

There is a lot of interest in the "box and whisker" plots at my 
daughters school, where their textbook shamelessly promotes the 
graphing features of the TI-82.

To install, just load onto the stack, and save with any name you 
choose  (BAW suggests itself rather quickly...).  This version will 
not create its own directory or a bunch of other programs.

To delete version 1.0, type 'BAW.DIR' or whatever name you gave it, 
then PGDIR. Version 2.0 or 2.1 will be overwritten if you set your I/O 
parameters to overwrite, otherwise just delete the old one.


Version 3.0 will now accept either one or two lists on the stack.  If 
there are 2 lists, it will plot them both for comparison, this means 
you shouldn't leave any leftover lists on the stack if you don't want 
it plotted! It expects to see a list of data on level one or levels 
one and two of the stack, as I believe that is the easiest and fastest 
method of input. Of course it could be easily modified to accept a 
matrix to more closely resemble other statistical analysis programs.  
It returns the 1st quartile, median, and 3rd quartile to the stack, 
with min and max values displayed on the picture display.  It plots 
"outliers" (data removed more than 1.5 times the interquartile range 
from the 1st to the 3rd quartile) as dots.  The quartiles are taken 
as the median of the values above and below the median.

Changes in version 3.0

1. Now in System RPL, it is about 1/3 smaller and 1/3 faster.
2. Accepts 2 lists for comparison.

Changes in version 2.0

1. Sub programs deleted.

2. Different algorithm for quartiles, based on PTILE but uses MOD 
   arithmetic to account for "median of the median" standard used by 
   some to calculate quartile ranges.

3. Eliminated PVIEW, you cant "watch while you draw".  It didn't seem 
   worth it, a 40 item list only takes 8 seconds to complete anyway.

Changes in version 2.1

1. Fixed "whiskers" that drew to fictional data location when 
   interquartile range exceeded a legitimate value in list but was 
   less than an outlier.

2. Fixed location of Y axis when outliers exist so it remains off 
   screen.

3. Installed tick marks at each 10% increment to better estimate other 
   percentiles of data.

4. Calibrated picture environment so (X,Y) scrolling will be accurate 
   (at least within one pixel!)

Let me know what you think or if you use it.  Thanks to all who have 
written so far!

Checksum: #25690d 672.5

Dennis Straley
dst12345@aol.com