You conduct a hypothesis test and obtain statistical significance for your research. However, you never consulted with a statistician. But that’s okay right? Well, not so fast…sigh. You want to make sure you know the power of the test. As I have described in other blogs, the power of the test will determine how many patients you need to obtain statistical significance. So you could have statistical significance and a lousy power of the test. Why? Because a low power of the test reflects a decreased probability of showing a true effect. Furthermore, a low power probability means you could run an experiment with more subjects than you actually need in the study.
Now, I know that grades make people confident. You passed all of your statistics courses with flying colors. You were the “Karate Kid” of experimental design class. There is not a statistical model that you can’t fix and I get it. But, there is a difference between taking a few courses in statistics and making stat your career. The term “statistician” implies that the person spends all of their workday estimating the best model possible. And each statistician has a particular specialty. Some people specialize in oncology studies and survival analysis. Others prefer longitudinal studies and psychological concentrations. The statistics field is a mixed bag. If it’s your first research project, they might be able to tell you what model will meet the objectives of your study.
Before you take a project and run with it, you might consider speaking with a trained professional in regards to your statistical methods.
If you enjoyed this blog there are more to follow-Amy