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

Quasiexperimental Design

In a classic experiment, a researcher randomly assigns a treatment to some subjects or units of analysis while assigning other subjects to a control or comparison group. Random assignment allows for probabilistic equivalency between the groups, permitting the researcher to assume the two groups are the same except for treatment. This in turn allows for a clear inference about the causal effect of a treatment after a comparison of results between groups.

Natural experiments and quasiexperiments attribute some causal property to an event (e.g., the introduction of a law) or an intervention of nature (e.g., a natural disaster) and then seek to demonstrate differences before and after the treatment or between the treated and the untreated. The distinction between natural and quasiexperiments generally turns on whether the intervention was a natural or a non-natural one. Other than the source of intervention, they are theoretically indistinct. Either way, the analyst has no control over assignment. Because the researcher merely observes the application of treatment, he or she most often cannot be sure that it was randomly assigned. Consequently, quasiexperiments differ fundamentally from classical experiments. Although quasiexperiments appear to be like experiments, this lack of random assignment of subjects to conditions is a fundamental difference.

Conceptual Overview and Discussion

The proper use of a quasiexperiment depends on establishing the equivalence between units. This is often achieved through the use of control variables in a multivariate regression framework, or through the use of extensive case histories. Analogous to the between- and within-subject design, some quasiexperiments will compare differences across regions or subjects. For example, one could examine the effects of smoking on community health in one state that has a smoking ban and another that does not have such a ban, provided the states are otherwise similar on key variables. By comparing community health measures, some inference could be made about the effects of smoking bans. Alternately, an analyst could compare changes over time in the same region, for example, gauging the effects of a smoking ban in an isolated community before and after a ban is introduced, as was performed by Richard Sargent, Robert Shepard, and Stanton Glantz in their study of a smoking ban in an isolated Montana town.

The difficulty of inference from quasiexperiments is clear. Unlike in the laboratory, the analyst here has to contend with endless sources of interference, lack of control, and multiple plausible explanations of observed changes. In their 1968 study of a crackdown on speeding in Connecticut, Donald T. Campbell and H. Laurence Ross provide a definitive analysis of a quasiexperiment. The fundamental credo in drawing causal inference from a quasiex-periment, they contend, is that because we lack control and randomization it becomes much more difficult to establish cause and effect, especially in the face of competing explanations. Causal inference is possible, then, only because analysts can use observations before and after the event as well as auxiliary data to establish trends, minimize noise, pick up regressions to the mean, and rule out other explanations.

Application

A prototypical and well-conducted quasiexperiment is Steven Levitt and John Donahue's analysis of the relationship between the legalization of abortion and reductions in crime in the United States. It makes a strong causal claim because it meets Campbell and Ross's standard. First, it lays bare a mechanism. This means that it makes clear what the theoretical relationship between a cause and an effect should look like. Second, it establishes its causal plausibility; that is, the authors demonstrate that the relationship is possible or even likely. Third, they leverage auxiliary data. This means that they take measure of and control for other factors that may be explaining the outcome they are interested in understanding. They make use of cross-sectional variation; they look at differences in their cause and effect across units and show that their differences are what would be expected according to theory. Finally, they consider and dismiss other possible explanations. Not only do they demonstrate support for their own theory, but also they show why their data make other explanations less plausible. The puzzle they confront is significant: What explains the drop in violent and property crime in the United States in the 1990s? Their argument is that the widespread availability of abortion following Roe v. Wade reduced the number of unwanted children brought into the world while also allowing mothers to more optimally time the birth of their children so as to reduce early hardship on their offspring. As a result, the population had a smaller share of individuals likely to commit crime.

...

  • 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