Everything you've said certainly makes the study skewed. And shows that it is only effective with a selected population (which, if the point of the study was to ultimately create a program similar to the study, is certainly wholly valid.)

But to say these factors show it specifically "isn't effective" is just as invalid a logical leap.