Selecting Variables into a Regression Equation/Summary of Recent Research
SECTION A
References Miller, N. A., Kirk, A., Kaiser, M. J., & Glos, L. (2014). The Relation Between Health Insurance and Health Care Disparities Among Adults With Disabilities. American Journal Of Public Health, 104(3), e85-e93. doi:10.2105/AJPH.2013.301478
1.Summary of Study
The study was aimed at understanding the disparities among US adults with disabilities. It aims to determine the extent to which health insurance reduces the disparities by race, ethnicity, and socioeconomic statuses. Measures of access and use was modelled as functions of predisposing, enabling, need and contextual factors. The effect of health insurance on reducing observed differences on the base of race, ethnicity and SES was noted. (Miller, Kirk, Kaiser and Glos, 2013)
2. Identify the null hypothesis and alternative hypotheses (you will need to write the null in your own words as this is not provided in most research reports).
The null hypothesis was that providing health insurance does not reduce the differences and disparities for access and use amongst people with disabilities on the basis of race, ethnicity and economic status significantly. The alternative hypothesis is that providing health insurance does reduce the disparities for access and use amongst people with disabilities, who are different on the basis of race, ethnicity and economic status. There were four different kinds of hypotheses tested here: (Miller, Kirk, Kaiser and Glos, 2013)
3. Share the regression analysis statistic that is reported (linear regression, multiple regression, hierarchical, stepwise, etc.).
The statistical analysis used was multivariate logistic regression analysis as study of the effect of many independent variables on the dependent variables was carried out. Multivariate regression is a technique that calculates estimations from a single regression model with more than one outcome variable. When there is more than one predictor variable in a multivariate regression model, the model is a multivariate multiple regression. (State Data Analysis examples, n.d.)
4. Address the type of data required and whether assumptions of the test were met.
Since this is a multivariate analysis, you need large number of observations from many predictor variables and this was available for this study where data was obtained from many populations under the effect of four or five predictor variables, whose effect on the dependent variables (reduction of disparites in the disabilities) was studied. Use of multivariate logistic regression model analysis requires use of large samples and there should be a moderate correlation between the outcome variables for the regression to make sense. This was met in this test. (State Data Analysis examples, n.d.) (regression, 2014)
5. Briefly discuss whether the finding in your study is statistically significant and what this means.
The findings in the study were statistically significant as p was <0.005 and it indicated that in general health insurance did not have much effect on reducing the disparities suffered by disabled people on the basis of various reasons.
SECTION B
1. Create an example of an imaginary quantitative nursing study you would design that would use regression analysis.
A nursing study one can design using multivariate regression analysis would be involving studying the impact of different environmental factors such as oxygen, nutrient, iron and vitamins, on the development of a child, under conditions of stress and absence of stress. This thus involves creation of a regression model and would generate values such as there would be severe, moderate or less impact, which can be measured by some measurable outcome such as increase in weight, height, haemoglobin levels and so on. Thus, here the independent variables (effect of different factors), dependent variable (development as measured by some measurable outcome) and unknown parameter would be the information that would be obtained.
2. Describe two methods for selecting variables into a regression equation for your imaginary study, and the rationale for using those methods.
Two methods which can be used for selecting variables into the regression equation are Stepwise regression and MSE. Stepwise regression is in which variables are added in steps, their significance is determined and if its significance is below accepted level, it is removed from the model. This helps in screening out the unnecessary or less important variables. MSE is similar to stepwise regression with the difference that instead of probabilities, the change in root mean square error is used to screen out the variables. Both these techniques are helpful in screening out a large number of variables (stepwise regression, 2014) (stepwise regression, 2014).
List of References:
Ats.ucla.edu,. (2014). Stata Data Analysis Examples: Multivariate Regression Analysis. Retrieved 15 May 2014, from http://www.ats.ucla.edu/stat/stata/dae/mvreg.htm
law.uchicago.edu,. (2014). regression. Retrieved 15 May 2014, from http://www.law.uchicago.edu/files/files/20.Sykes_.Regression.pdf
Miller, N., Kirk, A., Kaiser, M., & Glos, L. (2013). The Relation Between Health Insurance and Health Care Disparities Among Adults With Disabilities. American Journal Of Public Health, (0), 1--9.
ncss.com,. (2014). Stepwise regression. Retrieved 16 May 2014, from http://www.ncss.com/wp-content/themes/ncss/pdf/Procedures/NCSS/Stepwise_Regression.pdf
Onlinecourses.science.psu.edu,. (2014). Stepwise regression | STAT 501 - Regression Methods. Retrieved 15 May 2014, from https://onlinecourses.science.psu.edu/stat501/node/88
Ratings
