Day 6 - Multiple Regression | Data Analysis for Neuroscientists

Topics Covered

Fit models like body_mass_g ~ flipper_length_mm + sex and interpret partial effects while controlling for the other predictor.

Model mean differences with body_mass_g ~ species + sex, understand reference levels, and read dummy-coded coefficients.

Use anova() to test whether we need interaction terms like species * sex or flipper_length_mm * sex.

Tutorial for continuous outcome with one continuous predictor and one categorical predictor, including model comparison and adjusted predictions.

Tutorial for continuous outcome with two categorical predictors, including additive vs interaction models and predicted cell means.

Dedicated interaction-effects tutorial covering continuous * categorical and categorical * categorical models with nested comparisons.

Partial regression coefficients: each term is interpreted while holding other predictors constant.
Reference levels in factors: the intercept and factor coefficients are always relative to a baseline category.
Additive models vs interaction models: interaction terms test whether one predictor's effect changes across groups.
Nested model comparison with anova(): checks if extra terms provide a meaningful fit improvement.

Run mr_cont_cat.R and explain the coefficients for flipper_length_mm and sexmale.
Use anova() to compare additive and interaction models in both scripts. Report which model is preferred and why.
Run mr_two_cat.R and interpret one species coefficient and one interaction coefficient in plain language.