Topic: Have BMI and food consumption trends changed over time?
Statistical Concepts Covered:
In this applet, you’ll learn that in some research designs we are interested in looking at how data changes over time or trends. It is important to remember that in these types of studies we often cannot infer causality since there may be other variables contributing to the observed relationship between variables.
Introduction:
Hunger is the way our bodies motivate us to start and stop eating. But, as discussed in the chapter, our relationship with food is complex. Due to the increased health care costs and decreased quality of life for those who have obesity-related illnesses, the topic of obesity has become extremely relevant. This applet will allow you to explore data sets from two different sources regarding calorie consumption and obesity.
The first data set comes from a study by Wansink and Payne (2009) on calorie density and number of servings in recipes published in the “Joy of Cooking” from 1936 to 2006. The researchers analyzed the number of calories and servings for 18 recipes that appeared in each of the seven editions published in the following years: 1936, 1946, 1951, 1963, 1975, 1997, and 2006. The goal of this descriptive study was to determine whether there was a significant increase in calorie count and serving sizes over a 70 year period.
The second data set comes from the World Health Organization’s (WHO) global database on body mass index (BMI) and calories consumed daily. Due to what the WHO refers to as “globesity,” the global epidemic of overweight and obesity, data has been collected on BMI and the number of calories consumed on a daily basis to assess the relationship of obesity and calories to diet-related chronic diseases. The focus of this applet will be primarily on data collected within the United States between 1961 and 2001.
1) In Wansink and Payne’s (2009) study, the total number of calories for the 18 recipes evaluated were averaged for each of the seven years. What appears to be the trend in the data over time? (Pick “Joy of Cooking” as the group, and “Calories in recipe” as the option to display.)
| A. |
| B. |
| C. |
| D. |
2) Compare the trend observed in question 1 with the change over time for the calories in the recipe, calories per serving, and number of servings. How do these trends compare? (Pick “Joy of Cooking” as the group and “Calories in Recipe” for the option. Compare this trend to when “Calories per recipe serving” and “Number of recipe servings” are chosen separately for the option.)
| A. |
| B. |
| C. |
| D. |
3) In the previous two questions we evaluated the trends over time for three measurements – total calories, calories per serving, and number of servings. Based on the observed trends and question 2, what conclusion can you come to regarding how calorie density and servings have changed over time?
| A. |
| B. |
| C. |
| D. |
4) Now let’s take a look at trends in BMI data collected by the WHO. In this data set we have information on the percentage of adults who have a normal BMI (18.5 – 24.99), overweight BMI (25.00 – 29.99), and obese BMI (≥ 30) from 1961 to 2001. What appears to be the trend in the data over time for each of these categories? (Pick “BMI” for the group.)
| A. |
| B. |
| C. |
| D. |
5) In addition to providing information on BMI percentages, the data from the WHO has the total number of calories consumed per day and for each of the following four food groups: cereals; fruits, vegetables, pulses (dry beans and peas, lentils, etc.), and nuts; meat, fish, milk, and eggs; and oils, fats, and sugars. What appears to be the trend in the data over time for each of these categories? (Pick “Calories by Food Type” for the group.)
| A. |
| B. |
| C. |
| D. |
6) We want to look at the relationship between BMI and total number of calories consumed per day. Does it seem that the percentages of individuals who are normal, overweight, or obese increase or decrease as the total number of calories increases? (Cycle between “BMI” and “Calories by Food Type” for the groups.)
| A. |
| B. |
| C. |
| D. |
7) How might the relationship between BMI and calories change as we consider different food groups? Compare the calories of the cereal and oils, fats, and sugars food groups to those who are overweight. Is there any significant finding here? (Cycle between “BMI” and “Calories by Food Type” for the groups.)
| A. |
| B. |
| C. |
| D. |
Statistical Lesson. The important thing to remember about relationships between variables is that we cannot conclude that one of the variables caused the change we see in the other variable. This is especially true when we do not conduct an experiment that included manipulation or control over the variables or environment being studied. We cannot say conclusively that the variables affected each other. This is what statisticians are referring to when they say “do not infer causality.” Just because there is a relationship or pattern between two variables, such as they both increase, both decrease, or one increases while the other decreases, this does not mean that one causes the other. Most likely there is another third, or what we call confounding, variable that could be contributing to the pattern we observe.
A popular example is the relationship between ice cream sales and violent acts. A researcher in the 1980s found that in New York City as ice cream sales went up, so did violent acts. Though we might jump to the conclusion that ice cream causes people to be more violent, and subsequently ban ice cream everywhere, we cannot nor should not do this. When collecting naturally occurring data without the use of a true experiment, we cannot infer that one of the variables cause the others. What might be contributing to this relationship? Well, people tend to buy more ice cream when it is hot outside or in the summer time, and this is also the time when more crime occurs. Could it be that heat or being outside more is what links these two variables together? Keep this in mind when interpreting relationship – do not infer causality, and always look for confounding variables.
8) Based on what you now know about relationships and causality, what can we infer about our finding between calories and BMI? In other words, what role do each of these variables play in the data presented here?
| A. |
| B. |
| C. |
| D. |
9) Using question 8, which of the following is a possible confounding variable that could explain the relationship between calories and BMI? Think about which of the following would be related to both variables.
| A. |
| B. |
| C. |
| D. |
10) Based on what you’ve learned from the data explored in this applet and the readings in the chapter on dieting and obesity, what would be the best conclusion we can reach regarding the relationship between these variables?
| A. |
| B. |
| C. |
| D. |