Inferential Stats Flashcards
1
Q
Why do we use inferential stats tests?
A
They allow psychologists to make conclusions based on the probability that a particular pattern of results could have arisen by chance
2
Q
Which tests are OV < 0.5 to be significant?
A
- Sign
- Wilcoxon
- Mann Whitney
3
Q
Which tests are OV > 0.5 to be significant?
A
- Chi-square
- Spearman
4
Q
What are the 3 reasons why a particular test is carried out?
A
- Test of difference or association
- Level of data
- Study design
5
Q
Binomial sign
A
- Calculate differences between two conditions and allocate + or -
- Find the least frequent sign and how many people who had that
- Not including ps who showed no difference, find the critical value from the table and compare OV
6
Q
Wilcoxon test
A
- Calculate the differences between two scores by taking one from the other
- Rank the differences giving the smallest difference rank 1
- DO NOT RANK ANY DIFFERENCES OF 0 - when adding number of scores, do not count those with a difference of 0 and ignore the signs when calculating the difference
- Add up the ranks for positive differences
- Add up the ranks for negative differences
- T is the figure that is the smallest when the ranks are totalled (may be positive or negative)
- N is the number of scores left, ignore those with no difference
7
Q
Mann Whitney test
A
- Rank as whole cohort
- Put ranks back into separate groups
- Sum the ranks in each group
- Sub into formula
8
Q
Spearman test
A
- Rank each covariable separately
- Differences between the covariables
- Square the sum of the differences
- Sub into formula
9
Q
Chi Square test
A
- Sum columns and rows (and find grand total)
- Find expected frequency ((column x row)/grand total)
- For each cell find: ((O-E)^2/E)
- Sum them all together to find observed value
- Degrees of freedom = (row total - 1) x (column total - 1) for CV
