Talking about experiments
Observational studies passively collect data. We observe, record, or measure, but we don’t interfere. Experiments actively produce. Experimenters intentionally intervene by imposing some treatment in order to see what happens. All experiments and many observational studies are interested in the effect that one variable has on another variable. Here is the vocabulary we use to distinguish the variable that acts from the variable that is acted upon.
94
Vocabulary
A response variable is a variable that measures an outcome or result of a study.
An explanatory variable is a variable that we think explains or causes changes in the response variable.
The individuals studied in an experiment are often called subjects.
A treatment is any specific experimental condition applied to the subjects. If an experiment has several explanatory variables, a treatment is a combination of specific values of these variables.
EXAMPLE 1 Learning on the Web
An optimistic account of learning online reports a study at Nova Southeastern University, Fort Lauderdale, Florida. The authors of the study claim that students taking undergraduate courses online were “equal in learning” to students taking the same courses in class. Replacing college classes with websites saves colleges money, so this study seems to suggest we should all move online.
College students are the subjects in this study. The explanatory variable considered in the study is the setting for learning (in class or online). The response variable is a student’s score on a test at the end of the course. Other variables were also measured in the study, including the score on a test on the course material before the courses started. Although this was not used as an explanatory variable in the study, prior knowledge of the course material might affect the response, and the authors wished to make sure this was not the case.
EXAMPLE 2 The effects of a sexual assault resistance program
Young women attending universities may be at risk of being sexually assaulted, primarily by male acquaintances. In an attempt to develop an effective strategy to reduce this risk, three universities in Canada investigated the effectiveness of a sexual assault resistance program. The program consists of four 3-hour units in which information is provided and skills are taught and practiced, with the goal of being able to assess risk from acquaintances, overcome emotional barriers in acknowledging danger, and engage in effective verbal and physical self-defense.
95
First-year female students were randomly assigned to the program or to a session providing access to brochures on sexual assault (as was common university practice). The result was that the sexual assault resistance program significantly reduced rapes as reported during one year of follow-up.
The Canadian study is an experiment in which the subjects are the 893 first-year female students. The experiment compares two treatments. The explanatory variable is the treatment a student received. Several response variables were measured. The primary one was rape as reported by participants after a one-year follow-up period.
You will often see explanatory variables called independent variables and response variables called dependent variables. The idea is that the response variables depend on the explanatory variables. We avoid using these older terms, partly because “independent” has other and very different meanings in statistics.