in a correlational study on the relationship between caffeine consumption and heart disease in police officers, the fact that the officers could not be randomly assigned to high and low caffeine groups suggests the results may be due to: