|
The matching of participants across the different conditions of your experiment, if done properly, can considerably reduce the confounding impact of individual differences. Indeed, sometimes participants in experiments are actually paired off and direct comparisons are made between the scores of the two participants in a pair who are exposed to the different experimental treatments. When this level of matching is achieved some authors argue that you can treat the scores from any pair of participants as if they were scores from the same participant and analyse the data using techniques for related samples. (There are considerable statistical advantages to this.) However, as a student you are extremely unlikely to be involved in an experiment that involves matching participants at the level necessary to enable you to do this. So, for your purposes, if you have different participants in your conditions (even when matched) you are probably best advised to analyse your data using the appropriate analysis for unrelated samples.
|
