ranking based on numbers

real_jklee

New Member
Hi guys,

I have a set of numbers (photon counts based on infection severity) and I want to rank "low", "intermediate" "high" based on these numbers.

I've looked into Wilcoxon signed rank test but it requires two sets of data.

Is there any statistical analysis that I can rank "low, intermediate, high" these numbers based on photon counts?

obh

Member
Wilcoxon signed rank /Man Whitney U are tests to "compare" populations using samples (or the same population in different condition) . But this is not what you try to do.

You can try for example ranking based on the order:
Sort all values
<Q1 low, between Q1 and Q3 - intermediate >Q3 high (Q2 is the median, Q1 is the median of the first part ...)
[1,2,4,7,8,8,10,11,21,40,80,100] = > [1,2,4,7,8,8],[10,11,21,40,80,100]
Q2=(4+7)/2 Q3=(21+40)/2
Or maybe use third instead of the quarter. <third1 low, >third3 high ...

You can also use the standard deviation to choose the thresholds for example use thresholds of Z0.25 and Z0.75

All depend on what do you want to achieve. For example, do you want a similar count for each rank?