Quasi-experimental design refers to a type of experimental design that uses pre-existing groups of people rather than random groups.
Because the groups of research participants already exist, they cannot be randomly assigned to a cohort. This makes inferring a causal relationship between the treatment and observed/criterion variable difficult.
Quasi-experimental designs are generally considered inferior to true experimental designs.
Limitations of Quasi-Experimental Design
Since participants cannot be randomly assigned to the grouping variable (male/female; high education/low education), the internal validity of the study is questionable.
Extraneous variables may exist that explain the results. For example, with quasi-experimental studies involving gender, there are numerous cultural and biological variables that distinguish males and females other than gender alone.
Each one of those variables may be able to explain the results without the need to refer to gender.
Quasi-Experimental Design Examples
1. Smartboard Apps and Math
A school has decided to supplement their math resources with smartboard applications. The math teachers research the apps available and then choose two apps for each grade level. Before deciding on which apps to purchase, the school contacts the seller and asks for permission to demo/test the apps before purchasing the licenses.
The study involves having different teachers use the apps with their classes. Since there are two math teachers at each grade level, each teacher will use one of the apps in their classroom for three months.
At the end of three months, all students will take the same math exams. Then the school can simply compare which app improved the students’ math scores the most.
The reason this is called a quasi-experiment is because the school did not randomly assign students to one app or the other. The students were already in pre-existing groups/classes.
Although it was impractical to randomly assign students to use one version or the other of the apps, it creates difficulty interpreting the results.
For instance, if students in teacher A’s class did better than the students in teacher B’s class, then can we really say the difference was due to the app? There may be other differences between the two teachers that account for the results. This poses a serious threat to the study’s internal validity.
2. Leadership Training
There is reason to believe that teaching entrepreneurs modern leadership techniques will improve their performance and shorten how long it takes for them to reach profitability. Team members will feel better appreciated and work harder, which should translate to increased productivity and innovation.
This hypothetical study took place in a third-world country in a mid-sized city. The researchers marketed the training throughout the city and received interest from 5 start-ups in the tech sector and 5 in the textile industry. The leaders of each company then attended six weeks of workshops on employee motivation, leadership styles, and effective team management.
At the end of one year, the researchers returned. They conducted a standard assessment of each start-up’s growth trajectory and administered various surveys to employees.
The results indicated that tech start-ups were further along in their growth paths than textile start-ups. The data also showed that tech work teams reported greater job satisfaction and company loyalty than textile work teams.
Although the results appear straightforward, because the researchers used a quasi-experimental design, they cannot say that the training caused the results.
The two groups may differ in ways that could explain the results. For instance, perhaps there is less growth potential in the textile industry in that city, or perhaps tech leaders are more progressive and willing to accept new leadership strategies.
3. Parenting Styles and Academic Performance
Psychologists are very interested in factors that affect children’s academic performance. Since parenting styles affect a variety of children’s social and emotional profiles, it stands to reason that it may affect academic performance as well. The four parenting styles under study are: authoritarian, authoritative, permissive, and neglectful/uninvolved.
To examine this possible relationship, researchers assessed the parenting style of 120 families with third graders in a large metropolitan city. Trained raters made two-hour home visits to conduct observations of parent/child interactions. That data was later compared with the children’s grades.
The results revealed that children raised in authoritative households had the highest grades of all the groups.
However, because the researchers were not able to randomly assign children to one of the four parenting styles, the internal validity is called into question.
There may be other explanations for the results other than parenting style. For instance, maybe parents that practice authoritative parenting also come from a higher SES demographic than the other parents.
Because they have higher income and education levels, they may put more emphasis on their child’s academic performance. Or, because they have greater financial resources, their children attend STEM camps, co-curricular and other extracurricular academic-orientated classes.
4. Government Reforms and Economic Impact
Government policies can have a tremendous impact on economic development. Making it easier for small businesses to open and reducing bank loans are examples of policies that can have immediate results. So, a third-world country decides to test policy reforms in two mid-sized cities. One city receives reforms directed at small businesses, while the other receives reforms directed at luring foreign investment.
The government was careful to choose two cities that were similar in terms of size and population demographics.
Over the next five years, economic growth data were collected at the end of each fiscal year. The measures consisted of housing sells, local GDP, and unemployment rates.
At the end of five years the results indicated that small business reforms had a much larger impact on economic growth than foreign investment. The city which received small business reforms saw an increase in housing sells and GDP, but a drop in unemployment. The other city saw stagnant sells and GDP, and a slight increase in unemployment.
On the surface, it appears that small business reform is the better way to go. However, a more careful analysis revealed that the economic improvement observed in the one city was actually the result of two multinational real estate firms entering the market. The two firms specialize in converting dilapidated warehouses into shopping centers and residential properties.
5. Gender and Meditation
Meditation can help relieve stress and reduce symptoms of depression and anxiety. It is a simple and easy to use technique that just about anyone can try. However, are the benefits real or is it just that people believe it can help? To find out, a team of counselors designed a study to put it to a test.
Since they believe that women are more likely to benefit than men, they recruit both males and females to be in their study.
Both groups were trained in meditation by a licensed professional. The training took place over three weekends. Participants were instructed to practice at home at least four times a week for the next three months and keep a journal each time they meditate.
At the end of the three months, physical and psychological health data were collected on all participants. For physical health, participants’ blood pressure was measured. For psychological health, participants filled out a happiness scale and the emotional tone of their diaries were examined.
The results showed that meditation worked better for women than men. Women had lower blood pressure, scored higher on the happiness scale, and wrote more positive statements in their diaries.
Unfortunately, the researchers noticed that men apparently did not actually practice meditation as much as they should. They had very few journal entries and in post-study interviews, a vast majority of men admitted that they only practiced meditation about half the time.
The lack of practice is an extraneous variable. Perhaps if men had adhered to the study instructions, their scores on the physical and psychological measures would have been higher than women’s measures.
The quasi-experiment is used when researchers want to study the effects of a variable/treatment on different groups of people. Groups can be defined based on gender, parenting style, SES demographics, or any number of other variables.
The problem is that when interpreting the results, even clear differences between the groups cannot be attributed to the treatment.
The groups may differ in ways other than the grouping variables. For example, leadership training in the study above may have improved the textile start-ups’ performance if the techniques had been applied at all. Similarly, men may have benefited from meditation as much as women, if they had just tried.
Baumrind, D. (1991). Parenting styles and adolescent development. In R. M. Lerner, A. C. Peterson, & J. Brooks-Gunn (Eds.), Encyclopedia of Adolescence (pp. 746–758). New York: Garland Publishing, Inc.
Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design & analysis issues in field settings. Boston, MA: Houghton Mifflin.
Matthew L. Maciejewski (2020) Quasi-experimental design. Biostatistics & Epidemiology, 4(1), 38-47. https://doi.org/10.1080/24709360.2018.1477468
Thyer, Bruce. (2012). Quasi-Experimental Research Designs. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195387384.001.0001