Question

You are a researcher tasked with conducting an observational cross-sectional study that will examine the prevalence of food insecurity among Stanford students. Food security will be measured as a categorical variable with three levels: not insecure, insecure, very insecure. You will also gather data on age, students' income, gender, class rank, and weight.

a) Describe how you would collect data for this observational cross-sectional study. Explain how observational cross-sectional study differs from an experimental study?

b) Will your data collection strategy yield sample or population data? Explain the difference between sample and population data and why one or the other will be more suitable for your study?

C) What are the dependent and independent variables in this study?

Answer 1

a.

In order to collect data for the observational cross-sectional study, the below steps can be follows:

• Identify the different types of classes taught at Stanford University. Select one or more classes of each type to represent the class. In the usual sense, the students in a particular class would be of similar age.
• Divide the selected classes based on gender.
• For the students of each gender, identify the various income levels.
• Among the students of different income levels of a particular gender, consider the different weight categories, for example, extremely underweight, slightly underweight, normal weight, slightly overweight, overweight, obese, etc.
• From each of these subdivided groups, select a suitable number of students, and note their class ranks.
• Ask each selected student to choose their food insecurity level, as not insecure, insecure, and very insecure.

An observational cross-sectional study makes use of the already available diversity (based on age, income, gender, class rank, and weight) in the population (students of Stanford University), and collects data in such a way that each sub-population is well-represented. Then, it analyses the data to understand how the sub-populations compare with respect to each other, in terms of the question of interest (food insecurity).

On the other hand, an experimental study would actually assign the subjects to the various factors of interest (sub-populations), obtain the data from each factor, and analyse it to understand how the different factors (or their combinations) affect the question of interest.

In this case, it is practically impossible to conduct an experimental study, as the age, income, gender, class rank, and weight of the students are their intrinsic properties, and students cannot be “placed” into any of these factors to study their food insecurities.

b.

The data collection method described above yields a sample, and not a population.

A population study or a census includes all the relevant individual possessing a characteristic of interest into the study, whereas a sample study would include only a section of the population, which is preferably a good representative of the population on the whole.

Now, Stanford University is pretty large, with nearly 17,000 students. If data were to be collected from all these students to understand their food insecurities, it would be a very large study, that would take a very long time (maybe months) to complete. It would also provide data of very large dimensions, which is difficult to handle. Further, it would not be surprising to find some of the students refuse to take part in the study.

Thus, a sample survey is a much more feasible idea in this case.

c.

The dependent variable in the study is the variable of interest- the question that the researcher wants answered. In this case, the researcher wants to learn about the food insecurities of the students, classified as not insecure, insecure, and very insecure. Hence, the dependent variable in the study is the “level of food insecurity” of the students of Stanford University, which has 3 levels, as mentioned.

The independent variables in the study are the variables that potentially affect the dependent variable (food insecurity), and can be believed to be the reason behind the particular outcomes of the independent variable. Hence, the independent variables in the study are age, income, gender, class rank, and weight.