|
Abstract:
|
Earlier in 2015, the editors of Basic and Applied Psychology (BASP) announced "BASP is banning the NHSTP (null-hypothesis significance testing procedure)." Why? Because of "the problem of traversing the distance from the probability of the finding, given the null hypothesis, to the probability of the null hypothesis, given the probability of the finding." What? So what statistics should be used in papers? The answer? "Strong descriptive statistics, including effect size." This is because the BASP editors believe the state of the art of statistical inference making remains uncertain. Of course it is uncertain! That is one of the cornerstones of statistics and science in general. Remove uncertainty and you go from statistics to math. So the real problem appears to be not whether statistical inference is fraught with problems, but whether the statistical inferences are being properly described and used in published works. Let's talk about how we, as statisticians, can educate and assist our colleagues in properly interpreting and publishing the statistical analysis of their experiments.
|
