This is a very interesting article (even if I came across it rather late). I guess it is even more interesting to me because I have been puzzled by the use of what we could call NHST in high ranking journals such as Nature and Science over the years. I see this occurring often in neuropsychological articles, which makes me wonder whether the use of p-values is more of a relic of undergraduate training and routine than anything else. Most often than not, p values appear as stand-alone and unnecessary addenda in a rather inconsistent manner (basically, if those articles eliminated their p values, none of the articles' contributions would be lost.)

Also to add, the use of NHST may be a sign of poor statistical standards... but I wouldn't be so sure about it. NHST is a philosophical confused merging of Fisher's tests of significance and Neypam-Pearson's tests of acceptance and it should be rejected on this ground alone. I wouldn't extend this recommendation to the base theories of Fisher and Neyman-Pearson, though (I even proposed a solution to their confusion in Perezgonzalez, 2014). In any case, much use of NHST is rather harmless, as researchers do not seem to really care much about the philosophical background of NHST and, thus, it defaults to Fisher's more often than not.

However, I am not too sure confidence intervals (CI), effect sizes (ES), etc necessarily provide better statistical standards either. After all, as I discussed in another article in relation to CIs (Perezgonzalez, 2015), tests and CIs answer different questions and, thus, different standards. CIs and ES are more descriptive, tests seek only a yes/no answer to a question.

----
Perezgonzalez JD (2014). A reconceptualization of significance testing. Theory and Psychology. DOI: 10.1177/0959354314546157
Perezgonzalez JD (2015). Confidence intervals and tests are two sides of the same research question. Frontiers in Psychology. DOI: 10.3389/fpsyg.2015.00034