Intentionality in Modeling (with student comments in quotes)

I think students may start this class thinking that analysis is an arbitrary set of conventions that are carried forward without thought or preference. I think students leave the class feeling a bit more how much agency and responsibility they can have for shaping their analysis to the assumptions they understand to hold (or not) and what questions they want to ask

"That is, they really seem to catch on to the idea of how quantitative modeling is a way to phrase an argument, how there is a user bringing perspective and bias not just to the interpretation of a result but also the the application and design of a test."

"The assumptions underlying the modeling we do (i.e. linear models) frames the types of questions we can ask/how we interpret any results."

"I have learned to keep trying because there are so many more options for analysis than t-tests and ANOVAs. I had almost no experience of modelling before this class, so even just the fact that I now feel like I have more options is very exciting. I think the ability to use modelling to account for so many variables which can affect the model at different rates (or different times!) - that makes me excited to see how the literature in the field of psychology and neuroscience will grow and incorporate this more in the future."

"I've learned a deeper understanding of regression analysis. Being able to look at the world and think of actions as a sum of variables is really different but has already changed the way I think about the world. This skill allows me to ask questions in a new way that I haven't been able to conceptualize before."

"The most important thing I've learned is being familiar with some of the assumptions we are operating under every time we click those 3 simple clicks to run our t-tests/ANOVAs--most especially independence."

"Data collected intentionally can teach us many things, the greatest limit is the questions we do (not) ask. Going forward, this means I can find whatever questions I want so long as my models and questions are insightful."

Sheesh! This class does not has as much time as I'd sometimes like for discussions about what makes a problem computable or not, but clearly some of these students are already making their way to this question.

Back to What you Learn, back to Longitudinal & Time Series Analysis, or back to In the Classroom.