This discussion focuses on diabetes and intervention to impact patients’ health outcomes. According to Ford et al. (2021), a possible way to address the issue and reduce the spread of diabetes is to introduce an education program to increase physical activity among at-risk individuals. These researchers relied on a quantitative methodology to test the effectiveness of the intervention. Thus, 292 individuals were members of the control group, while the intervention one included 281 members, making a total of 573 participants (Ford et al., 2021). The practice question is: In people with prediabetes (P), does the education program to increase physical activity (I), as compared to current practice (C), lead to reduced incidence of diabetes (I) over 8 weeks (T)? The scientists relied on Student’s t, chi-square, and other parametric tests to determine how effective the education program was. However, it could also be appropriate to rely on non-parametric tests, including Mann-Whitney and Mood’s median tests.
If I were attempting to determine if the intervention was better than current practice, I would implement the Mann-Whitney test. This non-parametric test is typically used when it is necessary to assess the differences between two independent groups. There is a robust rationale behind implementing the given statistics to the situations under consideration. On the one hand, Noguchi et al. (2020) stipulate that the effectiveness of this test is close to 100% when it is applied to data that is not normally distributed. This information demonstrates that using Mann-Whitney statistics results in obtaining statistically significant and relevant outcomes. On the other hand, this test is suitable for post hoc comparisons when some statistically significant differences exist between groups (Noguchi et al., 2020). Ford et al. (2021) admitted that the groups showed notable variations regarding their waist circumference rates. That is why there is some reasoning behind implementing the Mann-Whitney test to assess the effectiveness of the proposed intervention.
In addition to that, Mood’s median test can produce valuable information. The given non-parametric statistics is typically used to identify whether the medians of two independent groups are similar or different. According to Alfaro et al. (2021), this test also produces credible and reliable outcomes when the level of significance is set at 0.05. The given approach is a non-parametric counterpart of a one-way ANOVA. It is reasonable to use the test under consideration to the proposed scenario because if it identifies that the medians are different, it is possible to stipulate that the intervention is effective. Consequently, Mood’s median test can help answer the research question that has been presented above.
In conclusion, it is necessary to explain why it is at all reasonable to implement non-parametric tests to analyze the intervention’s effectiveness. Even though the study under analysis does not offer a small sample size, other features advocate for using non-parametric tests. Firstly, such statistical values should be used when data is not normally distributed. In other words, if researchers are not sure that this condition is met, it is more appropriate to rely on non-parametric tests that focus on medians rather than means. Secondly, since Ford et al. (2021) did not explain whether they identified any outliers and what effort they made to minimize their impact, the proposed non-parametric tests are appropriate. The rationale behind this statement is that these statistical values are applied when the data has outliers that cannot be removed.
References
Alfaro, M., Munoz-Godoy, D., Vargas, M., Fuertes, G., Duran, C., Ternero, R., Sabatin, J., Gutierez, S., & Karstegi, N. (2021). National health systems and COVID-19 death toll doubling time.Frontiers in Public Health, 9, 669038.
Ford, C. N., Do, W. L., Weber, M. B., Narayan, K., Ranjani, H., & Anjana, R. M. (2021). Moderate-to-vigorous physical activity changes in a diabetes prevention intervention randomized trial among South Asians with prediabetes – The D-CLIP trial.Diabetes Research and Clinical Practice, 174.
Noguchi, K., Abel, R. S., Marmolejo-Ramos, F., & Konietschke, F. (2020). Nonparametric multiple comparisons.Behavior Research Methods, 52, 489-502.