Possible sources of attenuation

Problem in
Rind et al. ?

Random error of measurement in dependent variable

Partial

Random error of measurement in independent variable

Yes

Artificial dichotomization of continuous dependent
variable 
In some studies

Artificial dichotomization of continuous independent
variable 
Yes

Imperfect construct validity of the dependent variables

Yes

Imperfect construct validity of the independent
variables 
Yes

Range restriction on the independent variable

Yes

Range restriction on the dependent variables

Yes

Bias in the correlation coefficient

Minimal

Studycaused variation

Yes
