Lecture note on quasi-experiment with an example for MBA students

Lecture note on quasi-experiment with an example for MBA students

As a part-time MBA student learning to do my 4-month dissertation project, I am interested in the research method of quasi-experiment. I feel that it is more feasible and less cumbersome for me to apply as compared with the conventional method of experiment. Please highlight 4 main ideas of the research method of quasi-experiment as well as 4 how-to advices. Try to illustrate your ideas and advices with regard to the MBA research theme of "to evaluate the business value of using influencer marketing to promote UK MBA degree programs in Hong Kong to gen y"

Yes — for a 4-month MBA dissertation, a quasi-experiment is often a practical choice because it lets you study cause-and-effect in a real business setting without full random assignment. For your topic, it fits well if you want to test whether influencer marketing changes application interest, brand attitude, or enquiry intention for UK MBA programs in Hong Kong. Quasi-experiments are specifically useful when randomization is not feasible, and they often use pretest-posttest, control-group, or time-series structures.pmc.ncbi.nlm.nih+2

4 main ideas

1.    No random assignment, but still comparison.
The key difference from a true experiment is that participants are not randomly assigned, but you still compare a treatment group and a comparison group. For your topic, one group could see influencer-led content about UK MBA programs, while another group sees standard university advertising.pmc.ncbi.nlm.nih+2

2.    Focus on change over time or between groups.
Quasi-experiments usually measure outcomes before and after an intervention, or compare groups exposed to different conditions. In your MBA theme, you could measure Gen Y respondents’ perceived program value, trust, or application intention before and after exposure to influencer marketing.pmc.ncbi.nlm.nih+1

3.    Works well in real-world settings.
This design is valuable when the field setting is messy, such as marketing campaigns, policy changes, or education promotion. That makes it suitable for evaluating a live or simulated influencer campaign for UK MBA recruitment in Hong Kong, where you cannot control every consumer exposure.pmc.ncbi.nlm.nih+1

4.    Causal inference is possible, but weaker than a true experiment.
Because groups are not randomized, you must think carefully about selection bias, confounding variables, and baseline differences. In your case, age, prior interest in overseas study, brand familiarity, and income may affect responses, so you need to control or match these factors as much as possible.pmc.ncbi.nlm.nih+2

4 how-to advices

1.    Use a simple pretest-posttest with comparison group.
This is usually the most feasible design for an MBA dissertation. Example: compare one Gen Y group exposed to influencer content about UK MBA programs and another group exposed to non-influencer content, then measure intention to apply before and after.pmc.ncbi.nlm.nih+1

2.    Choose one clear outcome variable.
Do not try to measure too many things. For your topic, pick one main dependent variable such as perceived business value of the MBA, brand credibility, or application intention, and use 1–2 secondary outcomes only if needed.pmc.ncbi.nlm.nih+1

3.    Match or control for key differences.
Since you cannot randomize perfectly, try to make the two groups as similar as possible on important variables such as gender, income, work experience, prior overseas study intention, and prior awareness of UK universities. This strengthens the credibility of your findings.dimewiki.worldbank+1

4.    Use a realistic influencer stimulus.
Your treatment should look like a believable campaign, such as short Instagram or LinkedIn posts from a business influencer discussing UK MBA benefits, career outcomes, and Hong Kong relevance. Then compare reactions against a standard university brochure or neutral post so the effect of influencer marketing is easier to isolate.pmc.ncbi.nlm.nih+1

Example for your topic

A workable dissertation design could be:

• Treatment group: Gen Y respondents in Hong Kong view an influencer-endorsed campaign for a UK MBA program.
• Control group: Similar respondents view a standard university promotional post.
• Pretest and posttest: Measure perceived value, trust, and application intention before and after exposure.
• Analysis: Compare change scores or use ANCOVA to adjust for baseline differences.

 

 

Please highlight 4 main ideas of ANCOVA in the context of our discussion.

In our context, ANCOVA has four main ideas:

4 main ideas of ANCOVA

1.    It compares groups after adjustment.
ANCOVA tests whether the influencer-marketing group and the non-influencer group differ on the outcome, but it does so after statistically adjusting for a baseline variable such as pretest intention or prior interest in UK MBA study.pmc.ncbi.nlm.nih+1

2.    It controls for starting differences.
This is especially useful in quasi-experiments where groups are not randomly assigned and may differ before the intervention starts. For your topic, one group may already be more interested in studying abroad, so ANCOVA helps reduce that imbalance.pmc.ncbi.nlm.nih+2

3.    It usually improves precision.
By using the pretest or another covariate, ANCOVA can reduce error variance and make it easier to detect the effect of the influencer campaign on outcomes such as perceived value or application intention.pmc.ncbi.nlm.nih+1

4.    It produces adjusted means, not just raw means.
The result is an “adjusted” comparison, which is useful for dissertation reporting because you can say whether influencer marketing has an effect after controlling for baseline attitudes or prior exposure. In your study, that might mean comparing the adjusted post-campaign intention scores between the two groups.theanalysisfactor+1

Example in your MBA topic

If your dependent variable is intention to apply for a UK MBA, and your covariate is pretest interest in overseas study, ANCOVA would ask: after accounting for initial interest, did the influencer-marketing group show a higher posttest intention than the control group? That is the kind of question ANCOVA is designed to answer.pmc.ncbi.nlm.nih+1

 

** A collection of lecture notes on the subject of research methods for MBA students, 2026 June

A collection of blog notes on using chatgpt for research purpose.