Experiment
 

Home Introduction Methods Samples Review

 

 

 

 

Case study
Questionnaire
Correlation
Experiment
This section describes the research method known as the experiment. There are several sections on this page:

 

Purpose

Variables: independent and dependent

Conditions: treatment and control

Advantage & disadvantages

 

Purpose

The experiment is the method to use to demonstrate a cause and effect relationship between two variables. When researchers want to know about the cause of a behavior or mental process, they should do an experiment. In an experiment, the researcher manipulates or changes the environment in a controlled way, then measures the effect of that manipulation.

For example, it is through experiments that we know that drinking alcohol causes slower reaction times. The experimenter can give a set amount of alcohol to a group of participants, then measure their reaction times. If their time slows down after drinking the alcohol, we know the alcohol caused that effect.

Back to top

 

Variables

Independent variable

In every experiment, the researcher makes some kind of manipulation to the participants, then measures their behavior afterward. The independent variable (I.V.) is the part of the experiment that is manipulated. This is the aspect of the experiment that the researcher does to the participants. We hold everything else among the participants constant, then make one change to them. That change is the independent variable.

Dependent variable

After the researcher manipulates the participants in some way (the independent variable), the next step is to measure the participants’ behavior. This behavior is the dependent variable (D.V.). It is the part of the experiment that we measure, to see if there has been an effect of the independent variable.

Let’s do some examples. For each hypothesis listed below, I will label the independent variable and the dependent variable, then briefly describe the method.

1. Hypothesis: Drinking alcohol slows reaction times.

I.V.: Alcohol

D.V.: Reaction times

The experimenter would randomly divide a group of participants into two groups. He or she would give one group a drink with alcohol and the other group a drink with no alcohol. After everybody drinks the beverage, the experimenter measures their reaction time to a task. For example, the time to press a button in response to a sound. If the participants who drank alcohol had a  slower average time than those who did not drink alcohol, we can say that the alcohol caused the slower reaction times.

           

2. Hypothesis: Taking a test in a hot room causes lower grades

I.V.: Room temperature

D.V.: Test scores

The experimenter would randomly divide a group of participants into two groups, and place them in two separate classrooms. He or she would increase the temperature in one room (for example to 90 degrees) and would keep the other room at normal temperature (for example, 72 degrees). The experimenter would give the same test to every person in both rooms, then grade the tests when the participants finish. If the people in the hot room had a lower average score than the people in the normal room, we can say that the hot room caused the lower scores.

Back to top

 

Conditions

Every experiment has to compare groups of participants. The researcher creates groups by manipulating the I.V. Every participant then does the D.V. task, and the researcher compares the outcomes across the groups.

Another name for these groups of participants is conditions. There are always at least two conditions in an experiment.

One condition is called the treatment condition (or the experimental condition). This group of participants is having something done to them. They are being manipulated by the experimenter. In other words, they are getting the treatment.

The second condition is called the control condition. This group of participants is the comparison group, so they are not having any manipulation done to them. They are the “normal” group, so the experimenter can compare the treatment group to them.

An important concept to understand about these conditions is the method used to divide participants into the groups. The experimenter must use random assignment. He or she must randomly put the participants into the groups. This ensures that the groups start out the same. That way, the only difference between the groups will be that one receives the treatment and the other does not.

The experimenter can randomly assign participants into groups by flipping a coin, drawing marked slips of paper out of a hat, or using a random number table. The researcher must use a mechanical, unbiased process—not his or her own decision about which participants should be in which group. We want the groups to start out the same (same number of participants, same proportions of psychological variables such as intelligence or motivation, etc.).

In the best case experiment, neither the participants nor the experimenter will know which participants are getting the treatment. This is known as a double-blind design.

Let’s walk through some examples. For the following hypotheses, I will list the I.V., D.V., treatment condition, and control condition; then briefly describe the method.

1. Hypothesis: Drinking alcohol slows reaction times.

I.V.: Alcohol

D.V.: Reaction times

Treatment condition: Alcohol in a beverage

Control condition: No alcohol in a beverage

The experimenter would randomly divide a group of participants into two groups by a coin flip. Participants with “heads” get 1 ounce of alcohol mixed with orange juice (the treatment). Participants with “tails” get plain orange juice (the control). However, the participants will not be told which beverage they are getting. After everybody drinks the beverage, the experimenter measures their reaction time to a task. For example, the time to press a button in response to a sound. If the participants who drank alcohol had a  slower average time than those who did not drink alcohol, we can say that the alcohol caused the slower reaction times.

           

2. Hypothesis: Taking a test in a hot room causes lower grades

I.V.: Room temperature

D.V.: Test scores

Treatment condition: Hot room

Control condition: Normal temperature room

The experimenter would randomly divide a group of participants into two groups by rolling a die. People who roll an even number (2, 4, or 6) are placed in one classroom, and people who roll an odd number (1, 3, or 5) are placed in another classroom. He or she would increase the temperature in one room (the treatment) and would keep the other room at normal temperature (the control). The experimenter would give the same test to every person in both rooms, then grade the tests when the participants finish. If the people in the hot room had a lower average score than the people in the normal room, we can say that the hot room caused the lower scores.

Back to top

 

Advantage

1. The primary advantage of the experiment is that we can determine the cause of something. None of the other research methods allows us to know something about the cause of a behavior. If an experiment is done properly, using random assignment and participants blind to their condition, then we know that any difference in their behavior (D.V.) was because of the difference manipulated by the experimenter (I.V.).

Back to top

 

Disadvantages

1. No method is perfect for all research questions. There are some situations for which the experiment cannot work. When we are interested in the effects of variables that cannot be manipulated, then we can’t do a true experiment. For example, if we wanted to know the effects of gender differences, or racial differences, we can’t study them with an experiment. That is because we can’t randomly assign a person to a gender or a race. A participant already has a gender and a race.

2. Another feature of experiments that can be frustrating is that we can only study little bits of a research question at a time. Because we have to hold everything in the environment constant, except the IV, we have to break the larger research question into smaller pieces. We can’t do one experiment to find out whether watching violent TV shows causes aggression. We would have to break it into violent cartoons, violent dramas, violent news broadcasts—and even further, we would break it down into one person being violent and a group being violent. There are many types of violence on TV, so we would have to do experiments on them piece by piece.

Back to top

Read an experiment sample.

 

Last updated 03/29/2002

© 2002 Elizabeth Lanthier, Ph.D.

email elanthier@nvcc.edu