Final Exam Flashcards
Threats to validity in repeated measures design
Practice effects
order effects
fatigue effects
Power in repeated measures design
power will increase because the same participants across conditions
it reduces variability due to individual differences
less random error due to more observations per participant
Repeasted measures design aka
related samples
dependent smaples
correlated samples
within participants
within groups
advantages
- Elimination of permanent or chronic individual differences
- Fewer participants required
disadvantages
- Order effects
- Carry over effects
Between groups aka
unrelated samples
indepenant samples
uncorrelated samples
between participants
between subjects
Advantages
- There is no problems with order effects
- No need to duplicate and match materials
disadvantages
- Individual differences experimental groups
- Need more participants
Counterbalancing for order effects
Each condition in our experiment follows an is preceded by every other condition is equal number of times
Each condition variant occurs an equal amount of times
However adding more conditions becomes extremely complicated quickly
Number of order is calculated by
k X (k - 1) X (k - 2) X … 1
(3X2X1=6)
(4X3X2X1=24)
Latin squares (incomplete counterbalancing)
Sequences might not occur however each condition sequence only occurs once
However the 3 conditions is then just looped
Randomising the order of conditions
excel in B put conditions
In 1 put =random
Select both A and B and then sort smallest to largest and each time it will give you are new random order
Matched-subjects design
A compromise between groups and repeated measures design
Procedure
- Measure participant 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 the sensitivity to an effect by reducing variance due to the group differences
Matching in sets
- Set size is equal to the number of conditions
Matched Designs analysis
Analysed the same way as a repeated measures
- Data from matches is organised the same as data 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 designs advantages
Increases statistical power
No sequence effects
Can improve internal and external validity
Matched designs disadvantages
Participants may guess the purpose of the experiment (damaging construct validity)
If your matching variables are no good
Matching is difficult
- The number of matching variables increases
- Matching iodine on continuous variables
- The number of conditions increase
Increased time and energy to use a match design
Participants without appropriate matches cannot be used
Assumptions of RM ANOVA
Accuracy for the F-test used in the one way ANOVA depends on the assumption that scores in different conditions are independent
- Repeated measures designs violates this assumption
- The normal F-test will be wrong and biased
- Repeated measures use Sphericity
Assumes that the relationship between pairs of experimental conditions is similar, or, that the level of dependence between experimental conditions is roughly equal - Compound symmetry
Occurs when both the variances across conditions are equal and when the covariances between pairs of conditions are equal
Compound symmetry assumes that the variation within experimental conditions is fairly similar and that no 2 conditions are any more dependent or related than any other 2
Compound symmetry itself is not a condition of the one-way repeated measures ANOVA
Sphericity (less restrictive form of compound symmetry) is a necessary condition
Spss tests the severity of departures from sphericity using Mauchly’s test
Tests the hypothesis that the variances of the differences between conditions are equal
If test stats is significant then the assumption has been violated
If not significant then the assumption is cleared
Mauchly’s test
Affected by sample size
- Big samples then only small deviations can produce a significant test
- Very small samples, sometimes even large deviations from sphericity can produce a nonsignificant test
- When assumption has been violated there is a significant loss in the power of the F-test and inaccuracies in the F-ratio that is produced in the output
- If data violate the assumption there are 3 alternatives
Greenhouse-Geisser Correction
- Varies between 1/k-1 and 1 (where k is the number of RM Conditions)
- The closer the correction is to 1 the more homogeneous the variances of differences and the closer the data are to being spherical
- This correction is over conservative
Huynh-Feldt Correction
- Less conservative correction than the greenhouse-geisser correction
- But it can overestimate sphericity
Factorial ANOVA
Involves the manipulation (experiment) or measurement (quasi-experiment) of 2 or more independent variables
Investigate the separate effects of each independent variable
Aka - main effects
Investigate the combined effects of all independent variables
aka - interactions
How does the effect of an IV depend on the effect of the other
There is an F-statistic for every main effect and interaction
Factorial Designs
Between-groups
- At least 2 factors
- All factors are manipulated between subjects
- Each participant provides data in only 1 cell of the analysis
Within subjects
- At least 2 factors
- All factors are manipulated within subjects
- Each participant provides data in all cells of the analysis
Mixed
- At least 2 factors
- At least 1 is manipulated between subject and 1 is manipulated within subjects
What are factors
Independent variables
One-way ANOVA = 1 IV
1-factor ANOVA = 1 IV
2-factor ANOVA = 2 IV’s
3-factor ANOVA = 3 IV’s
Include as many factors as you like but the interpretation becomes more complicated
What are levels
The number of experimental conditions for each IV
IV = age group (children and adolescents) = 2 levels
IV = expertise (novice, intermediate, experienced) = 3 levels
What are cells
The number of individual treatment conditions
Calculated by multiplying the number of levels of each IV
2 x 3 x 5 design
