QUASI-EXPERIMENTAL AND NONRANDOMIZED STUDY DESIGNS
Despite the utility of randomized study designs in minimizing bias due to extraneous confounding factors, they are not always feasible or ethical. For example, randomized studies of group interventions aimed at changing laws and policies and improving healthcare systems may not be feasible.9 Nonrandomized study designs and quasi-experimental studies have also been used in CBPR and program evaluation. Although attention is required to control potential confounding factors and to minimize statistical bias, carefully designed, nonrandomized studies offer useful alternatives to studies that involve random assignment of research participants or groups of people. If the generalizability of results obtained in randomized controlled trials is of concern, results obtained in nonrandomized studies of less highly selected populations may be of particular interest. Alternatives to randomized controlled trials include use of multiple baseline measurements across settings or interrupted time-series designs in which the intervention is introduced sequentially at experimentally determined times in different set- tings.12 There are also examples of successful quasi-experimental studies involving a pre-test/post-test design with a comparison group. Similar pre-test measurement results between the intervention and comparison groups suggest that the groups were comparable before the beginning of the health intervention.4 One advantage of quasi-experimental and nonexperimental studies is that they may have greater external validity, that is, the results may be more likely to apply to other populations or contexts.9
As an example of a quasi-experimental study that was conducted using a CBPR approach, Ralston et al.13 undertook a church-based longitudinal intervention study aimed at reducing cardiovascular disease risk in mid-life and older African Americans. The study incorporated a longitudinal pre-test/post-test design with a comparison group. The intervention was guided by the transtheoretical model of behavior change.14 Community-based participatory research was used to discover research ideas, identify community advisors, and recruit churches (three intervention and three comparison) in two counties in North Florida. A community advisory committee that included pastors, government officials, and representatives of health agencies was established. Measurement data were collected at baseline through self-report questionnaires and clinical assessments. This study is currently underway.