Module 9- Complex Experimental Designs Flashcards
1
Q
Interaction Effect
A
- 2 or more IV’s
- work together to create a bigger change in the DV than either one could do alone
- study interaction effects by factorial designs
- Impact of one IV depends on the level of the other IV
ex. if 3x increase in icy roads and 2x increase in speeding, and observe 15x accident rate; tells this is an interaction
2
Q
Factor
A
- another way to say IV
3
Q
Problem with Simple experimental designs
A
- will tell about the impact of the IV on the Dv separately for each IV
- but does not say the interaction between the IV’s to produce a bigger DV
4
Q
Factorial Designs
A
- studies the combined effects of the IVs on a DV; interaction effects
- also studies each IV separately; main effect
- important bc many IV’s work together to produce human behaviour
- 2 simple between subjects designs combined into one
5
Q
Main Effects
A
- how each IV impacts the DV on its own
6
Q
each matrix cell represents
A
- the score of the DV formed by the intersection of 2 IV’s
7
Q
No interaction
A
- means the combination of the 2 IV:s did not produce a larger effect than either one could do alone
- observe a additive effect; what we would expect when combining the IVs
- ex. 3x increase in icy roads, 2x increase in speeding results in 6x increase accident rate
- Level of one IV does not depend on the level of the other IV
8
Q
when have 2 IVs, how many treatment conditions?
A
4
9
Q
To get the main effects
A
- calculate the column and row means
10
Q
to get the interaction
A
- examine the cell means
11
Q
Factorial designs test which 3 null hypotheses
A
- Null hyp related to factor A
- Null hyp related to factor B
- Null hyp related to the interaction of factors A and B
12
Q
Test each null hypothesis using
A
ANOVA/ analysis of variance
13
Q
In ANOVA reject whatever null hyp when
A
p value< alpha
14
Q
If ANOVA indicates a significant interaction
A
- *interpret the interaction first
- interaction is going to qualify the main effects
- cannot interpret the main effects bc the impact of the IV is going to be different at each level of the other IV (IVs are dependent on eachother)
- main effects aren’t as informative in this instance
15
Q
Mixed Design
A
- combination of bw subjects design and within subjects design
- one IV is between subjects
- other IV is within subjects
- same layout as a factorial design; testing all 3 null hypotheses
16
Q
Matched subjects design
A
- form of participant assignment to equate groups on specific confounds
- pre identify major confounds and match participants
- then assign a participant from each set to a group
- equates the groups on various confounds
- want the groups to be equal before the introduction of the IV
17
Q
Main characteristics of matched design
A
- each participant gives one DV measure and exposed to one level of the IV (bw subjects design)
- each participant has a matched participant in the other condition so groups are correlated
- comparison bw groups
18
Q
bw subjects design
A
- involves random assignment of different people therefore large error variance
- least sensitive to treatment effects
19
Q
Within subjects design
A
- involves the same group of participant therefore small error variance
- most sensitive to treatment effects
20
Q
Matched subject design and error variance
A
- bc groups are similar therefore small error variance in the denominator ^ increases the sensitivity to the treatment effect
- however, not as sensitive as the w/in subject design because it is still different groups and not the exact same people
- more sensitive to the bw subjects design
21
Q
we only match on the most
A
relevant variables
22
Q
tend to use matched designs when
A
- need a design sensitive to the treatment effect but cannot do a within subjects design
23
Q
Subject Variables
A
- not true IVs
- cannot be randomly assigned or manipulated
- can only be interpreted in a correlational manner
- cannot attribute cause to these variables
- ex. gender, eye colour (pre existing participant characteristics)
