Selecting Analyses Posted on October 24, 2007 Consider your hypothesis and the data when choosing analyses, states Wesley K. Thompson (bio).

So selecting the correct analyses, it’s important to consider two issues. You should consider the goals of your hypotheses, and you should also consider the way the data was generated, the type of study that it was generated from. So ideally in a well-designed study both of those should be kind of working together, and if they’re not, that’s a sign that you haven’t really designed the study very well.

Other things you might think about are the type of data that you’ve generated; is it continuous or discrete? Do you have covariates you want to include to adjust for? So what type of — is it longitudinal or cross-sectional? And all of those things go into determining which type of model you use when you analyze the data.

« Back to Article