We’ve been studying these two designs for a very long time. However, when it came to the exam, even though I knew everything, I still felt like I confused them. So, I thought it would be a good idea to write a blog on this topic.
Experimental design is a way of allocating participants to different groups in an experiment. In a research study there are usually two or more independent variables and the researcher must decide how they will allocate participants to these IVs. For example, if there is 10 participants, will all of them take part in all conditions or will they be split in half? The most commonly used types of designs are independent measures and repeated measures. With respect to these two designs ANOVA is very similar to t-tests. Independent-measures design (or between-subjects design) means that there is a separate sample for each of the treatments being compared. Repeated-measures design, means that the same sample is tested in all of the different treatments. The only big difference between an analysis of variance and a t-test in this case is that ANOVA can be used to evaluate the results of the study that has more than one factor. (Gravetter and Wallnau)
As in everything else in statistics, there are various pros and cons for both of these designs. Lets look at the independent measures first. The big advantage of this design is that it avoids order effects (e.g. practice, fatigue). Participants only take part in one condition only and therefore there is no real chance that they will get bored or wise to the requirements of the experiment. (http://www.simplypsychology.org/experimental-designs.html) The disadvantages of the between subjects design are: 1. More people are needed and therefore it is much more time-consuming; 2. No two people are the same and therefore participant variables such as age, sex, etc. may interfere.
Repeated measures uses the same participants in every treatment condition and therefore confounding variables such as participant variables are eliminated. Using the same people in every condition, means less participants needed. Of course, there are also a few problems associated with this design. Nothing is perfect in our world 😛 Order effects can become a problem here. Something that was a n advantage in an IMD becomes a downfall here. Order effects refer to the order of the conditions having an effect on participants’ performance. So, for example, performance in the second condition may become better because the participants already know what to do or it may become worse because they get tired. However, even though order effects interfere, there is a solution. Counterbalancing – a magic tool that helps researchers to combat the order effects. Counterbalancing means alternating the order of the conditions. This helps to make sure that practice effects are distributed equally across the conditions. (http://www.psychmet.com/id16.html)
So, as you can see there are pros and cons in each of the designs. However, I personally think that repeated measures is better in a way that there is fewer problems associated with it and there is solutions for some of them. Which one of the two designs do you, guys prefer? Or are you the type that goes for an alternative ( e.g. matched pairs)? See what you think 🙂
Statistics for the bahavioral sciences – Gravetter and Wallnau