Evaluating Research: Methodology for People Who Need to Read Research by Francis C. Dane

Author:Francis C. Dane [Dane, Francis C.]
Language: eng
Format: azw3
Publisher: SAGE Publications
Published: 2017-07-27T04:00:00+00:00

Of course, the nonequivalent groups basic pretest design, similar to the time-series and regression–discontinuity designs, is subject to selection effects because it lacks random assignment to conditions. When researchers use this design, critical readers must look for information in the report about how the researchers ruled out alternative explanations through logical analysis rather than through random assignment and control over variables.

Matching, also discussed in Chapter 6, provides an alternative to random assignment but is not as effective as random assignment for dealing with selection effects. Matching, you should recall, involves an attempt to equalize scores on variables of interest among groups. There is no end, however, to the number of potential variables that may require matching, and therefore, one can never be sure one has matched participants on all relevant characteristics. Cerdá et al. had the added advantage of having all of the pretest results available while they chose the control neighborhoods and so were able to use propensity scores, also discussed in Chapter 6, to match the intervention and control neighborhoods statistically on all of the variables of interest. On those posttest measures on which there were differences (including homicide rates), one would need to construct an alternative explanation based on a variable that was not measured in the study to discount the results.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.