In: Statistics and Probability
Explain the difference between a matched pairs experimental design and a randomized ex- perimental design. How does this difference affect the statistical tests that we perform on the data gathered?
>>Randomized Experimental design is the study of assigning two samples into groups randomly such that the randomization takes care of the confounding effect( The effect due to other variables). Then one of the groups is given treatment to measure the effect of the treatment. It is assumed that the samples between the two groups are similar in every respect
Example: Consider a medicine company is testing a potential vaccine for coronavirus. the sample contains 500 men and 500 women of age 21-70 years. the dependent variable is the proportion of people having coronavirus
A Randomized design randomly assigns these 1000 samples into two groups. it is assumed that the randomization will control the extraneous variables such as age and gender.
Control group | Vaccine( Teatment) |
500 | 500 |
>>In matched pair experimental design the samples are paired such that the pairs are almost similar. And then the pairs are randomly assigned to the treatment or control group.
Consider the above example, If the 1000 people are groups based upon their ages, then let the first group is a pair of youngest men and 2nd pair is the pair of youngest women and so on until 500 pairs and then these pairs are randomly assigned to the treatment of control groups
control group | vaccine( Treatment) |
1 | 1 |
1 | 1 |
1 | 1 |
.....499th | .....499th |
1 | 1 |
Both are designs offer different results. the matched pair design control the effect of both age and gender whereas the randomized design assumed to control the effect of extraneous variables age and gender.
Hence, we can say that a matched pair design is more sensitive measure than randomized design and hence produce better results