Ways by which a researcher can control confounding variables in a correlational study include all of the following, except:
a. Randomization
b. Using a homogeneous sample
c. Analysis of covariance
d. Matching subjects