# Ranking Performance

#### clemkonan

##### New Member
In statistics if 10 people take a test you get your mark along with your ranking relative to the rest of the class, I think this is done as percentile or Quartile. So if you score at the 75th percentile, 25 percent of your class got a better mark.

My problem is slightly different. I have a 800 points of data representative of testing conducted on 34 farms over a 2 year period some farms have as many as 24 data points others may have as few as 5. This is reflective of new and existing or long term participants.

I have the histogram and the cumulative distribution function but I need a simple way of illustrating to Mary how her average result ( in reality I have used the median) compares to that of her member in the other town.

In addition in my case a low score is indicative of high performance, so on a scale of 100 to 800 , 100 is excellent and 800 is in the hot zone.
Is the percentrank function a good idea or is there something better? I am using Excel. This ranking system may also be used for incentive purposes,

#### ephialtes

##### New Member
Some clarification needed

How are the scores for each farm calculated? Raw score added up? A mean based on score/possible score? I'm assuming the latter since you have varying numbers of data points from each subject.

Percentile rank will tell you where each farm falls with respect to another. What it won't really tell you is if Mary's farm is performing significantly better/worse than Bill's farm.

I'm not sure how many categories you need, but I might suggest looking at standard deviations. Say those farms that are 2 or more SDs above the mean get a bonus of \$X. This would have the advantage of taking into account the variability of the scores.

-E