Students struggled with…

  • Using “predicted” in interpretations.  It is so important that students explain that the slope and y-intercept of the LSRL are predictions, not guarantees.
  • Describing why a residual plot was evidence that a linear model was not appropriate.  They said things like, “It is random” and “the distribution shows no pattern.”
  • The interpretation of the standard deviation of the residuals.

Remedies…

  • It may seems harsh, but take off partial credit for leaving out “predicted”.  It’s important to break this early and often.  Assuming tests are worth more than quizzes anyways, better they get points off on quizzes than on the test.  This is always are motto with quizzes.  We tend to grade very precisely.
  • Again, it may seem stringent, but students must be precise in their language.  They need to explicitly reference the residual plot.  It is not the distribution that shows no pattern, but the residuals.
  • Encourage students to make flash cards for the interpretations.  Then practice them often! We used the Big Ideas page to keep all the interpretations in one place.

 

Previous DayNext Day