vs repeated measures designs
The distinction between a "Between groups" comparison and a ‘Repeated measures’ comparison is a very important one. Ray devotes two chapters to discussing the various design features of these two approaches. The decision to use a between groups design rather than a repeated measures design has major ramifications for how participants are selected and allocated, how data is entered into SPSS, and what analyses are conducted. In fact, the whole economics of the research design is affected.
A "Between groups research design" is defined as:
"a design that uses a separate sample of individuals for each treatment condition."
The main thing to look for is whether the data come from the same sample of people or whether different people provided the data. The definition assumes that exactly the same dependent variable (DV) is being measured in each treatment condition.
A "Repeated measures research design" is defined as:
"a design in which a single sample of subjects is used for each treatment condition."
This definition is again only meaningful if the two sets of scores represent measures or observations of exactly the same thing. Therefore exactly the same test needs to be given at both times or under both conditions. Sometimes this is easy with a task for which practice has no effect (perhaps reaction time, physiological responses) but obviously has problems on most psychological measures. Fatigue, boredom, and practise effects have to be overcome if one wants to use a repeated measures design. Ray discusses some advantages and disadvantages of these designs on pages 220-222.
Repeated measures designs come in three main forms. One is the situation in which each person is assessed at two (or more) different time points. For example, measuring sensation seeking at age 12, at age 20, and again at age 28. Another situation is when each person is assessed under different tasks. For example, the same group of people might be assessed for reaction time when the stimulus is a real word versus when the stimulus is a nonsense word, versus when the stimulus is a picture.
The third situation is discussed in more detail in a later section and in Howell (p. 187) and Ray (pp. 225-230). It refers to the special case of using two separate groups of individuals but matching pairs of individuals so that each pair is as similar as possible on some dimension. In this case we treat the design as if it were a repeated measures design even though separate groups of individuals are used.
A "Matched samples design" is defined as:
"a design in which each individual in one sample is matched with an individual in the other sample. The matching is done so that the two individuals are equivalent (or nearly equivalent) with respect to a specific variable that the researcher would like to control."