We need to make a logic tree for statistical tests, they said. It will be fun, they said.
Choosing the right statistical test for your analysis can be extremely confusing and challenging. To help you out, I've complied this decision tree. Here are some quick tips to get you started:
1. Figure out what sort of data you have. Is it numerical? Categorical? Ordinal (categories that are ordered, such as small, medium, and large)?
2. Identify your dependent and independent variables. Your independent variables are the factors that influence the dependent variable in your analysis. The dependent variable is the one you're trying to predict or explain.
3. Work out how many dependent and independent variables you have. Usually, we set up our experiments so that one dependent variable is predicted by one or more independent variables, but it's surprisingly easy for the setup of the experiment or other circumstances to end up making this much more complicated!
4. Examine the nature of your variables. If they are categorical, how many levels do they have? Do they meet statistical test assumptions such as normality (shaped like a bell curve)? Are there a lot of missing or zero values that might influence the outcome of your test?
These steps and the logic tree above are just a simple starting point, but hopefully it will get you started!