In research, the p-value is a statistic that shows if data is statistically significant. A p-value that is less that .05 shows strong evidence to reject the null hypothesis. Lately, the p-value has been under scrutiny. When studies that are less impactful or less evidence based have an acceptable p-value, and studies with good evidence and an 'unacceptable' p-value are not published, that's when things get heated.

It is not a replacement for scientific reasoning, but both sides need to be present in order for accurate conclusions to be made. It is reassuring that people believe scientists should publish their research even if the p value significance level is not met. We should not emphasize smaller p-values over other more important scientific reasoning.

Basic and Applied Social psychology says their research will no longer be accepting the “null hypothesis significance testing procedure," and authors now have to remove them in order for publication. They say the problem is that p values are “traversing the distance from the probability of the finding, given the null hypothesis, to the probability of the null hypothesis, given the finding.” They say banning the NHSTP will allow more quality manuscripts to published, rather than low quality research that has an acceptable p-value.

In an article titled "Statistical Errors," author Regina Nuzzo says, “The P value is neither as reliable nor as objective as most scientists assume.” Why doesn’t it do its job? It creates problems when someone is trying to replicate a study. It also was meant to be something that is part of a bigger revision of scientists with their work, but I think because it is easy to skim over and numeric, more importance was placed on it. I think we should make readers aware that all the p value does is “summarize the data assuming a specific null hypothesis.” There are limits to what the p-value can do.

I think as a collection of scientists, we need to create a new and improved system for determining significance more so than just relying on one statistic.