Hi, this tutorial covers a type of experimental design called the matched-pair design. So let's first start with the definition. A matched-pair design is a method of experimental design where participants/subjects are found in closely matched pairs, and one is put in the treatment group and one is put in the control group. And this is the case when there's only one treatment. This design eliminates many extraneous factors. Sometimes your control group will also be a placebo group.
So let's take a look at an experiment here. So 30 police trainees were seated in a darkened room facing a projector screen. 10 different license plates were projected on the screen, one at a time, for 5 seconds each, separated by 15 second intervals. After the last 15 second interval, the lights were turned on and police trainees were asked to write down as many of the 10 license plate numbers as possible in any order at all. After tabulating how many plates each trainee identified correctly, each trainee was paired with another trainee with the same or a very similar score.
One member of each pair, then, was asked to flip a coin. If the trainee flipped a head, he or she was placed in the placebo group. If the trainee flipped a tail, he or she was placed in the treatment group. The second member of the pair was placed in the opposite group. So if the first trainee in the pair got a tail, they would go in the treatment group, which means the other person in the pair would go into the placebo group. The trainees were not told which group they were in, so this was a blinded experiment. So really only the researcher knew which group each of the trainees were in.
Trainees that were designated as the placebo group were given police training that they were told would improve their memory. So they were just given some training on something else, and they were just told it would improve their memory but really it didn't have any effect on their memory. Now, trainees in the treatment group were given a memory training course. So this was something that was specifically designed to help improve their memory.
After two weeks, all the trainees were given another license plate test. The score of the trainee in the treatment group was compared to the score of the trainee in the placebo group within each pair. So really, then, within each pair, the score of the treatment group was compared to the score of the placebo group.
So the question is, is this a matched-pair design? And I would argue yes, this is a matched-pair design. So each of the 30 trainees was paired with another trainee with a similar memory level, and then the treatment, of course, would be the memory course, and then there was also the placebo that was involved. So the difference of the treatment score-- if the difference of the treatment score and the placebo score is positive, which would mean that the treatment score was bigger than the placebo score, it would seem that the memory course worked.
So if we see that positive difference-- so if we can see that within each pair, or at least most of the pairs, the treatment score is bigger than the placebo score, this matched-pairs design would allow us to conclude that the memory course worked. So this has been your tutorial on matched-pair design. Thanks for watching.
