Why is the prediction interval (individual) getting larger with larger sample size?

rs

Hi All,

 

I am using JMP14 to calculate the prediction  interval of individuals.  On the figure below you will see:

 

n=1             PI  (     74.5  - 125.5)

n=10           PI  (  63.5     -    136.5)

n=100,000   PI  (34.6       -        165.4)

 

Should the PI (individual) interval get narrower  as the PI (mean)?

 

Thx

 

Phil_Kay

Hi,

 

I think this is a misunderstanding about prediction intervals. The intervals get wider with higher N because these are intervals within which you expect to find all N observations. I found this white paper really useful in understanding the differences between confidence, prediction and tolerance intervals.

 

Hope that helps.

Phil