Skip to main content
. 2023 Nov 15;21:440. doi: 10.1186/s12916-023-03146-5

Table 5.

Results of robustness checks

Task Better score Max. skew p (Supplement)
Normal 20% trim 5% winsorisation 20% winsorisation Bootstrap and 20% trim
RAPM creatine  − 059 0.327 0.412 0.361 0.672 0.354
BDS creatine 1.14 0.064 0.17 0.009 0.05 0.37

Creatine effect p-values (two-tailed) for different ANOVAs. The given trim and winsorisation percentages are applied to each side. Better score based on estimated marginal means. “Max. skew” gives the highest skewness statistic in any combination of conditions (supplement and supplement order)