Assign 1 to lowest value, 2 to next lowest etc and average a score for same values (ties)
1 of 12
Why do you need to perform a rank transformation?
1. to flatten any distribution and remove gaps between outcome value ranges 2. to move outliers closer to the centre of the data
2 of 12
What are the two non-parametric correlation methods?
1. Spearman's rank 2. Kendall's tau
3 of 12
What is spearman's rank used for?
for non-normal but continuous data (interval and ratio) / can also be used for ordinal data for many levels
4 of 12
What is Kendall's tau used for?
for discrete and ordinal data
5 of 12
What do Spearman's rho and Pearson's r assess?
S - monotonic relationship / P - the linear relationship
6 of 12
What are ties
pairs which are neither concordant nor discordant
7 of 12
Tau-A
no adjustment for ties / this can keep true range of tau smaller than (-1,1)
8 of 12
Tau-B
adjustment for ties to keep range of tau in (-1,1) / easier to interpret
9 of 12
Tau-C
not frequently used
10 of 12
Kendall's tau vs Spearman's rho
they typically yield smaller results but Tau is more robust for significance testing in smaller samples and particularly suitable for discrete and ordinal data
11 of 12
Non-parametric tests for group differences
Wilcoxon signed-rank test: for paired samples (alternate to paired t test) / Mann-Whitney U: for independent samples / Binomial test: for binomial data
12 of 12
Other cards in this set
Card 2
Front
Why do you need to perform a rank transformation?
Back
1. to flatten any distribution and remove gaps between outcome value ranges 2. to move outliers closer to the centre of the data
Card 3
Front
What are the two non-parametric correlation methods?
Comments
No comments have yet been made