It can be easy to get distracted during your research by exciting, unexpected findings or outcomes that are concerning. But you have a timeline to stick to and running after every shiny object that comes your way can throw you way off your original, intended path. Instead of splitting your attention between your actual research experiment and other factors deserving of further attention, it’s a better choice to sideline those unintentional results by documenting them and scheduling them for a post-hoc analysis once you are done with your current project.
A post-hoc analysis isn’t meant for every random outcome in an experiment. It’s a way to look for patterns when you see something that could warrant its own deeper look and possible research study to further the knowledge of this interesting concept. When you implement post-hoc analysis in your research findings, it is a great opportunity for you to delve deeper into your subject base and start your next experiment based on a foundation of information that you’ve already established.
What is a Post-Hoc Analysis?
Sometimes a post-hoc analysis can be just as revealing and instrumental as the original research study. In this type of deep analysis, data that was made apparent during the research is evaluated for patterns that could have relevance in your field or somewhere else. By focusing solely on your original study’s primary objective, you can get the main research done in the detail that is necessary and then come back to anything that was not in your main plan and make it subsequent, additional research.
Post-hoc studies use data that was already collected. You can use that information to analyze the same data for other objectives you hadn’t intended to originally. The pooled data from your original study’s trials can be studied and evaluated for post-hoc analyses that may call for an entirely new research experiment.
When is a Post-Hoc Analysis Necessary?
A post-hoc analysis has lots of benefits, but it’s not always the most effective use of a researcher’s time. You have to be cautious not to place more potential value in an unexpected outcome than it may be worth; then again, you also need to use your judgment to see if there might be a diamond’s worth of knowledge hiding under the rough data you’ve inadvertently uncovered. Is it a coincidence, or is it an entirely new development that could have substantial impacts?
There are limitations, too. For instance, in a clinical trial where all the data was ascribed to a specific parameter, it can’t be used in a cross-study or a post-hoc analysis for future use without consent.
A good example of a time when a post hoc analysis would be warranted is when there is reason to check for error rates, probabilities of a potential hypothesis being significant or null, and to look for false positives or statistical significance. When you’re conducting multiple trials or going in different directions with your research, the chances of errors and false positives increase.
However, as you continue to become experienced in the regular patterns of research experiments, it is easier to notice an unexpected finding, document it accordingly, and stay focused on your original project. You can come back to it when you’re done, and if it’s relevant to the original work, you can even include your findings in your research article, add a section to address the upcoming post-hoc analysis based on the sidelined data you plan on studying further, or make an addendum later after you’ve published your original findings.