The Article by Raj Makkar and colleagues has important potential. However, the statistical analyses would benefit from a better description, since there is no mention of any of the specific analyses used to obtain the p values. The sample sizes are small for each of the groups, and many of the continuous variables have skewed distributions, as suggested by the SDs being greater than the means. There are special statistical concerns when analysing data with a small sample size and a skewed distribution. For example, the analysis shown in figure 4A has a p value of 0.02 for comparing the cardiosphere-derived cell group versus control group. If the data had a normal distribution (ie, not skewed), a t test or ANOVA analysis could be used. However, there are problems with the use of these tests without the presence of skewness being addressed. Since Makkar and colleagues did not mention the issue of skewness, one could be concerned that this issue was not addressed. There are many other reported p values that are also subject to this potential concern.
展开▼