Between subjects design
Different groups of people in each experimental condition
Otherwise called – unrelated samples, independent-samples/measures, uncorrelated-samples/measures, between-participants, between-groups
Adv:
- There are no problems with order effects
- No need to duplicate and match materials
Dis:
- Individual differences experimental groups
- Need more participants
Within subjects design
Same group of participants performing in all experimental conditions
Otherwise called – related samples, dependent-samples/measures, correlated-samples/measures, repeated-measures, within-participants, within-groups
Adv:
- Elimination of permanent or chronic individual differences
- Fewer participants required
Dis:
- Order effects
- Carry-over effects
Within subjects advantages
- Experimental research
- Longitudinal (sequence-based) research
- increased efficiency
A repeated measures design is more powerful than an independent groups design
Power is enhanced in two ways: - Reduced variability due to individual differences
- More observations per participant
Experimental research -> within subjects advantage
- Each level of the independent variable is experimentally manipulated
- The sequence in which the levels of the independent variable are given can be varied
Longitudinal research -> within subjects advantage
- Each level of the independent variable normally represents a quantitative change
- The order of the sequence is fixed
- Examples: grade level of child, number of practice trials, number of treatment sessions
Reduced variability due to individual differences -> within subjects power advantage
As the participant is the same at each measurement, it follows that individual differences will be the same, thus error from this source is reduced
More observations per participant -> within subjects power advantage
More observations will help balance out random error
Increased efficiency -> within subjects advantage
Fewer participants are needed for the same level of power
Why use a within subjects design
- Many variables in psychology can only be studied in a repeated measures design
- Longitudinal or sequence-based research
- Practice/training
- Age-related changes
- Treatment effect over time
Order effects -> within subjects disadvantage
- can threaten internal validity
- The participant’s behaviour may change over time, independently of the level of the independent variable
Example:
Practice effects
Fatigue effects
Other psychological effects
Treatment carryover effects
Sensitisation effects
Practice effects
- Improved performance in later conditions due to being more familiar with task and having done task before
- Difficult to avoid; can use shorter tasks; give practice beforehand to get a post-practice level of performance
Fatigue effects
- Reduced performance in later conditions due to lowered motivation or reduced energy
- May avoid with rest periods; shorten testing time; increase interest in participant
Other psychological effects - order effects
- Others, sometimes difficult to predict, psychological consequences of doing a task again
E.g., reduced anxiety on later testing conditions due to familiarity with the task, experiment, etc. - Can be difficult to anticipate; might not be consistent across all participants
Treatment carryover effects
- The effects of the earlier condition will affect performance in later conditions
E.g., lingering effects of alcohol or other drugs in the system; increased anxiety due to previous condition; using strategies learned in previous conditions - Can test for lingering effects; give ample time between conditions; use a filler task
Sensitisation effects
- The participant’s behaviour changes because they learn (correctly or not) the hypothesis being tested in the experiment
- Can produce participant bias
- Will threaten both internal validity and construct validity
- Can be difficult to avoid but can use general strategies to avoid participant bias (e.g., placebo conditions); use more subtle variations of the conditions
Matched designs procedure
- > Measure participants on variables that are known to affect the DV
- > Match participants on these key variables (create equivalent pairs)
- > Randomly assign matched participants to experimental groups
- > Increases sensitivity to an effect by reducing variance due to group differences
Matched design
Matching in “sets”
- Set size is equal to the number of conditions
E.g., if we have 4 experimental conditions we need to recruit in sets of 4 matched on the desired attributes
Matched design analysis
- Analyse as if it were a repeated measures study
- Data from matched participants are organised as though the data from a matched pair came from a single participant
- The number of participants is equal to the actual number of participants divided by the number of conditions
6 participants and 2 conditions = 3 rows (1 row for each pair) - Paired-samples t-test
Matched design advantages
- Increases statistical power (i.e., sensitivity to group differences)
i. e. by trying to minimise differences within group error - No sequence effects
- Can improve internal and external validity
Matched design disadvantages
- Participants may guess the purpose of the experiment (damaging construct validity)
- If your matching variables are no good
- Matching is difficult, more difficult as:
- > The number of matching variables increases
- > Matching is done on continuous variables
- > The number of conditions increase
- Time and energy
- Participants without appropriate matches cannot be used
