Skip to main content icon/video/no-internet
icon backReturn to Entries

Nonexperimental Designs

Nonexperimental designs include research designs in which an experimenter simply either describes a group or examines relationships between preexisting groups. The members of the groups are not randomly assigned and an independent variable is not manipulated by the experimenter, thus, no conclusions about causal relationships between variables in the study can be drawn. Generally, little attempt is made to control for threats to internal validity in nonexperimental designs. Nonexperimental designs are used simply to answer questions about groups or about whether group differences exist. The conclusions drawn from nonexperimental research are primarily descriptive in nature. Any attempts to draw conclusions about causal relationships based on nonexperimental research are done so post hoc.

This entry begins by detailing the differences between nonexperimental and other research designs. Next, this entry discusses types of nonexperimental designs and the potential threats to internal validity that nonexperimental designs present. Last, this entry examines the benefits of using nonexperimental designs.

Differences among Experimental, Quasi-Experimental, and Nonexperimental Designs

The crucial differences between the three main categories of research design lie in the assignment of participants to groups and in the manipulation of an independent variable. In experimental designs, members are randomly assigned to groups and the experimenter manipulates the values of the independent variable so that causal relationships might be established or denied. In quasi-experimental and nonexperimental designs, the groups already exist. The experimenter cannot randomly assign the participants to groups because either the groups were already established before the experimenter began his or her research or the groups are being established by someone other than the researcher for a purpose other than the experiment. In quasi-experimental designs, the experimenter can still manipulate the value of the independent variable, even though the groups to be compared are already established. In nonexperimental designs, the groups already exist and the experimenter cannot or does not attempt to manipulate an independent variable. The experimenter is simply comparing the existing groups based on a variable that the researcher did not manipulate. The researcher simply compares what is already established. Because he or she cannot manipulate the independent variable, it is impossible to establish a causal relationship between the variables measured in a nonexperimental design.

A nonexperimental design might be used when an experimenter would like to know about the relationship between two variables, like the frequency of doctor visits for people who are obese compared with those who are of healthy weight or are underweight. Clearly, from both an ethical and logistical standpoint, an experimenter could not simply select three groups of people randomly from a population and make one of the groups obese, one of the groups healthy weight, and one of the groups underweight. The experimenter could, however, find obese, healthy weight, and underweight people and record the number of doctor visits the members of each of these groups have to look at the relationship between the variables of interest. This nonexperimental design might yield important conclusions even though a causal relationship could not clearly be established between the variables.

Types of Nonexperimental Designs

Although the researcher does not assign participants to groups in nonexperimental design, he or she can usually still determine what is measured and when it will be measured. So despite the lack of control in aspects of the experiment that are generally important to researchers, there are still ways in which the experimenter can control the data collection process to obtain interesting and useful data. Various authors classify nonexperimental designs in a variety of ways. In the subsequent section, six types of frequently used nonexperimental designs are discussed: comparative designs, causal-comparative designs (which are also referred to as differential or ex post facto designs), correlational designs, developmental designs, one-group pretest–posttest designs, and finally posttest only nonequivalent group designs.

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Read next

More like this

Sage Recommends

We found other relevant content for you on other Sage platforms.

close

Have you created a personal profile? Login or create a profile so that you can save clips, playlists and searches

close
close
close Signed In to profile You are signed in as:
close
Logged in Institution Institution