Session Ready
Exercise

From experimental results to a prediction

In the previous exercise, you were directed to follow a common strategy: establish a baseline, then change one input at a time to estimate the effect of that input on the output.

In this exercise, you'll be working with a new (and different) test_scores() mathematical model. You'll follow the same procedure as before to estimate the effect of each variable.

Instructions
100 XP
  • Make a baseline of a "public" school with academic and religious motivation both set to zero.
  • Make the following changes one at a time, leaving the other inputs at their baseline values:
  • Change school to "private".
  • Change academic motivation to 1.
  • Change religious motivation to 1.
  • Using just the results you get from these four model runs (be honest now!), make a prediction of what the model output will be for a "private" school where the student has religious motivation of 2 and academic motivation of 2. Save your prediction in the variable my_prediction. Use this very common linear strategy for making your prediction:
  • If the change in output going from 0 to 1 is X, then the change in output going from 1 to 2 will also be X.
  • Find the change in output for each of the inputs in turn. Add up the individual changes to get the anticipated change when more than one input is changed.
  • Once you have made your prediction, test it out by making a new run of test_scores() with those new inputs.