Skip navigation

Models & statistical models

Learning outcomes

On watching this video, students should be able to:

  1. Sketch/plot simple theoretical models for continuous outcomes that involve up to 2 explanatory variables, including being able to illustrate confounding and interaction
  2. List the questions a regression model can answer: Is there a relationship? What it the (nature of the) relationship?
  3. Explain the terms: simple regression, simple linear model, multivariable regression, and adjusted measure of effect.
  4. Identify a regression coefficient as an estimate of a measure of effect that is of interest to a study, and hence realise that a regression coefficient is just a point estimate of a population parameter exactly like we’ve encountered in statistical inference with as sample means and proportions.

Corrections:

04:00: The green double headed arrow to illustrate the difference between the reference category and 2nd sample mean for the “yes” category should come all the way down to the reference group line

7:30: I talk about effect modification as the association being dependent on a 3rd variable. You don’t need to think about the modifier as the “3rd variable”, I just stated that because I was emphasising that there are 3 variables in the model.