1/20
Looks like no tags are added yet.
learning outcomes
-understand the key benefits + limitations of complex within-subjects designs
-understand the concept of an interaction
-recognise when post-hoc tests are appropriate
-understand different ways of breaking down an interaction
what is complex anova?
-used when a study has two or more independent variable
also referred to as:
-factorial ANOVA.
-X-way ANOVA (where X is the number of IVs)
what is difference between simple + complex ANOVA?
simple ANOVA: one IV (one-way ANOVA)
complex ANOVA: ywo or more IVs
→ biggest difference is that complex ANOVA allows us to examine interactions between IVs
what interaction in complex ANOVA?
-interaction occurs when the effect of one IV on the DV is dependent on another IV
→ the effect of one variable changes depending on the level of another variable
why can't we simply run separate one-way ANOVAs for each IV?
-this ignores the potential interaction between the IVs.
-running separate ANOVAs treats the IVs as independent, whereas an interaction examines their dependence
what is a main effect in complex ANOVA?
-tells us whether there is a difference between conditions for one IV, averaging over all other IVs
-there is one main effect for each IV
what are the two types of effects examined in complex ANOVA?
Main Effects (one for each IV):
-these can be viewed almost like a simple ANOVA in their own right
-you use the same analysis logic that you would for a simple ANOVA
-if significant + IV has more than two levels, we need to do post-hoc tests to identify where the difference(s) is/are (ie which levels differ from which)
Interaction between the IVs:
-unique to complex ANOVA
-only tells you that the effect of these IVs on the DV depend on one another
-if significant, we need to do post-hoc tests to break-down the interaction + identify how exactly the IVs are interacting, regardless of how many levels
what is the key benefit of complex within-subjects designs?
increased power
→ as we add more IVs, we (mostly) don’t need to add more participants
→ we maintain N people per cell of the design (ie in every condition)
→ this is because all participants perform all conditions
→ 1 IV with 3 levels: each person contributes to 3 conditions
→ 3 IVs, each with 3 levels: each person contributes to 3*3*3 = 27 conditions
what is the key limitation of complex repeated-measures designs?
participants need to be tested many times, which can introduce order effects
The three main order effects are:
→ carry-over effects
→ practice effects
→ fatigue effects
what are carry-over effects?
-when one condition 'bleeds' into another
-example: participant completes a high-stress condition followed by a control condition, they may still feel stressed during the control condition, affecting their results
how can carry over effects be controlled? (1/2)
Trying to make the carry over effects equal between conditions
→ using counterbalancing and other forms of condition order randomization
→ if possible, randomizing conditions between trials or blocks, instead of doing one entire condition before another entire condition
how can carry over effects be controlled? (2/2)
Accounting for carry over effects in the statistical analysis
→ eg including this information as another variable in an ANOVA
→ this won’t “fix” the carry over effects per se, when used in combination with counterbalancing, can at least account for this variability (ie so it’s not “bad”)
(however, in some cases, carry over effects are what we’re interested in (eg priming research), so how you deal with them depends on your research question)
what are practice effects?
people get better at tasks with more practice
why can practice effects occur?
1) the cognitive processes that underpins tasks can change with practice:
→ if the same or very similar questions are repeated, the memory retrieval process may take over from the process used to solve the question
2) people may develop strategies to improve performance:
→ you can be quicker overall in a stroop task by defocusing to ignore the word
3) some tasks have a rule that can be learnt, which changes performance:
→ in the Iowa gambling task, people often learn that low risk low pay off decks are better than high risk high pay off decks
how can practice effects be controlled? (1/2)
Trying to make the practice effects equal between conditions
→ using counterbalancing + other forms of condition order randomization
→ if possible, randomizing conditions between trials or blocks, instead of doing one entire condition before another entire condition
→ if not possible, having different sessions for different conditions, which are far enough apart (eg a week) that practice effects have hopefully disappeared
how can practice effects be controlled? (2/2)
Giving people time to achieve “peak performance”
→ for example, practice trials or blocks (or sessions…)
(however, in some cases, practice effects are what we’re interested in (eg learning research), so how you deal with them depends on your research question)
what are fatigue effects?
-people can get worse at tasks if they spend too long on them
-they are particularly problematic in long studies with many trials or items
what problems can fatigue effects cause?
→ reduced performance.
→ participants becoming tired during the study
→ attrition, where participants fail to complete the study
how can fatigue effects be controlled (1/2)?
Trying to make the practice effects equal between conditions
→ using counterbalancing and other forms of condition order randomization
→ if possible, randomizing conditions between trials or blocks, instead of doing one entire condition before another entire condition
→ if not possible, having different sessions for different conditions
how can fatigue effects be controlled (2/2)?
Giving people adequate breaks
→ for example, breaks between blocks of trials
→ splitting longer experiments into multiple sessions to reduce fatigue
summary
-“complex” designs, in this context, refer to designs with 2 or more IVs
-complex within-subjects designs are useful for creating a powerful test
-however, they also create a potential for confounding order effects
-complex “ANOVAs” (Linear Mixed Models) have “main effects” + “interactions”
-main effects are the effect of an IV averaged over the other IVs
-interactions are where the effect of one IV on the DV is dependent on another IV
-alone, the interaction tells us only that a dependency exists. We have to break-down the interaction (usually with post-hoc tests) to understand what the dependency is