Topic 6 Flashcards
Between subject design
A group
B group
No mix group
- must be used when IV is a subject variable (unless it’s the same person over time)
- subjects are naïve with respect to the hypothesis at first
- large samples needed
- more likely to have difference among groups other than the IV
- must make equivalent groups
Within subject design
Participant receives levels of A and B
- fewer recruits
- use tasks that tests all levels of variables and repeat
- eliminates equivalent groups issues
- experience could effect result when repeated (order effect)
- progressive effects is when performance changes steadily
- carryover effect is when some sequences produce different results than other sequences
Ways of creating equivalent groups
- Random assignment
- placing participants into different groups randomly
- all participants have equal chance of being in either group
- better for larger sample sizes
- blocked random assignment assigns subjects to each condition before repeating the conditions
- lab research - Matching
- participants are grouped together with subject variables (matching variables)
- separate matches to get equal groups
- matching variable must be correlated with the dependent variable
- have reasonable ways of determining scale of matching variable
- lots of options for matching variables
- smaller sample sizes
-
How to control order effects
Counterbalancing
Test once per condition
- complete counterbalancing( every possible sequence is used at least once) calculate with X!
- partial counterbalancing (using a subset of total number orders)
- Latin square
Testing more than once per condition
- reverse counterbalancing (presenting conditions in one order and flip flopping with the reverse order)
- block randomization ( every condition must occur once before repeating in different orders)
Error bars
Indicate the amount of variability within each condition
Cross sectional study
- between-subjects approach
- study different participants of different groups
-quicker - cohort effects ( people born at the same time but in different environments)
Longitudinal study
- within-subject approach
- study same participants as they change over time
- longer
- attrition threatens because it changes the already small sample size
- informed consent is ongoing
Cohort sequential design
- Combination of cross-sectional and longitudinal
- one group is tested longitudinally while additional cohort groups are tested and retested
Experimenter bias
Emotionally involved experimenters might behave in such a way to influence results
- might treat groups differently
- unawarely communicate expectations (animals too)
How to control experimenter bias
- mechanize procedures as much as possible (reduce human interaction)
- understand experimenter protocols
- double blind procedure (neither the experimenter or participants know what to expect)
Subject bias
Can occur based on what participants expect or what they believe their role is/should be in the experiment
Hawthorne effect
The knowledge that one is in an experiment/ has a role may lead to altered behaviours
What makes a good subject
- cooperative
- persevere through repetitive and boring tasks
Demand characteristics
Aspects of the study that reveal the hypothesis being tested, causing participants to behave differently
- affects if a between or within study should be done (introduce confounds)
Evaluation apprehension
Participants want to be evaluated positively so act as an ideal participant
