23rd March 2014 at 10:30 am #1137MajddiMember
I am doing an experiment research that I have two groups of students of the same size but have different tests. I am using quasi-experiment but confused between the controlled group and the experimental group. I have found different studies about this, some studies put the two groups in experimental study A and B, the other studies put them as one experiment, and the other is as control group.
Your response is highly appreciable23rd March 2014 at 3:01 pm #1143
I am not sure what the question is, but you have some leeway in defining what the treatment is and thus the treatment/experimental group, and what the control group is. I would say both groups together are part of one experimental study.
Why do you call it quasi-experiment? It seems that the design is under your control which would render it an experiment.24th March 2014 at 5:41 am #1142MajddiMember
Thanks Ingo for your answer
I call it quasi-experiment because my participants are almost equal and they are not randomly assigned to the research.25th March 2014 at 6:55 pm #1141
Okay, makes sense not to call it experiment. But what a “quasi-experiment” is is clearly defined: it is an observational design that produces the same results as an experiment when the former’s assumptions are met. The three designs that are subsumed under quasi-experiments are interrupted time-series designs, regression-discontinuity designs, and differences-in-differences designs. I found this text useful as an intro: Cook, Thomas D. and Vivian C. Wong (2008): Better Quasi-Experimental Practice. Alasuutari, Pertti, Leonard Bickman and Julia Brannen (ed.): The Sage Handbook of Social Research Methods. Thousand Oaks: Sage: 134-165.28th March 2014 at 11:28 am #1140Patricia RogersMember
You can find descriptions of other types of quasi-experiments on the BetterEvaluation site http://betterevaluation.org/plan/understandcauses/compare_results_to_counterfactual.
This lists a number of different ways that you can create a comparison group (which does not involve random assignment). It would be important to be clear which methods you have used to create the comparison group and to be confident it is really comparable to the treatment group.28th March 2014 at 12:11 pm #1139
This is a useful overview, but some designs such as ‘ordinary’ matching are not quasi-experimental designs. Matching in itself does not ensure that the results are similar to those of an experiment.29th March 2014 at 12:10 pm #1138Stephen GorardParticipant
This query does not make sense, so I am not going to offer an answer. Two bits of advice. Be much clearer in describing your design, perhaps presenting in standard design notation. This would help garner appropriate advice across this website in fact. Also, do not worry much about what your design is called. Get the design right for the research question(s) you are curious about. Terminology wars do not matter.
- You must be logged in to reply to this topic.