Limited access

Upgrade to access all content for this subject

Suppose a teacher wishes to determine if his students learn better through in-class instruction or computer program.

The teacher has only $10\text{ students}$, so he randomly divides them into two groups of $5$, one group of which learns through in-class instruction while the other group learns via computer program.

Why could the results be better trusted if the experiment were done in a matched pairs fashion?


The teacher would have $10\text{ data points}$ to work with instead of $5$.


Even if the subjects were randomly divided, such a small sample size makes it very likely that one group could have more intelligent students than the other simply by chance.


Students would feel more comfortable doing an experiment such as this with a pair.


A matched pairs design in this situation allows the teacher to use a control group.


A matched pairs design ensures the correct teaching method would be found in this situation.

Select an assignment template