You are here

Statistical Relationships Between Variables

17 November, 2015 - 16:48

Some research questions in psychology are about one variable. How accurate are children’s memories for being touched? How talkative are American college students? How common is it for people to be diagnosed with major depressive disorder? Answering such questions requires operationally defining the variable, measuring it for a sample, analyzing the results, and drawing conclusions about the population.

For a quantitative variable, this would typically involve computing the mean and standard deviation of the scores. For a categorical variable, it would typically involve computing the percentage of scores at each level of the variable.

However, research questions in psychology are more likely to be about statistical relationships between variables. There is a statisticarelationshibetween two variables when the average score on one differs systematically across the levels of the other. Studying statistical relationships is important because instead of telling us about behaviors and psychological characteristics in isolation, it tells us about the causes, consequences, development, and organization of those behaviors and characteristics.

There are two basic forms of statistical relationship: differences between groups and correlations between quantitative variables. Although both are consistent with the general definition of a statistical relationship—the average score on one variable differs across levels of the other—they are usually described and analyzed somewhat differently. For this reason it is important to distinguish them clearly.