Before attempting intellectual skills training, there are some important matters to consider. These matters take the form of a set of criteria, or prerequisites, for intellectual skills training programs. There are at least five reasons why a set of prerequisites is needed for skills training programs. First, the programs that have been implemented in the past have trained different skills in different ways, and, where they have been evaluated, they have been evaluated by different means (e.g., see Detterman & Sternberg, 1982). The variety in program evaluations has made comparison of success across training programs difficult. A common set of criteria for evaluating the programs and their evaluations should be useful in achieving some basic uniformity of standards in the face of great diversity of implementations. Second, it must be clear what it is we are supposed to train and what it is we are supposed to evaluate. Third, we seem to be emerging from a quiescent period for intellectual skills training programs that followed the most optimistic years of Head Start. After this period of retrenchment, we are witnessing a burgeoning of new programs. If there has ever been a time for some setting of guidelines, now is that time. Fourth, the costs of such programs--measured in terms of monetary, physical, and human resources (both the students and trainers)-can be monumental. Before making a massive investment of any kind, it would be helpful to know that a proposed program meets at least some basic guidelines. Finally, intellectual skills training programs may do considerable good, but they also may do considerable harm: A review of the history of educational interventions in this country would show that many of the innovations that were viewed with great favor when they were introduced were viewed with disfavor and dismay shortly thereafter. Anything that can be done to spare us the errors before they are committed will perform a service indeed. The eight prerequisites that I propose may seem obvious but they are important. What is less obvious is why few, if any, available programs have adhered to the full set of recommendations. Therefore, although these are not the only recommendations that could be listed, and although adherence to them will not necessarily guarantee a successful program, collectively taken they should enable investigators to avoid those errors that are avoidable.