Page   5

Test Yourself

Chapter 4:Two-Way Analysis Questions

Using the Kenya eastern and coastal data set kceneast2.sav complete the following tasks to test your understanding of the material presented in this chapter. If you have any questions, use the analysis package for assistance. Answers to these questions can be obtained through the link at the end of the testing questions.

1)  Answer the following research question: Does latrine (access) have an association with prevalence of underweight (nutritional status) independent of education?


Use the compare means routine here and create a 2x2 dummy table. Use the following dummy table as a format for your results, making sure to include both the prevalence of underweight and the N for each group.


Prevalence Underweight

Educational Attainment

(none/incomplete primary)

(completed primary +)



Poor (bush/no facility)

Good (flush/pit toilet)



2)  Continuing to explore the research question posed in #1, is there an interaction between these two variables? If there is an interaction, is it significant?

Plot results from #2 to see if there is interaction and check both the size of the difference between groups and the significance (p-value).

Note: for questions 3 and 4 please use the Kenya eastern province data set (keast4j.sav).

3)  Using linear regression, explore the following research question: Is education significantly associated with weight for age Z score (nutritional status) independent of SES? Is there significant interaction between these two variables?

Use the regression table provided below to summarize your results, it makes it easier to see what is happening between the different models. Perform a linear regression and study the data to see if there is confounding and/or interaction between these two variables. Create the interaction variable and include it in the regression if it is significant, then calculate, by hand, the possible values (using the regression eqn.) and plot them

Regression Table: 
Does educational attainment have an effect on weight/age z score controlling for SES?
Dependent Variable: WAZ
Independent Variable

Model Number: coefficient (t,p)





Low education
Interaction (low edn * roof)
Adjusted R2


Compare your results with the Chapter Four Testing Questions answer key.