Working as statistical consultants in business and industry, we find p-values are often a hot topic of conversation.
Essentially, p-values are a widely-accepted convention for reporting on the statistical significance of evidence. And when deliberately or unintentionally omitted from the results, customers often ask for them. It’s as if p-values’ familiarity and ease-of-use bring people comfort in the complex world of statistics. But isn’t it rather strange, if not downright inappropriate, to report results by summarizing them in just one value?
We see it as our challenge to ensure that the essence of experimental results doesn’t get lost by focussing solely on p-values. And in this context, we’d recommend Regine Nuzzo’s fascinating article on this very topic: http://www.nature.com/news/scientific-method-statistical-errors-1.14700