isaac wrote in message <[EMAIL PROTECTED]>... >Hi, I'm a newbie at stats. If I have a normal distribution defined by >mean x and stddev s, I know that x + s defines a threshold at the 86th >percentile. I know of a chart that dumps out a relationship between >threshold ( expressed as # of s from x ) and percentile. However, is >there a formula for computing the threshold, given a percentile, x, >and s? For example, how about the 93rd percentile? I'd rather not >have to use interpolation. > >Thanks, > >Isaac
There is no simple analytic formula. Most people use either tables, or their favourite statistical package. There are several approximations around. Probably the most widely used are those based upon Applied Statistics algorithms AS 66 and AS 241. AS 66 gives the area under the normal curve up to a user-specified number of standard deviations from the mean, while AS 241 gives the converse - i.e. you give it the area and it tells you how many standard deviations from the mean. The Applied Statistics algorithms were published by the Royal Statistical Society in its journal Applied Statistics. Most of them were in either Algol or Fortran. They can be found at statlib: http://lib.stat.cmu.edu Some of them, including 66 and 241, are also available at my ozemail web site (see signature below). Cheers -- Alan Miller (Retired from CSIRO Mathematical & Information Sciences) http://www.ozemail.com.au/~milleraj http://users.bigpond.net.au/amiller/ . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
