Evaluating Design Choice and Threats to Validity in a Quasi-Experimental Design

¶ … Threats to Validity in a Quasi-Experimental Design

Evaluating design choice: Walk Texas!

The research study by Bartholomew (et al. 2008) entitled "Walk Texas! 5-A-Day intervention for women, infant, and children (WIC) clients: A quasi-experimental study" is defined as quasi-experimental because it lacks a formal control group. The purpose of the study was to determine an intervention designed to improve the eating habits of low-income WIC clients. The participants were "primarily native Spanish speaking, Hispanic women, of low educational level" (Bartholomew 2008: 297). The study "utilized a pre-test post-test quasi-experimental design, with two intervention and two comparison clinics that were matched for size and ethnicity" (Bartholomew 2008: 297). The comparison clinics served as an informal control although participation in the experimental and control groups was not randomized, as would be the case in a true experimental study.

The purpose of the experiment was to see if low-income women who make use of the WIC program would improve their consumption of fruits and vegetables with the dietary intervention designed by the study's authors. In general, low-income people consume fewer fruits and vegetables than higher-income persons because of perceived barriers of cost and preparation difficulty. To create a comparison with a non-intervention group, two similar clinics to those in the study were selected. Also, a follow-up period was undertaken. "Primary measures included: stage of change, barriers to change, attitudes toward F&V, self-efficacy, and self reports of F&V consumption. These were collected at baseline and during follow-up visits to the WIC clinic" (Bartholomew 2008: 297). The follow-up visit was used to see if the nutritional counseling had a lasting effect on participant's eating habits.

Quasi-experimental studies such as these lack the same level of rigor as experimental studies because different factors other than ethnicity and size could impact the results. The lack of randomization could account for the fact that the intervention group showed enhanced willingness to change and also enhanced fruit and vegetable consumption (although it did not show enhanced changes in cognitive constructs linked with a healthy diet such as self-efficacy). Particularly given subjective measurements such as attitudes were said to be enhanced, it is difficult to draw sweeping conclusions from the study. There was also a high degree of attrition from the program because of the instability of the women's lives. Furthermore, merely because they were subjected to an intervention (regardless of whether it was the optimum type of intervention), this could have also caused the change in behavior. At best, this study is a preliminary indication that the intervention has some potential value and could be used to justify a larger and more expensive genuine experimental study.