Emma’s research topic for the BD3-REAP program involves the investigation of the mean differences in Nutrition Quality Index (NQI) among different types of carbonated beverage drinkers by weight categories. My hypothesis was that individuals who do not consume carbonated beverages would have higher NQIs than individuals who consumes carbonated beverages. The study used the data from two non-consecutive days of dietary intake using 24-hour recalls of 36,877 subjects from “What We Eat In America (WWEIA)” dataset from the National Health and Nutrition Examination Survey (NHANES) from 2005 to 2012. Three types of carbonated drinkers were created: regular carbonated drinkers, diet-carbonated drinkers, and non-carbonated drinkers. For each individual, the intake quality score (IQS) of each of 13 nutrients was calculated, which included protein, Vitamin A, Vitamin C, calcium, iron, zinc, folate, thiamine, riboflavin, vitamin B-12, vitamin E, and potassium. To calculate an individual’s IQS of each nutrient intake, the average intake over 2 days was divided by the reference value of the corresponding nutrient.
Subsequently, the nutrient-based NQI was used to estimate the total diet quality of each individual. The individual’s NQI was calculated using the harmonic mean of the IQS values of the 13 nutrients in order to avoid any mathematical compensation of deficient intakes of one nutrient by exceeding intakes of other nutrients. To analyze the data, the IBM SPSS Version 24 was used to run the descriptive statistics and a two-population t- test (using independent samples). We divided individuals into different categories based on their BMIs and compared the NQI’s among the different types of beverage drinker in each weight category. Our results showed that underweight individuals had the highest NQI, followed by normal weight, overweight, obese and morbidly obese individuals, respectively.
The preliminary findings demonstrated that the mean differences in the NQI between non-carbonated drinkers and regular carbonated drinkers are positive and statistically significant across all weight categories. To compare regular carbonated drinkers with diet-carbonated drinkers, the mean differences in the NQI are negative and statistically significant in overweight, obese, and morbidly obese groups. Last, the mean differences in the NQI between non- carbonated drinkers and diet-carbonated drinkers are positive and statistically significant only in underweight and normal weight groups.
In summary, across all weight categories, non-carbonated drinkers have the highest average NQI followed by diet-carbonated drinkers and regular carbonated drinkers, respectively. The effect size was calculated to check for the validity of the significant results from the t-tests because of the large sample sizes in our dataset. Interestingly, the results showed that almost all the effects of the mean differences in the NQI are small to non-existent across all weight categories. Only the effect of the mean difference in the NQI between non- carbonated drinkers and diet-carbonated drinkers in underweight category is moderately significant. Continuing efforts will explore more on the association between these two variables via regression methods.
