Counterbalanced quasi-experimental design

Definition:
Counterbalanced quasi-experimental design: A type of quasi-experimental research design that establishes control by using a single group to test all treatment and control conditions. It is also called a repeated measures or within-subjects design.

Notes
In a counterbalanced quasi-experimental design, all participants receive each of the treatment and control conditions; (Suter, 2012, pg. 26)

Example
Here is an example: A researcher posed the question “What type of background noise results in the greatest reading comprehension—silence, dull hum of a motor, or nature sounds (ocean, river, birds, etc.)?” To answer this question, 100 students read a 500-word excerpt from a book on one of three relatively obscure topics: histories of Belize, Brunei, and Micronesia. All three excerpts were judged to be equally interesting and equally difficult.

All the subjects read the excerpts in all three background noise conditions. The researcher, though, was careful to avoid confounding noise conditions with excerpts and with the order in which they were delivered. To accomplish this, each noise condition was determined randomly to be first, second, or third for each subject. Also, for each subject, a random excerpt was selected for each noise condition, creating a type of double random assignment. For example, the first participant, Jacob, received the following randomized order of noise conditions: motor noise, nature sounds, silence. Furthermore, in the motor noise condition, he read about Brunei; in the nature sounds condition, Belize; in the silence condition, Micronesia. The second participant, Ashley, received the following randomly determined noises and topics: nature sounds while reading about Micronesia, silence while reading about Brunei, and motor noises while reading about Belize. This design is said to be counterbalanced, a term you will recall from Chapter 7, where the concept of control was discussed. The study was counterbalanced in the sense that any advantage motor noise might have by being first (for Jacob) is offset by its being second for, say, Tyler, and third for Ashley. Similarly, any advantage that might exist for Brunei in being first and being paired with motor noise, as in Jacob's case, is offset by its being second or third for other participants and being paired with silence and nature sounds.

This type of randomized counterbalancing establishes control by having each noise condition occur first, second, or third an equal number of times on average and having each noise condition preceded and followed by the other two conditions an equal number of times on average. The same principle holds for the three different excerpts. Over many participants, the random process equalizes order effects (the advantage of being first, the disadvantage of being last, etc.) and carryover effects (the advantage Micronesia might have by following Belize, the disadvantage Micronesia might have by following Brunei, etc.). You can probably understand these order and carryover effects better by imagining being in a market research study providing taste ratings after sampling chocolate Brands A, B, and C, in that order. The ratings of Brand C chocolate might suffer because it was tasted last and subjects' overstimulated taste buds begin to tire out (an order effect) or as a result of contamination from the two preceding samplings (a carryover effect).