ERIC Number: ED224832
Record Type: RIE
Publication Date: 1982-Oct
Reference Count: 0
When Groups Aren't Random: Using the Analysis of Covariance in Family Studies Research.
Brown, Lynne Harrington
Among the many problems faced by family researchers in conducting experimental research in field settings is that posed by nonequivalent and intact treatment and control groups. Nonequivalence refers to the assignment of individuals to treatments on a nonrandom basis, thus generating treatment groups with different expected values on one or more relevant variables. Treatment groups formed independently of either experimenter or experiment may be viewed as both nonequivalent and intact. Given that practitioners and researcher increasingly view treatment efficacy as an important goal in treatment delivery, the question arises as to how variations in outcomes due to the treatments of prevention programs may be distinguished from variations in treatment outcomes due to nonrandom assignment. The use of an indirect or statistical control technique, the analysis of covariance (ANCOVA), in place of or in addition to experimental control is examined. An overview and explanation of the ANCOVA includes a demonstration of the procedure using outcome data from intact groups. Possible applications of ANCOVA for family studies research with illustrative examples from current literature are presented. This paper shows the utility of the basic ANCOVA model as a statistical control in a broad range of research settings. (CM)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Nonrandom Selection
Note: Paper presented at the Pre-Conference Theory Construction and Research Methodology Workshop of the National Council on Family Relations (Washington, DC, October 13-15, 1982).