Experimental Designs

Repeated Measures

The same participants are used across all the conditions in a research. 

+ Eliminates any effect of participant variables as all participants take part in all conditions; therefore, it is more controlled. The results are less likely to be affected individual differences.

+ It requires fewer participants, making it easier to carry out research and possibly cheaper.

– There is a chance of demand characteristics affecting the study since all participants take part in all conditions, might find out the aim and behave in a certain way in order to please the experimenter. 

– Validity of studies can be reduced due to other effects that could affect the results; for example practice effect, when participants get better at a task when they complete a similar one for the second time. 

Independent Groups

 Different participants take part in each condition. If the independent variable is naturally occurring (e.g. age or gender), the experimenter must use this design. 

+ Participants are less likely to guess the aim as they only take part in one condition; therefore, the potential effects of demand characteristics are reduced.

+ There are no order effects since participants take part in one condition, therefore it makes the results more reliable and valid.

– There may be a problem with participant variables affecting the dependent variable rather than the independent variable; therefore, there is a higher chance of individual differences affecting the results.

– A larger sample is required for this type of design if the study wants to be generalisable; therefore, it makes the research more complicated and possible more expensive.

Matched Pairs

Different but similar participants are used in each condition. Participants are matched on variable the researcher wishes to control for the study. Once participants are matched, each member of the pair will take part in one condition. A good source of participants for this design are identical twins.

+ Participants variables are controlled and eliminated as pairs have been matched on them; therefore, the researcher can be more confident that the independent variable is affecting the dependent variable.

– It is very time consuming to find participants that match.

– There may be one/two participants variables that are overlooked, which could affect the dependent variable rather than the independent variable.