Obesity in the United States

ABSTRACT

This paper discusses an article about Obesity and the statistical analysis that the authors had used to analyze the data information in their experiment. A brief description of the t-test and ANOVA is discussed in the paper. A discussion on how the author had done the computation is also provided in this text.

The article under study is about Obesity in America. The proponent of the paper had chosen the article of Larsen, et al (2004), which studies the five year transition period   obesity incidence between adololesence and adulthood. This does not only compare the transition period but it also compares the race of the subject. The article features two tests. First is the comparison between the body mass indexes during the five year period using the software Stata version 8.0. Statistical tests done was t-test. The subjects were grouped according to the ranges in age and there are two control groups Wave I and Wave II. The Wave I would be included in Wave II after five years and the Wave II would become Wave III after five years. Null hypothesis would be “Body Mass Index five years ago would be the same with the current Body Mass Index of the group observed”. The Alternative Hypothesis would be “Body Mass Index five years ago is less than the recent Body Mass Index of the group observed.”

Inferential statistics are simple computations that helps determine if there is really a significant difference between two samples. This particular computation is helpful for experiments such as this one and it could also be use as a comparison for a same product but a different brand. Significant differences would be shown if there is really a difference between an expensive brand and the ordinary brand. T-test can only be used to compare two groups however, it is not an effective tool for comparing three or more samples as per QMSS (2006)

The Independent Variables for the first test is age. The Dependent Variable in the test is the Body Mass Index. Level of Confidence was not indicated in the problem. It can be therefore assumed that the confidence level is 95%. The results of the statistical computations show that 22.1% from of Wave III young adults with a mean of 21.4 has a BMI of less than 30. As compared to that of Wave II with a mean of 16. The prevalence of overweight is shown to be 10.9% in 1996.Based on the computation, the Null hypothesis is rejected.

According to QMSS (2006), Analysis of Variance or ANOVA is used to test the hypothesis about the differences in mean values of some outcome between several groups at once. ANOVA can show then influence of an independent variable to a dependent variable. There are assumptions when doing the ANOVA. It is assumed that the standard deviations of the populations for all groups are equal. The samples are randomly selected from the population. Before starting an ANOVA test, the null and alternative hypothesis must be first formulated. Standard Deviation is then computed. After computing for the standard deviation, calculation of the variation between groups is done. This include the measure of between group variance which can be referred to as “between sum of squares”. Next step is to calculate the variation within groups, which is the “within sum of squares” to get the Within Mean Squares and then  the between sum of squares is divided by within sum of squares. The quotient would be the F statistic that would be compared to the Fvalue.

The second test done is ANOVA testing. This was used to compare the gender, the racial differences per wave. Null hypothesis for the gender is that Male Group BMI is equal to the Female Group. The Alternative Hypothesis is that the Make Group BMI is not equal to the Female Group BMI. The Null hypothesis for the race difference is that the BMI of the White, Black, Asian and Hispanic are all equal. The alternative hypothesis is that the BMI for the White, Black, Asian and Hispanic are all different. Confidence Interval is at 95%. Pairing was done according to the situation if the person had become obese or nonobese and the transition of age. Independent Variables in Anova is the ethnicity, gender and age. Dependent Variables is the BMI.

As per the computation, there is a significant difference in race ethnicity, race ethnicity and gender, and gender and age. However, there is no significant difference in the interaction of race, gender and age. This is evident for all pairings. As a conclusion, obesity incidence over the 5 year period amounts to 12.7%. Population that had remained obese is 9.4% and 1.6% had become non-obese.  There is a significant level of obesity incidence for non-hispanic black females which is 18.4% of the population relative to the white females. For the age group, prevalence of obesity  increased by 10.9% in Wave II and 22.1% in wave III. Extreme Obesity only happened in wave III which is 4.3%.

The research show a very thorough analysis  and detailed computation of the statistical tests. The use of a control in the ages had made it possible to show the transition of the Body Mass Index through the years.  If I am given a chance to ask the researcher, I would ask why he did not include the factors that had caused the subjects to be obese. There would be more significant findings if they had done this.

REFERENCES

QMSS e-Lessons. (2008). T-test. Retreived last February 29, 2008 from Columbia ISERP. Website: http://www.columbia.edu/ccnmtl/projects/qmss/t_about.html

QMSS e-Lessons. (2008). ANOVA. Retreived last February 29, 2008 from Columbia ISERP. Website http://www.columbia.edu/ccnmtl/projects/qmss/anova_about.html

QMSS e-Lessons. (2008). T-test –  Lesson One. Retreived last February 29, 2008 from Columbia ISERP. Website http://www.columbia.edu/ccnmtl/projects/qmss/t_one.html

Gordon-Larsen,P.  Adair L., Nelson M.,and  Popkin B. (2004) Five-year obesity incidence in the transition period between adolescence and adulthood: the National Longitudinal Study of Adolescent Health. Carolina Population Center.  Retrieved last February 29, 2008 from American Society for Clinical Nutrition.Website: <http://www.cpc.unc.edu/projects/debe/Papers/Gordon-Larsen%20Five%20Year%20obes%20incidence%20AJCN.pdf>