Hoerster, et al. (2011) explore the individual, policy, and environmental factors that influence the utilization of healthcare among farmworkers in the United States. They used secondary data from the National Agricultural Workers Survey and the Uniform Data System. The researchers used logistic regression with weighted hierarchical linear modelling to determine the impact of the various variables on the dependent factor. The study illustrates the effectiveness of logistic linear regression in exploring the relationships between variables and developing explanatory models.
The main goal of Hoerster, et al.’s study was to determine the various factors that influence healthcare access among the farm workers. They identified three main factors including individual, policy, and environmental factors as the main determinants of healthcare access among the target population. The individual level factors included gender, immigration and health status, proficiency in English language, access to transportation, and access to care outside the United States. The environmental factor included the proximity to the Mexico-US border, while the policy level factors included access to insurance and payment structure of the workplace. These factors had binary outcomes, which could only be analyzed using logistic regression modelling.
The research finds that limited access was the most significant hindrance to farmworkers healthcare. Hoerster, et al (2011) recommend the improvement of access at each of the three key factors. In addition, the researchers recommend the establishment of a federally Qualified Health Center close to the farms to enhance access. The researchers hope these recommendations will enhance both the access and quality of healthcare accessible to the vulnerable farm workers.
The study could have used alternative methods of analysis to answer the research question. Firstly, the researchers could utilize analysis of variance to identify the factors that exhibit significant associations. Analysis of variance is equivalent to linear regression if models test similar hypotheses and have same encoding (Jackson, 2016). Alternatively, the researchers would have used the nearest neighbor analysis to determine the distribution of variables across various geographic areas featured in the study (Wagner & Gillespie, 2018; Wisniewski, 2016). The nearest neighbor analysis would enable the researchers to determine the clustering or uniform spacing of the variables. In this case, close clustering of sampled population within specified geographical areas could have provided insights as to the underlying factors that drive access to healthcare.
The main weaknesses of Hoerster, et al.’s study is its reliance on secondary data and the assumption of linearity between the variables. The secondary data used in this study included only 39 US states. Therefore, it excluded 11 states that could have provided important insights on the population. Secondly, the research assumed a linear relationship between the dependent and independent variables. These variables could not have been linearly related given their nature and magnitude of variation. Lastly, the researchers used dichotomous outcomes for all the variables. Some of the variables such as English proficiency could have been coded into a continuous dataset. Nonetheless, the expression of these variables as dichotomous facilitated their processing using the logistic regression model.
The above weaknesses could have been addressed by collecting primary data. This primary data could be collected from a representative sample of the local population. The secondary data used in the research was not representative of all US states, which led to unreliable conclusions. In addition, it would use alternative statistical methods to process the data such as analysis of variance to make the findings more generalizable to the overall population.
References
Hoerster, K. D., Mayer, J. A., Gabbard, S., Kronick, R. G., Roesch, S. C., Malcarne, V. L., & Zuniga, M. L. (2011). Impact of individual-, environmental-, and policy-level factors on health care utilization among US farmworkers. American Journal of Public Health, 101(4), 685-692.
Jackson, S. L. (2016). Research methods and statistics: A critical thinking approach. Cengage Learning.
Wagner, W.E., & Gillespie, B.J. (2018). Using and interpreting statistics in the social, behavioral, and health sciences. New York, NY: Sage Publications
Wisniewski, M. (2016). Quantitative methods for decision makers. Harlow: Pearson.