Download Applied Research Methods: Understanding Causality in Biking Levels - Prof. Susan L. Handy and more Exams Environmental Science in PDF only on Docsity! ESP 178 Applied Research Methods 1/29 Class Exercise: Causality Introduction It’s not so hard to come up with a list factors (independent variables) that might influence biking levels (dependent variable). Designing a study that can prove that any one of these factors actually causes more biking is much trickier. In fact, we can never prove causality definitively in social science and we can never understand causal relationships completely. But if we design studies carefully we can be reasonably confident in concluding that a causal relationship exists. Let’s start with a conceptual model of biking in which living in Davis is positively associated with biking and income is negatively associated with biking. Questions 1. There are two basic types of causality we can talk about: nomothetic causality and idiographic causality. For the Davis biking question, what would be a nomothetic causal explanation? What would be an idiographic causal explanation? 2. Finding an empirical association (such as a statistical correlation) between the independent variable and the dependent variable is necessary but not sufficient for establishing causality. Another important criterion in establishing causality is time order: did the variation in the independent variable come before the variation in the dependent variable, or more simply, did the cause come before the effect? a. First, how would you use a cross-section design to test the association between living in Davis and level of biking for individuals? Hint: How do you make “living in Davis” vary? b. Say you use a cross-sectional design and find a positive association between living in Davis and level of biking for individuals. Can you conclude that living in Davis causes more biking? What else could explain this relationship? c. What kind of longitudinal design could you use to establish the time order of living in Davis and level of biking? 3. Another important criterion for causality is “nonspuriousness,” that is, that the association between the independent variable and the dependent variable is not due to variation in some third variable extraneous to the conceptual model. a. Say you find a negative association between income and level of biking. Can you conclude that having a higher income causes less biking? What other variables that are correlated with both income and biking could explain this relationship? b. What are some of the ways you could account for this possibility in your research design?