Researching Public Health Issues via Quantitative Data

Childhood obesity has become a public health issue as it has been unattended for long. Children are becoming obese which is a health risk. Some of the health risks associated with obesity is diabetes, high cholesterol and hypertension. Obesity can affect children and obesity in varied ways such as depression and low-self esteem. Children body composition vary by sex and age, therefore obesity in children is not measured using height but rather sex and age specific percentile for BMI. The factors that affect the prevalence in childhood obesity are socio-economic, socio-demographic factors, environmental factors, and certain medical conditions. Environmental behaviors are associated with individual behavior which may lead to high caloric intake, less physical activity and unhealthy diets (Sørensen, 2015). Other environmental factors include use of labor-saving technology, exposure to advertising content for unhealthy diets, and increased instances where one has to eat away from home. Socio-demographic factors include; age, sex and ethnicity. For instance, obesity is more prevalent in Hispanic men compared to the White, Asians and Blacks, on the other hand obesity is more prevalent in female who are black compared to the other ethnicities (Moreno-Black and Stockard, 2015). Obesity in 12 to 19 years old children is more prevalent compared to the younger populace. The last factor is socio-economic status. Childhood obesity is more prevalent in low income households compared to middle and high income households. The prevalence of childhood obesity is also high in households with a head adult who has not completed college education. 

Measures of central tendency are used to summarize data scores into a single number. Examples of measures of central tendency are mean, median and mode. Measures of dispersion are used to determine the degree of variation in a data set. The measures of dispersion are; range, variance and standard deviation. Measuring the data using for instance mean can be misleading because to data sets may be having the same mean but different therefore it is important to determine the variability in the data set for a more objective result. Range can be used to determine variability but it can only be informative in determining the maximum and minimum values in a data set. Range can be misleading in analyzing data set because it only utilizes the maximum and minimum values, therefore leaving the other values. The other disadvantage of using range is that it is very sensitive to outliers i.e. values that are “way off” the pattern in a data set. Interquartile range can be used to remedy the disadvantages presented by range such as sensitivity to outliers. Interquartile range is used to measure the difference between the 25th and 75th percentile, this provides an objective representation of the variability of data set at the center i.e. between the 1st and 3rd quarter of the data set. A large interquartile range signifies a more spread data set between 1st and 3rd quarters. Because the mean cannot be objective enough in measuring a data set, Standard Deviation (SD) can be used to determine the degree of deviation or spread about the mean. SD can be used to measure skewness in a data set which cannot be measured by measures of central tendency. SD is calculated by obtaining the square root of the sum of squared deviations from the mean divided by the population. If the data set is from a normal distribution then using Standard Deviation, then 95% observation would lie between mean ± 2 SD. 

Mahajan et al (2011) have provided statistics on childhood obesity in children between 6 to 12 years.  The table below provides a sample data of the obesity between the different ages 

Table 1

Age Population (n) Obesity (quantity) Obesity (%)
6 393 6 1.53
7 373 8 2.14
8 403 8 1.99
9 419 10 2.39
10 427 5 1.17
11 342 13 3.8
12 185 4 2.16
Total 2542 54 2.12%

 

Confidence level 

The mean for children with obesity is

54/7 = 7.714

Standard Deviation, Table 2

Obesity (o) X (mean) o-x (o-x)²
6 7.714 -1.714 2.94
8 7.714 0.286 0.082
8 7.714 0.286 0.082
10 7.714 2.286 5.23
5 7.714 -2.714 7.37
13 7.714 5.286 27.94
4 7.714 -3.714 13.80
57.43

 

Mean of squared difference = 57.43/7 = 8.20

SD =√8.20 = 2.86

Standard error = SD/√54

=2.86/7.35 = 0.39

Margin of error = 0.39 x 2 = 0.78

Confidence interval = 7.714 + 0.78 = 8.494 and 7.714 – 0.78 = 6.934

This is now a 95% confidence interval of 6.934 to 8.494

In calculation of the probability, the study will test the claim more that 3% of children have obesity. Therefore the sample proportion p= 0.212

Standard error of sample = p (1-p)/n = 0.212(1-0.212)/2542 = 0.000065

Test statistic = (0.212 – 0.3)/0.000065 = -1353

If the p-value is less than the significant level, the null hypothesis should be rejected. 

This means the sample results are 1353 standard errors below the claim. The confidence level is 0.05 therefore the claim should be rejected.

The potential gaps that are likely to be identified between inferences from example research and researching data will be in terms of objectivity. The sampling method used by Mahajan et al (2011) was random sampling which has a high probability of being subjective if the random selected are skewed on a particular direction e.g. ending up with a significant number of females and small number of male. In my own research I would use stratified sampling as a criterion for choosing the population. The population will be divided into different important groups such as sex, income, ethnicity and other socio-demographic factors. 

For proper comparison of the data between children with obesity and the rest of the children population, bar graphs can be used. The percentage values can be used to compare the prevalence of obesity between the different age groups. 

Figure 1

Figure 2

From figure 1 it can be identified that obesity is more prevalent in children who are 11 years and less prevalent in 10 year olds. There is no specific trend in the graph but there is a general indicator that obesity generally increases with age. The lack of a specific trend in the data may be due to bias in the data collection or this may be actual situation on the ground, but either would be difficult to verify. Figure 2 compares children who are obese and those who are not. 2.12 % of the children are obese in the population; this may look as a small number but when it is converted to actual situation in the ground, it would reflect how large this number is. According to the United States Census Bureau (2013) there were 41,844,000 youth aged 10-19 years, therefore 2.12% of this population would be 887,840 youth who are obese; this is actually a huge number. 

Leave a Reply