Exercise 11: Exploring a simple dataset step by step...

Use these simulated data. They represent...

avgprevgrades: A student's average grade in all previously recorded courses
midterm: A student's midterm exam score in a current class.
final: A student's final exam grade for the course relevant to the midterm variable. 

(1) Answer in one sentence, and report b, beta, and p: Does midterm grade predict grade on final by itself in a simple regression model?

Now you might ask, in a research context, "It appears midterm predicts grade on final, but perhaps a better explanation comes from overall performance academically. Does their midterm grade continue to predict the final grade if we control for their average previous academic performance?..."

(2) Answer in one sentence, and report b, beta, and p: Does midterm grade still predict grade on final when you've controlled for average previous grades?

(3) Answer in one brief complete sentence: What is the interpreted effect size of the predictors in the full model, in Keith's basic terminology? (See p. 62, or lecture slides.)

(4) Briefly: How much variance in percentage does the overall, two-predictor, model account for?

(5) Compared to just the model with midterm grade as the predictor, by how much did the variance accounted for increase when you add average academic performance?

(6) Imagine you're a researcher and you just ran this short study testing whether previous academic performance is a better predictor of final term grade than midterm grade. Run the full model and write it up in the manner describe in Keith (pp. 31-32) or as demonstrated in the slides.

Send answers to psyc7302@gmail.com with subject line "EXERCISE 11."