+
Related Flashcards
Related Topics
Cards In This Set
| Front | Back |
|
An alternative to the omnibus approach
|
Planned comparisons
|
|
What do planned comparisons test?
|
the specific predictions or hypotheses the researcher has advanced
|
|
Planned comparisons have Have both__ l and statistical advantages over the __ F test
|
Conceptual
omnibus |
|
The mean(s) for the group(s) with positive contrast coefficients are compared with the mean(s) for the group(s) with negative contrast coefficients
|
Read
|
|
The Groups with a __ are excluded from the comparison
|
0
|
|
The particular values we use for the contrast coefficients (1, -1 and 0) are ___
|
Arbitrary
|
|
What does it mean that particular values we use for the contrast coefficients (1, -1 and 0) are arbitrary?
|
It means we can use any numbers we like, provided they sum to 0 across rows.
|
|
As the number of groups (or conditions) in an experiment increases, the number of possible comparisons __ right along with it
|
Increases
|
|
Why is it Important to set limits on the number of contrasts you plan to perform?
|
As the number of groups (or conditions) in an experiment increases, the number of possible comparisons increases right along with it
|
|
What is the advantage of planned comparisons?
|
- we can do a single df test (more poweful than tests with more df in numerator)
- we can do a directional test (t-test) (gaining more power) |
|
How do you increase the advantage of planned comparisons?
|
By increasing the number of groups in the design
|
|
What is the disadvantage of planned comparisons?
|
- based on the full sample, MSRes from the omnibus is smaller than MSres with the reduced sample
|
|
How can we solve the disadvantage of the planned comparisons?
|
As long as the variances for the groups are roughly equal, we can use the pooled error term as th error term for our planned comparison
|
|
What is the pooled error term?
|
It is simply the error term (MSRes) from the omnibus analysis
|
