Variables into a Regression Equation and Summary of Recent Research
Name of the Student
Name of the Institution
Example of a Quantitative Study using Regression Analysis
Childhood obesity is currently one of the major health concerns in the US. The high rate of obesity among children aged below 12 years has been attributed to the consumption of energy-dense foods (Maher et al., 2007). A study can be conducted to confirm whether the high rate of obesity among children who are below 12 years is associated with the consumption of energy-dense foods. The National Health and Nutrition Examination has been conducting surveys on the consumption behaviors and the causes and rates of obesity and other health-related problems. I would use the national representative data collected by the organization to determine whether there is significant association between the consumption of energy-dense foods and the high rate of obesity among children who are below 12 years. I would use quantitative data on dietary energy density and the predictors of obesity (measures of Body Mass Index and waist circumference). I would apply regression models to the variables to determine whether there is significant association between dietary energy density and the predictors of obesity.
Methods of Selecting Variables into the Regression Equation
During the process of selecting variables into the regression equation, I would use either the forward inclusion method or the stepwise-selection multiple regression. The forward inclusion method involves addition one independent variable at a time to the regression (Sawilowsky, 2007). I would select one predictor of obesity that I expect to have the highest correlation with the dietary energy density and put it first into the equation. Then, I would then conduct regression analysis with just the selected variable. I would check whether dietary energy density has significant association with the variable. I would then repeat the same process with the other variables, each at a time. Alternatively, I can use the stepwise-selection multiple method. The stepwise-selection multiple regression is almost similar to the forward inclusion method (Sawilowsky, 2007). In the forward inclusion method, a variable that is added to the regression equation remains there during the subsequent regressions. However, in the stepwise-selection multiple regression, a variable that has no significant contribution to the equation is not included in the subsequent calculations. The two methods would help to determine whether each of the three predictors of obesity has significance association with dietary energy density (Veney, Kros & Rosenthal, 2009).
A critical analysis and application of this study
Over the past one decade, the rate of obesity among children has been rising rapidly. Studies have linked the problem to lifestyle behaviors (Maher et al., 2007). In particular, studies have linked the high rate of obesity among the children to the consumption of energy-dense foods. In most cases, energy-dense foods, high in added fats, added sugars and refined grains are inexpensive and palatable. However, they are associated with poor diet quality and high level of energy intake. Such foods have been found to contribute to obesity (Maher et al., 2007). However, the link between the consumption of energy-dense foods and the development of obesity is not well understood. Whereas some studies have found significant association between the two, others have not. The findings of the study would be useful to add more evidence on the existing knowledge. If the study would show significant association between the independent and dependent variables, the findings would be useful to the public, health professionals and organizations that campaign against lifestyle behaviors that may lead to obesity.
Summary of Recent Research
Mongkolsomlit, S., Patumanond, J., Tawichasri, C., Komoltri, C. & Rawdaree, P. (2012). Meta
Regression of Risk Factors for Microalbuminuria in Type 2 Diabetes. Southeast Asian J
Trop Med Public health, 43(2), 445-466
Mongkolsomlit et al. (2012) conducted a study to determine the risk factors that are associated with microalbuminuria in patients with type 2 diabetes. Mongkolsomlit et al. (2012) analyzed 22 previous empirical studies related to the study topic and conducted a meta-regression analysis on the risk factors and microalbuminuria. As well, the researchers applied the random effect model to obtain pooled odd ration estimates. Mongkolsomlit et al., (2012) found four risk factors to have significant association with microalbuminuria, namely smoking, uncontrolled hypertension, poor glycemic control and central obesity. Mongkolsomlit et al. (2012) concluded that there is need for establishment of health promotion programs in order to mitigate the risk factors in patients with type 2 diabetes.
The null hypothesis and alternative hypotheses
The null hypothesis (H0): Smoking, age, gender, uncontrolled blood pressure, dylispidemia, uncontrolled hypertension, duration of diabetes, poor glycemic control, and central obesity (body mass index) do not have significant association with microalbuminuria in patients with type 2 diabetes.
Alternative hypothesis (H1): Smoking, age, gender, uncontrolled blood pressure, dylispidemia, uncontrolled hypertension, duration of diabetes, poor glycemic control, and central obesity (body mass index) have significant association with microalbuminuria in patients with type 2 diabetes.
Regression Analysis Results
Mongkolsomlit et al. (2012) applied meta-regression analysis to the identified risk factors and microalbuminuria. The analysis found that out of the risk factors identified in the previous studies, smoking, uncontrolled hypertension, poor glycemic control and central obesity were the only risk factors that were significantly associated with microalbuminuria in patients with type 2 diabetes. The regression results for smoking, uncontrolled hypertension, poor glycemic control and central obesity were; 1.37, 95% CI 0.95-1.98; OR 1.26, 95% CI 1.08-1.46; 0.79, 95% CI, 0.63-0.99; and 1.43, 95% CI, 1.14-1.80, respectively (Mongkolsomlit et al., 2012).
Type of Data Required and the Assumptions of the Test
The study needed to use quantitative data on the results of the previous studies examined. The researchers needed to extract data from previous cohort studies, case-control studies and analytical cross-sectional studies. The main assumption of the study was that the there was a high level of heterogeneity in the findings derived from previous studies. The researchers found significant level of heterogeneity inn the results of the previous studies.
Whether the Study is Statistically Significant
Mongkolsomlit et al. (2012) conducted tests for statistical significance of the level of heterogeneity in the studies from which the data was derived. The results indicated that the heterogeneity in the previous studies, with regard to smoking and central obesity, were not statistically significant (p< .05). The study found a statistically significant level of heterogeneity with regard in the results for uncontrolled hypertension and poor glycemic control (p< .05).
Statistical significance refers to the probability that an outcome has a high probability of being true and that it has not occurred due to chance (Veney et al., 2009).
The Possible Implications of the Study
The previous studies on the risk factors associated with microalbuminuria have found mixed and controversial results. The meta-regression analysis by Mongkolsomlit et al. (2012) highlighted the overall effect specific risk factors examined by previous researchers. The study harmonized the findings from previous studies and it identified the specific risk factors that have significant impact on the development of microalbuminuria in patients with type 2 diabetes. In addition, previous studies have shown that microalbuminuria causes cardiovascular and nephropathic complications in persons with type 2 diabetes. Mongkolsomlit et al. (2012) highlighted the importance of avoiding health behaviors that may lead to the development of microalbuminuria and the associated complications.
References
Jekel, J. F. (2007). Epidemiology, Biostatistics, and Preventive Medicine. London: Elsevier
Health Sciences
Maher, E. J., Li, G., Carter, L. & Johnson, B. D. (2007). Preschool Child Care Participation and
Obesity at the Start of Kindergarten. Pediatrics, 122 (2), 322 -330
Sawilowsky, S. S. (2007). Real Data Analysis. New York, NY: IAP
Veney, J. E., Kros, J. F. & Rosenthal, D. A. (2009). Statistics for Health Care Professionals:
Working With Excel. California: John Wiley & Sons
