Strengthen Internal Validity
Moreover, the nurses in the experimental group worked a fixed schedule whereas nurses in the control group worked alternating shifts. Secondly, the lack of randomization coupled with the fact that the participants worked for the same organization could have contaminated the results. Thirdly, while the results of the research showed the exercise intervention improved the physical fitness of the participants in the experimental group, participants were not evaluated for musculoskeletal improvements.
Recommendations to Strengthen Internal Validity
A different research design would have strengthened internal validity. Randomization is the most effective way to control individual characteristics of participants. Randomization also eliminates for the Hawthorne Effect, which occurs when participants behave differently because they know they are being studied. Moreover, a cross-over design is highly effective when groups are being compared to one another. Although, this design is subject to carryover bias, in which an effect carries over from one experimental condition to another (Polit & Beck, 2017).
I contend that a randomized control trial with a cross over design would have increased the strength of the internal validity in this study. In a cross-over design participants serve as their own control group, which would negate the convoluting variables that influenced the results of this study, and would more accurately gauge changes resulting from the exercise intervention. I would also add a metric to assess the musculoskeletal status of the participants. To limit the effects of carryover bias, the health metrics of the participants would be obtained before the exercise intervention to establish a baseline, then after the exercise intervention, and finally, after a wash-out period, the metrics should be re-recorded.
The Impact of Changes on Other types of Validity
In contrast to internal validity, statistical validity is not concerned with the causal relationship between variables, but rather measures the mathematical correlation of all relationships that occur between the variables (Polit & Beck, 2017). The randomized control, crossover design would improve statistical validity because the participants would serve as their own control group making statistical analysis more powerful. Construct validity determines if the outcome measured corresponds to the theoretical construct of the study (Polit & Beck, 2017). In this research, the theoretical construct was Pender’s health promotion model. Construct validity also would have been improved by changing the design of the study. The same health promotion strategy yields different outcomes for different participants based on individual differences. The modification of the study’s design would have negated these individual differences. External validity indicates if the results of the research will remain the same when applied to other people or settings (Polit & Beck, 2017). Again, a change in the design of this research would optimize external validity which would increase the likelihood of the results influencing evidence-based practice.
Failure to Consider Validity in Research
