Reflection on p-hacking (1 hour)

↵ Back to module homepage

Here is a list of the papers from the previous task, along with brief descriptions of what ridiculous things they demonstrated:

What all those papers have in common is that they demonstrate that a certain way of data analysis is bad because it might lead you to making crazy conclusions. For example, what the Bennett paper shows is: if you don't do proper correction for multiple comparisons when analyzing fMRI data, you might wrongly conclude that the brain of a dead fish shows different brain responses when it is shown happy faces vs. angry faces. (This is obviously insane.) What the Simmons paper shows is: if you perform lots of different kinds of data analyses without any analysis plan, you might wrongly conclude that participants who listen to a Beatles song literally become younger. etc. The Simmons paper and the Gelman paper included at the above website both give detailed and comprehensive explanations of what p-hacking is and how it can cause you to get junk results which lead to junk conclusions; if you don't feel like you understand the concept of p-hacking, I recommend reading first the Simmons paper and then the Gelman paper. Or, for a funny introduction, see the video below (this touches on a lot of things other than just p-hacking, but p-hacking is briefly mentioned).

Continue to the questions below to reflect on your understanding of p-hacking.

Are there any common research practices (including research practices that you often see in papers in your field, or research practices you've done yourself on previous research projects you did) that you thought were ok but that you now learned are a form of p-hacking? If so, what are they?

Think about your own research (including studies you have done before and/or studies you plan to carry out in the future). What is one way that you might accidentally p-hack—in other words, what is something you might be tempted to do, that would be p-hacking? (The reason I'm asking this is of course not to encourage you to p-hack, but to encourage you to identify potential research practices that you need to be wary of and need to be careful to avoid.)

When you have finished these activities, continue to the next section of the module: "Pre-registration".


by Stephen Politzer-Ahles. Last modified on 2021-05-17. CC-BY-4.0.