Math Forum Discussions

Bob Hayden

Posts: 2,384
Registered: 12/6/04

Stacked boxplots in Minitab
Posted: Nov 1, 1996 1:10 PM

Plain Text

----- Forwarded message from Louise Hilton -----

I'm trying to prepare a stacked boxplot in minitab with two 5-number
summaries:

Men's Salaries (408,520,589,846, 1040)
Women's Salaries (209, 312, 480, 650, 1060)

When I go to the boxplot menu and ask for the stacked boxplot option I
get the following error message:

MTB > BoxPlot 'Men';
SUBC> By 'Women'.
NOTE -- 1 observations with other subscripts.
* ERROR * Illegal value in subscript column

----- End of forwarded message from Louise Hilton -----

I'm replying to this to the whole list because it raises a number of
issues of general importance.

First, Minitab is designed to analyze RAW DATA. There have been a
number of questions/problems on this list that arise from trying to
feed it a data summary prepared elsewhere or given in a textbook.
Generally, this does not work, and I seee that as a FEATURE rather
than a bug in Minitab, because I don't think we should be assigning
problems that do not give the raw data.

For the particular data in question, Minitab gives the same boxplot
whether you enter the five number summary or the raw data. This does
not always happen, and it only happens at all because Minitab uses
Tukey hinges instead of quartiles or some other objects for the ends
of the box. In particular, if you enter the five number summary into
the TI-82, you get a very different looking boxplot.

Finally, the 'stack' command is a source of endless confusion on
Minitab. Here it is applied to the data in question. I'm using the
command line. Menus add an additional level of compexity and I tend
to avoid them!-)

MTB > set 'men' c1
Entering data from file: men.DAT
MTB > set 'women' c2
Entering data from file: women.DAT
MTB > name c1 'men' c2 'women'
MTB > print c1 c2

ROW men women

1 408 209
2 520 312
3 589 480
4 846 650
5 1040 1060

MTB > stack c1 c2 into c3;
SUBC> subscripts in c4.
MTB > print c3 c4

ROW C3 C4

1 408 1
2 520 1
3 589 1
4 846 1
5 1040 1
6 209 2
7 312 2
8 480 2
9 650 2
10 1060 2

MTB > boxplot c3;
SUBC> by c4.

C4

---------------------
1 -------I + I------------
---------------------

----------------------
2 -------I + I-------------------------
----------------------
--------+---------+---------+---------+---------+--------C3
320 480 640 800 960

I would have received an error message like

MTB > BoxPlot 'Men';
SUBC> By 'Women'.
NOTE -- 1 observations with other subscripts.
* ERROR * Illegal value in subscript column

if I had typed

MTB> boxplot c1;
SUBC> by c2.

Then Minitab would have tried to make five boxplots -- one for each
value in C2. It would have encountered the 209 in C2 and tried to
find all the occurences of (Category #) 209 in c2. Then it would take
the corresponding numbers in c1 and try to make a boxplot of them.
Here it would try to make a boxplot of the single 408 in C1. This
does not work so it gives an error message. The column folowing "by"
must contain CATEGORICAL data flagging the different groups for which
you want boxplots.

_
| | Robert W. Hayden
| | Department of Mathematics
/ | Plymouth State College MSC#29
| | Plymouth, New Hampshire 03264 USA
| * | Rural Route 1, Box 10
/ | Ashland, NH 03217-9702
| ) (603) 968-9914 (home)
L_____/ hayden@oz.plymouth.edu
fax (603) 535-2943 (work)