Key Info and Announcements

Modified

March 3, 2025

Announcements

  • I mostly caught up on muddy points!

  • Today is last day of new material

  • Next class, I will cover some good ways to communicate and visualize results

    • Will be helpful in final project!

  • Class on Wednesday is INFORMAL

    • I will be here in person

  • Classes on Monday (3/10) and Wednesday (3/12) will be virtual!

    • Basically like office hours for your project and final homework

  • Lab 4 up!

  • Homework 5 is up!

Key Dates

  • Lab 4 due 3/7
  • HW 5 due 3/14
  • Project poster due 3/17

Last class

We went through the steps of purposeful selection:

  1. Exploratory data analysis

  2. Check unadjusted associations in simple linear regression

  3. Enter all covariates in model that meet some threshold

    • One textbook suggest \(p<0.2\) or \(p<0.25\): great for modest sized datasets
    • PLEASE keep in mind sample size in your study
    • Can also use magnitude of association rather than, or along with, p-value

  4. Remove those that no longer reach some threshold

    • Compare magnitude of associations to unadjusted version (univariable)

  5. Check scaling of continuous and coding of categorical covariates

  6. Check for interactions

  7. Assess model fit

    • Model assumptions, diagnostics, overall fit