03. Experimental Design

When we look at traditional experimentation, the way it is designed has a significant impact on the validity and reliability of the results. In some situations a specific experimental design may be better suited than in others. But in all instances we want to select the design that allows us to isolate the relationship between the IV and the DV by means of removing any extraneous variables.

1. Between-participants design (Also known as between-subjects or independent measures or matched participants design): In a between-participants design, each participant is only exposed to one level of each independent variable. For instance, if you're studying the effects of two different types of teaching methods on test scores, you might have one group of students learn using the first method and a different group learn using the second. This design is often used when exposure to more than one condition might influence the participant's response to the other conditions. The major disadvantage is that individual differences between groups can confound the results.

2. Within-participants design (Also known as within-subjects or repeated measures design): In a within-participants design, each participant is exposed to all levels of each independent variable. To continue the previous example, this would mean having the same group of students learn using both teaching methods, likely with some time in between to avoid immediate carryover effects. The advantage of this design is that it controls for individual differences, as each participant serves as their own control. The primary disadvantage is that it may introduce carryover or order effects, where the experience of one condition affects responses to the next.

3. Mixed design (Also known as split-plot or mixed-model design): A mixed design combines the two previous designs, using between-participants conditions for some independent variables and within-participants conditions for others. This can be useful when researchers want to assess the impact of multiple variables that can't all be effectively tested within the same group. For example, you might test the effects of different teaching methods (a within-participants variable) on students of different ages (a between-participants variable). The mixed design can control for individual differences and examine interactions between different types of variables, but it can also introduce more complexity and potential for confounding factors.