Independent Groups Design
An independent groups design is where the Independent Variable is operationalised by having different groups of participants take part in the experiment in each condition.
For example, in an experiment on the effects of drinking and driving the experimental group may drive around the course after having a specified number of drinks, whereas the control group might drive around the course having not consumed any alcohol. Independent groups is an example of an unrelated design, the values in one condition are not related to values in the other condition.
A weakness of the independent groups design is that there may be important differences between the groups, other than those deliberately manipulated by the experimenter. These differences could consist of uncontrolled participant variables.
Participant variables is a collective term for any extraneous variables that are due to the individual differences between participants; for example, personality, cognitive ability, sex, attention span, etc. Participant variables can be reduced in an independent groups design by having a large sample and randomly allocating participants to the experimental and control groups.
In a Matched Pairs design, participant variables are controlled by matching pairs of participants on particular variables. E.g., If participants are matched for IQ one person with What are participant variables? an IQ of 120 could be paired with another person with the same IQ. One member of the pair is allocated to the experimental group and the other is allocated to the control group. In the statistical analysis that follows the scores are treated as pairs. Matched pairs is an example of a related design, as pairs of scores are related to each other when the data is analysed.
A problem with the matched pairs design is that it is very time consuming to consider all of the participant variables that could potential confound the results. It may also be very difficult to find an appropriate combination of participants to match on all important characteristics.
This is where one group of participants takes part in the experiment in all conditions of the IV. For example, the participants all drive around the course twice, once after consuming the specified amount of alcohol and once having not consumed any alcohol. Repeated Measures is an example of a related design — a value in one condition is related to a value in the other condition. Repeated measures designs ensure that all participant variables are the same in each condition; however one problem with this design is order effects. After completing the task once, the participant may do better or worse on the second attempt. Examples of order effects include: learning, boredom and fatigue.
Order effects can be controlled by counterbalancing. This is where, although all participants take part in both conditions, half the participants do the experimental condition first, then the control condition, and half of the participants do it the other way round. The learning effects are, thus, equal for each condition and do not affect the results of the experiment.
Useful teaching resource: Introduction to Designing Experiments - Ideal for research methods and statistics