Re: Draw approximated normal distribution based on mean, median, percentiles?
- To: mathgroup at smc.vnet.net
- Subject: [mg70104] Re: [mg70092] Draw approximated normal distribution based on mean, median, percentiles?
- From: Stern <nycstern at gmail.com>
- Date: Wed, 4 Oct 2006 05:58:37 -0400 (EDT)
If you knew for certain that it was a normal distribution, then you would not need the median or the 25th and 75th percentiles; the mean and standard deviation would suffice to completely describe the curve. The fact that these are provided suggests that the curve is *not* normal, and that it displays skewness or excess kurtosis, for example. You will have to fit various distributions to the data and decide which seems best. Using Mathematica to determine the parameters of the curve (alpha and beta of a beta distribution, for example), should be possible, and to graph the resulting curve would be easy. Any statistical gurus care to correct or elaborate? MS On 10/3/06, rflaaten at start.no <rflaaten at start.no> wrote: > > Hello all, > > I searched the forum for any info on this, so here goes. > > I need to reference some statistical data from a study. The information > given there for an age distribution is: mean, standard deviation, > median plus the 25th and 75th percentiles. Now, I have tried in vain to > get Excel or SPSS to draw this for me. Is there any way of inputting > percentiles and median into Mathematica? The normal distribution > function only takes mean and sd as far as I know. > > > Any help would be much appreciated. > Richard > >