Developing a Research Hypothesis Continued
Hypothesis 3: There is a positive correlation between walking speed and mood.
What I’ve done now is provide a hypothesis that is much less specific. I’m saying that as walking speed increases, mood will become more positive. I’m also saying that a positive mood will be related to faster walking speed. In this case I’ve removed any predictions about the direction of causality and I’m only saying that these two variables will be related in a specific way (positively). It’s typically a good idea to think in terms of a predictor variable and an outcome variable because it adds specificity to the hypothesis. On the other hand, it’s perfectly plausible that the relationship between these two variables really does work both ways.
As your textbook points out we use the terms predictor and outcome variables mostly when we are talking about a correlational design. For our example this might mean measuring people’s mood and measuring how fast they walk. Another way to test a hypothesis is by using an experimental design. The key to doing that would be to assign one of my variables to be the predictor (independent) variable, and the other variable to be the outcome (dependent) variable. Instead of just measuring the predictor variable I want to create levels. (Huh?) For example, let’s say that I want to see if mood can cause a change in walking speed. I could put half of my participants into a negative mood and the other half of my participants into a positive mood (this is usually done by having people watch either very sad or very funny movies). My independent variable (IV) is mood. It has two levels; positive mood vs. negative mood. My dependent variable (DV) is walking speed. I can now compare how fast the “positive mood people” walked with how fast the “negative mood people” walked. If there is a difference in walking speed then it must have been caused by mood. We will be returning to these issues a lot during this course, but for now remember that Independent and Dependent variables go with experimental designs, and predictor and outcome variables go with correlational designs.
Go to Assignment 2 and fill out Part 1.