ScWk 240 Edward CohenPage 1ScWk 240 Research Methods and Design Week 9 October 24, 2008 San José State UniversityEdward Cohen, Ph.D.Today: •For next week: Readings for Week 9 (10/24) •Lit Review check in •Quasi-experimental time series designs •Single case design •Introduction to qualitative research Quasi-Experimental Time Series DesignsA.Review your list of group designs to answer the following: 1.A researcher gathers a convenience sample of adolescents in a group home to try an intervention designed to reduce behavioral outbursts—a new group therapy focused on anger management. First the researcher measures the number of outbursts for 10 weeks before the intervention, then begins the group, and then measures the number of outbursts for 10 weeks after the group starts. 2.In the same group home, the researcher divides the adolescents into two groups—the first group gets the new group therapy, and the second group receives the standard counseling already provided. Both groups are matched on length of time in the group, severity of behavioral problems, gender, and ethnicity. The measurement strategy is the same as in #1. 3.Name two threats to internal validity that can be addressed by #2 to improve the research design from #1. B.Simple and Interrupted Time Series (Quasi-Experimental Design) 1.A simple time seriesis a longitudinal design with multiple, repeated observations or measurements. Example: one-year stock price chart. 2.An interrupted time series designis used to evaluate treatment or the effects policy or environmental phenomena on the dependent variable 3.For the interrupted time series design, the main question is “is the trend of the data line affected by the intervention, or by other rival hypotheses?” 4.The quasi-experimental time series design allows you to analyze trends in observations that can help rule out effects of history, maturation, or other reasons why the “trend line” might have patterns that have nothing to do with the treatment
has intentionally blurred sections.
Sign up to view the full version.