Chapter 8 Flashcards
2 types of experimental studies
- between subjects
- within subjects
between subjects design
- AKA: independent groups design
- Uses random assignment
- 2 types: Post-test-only design; Pre-test-Post-test design
- good because demand characteristics aren’t as much of an issue as with within subject designs
post-test only design
- test only after manipulation
- Pros: saves time over pre-test-post-test; cuts down on demand characteristics
- Cons: not able to see which condition is causing the change, not able to be sure that participants across conditions were equivalent in DV to start with
pre-test-post-test design
- test before manipulation to ensure that 2 groups were equivalent, and test again after manipulation
- pros: good if there’s a small sample size, helps select appropriate participants, and good if there’s a risk of participants dropping out of the study (AKA: mortality)
- cons: time-consuming, and may alert participants to what you’re studying, therefore creating a demand characteristic
within subjects design
- 2 types - concurrent measures and repeated measures designs
- good because you can use weaker manipulation and a less sensitive DV, and this design helps eliminate random error
concurrent measures design
- Participants exposed to both conditions almost at the same time, then asked preference-related questions
- Associated with single attitudinal or behavioural preference as DV
- Ex. Participants exposed to both violent and non-violent video games, then ask questions about both
- Problems: severe restriction of research questions and DV’s, strong demand characteristics -> obvious what study is trying to test
- Advantages: very simple to administer, very quick to run
repeated measures design
- Participants are exposed to all experimental conditions
- DV measured after each exposure
- Ex. Play violent video games -> DV -> Play non-violent video games -> DV
- Problems: demand characteristics, carryover effects/order effects
- Advantages: people serve as their own control, require fewer participants
carryover/order effects
- Participant’s response in one condition is affected by having been in other conditions
- The effect of one condition carry over to others
- 3 types of effects: contrast effects, fatigue effects, practice effects
Contrast Effects
when the participants’ perceptions of variable 2 are affected by their perceptions of variable 1
Fatigue Effects
when the participant gets tired or bored and performance decreases
Practice Effects
when participants’ performance improves because they’ve done the task enough to get better at it
counterbalancing
- switching up order of conditions in order to avoid carryover effects
- Ex. Having someone play violent first, then non-violent; having someone else play non-violent first, then violent
sources of variability in DV
- random error: sources of variability caused by things other than the IV (ex. selection differences)
- the ratio: what proportion of variability in DV is attributable to IV vs. Error?
- Conceptual variable (IV) + Random error = Score on DV
- The more random the variability in DV, the harder it is to detect effect of IV
sources of variability in within-subjects design
- Conceptual variable (IV) + Random Error + Systematic Error = Score on DV
- Systematic error: systematic pattern identified within the random error
- We can measure and separate systematic error only with repeated measures designs
- In repeated measures we can explain 75% of variability in DV using IV, otherwise we can explain 60%
3 steps in any experimental design
1) Obtaining 2 equivalent groups of participants
2) Introducing the independent variable
3) Measure effect of the independent variable on the dependent variable
