Section 9.13 Checklist
When running a logistic regression, have you:
-
Identified your outcome variable as binary (two categories only)?
-
Verified your predictors are correctly coded (numeric or factor as needed)?
-
Explored the data first (EDA, visualizations, proportions, odds)?
-
Split your data into training and testing sets?
-
Fit your model using
glm(..., family = "binomial")? -
Checked model summary for significant predictors?
-
Converted coefficients to odds ratios using
exp()? -
Checked for multicollinearity using VIF (< 5 recommended)?
-
Calculated a Pseudo-R² (e.g., McFadden’s R²) to assess model fit?
-
Generated predicted probabilities (
type = "response") for the test set? -
Chosen a classification threshold (e.g., 0.5) and created predicted classes?
-
Used a confusion matrix to evaluate accuracy, sensitivity, specificity?
-
Created an ROC curve and reported the AUC value?
-
Written a clear interpretation that answers your original research question?
