Charting question: Line or Scatter or ... ?

[David Thompson]

>  >Use a line chart with a second line to plot the scores of an
>>invisible student. Use a fill pattern of "none" for the invisible
>>student's line, and give the invisible student a score of 100 for the
>>first month. Don't have any records for months where testing has not
>>yet occurred, and have a procedure use ChangeObjects to adjust the
>>width of the chart based on the number of records
>
>Adjusting the width of the chart only widens the distance between
>each value as the highest value (less than 100) is just further from
>0 than before the widening.

Adjusting the width was intended to fix the horizontal scale, not the 
vertical. When you have 10 records, you want the chart to be wider 
than when you have just one record, so that the horizontal space 
devoted to each record remains the same. The invisible student would 
have a score of 100 in the *first* month, and after that it wouldn't 
matter what scores the visible or invisible students had, as long as 
none of them were greater than 100. With a fill pattern of none, you 
wouldn't see the line for the invisible student, but its presence 
would still cause the vertical scale to be dimensioned to accommodate 
it.

>
>>, or alternatively, place a white rectangle to the right of the
>>chart and bring to front.
>>Change its dimensions with ChangeObjects to cover the untested
>>months. Use more white rectangles to hide the legends.
>
>Imagine a piece of graph paper with all the grid lines showing and
>also multiple 'normal' curves preprinted on it. Now place a few white
>rectangles to cover up the incorrect plots. It just loses something
>in the aesthetics.

Maybe my imagination isn't that good. White rectangles don't print. 
The grid lines and normal curves are preprinted, so they aren't going 
away. Other than the legends and incorrect plots that you don't want 
to print, what's being lost?

Changing the width of the chart was tied to the idea of varying the 
number of records, so that there wouldn't be any records for tests 
that hadn't been taken yet.

Using white rectangles to cover the incorrect plots, was tied to the 
idea that the number of records would be fixed, and there would be 
records for tests that haven't been taken yet, which would need to be 
hidden.

Dave