The Table 1 Fail

 

Inferential statistics, particularly null hypothesis significance testing, are used to make population inferences. It is inappropriate to test for sex ratio differences between groups unless you're asking if there tend to be more men than women in the patient population than the control one.

 

Consider H0 for this test. H0, in words, H0 is that the patient and control groups are drawn from the same population.

 

The experimenter explicitly draws samples from two populations.

 

To assess the samples of a control and patient population, it is sufficient to provide graphical representations and summary statistics.

 

Here are a couple of references about the problem of balance tests, and a reference that provides a graphic solution that helps correct the Table 1 failure.

 

  1. Sassenhagen, J., & Alday, P. M. (2016). A common misapplication of statistical inference: Nuisance control with null-hypothesis significance tests. Brain and language, 162, 42-45. [link]
  2. Mutz, D. C., Pemantle, R., & Pham, P. (2019). The perils of balance testing in experimental design: Messy analyses of clean data. The American Statistician, 73(1), 32-42. [link]

 

Solution:

 

A 1D scatterplot representing the data of “Table 1” gives readers sufficient information to assess the group differences and individual differences but does not misapply NHST and inferential statistics.

 

Rousselet, G. A., Foxe, J. J., & Bolam, J. P. (2016). A few simple steps to improve the description of group results in neuroscience. European Journal of Neuroscience, 44(9), 2647-2651. [link]