No abstract

This investigation was designed to identify the risk factors associated with different stages of cigarette use in a large biracial adolescent sample. A questionnaire assessing smoking habits and variables thought to be related to smoking was administered to 6,967 7th graders. Analysis revealed that the best predictor of experimentation with cigarettes was the perception that they were easily available. Regular smoking appeared to be heavily influenced by cost. Social influences contributed to both experimental and regular smoking, but the impact of social models varied with ethnicity and gender. Analysis further revealed that weight-related variables were closely tied to regular smoking. Implications of the findings for smoking prevention programs are discussed.

This article reviews the major findings from a multiproject meta‐analysis of the effects of marital and family therapy (MFT). Across 163 randomized trials, MFT demonstrates moderate, statistically significant, and often clinically significant effects. No orientation is yet demonstrably superior to any other, nor is MFT superior to individual therapy. Cost effectiveness information is scant in these 163 studies, but supportive. Randomized experiments yield very different answers from nonrandomized experimental studies of the effects of MFT, calling into question whether we should mix the two in reviews. We have also found several new differences in the ways that marital therapy (MT) and family therapy (FT) studies are conducted, making them harder to compare. Finally, important questions still exist about whether any psychotherapy, including MFT, yet has sufficient information about how well research generalizes to everyday clinical practice.

Attrition from conditions in randomized experiments is common. Yet it is difficult to assess the possible effects of attrition because the outcome status of the dropouts is usually unknown. This article develops methods to assess those effects in studies with dichotomous outcomes, illustrating the methods with randomized experiments in drug abuse treatment, smoking cessation treatment, and alcoholism treatment. The methods include computing the lowest and highest possible effect sizes that could have been observed, enumerating the percent of possible study outcomes below a given threshold, estimating the probability that an outcome beyond any given threshold would be observed if all participants were measured, and constructing attrition analysis plots showing the effects of attrition under varied assumptions. For the kind of study to which they apply, these methods should replace the treatment of missing participants as failures in an "intent-to-treat" analysis. A user-friendly personal computer program is available to implement all of these analyses.Postassignment attrition of participants from conditions has always been a widely acknowledged Achilles' Heel of the randomized experiment. Cook and Campbell (1979), for example, noted that it is a key threat to internal validity (an inference that the treatment caused the outcome) that random assignment does not control. Attrition is widely known to occur, particularly in field experimentation; it is often systematic rather than random; and it can take many

No abstract