Too bad you solved it, because I have another approach. It's actually
related, but allows for multiple classes to be lined up above a
different X axis position. A single occurrence of a score is plotted
above the axis label. If there are two occurrences, they are spread a
bit left and right. If there are three, one is centered and the other
two are spread a bit further.
Sample data, starting in A2, explanation below:
Class Cumul. Total Delta X Name Score 0.07
1 1 4 -3 0.79 AA 22
1 1 2 -1 0.93 BB 24
1 2 2 1 1.07 CC 24
1 1 3 -2 0.86 DD 18
1 2 4 -1 0.93 EE 22
1 2 3 0 1.00 FF 18
1 1 1 0 1.00 GG 15
1 3 3 2 1.14 HH 18
1 1 1 0 1.00 II 19
1 3 4 1 1.07 JJ 22
1 4 4 3 1.21 KK 22
1 1 1 0 1.00 LL 16
2 1 4 -3 1.79 MM 17
2 1 2 -1 1.93 NN 18
2 1 2 -1 1.93 OO 19
2 2 4 -1 1.93 PP 17
2 3 4 1 2.07 QQ 17
2 2 2 1 2.07 RR 18
2 1 2 -1 1.93 SS 25
2 2 2 1 2.07 TT 25
2 1 1 0 2.00 UU 24
2 2 2 1 2.07 VV 19
2 4 4 3 2.21 WW 17
2 1 1 0 2.00 XX 20
This has scores for two classes (1 & 2). The Cumul. column (b) has the
number so far with the same score in the same class. B3 has this array
formula
{=SUM(($G3:$G$3=G3)*($A3:$A$3=A3))}
Don't type in the curly brackets; type the formula, then hold Ctrl-Shift
while pressing Enter, and Excel adds the brackets. Drag this down to
B26. The Total column (C) has the total number in the same class with
the same score. C3 has this array formula
{=SUM(($G$3:$G$26=G3)*($A$3:$A$26=A3))}
dragged down to C26. Delta is just a column with an intermediate
computation which will lead to the offset of the points from the center.
The formula in D3 is
=2*B3-C3-1
and it's dragged down to D26. Finally, the X column has the actual X
value for the student's plotted point. The formula in E3 is:
=A3+D3*H$2
H2 holds the standard horizontal offset (0.07 in my case, but it's
adjustable) for points with the same score. In the histogram example
Debra cited, the offset was 1, but was derived differently.
All the formulas in columns B:E could be combined into a single column,
but it's always easier to set these things up piecewise.
Finally, Name and Score are your raw data. I used initials in place of
names, because they are shorter, and less prone to being obscured by
each other.
Almost ready to chart the data. I set up a dummy range off to the side:
Class
A 0
B 0
A and B are the designations for the two classes being plotted in this
example. Select this range and construct a column chart. You get A and
B for category axis labels, and no columns appear because the values are
0. You could put the class averages into the cells instead of zeros,
and these will show up on the chart. It would show that Class A (1 in
the main table) did slightly better than class B.
Now select the column with the X values, and hold the Ctrl key and
select the corresponding scores. Copy this discontiguous range, click
on the chart, and select Paste Special from the Edit menu. Add this as
a new series, with categories in the first column.
You get another column series added to the chart. Right click on the
new series, choose Chart Type from the pop up menu, and pick out the
Scatter chart type with markers and no lines. Okay your way back to the
chart. Right click on the chart, and choose Chart Options from the pop
up menu. Click on the Axes tab, and uncheck both secondary axes.
Now use Rob Bovey's Chart Labeler (
http://appspro.com) to put the
initials onto the chart. What I did was center the labels right on the
charted points, then formatted the points to have no markers; the labels
are now the markers.
- Jon
