Best practices and evaluating evaluations

The August issue of the journal Family Relations includes an article for practitioners on "best practice" reports. Many best practices summaries may be based on evaluations that could be flawed; the evidence suggesting that a practice is a "best practice" could be weak. Even a randomized sample, a relative rarity in evaluation research but often considered the gold standard, may result in findings which are not ultimately applicable to the real world of programming. The article's author, Jonathan Olson of Penn State-Altoona, suggests considering the evaluations as well as the population used in the evaluation, as well as considering the source of the program and evaluation, to decide whether the results will be applicable to the population under consideration for programming.