The word experiment is common in everyday use. Colloquially, “to experiment” means to try something out. A cook might say they experimented with a recipe by replacing the eggs with applesauce. A friend might say they experimented with a different driving route to the beach. However, as you learned in Chapter 3, in psychological science, the term experiment specifically means that the researchers manipulated at least one variable and measured another. Experiments can take place in a laboratory, a school, or just about anywhere a researcher can manipulate one variable and measure another.
A manipulated variable is one whose levels are managed by the researcher, such as when the researcher assigns participants to a particular level (value) of the variable. For example, Murrar and her team (2020) manipulated the poster condition by flipping a coin to determine whether a person would be exposed to the norms poster or the flu shot poster in the waiting room. (In other words, the participants did not get to choose which poster they would see.) Poster condition was a variable because it had more than one level (norms and flu shot), and it was a manipulated variable because the experimenter assigned each participant to a particular level. The Leonard team (2017) similarly manipulated which model the babies watched by flipping a coin ahead of time to determine which session participants were in. (Parents did not choose which type of model their babies would see.)
Measured variables take the form of records of behavior or attitudes, such as self-reports, behavioral observations, or physiological measures (see Chapter 5). After an experimental situation is set up, the researchers simply record what happens. In their first study, Murrar’s team measured people’s attitudes and perceptions about stereotypes, prejudice, and inclusion. After manipulating the waiting room poster, they watched and recorded—that is, they measured—how people answered questions about inclusive climate. The Leonard team manipulated the adult model’s effort behavior and then measured how many times each baby pressed the inert button.
In an experiment, the manipulated (causal) variable is the independent variable. The name comes from the fact that the researcher has some “independence” in assigning people to different levels of this variable. A study’s independent variable should not be confused with its levels, which are also referred to as conditions. The independent variable in the Leonard study was the adult model’s effort behavior, which had two levels (two conditions): effort and no effort.
The measured variable is the dependent variable, or outcome variable. How a participant acts on the measured variable depends on the level of the independent variable. Researchers have less control over the dependent variable; they manipulate the independent variable and then watch what happens to people’s self-reports, behaviors, or physiological responses. A dependent variable is not the same as its levels, either. The dependent variable in the Leonard study was the number of button presses (not “25 presses”).
Experiments must have at least one independent variable and one dependent variable, but they often have more than one dependent variable. For example, the baby persistence study’s main dependent variable was the number of button presses, but the researchers also measured how long each baby played with the toy during a 2-minute interval. When the dependent variables are measured on different scales (e.g., button presses and seconds), they are usually presented on separate graphs. (Experiments can also have more than one independent variable; Chapter 12 introduces this type of experiment.)
Here’s a way to tell the two kinds of variables apart. When researchers graph their results, the independent variable is almost always on the x-axis, and the dependent variable is almost always on the y-axis (see Figures 10.2 and 10.4 for examples). A mnemonic for remembering the two types of variables is that the independent variable comes first in time (and the letter I looks like the number 1), and the dependent variable is measured afterward (or second).
When researchers are manipulating an independent variable, they need to make sure they are varying only one thing at a time—the potential causal force or proposed “active ingredient” (e.g., only the type of poster or only the amount of effort the adult model displays). Therefore, besides the independent variable, researchers also manage potential third variables (or nuisance variables) in their studies by holding all other factors constant between the levels of the independent variable. For example, Murrar’s team manipulated the poster people saw, but they held constant several other potential variables: People in both groups sat in the same waiting room, performed the same filler memory task, and had the same experimenter. Any variable that an experimenter holds constant on purpose is called a control variable.
In the Leonard et al. study (2017), one control variable was the toys the model was using: In both conditions, she played with the same frog-in-the-box and carabiner toys. The model also used the same cheerful, enthusiastic voice. The researchers kept constant how long the model demonstrated each toy (30 seconds each), the gender of the model (always female), the chair the infant sat in, the cubicle where the experiment took place, and so on.
Control variables are not really variables at all because they do not vary; experimenters keep the levels the same for all participants. Clearly, control variables are essential in experiments. They allow researchers to eliminate alternative explanations for results. Control variables are therefore important for establishing internal validity.
TRY THIS!1. An experiment has at least one manipulated variable and one measured variable, and a correlational study has at least two measured variables; see p. 277.
2. See pp. 277–279.