data set for general single case research, but struggling to find anything specifically for TAU-U.
Thanks in advance.
Hello Rebecca,
The question actually has two parts: (a) how many cases are needed in the baseline (A) phase in order to detect a trend, if one exists; and (b) for the adjusted scores, how many cases (in each phase) are needed to have a power of (whatever target value) to detect a difference (of whatever magnitude), if it exists?
Tarlow (see below) suggested at least 10 cases are needed in baseline to have decent power to detect a trend (of r >= .40) for his proposed method that is claimed superior to tau-u. Unfortunately, there's less guidance on the second part of the question. Running simulations, based on your target ES would be the best approach. But, that should give you a lower bound from which to work.
Good luck with your work.
Tarlow, K. R. (2017). An improved rank correlation effect size statistic for single case designs: Baseline corrected tau. Behavior Modification, 41, 427-467. doi: 10.1177/0145445516676750
- Badges
- Science topic