Statistics is one of the most important yet difficult subjects for many ecology and wildlife graduate students to learn. Insufficient knowledge about how to conduct quality science and the ongoing debate about the relative value of competing statistical ideologies contribute to uncertainties among graduate students regarding which statistical tests are most appropriate. Herein, we argue that increased education of the available statistical tests alone is unlikely to ameliorate the problem. Instead, we suggest that statistical uncertainties among graduate students are a secondary symptom of a larger problem. We believe the root cause lies in the lack of education on how to conduct science as an integrated process from hypothesis creation through statistical analysis. We argue that if students are taught to think about how each step of the process will affect all other steps, many statistical uncertainties will be avoided.
- Experimental design
- Graduate education
- Null hypothesis statistical testing