Differences between testing of within subjects design and between subjects design.

Hey! So I didn’t really know what to do my blog on this week but as we’ve been doing within and between subjects designs in lectures recently – and I barely understand them myself :p – it seemed appropriate. Enjoy. J

Experimental research designs are used to demonstrate a cause-and-effect relationship between two variables. For this several basic characteristics are required:

  • Manipulation on one variable to create set of two/more conditions
  • Measurement of second variable to obtain scores within treatment conditions
  • Comparison of scores between treatments
  • Control of other variables to prevent them becoming a confounding variable

Between subjects designs are also known as independent measures or unrelated measures. Between subjects design is a research design in which each of the different groups of scores are obtained from a separate group of participants.

A between subjects experimental design uses separate groups of individuals for each treatment condition being compared. For example, this design could be used to research the effectiveness of counselling or medication as a treatment for depression by testing the two conditions and determine which participants were less depressed after trials had taken place.

Advantages of a between subjects design:

  • Individual score is independent of other scores
  • Can be used for a wide variety of research questions/hypothesis – any experiment comparing two/more treatment conditions

Disadvantages of a between subject design:

  • Require large number of participants (as each participant gives only one score)
  • Individual differences can make the score less valid/reliable

Within subjects designs are also known as a repeated measures design. A within subjects design is a research design in which different groups of scores are obtained from the same groups of participants.

A within subjects experimental design is a design in which the same group of individuals participates in all of the different treatment conditions. For example, a within subjects experimental design can be used to measure reaction times of people.

Advantages of a within subjects design:

  • Can use a small sample of participants
  • Eliminates problems based on individual differences – can determine variance in data from individual differences

Disadvantages of a within subjects design:

  • Each participant goes through series of treatment conditions administered at different times – opportunities for time related factors (e.g. fatigue, illness) to influence participants score
  • Participant attrition – some participants may withdraw before study is complete

References

Gravetter & Forzano (2009) Research Methods for the Behavioural Sciences

 

About these ads

7 thoughts on “Differences between testing of within subjects design and between subjects design.

  1. Great analysis of the differences- but that said I think there was the potential to look at the application of the different styles in action. For example the theme of my own blog this week was research in health psychology (psud77.wordpress.com) so I’ll link it to that. In health psychology we are looking at abnormal people, that is people who have something that is affecting their physical health in some way. In much the same way as other disciplines, in application we are looking to create and deliver interventions that will improve health or coping, self-efficacy…. and the list continues. We can use both styles here.

    A between subjects experiment is particularly good at comparing the effects of an intervention on a test group and a control group. For example, blind drug trials allow us to test the effectiveness of a drug in treating physical symptoms. This is common practice in the medical sciences (Helgeland, 1980) in which a control group may be given a placebo and another group the active agent. This will allow for us to accurately test the effects of our intervention as in theory the two groups have as few individual differences as possible. However, a minor issue here is the ethical implications of with-holding a potential treatment from participants if we have the duty to protect them from harm….

    Conversely if we used a within-subjects design particularly in our example (Helgeland, 1980) then the effectiveness of intervention could never be accurately determined. If we take one group of participants and give them a placebo for six weeks and then give them the active agent for the following six weeks it would be impossible to determine which drug had caused any change to health.

    This said, in terms of therapeutic interventions such as CBT a within-subjects design can work at a more appropriate level. We can apply two forms of treatment (or just one matched against self-treatment) as long as we control for individual differences and time effects, and determine which is the best for our participant’s health (Propst, Ostrom, Watkins, Dean, & Mashburn, 1992).

    My suggestion is that your blog copes very well with theoretical statistics but maybe it should look at what happens when we use these statistical methods in applied psychology, as the advantages and disadvantages of each design are much more subjective to each study than the lists you have provided, but in essence it is explained well and perfectly sufficient for theoretical basis.

    I look forward to your next blog! : ).

  2. References for my previous comment:

    Helgeland, A. (1980). Treatment of mild hypertension: A five year controlled drug trial. The American Journal of Medicine, Volume 69, Issue 5 , Pages 725-732

    Comparative efficacy of religious and nonreligious cognitive-behavioral therapy for the treatment of clinical depression in religious individuals.
    Propst, L. Rebecca;Ostrom, Richard;Watkins, Philip;Dean, Terri;Mashburn, David
    Journal of Consulting and Clinical Psychology, Vol 60(1), Feb 1992, 94-103. doi: 10.1037/0022-006X.60.1.94

  3. A very good blog, and nicely laid out!
    I’d like to add that another advantage of between subjects design is that there are no order effects, which are changes in behaviour due to the order of the conditions, such as fatigue and boredom.
    And although repeated measures does suffer from order effects, this can be countered using counterbalancing! This basically means that the order of the conditions are mixed up for different sets of participants, for example, if there were two conditions, half the participants would do one condition first, and the other half would do that condition second. This way, any error that occurs is randomised.
    The only thing I think you could’ve added was a conclusion? If you compared the two and then decided which was better, but then the aim of the blog is to describe the differences, so a conclusion might not be necessary. :)

  4. Pingback: Homework for my TA – due 08/02/12 « scarlettrose23

  5. Pingback: Comments for Marking – Semester 2 Blog 1 | psud77

  6. Many a blog has focussed on this topic, and many a blog has forgotten another of the research designs. That of Mixed Factorial Design, although this design is merely a combination of the previous two (Keppel &Wickens, 2004), it does have some benefits. For example with two independent variables and two dependent variables we can see interactions between the variables (Gravetter & Forzano, 2009). For example if you wanted to test whether attractiveness and computer skills influenced whether someone got a job or not, and you wanted to test the difference between males and females, then this would be a mixed factorial design (Gravetter & Forzano, 2009). The results from the study would show you each variables influence on whether they got the job and their combined influence (Gravetter & Forzano, 2009; Keppel &Wickens, 2004).

    References
    Keppel, G., &Wickens, T. D. (2004). Design and analysis: a researcher’s handbook (4th ed.). Upper Saddle River, N.J.: Pearson/Prentice Hall.
    Gravetter, F. J., & Forzano, L. (2009). Research methods for the behavioral sciences (3rd ed.). Belmont, CA: Wadsworth Cengage Learning.

  7. Pingback: Homework for My TA – 10/02/12 « psucd3

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s