Computing overlaps in a list of variables

[Emmanuel Mutale]

I am working with SPSS v14 and I have a very basic understanding of SPSS
syntax. I have seven variables v1.....v7 each with two valid responses
(1=YES and 2=NO)as well as seven user-defined missing values for each of
the seven variables.
I want to work out the numbers/percentages of all possible two-way, three-
way, four-way, five-way... etc combinations in my data set of 15,000 cases
for which the response in each of the seven variables (v1...v7)is YES. Is
there an efficient way of doing this?

Emmanuel

[Maguin, Eugene]

Emmanuel,

I'm a little confused by this part of your message: "... seven user-defined
missing values for each of the seven variables." It sounds like your are
saying that each of your seven variables can have one of nine possible
values, i.e., 1=yes, 2=no, one of the seven user defined missing values.
Maybe it been changed in 14 but I have always understood that spss allowed
each variable to have a maximum of three values defined as missing.

This too is confusing. "I want to work out the numbers/percentages of all
possible two-way, three-way, four-way, five-way... etc combinations in my
data set of 15,000 cases
for which the response in each of the seven variables (v1...v7)is YES."

If you want to know how many cases have a value of yes for all seven
variables then there can be no two-way, etc combinations.

It may be that the Count command is what you need but I don't know. It is on
the menu I would expect (as well as being documented in the help system) but
the format is simply.

Count sevens=v1 to v7(1).

This will tell you the number of cases having exactly one yes, two yeses,
etc.

Gene Maguin

[Mark A Davenport MADAVENP]

Can someone confirm for me that GenLim is available only in (new to) 15.

***************************************************************************************************************************************************************
Mark A. Davenport Ph.D.
Senior Research Analyst
Office of Institutional Research
The University of North Carolina at Greensboro
336.256.0395
[hidden email]

'An approximate answer to the right question is worth a good deal more
than an exact answer to an approximate question.' --a paraphrase of J. W.
Tukey (1962)

[Max Jasper]

GenLim is not available in SPSS 15 but if you recommend it then it might be
implemented in SPSS 16.

|-----Original Message-----
|From: SPSSX(r) Discussion [mailto:[hidden email]] On
|Behalf Of Mark A Davenport MADAVENP
|Sent: Tuesday, May 01, 2007 8:24 AM
|To: [hidden email]
|Subject: GenLim
|
|
|Can someone confirm for me that GenLim is available only in
|(new to) 15.
|
|***************************************************************
|***************************************************************
|*********************************
|Mark A. Davenport Ph.D.
|Senior Research Analyst
|Office of Institutional Research
|The University of North Carolina at Greensboro
|336.256.0395
|[hidden email]
|
|'An approximate answer to the right question is worth a good
|deal more than an exact answer to an approximate question.'
|--a paraphrase of J. W. Tukey (1962)

[Richard Ristow]

A side point.

At 10:15 AM 5/1/2007, Gene Maguin wrote:

>I'm a little confused by this part of your message: "... seven
>user-defined missing values for each of the seven variables." Maybe it
>been changed in 14 but I have always understood that spss allowed each
>variable to have a maximum of three values defined as missing.

Actually, you can have many user-missing values, as long as they're in
a contiguous range with no non-missing values interpolated; that's been
around in most releases any of us have ever seen. It's a useful trick,
sometimes, when you want to distinguish among a lot of different
reasons a value is 'missing'. From the Command Syntax Reference (SPSS
15):


Specifying Ranges of Missing Values
-----------------------------------
A range of values can be specified as missing for numeric variables but
not for string variables.

.. The keyword THRU indicates an inclusive list of values. Values must
be separated from THRU by at least one blank space.

.. The keywords HIGHEST and LOWEST with THRU indicate the highest and
lowest values of a variable. HIGHEST and LOWEST can be abbreviated to
HI and LO.

.. Only one THRU specification can be used for each variable or
variable list. Each THRU specification can be combined with one
additional missing value.

Example
MISSING VALUES V1 (LOWEST THRU 0).
.. All negative values and 0 are declared missing for the variable V1.

[Weeks, Kyle]

Mark, do you mean Generalized Linear Models (GENLIN)?  If so the was added in SPSS 15.  To see a complete list of what's new in SPSS 15, see http://www.spss.com/spss/whats_new_base.htm and http://www.spss.com/spss/whats_new_modules.htm.

Regards.


Kyle Weeks, Ph.D.
Director of Product Management, SPSS Product Line
Product Management
SPSS Inc.
Phone: [phone] / Fax: [phone]
[hidden email]
www.spss.com
SPSS Inc. helps organizations turn data into insight through predictive analytics.


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Max Jasper
Sent: Tuesday, May 01, 2007 10:41 AM
To: [hidden email]
Subject: Re: GenLim

GenLim is not available in SPSS 15 but if you recommend it then it might be
implemented in SPSS 16.

|-----Original Message-----
|From: SPSSX(r) Discussion [mailto:[hidden email]] On
|Behalf Of Mark A Davenport MADAVENP
|Sent: Tuesday, May 01, 2007 8:24 AM
|To: [hidden email]
|Subject: GenLim
|
|
|Can someone confirm for me that GenLim is available only in
|(new to) 15.
|
|***************************************************************
|***************************************************************
|*********************************
|Mark A. Davenport Ph.D.
|Senior Research Analyst
|Office of Institutional Research
|The University of North Carolina at Greensboro
|336.256.0395
|[hidden email]
|
|'An approximate answer to the right question is worth a good
|deal more than an exact answer to an approximate question.'
|--a paraphrase of J. W. Tukey (1962)

[Jim Moffitt]

Emmanuel:
I'm not sure I understand your needs, but the following syntax may help.
Run it in a blank SPSS data editor and take a look at what it does. It
will create a combination variable that will identify patterns of
responses across 7 variables. Just run a simple frequency count on the
combo variable to count the numberof respondents whose responses
caonatin a certain pattern. I've assumed that none of your variables
contain blank cells which would be coded system missing. If this
presumption is incorrect, you may want to consider creating a new code
for the system missing values so all cells are populated. Good luck.

DATA LIST LIST (",") / id X1 to X7.
BEGIN DATA
001,1,2,2,1,1,2,2,
002,1,1,1,1,1,1,1,
003,2,2,1,2,1,1,2,
004,1,1,1,1,1,1,1,
005,2,2,1,2,1,1,2,
006,1,1,1,2,1,2,1,
END DATA.
COMPUTE Combo =
(X7*10**0)+(X6*10**1)+(X5*10**2)+(X4*10**3)+(X3*10**4)+(X2*10
 **5)+(X1*10**6) .
FORMAT Combo (N7.0).
FORMAT id (N3.0).
EXECUTE .

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Emmanuel Mutale
Sent: Monday, April 30, 2007 9:58 AM
To: [hidden email]
Subject: Computing overlaps in a list of variables

I am working with SPSS v14 and I have a very basic understanding of SPSS
syntax. I have seven variables v1.....v7 each with two valid responses
(1=YES and 2=NO)as well as seven user-defined missing values for each of
the seven variables.
I want to work out the numbers/percentages of all possible two-way,
three- way, four-way, five-way... etc combinations in my data set of
15,000 cases for which the response in each of the seven variables
(v1...v7)is YES. Is there an efficient way of doing this?

Emmanuel

[Eric Janssen]

It MUST be implemented, no asking.

Actually I don't really understand why a company like SPSS does not provide the
widest possible range of statistical techniques in one shot. Is it for
commercial purposes? Why waiting for so long to create a GLM unit for instance?
To make people buy the new releases?

This is non-sense, instead of waiting SPSS' good-will, customers will find a
solution using other statistical packages.


> GenLim is not available in SPSS 15 but if you recommend it then it might be
> implemented in SPSS 16.

[Weeks, Kyle]

Eric, Generalized Linear Models and Generalzed Estinating Equations IS available in SPSS 15.
 
Regards.
 
Kyle Weeks, Ph.D.
Director of Product Management, SPSS Product Line
Product Management
SPSS Inc.
Phone: [phone] / Fax: [phone]
[hidden email]
www.spss.com
SPSS Inc. helps organizations turn data into insight through predictive analytics.




________________________________

From: SPSSX(r) Discussion on behalf of Eric Janssen
Sent: Fri 5/4/2007 4:12 AM
To: [hidden email]
Subject: Re: GenLim



It MUST be implemented, no asking.

Actually I don't really understand why a company like SPSS does not provide the
widest possible range of statistical techniques in one shot. Is it for
commercial purposes? Why waiting for so long to create a GLM unit for instance?
To make people buy the new releases?

This is non-sense, instead of waiting SPSS' good-will, customers will find a
solution using other statistical packages.


> GenLim is not available in SPSS 15 but if you recommend it then it might be
> implemented in SPSS 16.