Answer to ANCOVA lab question.

Coefficients:
                     Estimate Std. Error t value Pr(>|t|)  
(Intercept)          0.460795   0.400080   1.152   0.2523  
age                 -0.003374   0.007252  -0.465   0.6428  
diagnosisschizo      0.443384   0.539462   0.822   0.4132  
age:diagnosisschizo -0.025433   0.010031  -2.535   0.0129 *

Remember, in multiple regression, the coefficient of one variable tells how the
predicted value changes as that variable is changed. When an interaction is
present, the interaction coefficient tells you how the effect (coefficient) of
one variable changes as the other changes.

Thus, if age changes by 1, the coefficient of diagnosis changes by an amount
equal to the interaction coefficient, and if diagnosis changes by 1, the age
coefficient changes by an amount equal to the interaction coefficient.

In this case, diagnosis is a 0/1 coded variable with schizo=1. Thus, as diagnosis
changes from 0 to 1 (normal to schizo), the age coefficient (effect) would change
as follows:

-0.003374 + (-0.025433) = -0.028807