how do i create sub groups in SPSS?

[CedricC...@gmail.com]

I need to know how to create sub groups in spss to be able to plot the
best fit line to sub groups rather than to total ....

I need a sub group to point to var1 and another sub group to point to
var2

Please tell me if im not being clear

[neila...@msn.com]

You are not being clear!

[CedricC...@gmail.com]

When plotting a simple scatter plot graph with two variables on the Y
axis I would like to apply the feature "apply best fit line to sub
group" in the chart editor which however is unavailable (grayed out).
What do I have to do to be able to make use of the "apply best fit line
to sub group" feature?

I have tried selecting 1 group of varibales only in the chart editor
but it stil does not allow me to "apply the best fit line to sub
groups"

Please tell me if im not being clear

[CedricC...@gmail.com]

When plotting a simple scatter plot graph with two variables on the Y
axis I would like to apply the feature "apply best fit line to sub
group" in the chart editor which however is unavailable (grayed out).
What do I have to do to be able to make use of the "apply best fit line
to sub group" feature?

I have tried selecting 1 group of varibales only in the chart editor

but it stil does not allow me to "apply the best fit line to sub
groups"

Please tell me if im not being clear

[Art Kendall]

"apply best fit line to subgroups" means that you have subgroups of cases.
For example (males and females), or (farmers, ranchers, greenhouses).
Did you make such a specification. After you exited the menu with
<paste> what did the syntax look like?
After you highlighted the procedure name in the syntax window, what did
additional options for the syntax look like?

Please give more specific information: what are the two variables you
are trying to scatterplot? How many values are legitimate for each?

How many cases do you have? So many that the scatterplot is a blur?

What is your group variable? How many values does it take?
How many cases are there in each group?

Art
A...@DrKendall.org
Social Research Consultants
University Park, MD USA
(301) 864-5570

[klange]

If you're using the standard scatterplot then the editing options
depend on what version of SPSS you're using, but in general you should
be able to include fit lines for sub-groups as long as you have
included the categorical grouping variable (eg, gender, age groups) in
the graph as a legend variable. In SPSS 12 this means the grouping
variable should go into the 'set markers by' box.

When editing the graph (in version 12 or higher) the options you have
available to you will also depend on what element of the graph is
currently highlighted. So try selecting all data points, or just one
sub-group of data points (by selecting that category from the legend),
to see if that gives you the option you're after.

Hope this helps.
Kylie.

[CedricC...@gmail.com]

I made no specification for sub groups of cases.?
where do i check the syntax?
I am trying to plot 2 scale variables on the y axis, var 1 which is the
actual result and var 2 the predicted result (from an earlier linear
regression) and I would like to have seperate best fit lines for every
variable on the same graph, is this possible without having them as
subgroups of cases in the same variable? Is there a way to present it
the way I am saying?
I have not more than 20 cases for each variable and all the values are
legitimate
I dont have a goup variable. I need to plot two different variables VAR
1 (actual reading) and VAR 2 (predicted reading)

[CedricC...@gmail.com]

Yes that does help, and thank u all for your generous replies. What I
m trying to do is a bit different now that i understand a bit better
how best fit to line to sub groups works ( group variable which has
cases defined into subgroups but belonging to the same variable) . I
wanted to plot a best fit line for two seperate variables on the same
graph to demonstate the difference in the best fit line for each
variable VAR 1 ( actual reading) VAR 2 (predicted reading) I am
plotting the results of a linear regression analysis as VAR2 (ZPRED)
with actual readings as VAR 1 against an x axis of YEAR plotted as a
scale variable. The readings/ no of cases for VAR 2 (ZPRED) is more
than for the actual readings/ cases for VAR1 since I have about 3
cases that are just predicted with the compute function for VAR 2.