Let me begin by confessing that I wrote this in part as an exercise in

programming the TI-92.  It's based on a similar program that I wrote for the

TI-85 which in turn was written out of a need to have some way to analyze final

exam scores for large classes (e.g. 500+ students) when there was no computer

available at the grading site (when is there ever one??).  When I finished with

this program I decided that it was worth sharing with others.



  There are three items contained in the group file grdanlze.92g which when

transferred to the TI-92 will all be in the main folder (see note 3 below if you

wish to either use them from or move them to another folder ). The first item is

the primary program, ganalyze().  This is the program you will run to do all the

work for you -- drawing histograms, finding the stats, grouping different classes

scores together, removing outlying scores, and so on.  The second is an ancillary

program, condense(), which is used by ganalyze() and is also of a slight amount

of use in its own right: it takes a list of repeated scores and rewrites them as

an equivalent data variable.  The third is a function, concat(), which is used by

ganalyze() and is also very useful by itself: it takes two given lists and

concatenates them.





  RUNNING ganalyze():



  The program is straightforward: choices are made via the toolbar and menus, and

input is given thru dialog boxes.  The main thing to be aware of is that the

scores you wish to work with should already be saved as lists BEFORE you begin to

run the program (e.g., {20, 34, 42, 48} -> exam2).  You will not be given the

opportunity to enter any actual scores while you are running ganalyze(), however,

you WILL be able to combine different sets of scores, as well as to eliminate

outlying scores that you think should not be included in any statistical

analysis.



  Once you have started the program you will need to enter the name of the list

of scores that you will be initially working with (e.g., exam2 from the above

note).  If you make a mistake you can enter the name again (error checking to

look for accidental and/or careless mistakes is present throughout the program).

After the name has been entered you will be presented with the main toolbar which

has five selections, two of which have menu items beneath.  From left to right,

here is what each does:



     Stat:  Does a one variable statistical analysis of the scores.  Use the

up/down arrow to see all the data.



     Plot:  Displays a menu of five items, all related to graphical plots of the

data:



          Histogram:  Graphs a histogram of the scores.  You will be asked to

give a bucket/bar width, with the default being the minimal possible for a graph

to be constructed



          Box:  Draws a box plot of the current scores.



          Trace:  Allows you to trace the last drawn plot using the left/right

arrow keys. Use ESC when done.



          Alter window height:  The algorithm I use for the window height (ymax)

can often give a display which is inadequate (i.e. not tall enough).  This item

will present you with a dialog box to change the height, with the default being

the current value for ymax.



          Save as pic:  Allows you to give a name to save the current plot under

so that you can recall it later.  The picture is saved in whatever folder you

were in when you started ganalyze().



     Modify:  This allows you to specify two values, a lower bound and an upper

bound, and then the program proceeds to toss out all scores that don't lie in the

range [lower bound, upper bound].  The original set of scores is NOT affected by

this --- just the internal set the program is using.



     New Grades:  Has a menu with two choices beneath it:



          Single Set:  Lets you choose a new set of grades to work with.



          Multiple Sets:  Lets you specify several lists of grades (no limit on

how many) that are to be combined into a new set to work with.  This new list

will be available to you when ganalyze() finishes (it's a global variable) so

that you can continue to work with it in the future.  Because of this it is

important to think about what name you give this new list --- you don't want to

replace a list that already exists.



     Quit:  This should be self-explanatory.





  NOTES for ganalyze():

    1. There is no upper limit (other than memory) on how large a list you can

use in ganalyze(). However, the larger the list the longer you have to wait for

some operations.  I have tried to speed things up by using a data variable (see

condense() below) but that can take time to set up.  For example, a list of 650

items takes one minute to process.  The positive side is that histograms etc. go

much faster (without processing, each histogram for the 650 item list would take

over a minute).

    2.  There is no limit on what values the grades can take.  You can use

ganalyze() for 100 pt exams or 500 point exams --- it makes no difference.

    3.  If you are going to use ganalyze() from a different folder than the one

it is stored in, you will need to modify the code (where it calls condense and

concat) to include the proper path name.  All this really amounts to is placing

<folder name\> in front of each call of condense and concat





  RUNNING condense(origlist,datastr):



  Origlist is a list of numeric values, datastr the name to give to the data

variable to place the result in.  The list of values is sorted and repeats

removed with the result being placed in the first column of the data variable.  A

count of how often each value occurs is kept and placed in the second column of

the data variable, so that when both the columns are read as (value, freq) they

end up describing the original list.





  RUNNING concat(list1, list2)



  list1 and list2 are lists (no limits on types of entries). They are

concatenated (list1 first) and the resulting list is returned.





  QUESTIONS?

  My e-mail address is rlittle@math.utk.edu; I also subscribe to the graph-ti

mailing list.