Is within-subjects design better than between-subjects design?

Within-subject design is a research method used to establish the difference within participants in differing scenarios. This method uses the same participants in both conditions. This can be advantageous as fewer participants are needed to be used to create a statistically significant result. However, exposing participants to both conditions can cause several negative consequences. Fatigue or boredom can be a common problem and affect the outcome of the study. With an increase of boredom a participant may be more likely to make more mistakes in the latter study irrespective of the independent variable. For example, the participant may become slower in the second trial testing reaction speed due to feeling tired not due to the IV. Another negative effect of using participants in the same study is the impact of the practice effect. This explains that participants get better at a particular test for example, and will therefore get improve their score in next test. This would lead to the wrong conclusions being drawn about the effect of a variable. Another effect cause by using the same participants multiple times is the same study is the carry over effect. This explains that the effect of an earlier condition affects the participant in the subsequent one. One example of this is using a within-subjects design in a study to investigate the effect of alcohol on reaction speeds. In the first condition, participants could be subjected to 8ml of alcohol and later 16ml (or vice versa), in the later condition they have receive a total of 21ml therefore this may be investigated investigate and not the 8 or 16ml (Howitt and Crammer, 2009). The obvious alternative to this is a between-subjects design. This method researches the difference between groups of participants. Participants are randomly assigned into differing conditions. For example, Loftus (1979) had two separate conditions to evaluate the effects of anxiety on memory. Participants were assigned to either a condition of high anxiety, where they overheard a heated argument and observed somebody leaving the room holding a paper knife and blood on their shirt. The other participants were exposed to a situation of low anxiety, they head a discussion and man leaving with a pen and grease on his shirt. The comparison of these groups indicates that high anxiety situations would decrease correct recall. Between subject-design are not affected by the practice effect, carryover effect and fatigue however there are some disadvantages. There may be variance in participants in different group, in the Loftus study example those in the high anxiety group may just have poorer memory skills. However, this can be counterbalanced by the use of random allocation and matched pairs. In conclusion both designs are useful, if only one could be chosen I would argue within subjects design is more likely to truly investigate whether there is an independent variable affects the dependent variable due to continuity of participant involvement across conditions. Nevertheless, a research method which incorporates both designs would be the better choice. The experiment we took part in using SAFMEDS used both within subject and between subjects design. This balances out the disadvantages of both designs and the findings collected could show the improvement individually and collectively.


4 thoughts on “Is within-subjects design better than between-subjects design?

  1. racewinner says:

    I feel that choosing the right design to use depends mainly on the research currently being conducted. As you mentioned, boredom or the practice effect can affect the results in a within-subject design. However, I feel that this can be easily countered with a counterbalance design, where depending on the number of conditions you want to test, the participants can be split to start either condition first. (eg. Grp 1 starts with condition A to B whereas Grp 2 starts with condition B to A) This is especially true for medical testing where researchers can test the effects of a drug by alternating the conditions between a placebo drug or the real one to see if there is really any effect caused by the drug itself. I feel that it is good to use the within subject design for such research as everyone is different, at such, comparing two different groups of people who use the drug will not really give a very reliable effect. Moreover, I believe that a within-subject design also allow us to find correlations more easily through using a A-B-A-B design to see which factor, the independent or dependent actually affects the results. However, I feel that the between subjects design though being more advantageous as it will cause fatigue/the practice effect to occur to the participants, should only be used for general research such as the testing of alcohol or anxiety as you mentioned.

    Tracy Yang

  2. When choosing which research design to use, it ultimately comes down to the type of research being conducted. A within-subjects design can cause practice effects, as the same participants are used in both conditions. However, using this design ensures results are not down to individual differences. The practice effect can easily be dealt with by using a counterbalance design, where the same participants are used and half carry out the experiment by going from A to B and the other condition carry out the reverse process.
    I also feel that a research design which incorporates both designs would be extremely useful. Between-subjects design investigates the differences between people, so different people are assigned randomly to different conditions. By combining this with a within-subjects design the disadvantages would be counterbalanced, and findings could be broader.

  3. psue8c says:

    Before deciding on what experimental design to choose, between or within subject, a number of issues need to be addressed. It may be that the subject of study results in a certain design being used, but commonly, within subject design is used. The reasons for this being, fewer participants are needed as each individual who participates in the study is tested in both conditions of the study. Secondly, there is less variance due to participant disposition, as there are fewer participating in the study. Although range effect is a warning found when using within subjects, as there is exposure to different conditions within the study to the same participant, within subjects still carries great validity and is the more commonly used design of the two experimental designs.

    Greenwald, G.A. (1976). Within-subject designs: to use or not to use?. Psychological Bulletin. 83, 2, 314-320.

  4. psued6 says:

    Within-subjects do elicit many problems with order effects such as fatigue and boredom, using the participant for more than one condition. However, because the same person is used, participant varibales can be maintained throughout the experiment such as IQ. On the other hand, between-subjects design have issues with maintaining/controlling participant variables, as many characteristics between each of the participants may differentiate. This means that another variable that is not being controlled may be influencing the dependent variable, and so the researcher cannot be sure that there is a cause and effect between the IV and DV. However, a solution to this could be matched-pairs design where the participants are matched based upon certain similar characteristics such as a similar IQ, but still uses different participants so that there is not an issue with order and practice effects.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s