SPSSX Discussion

Finding variables with the GUI

Classic

List

Threaded

15 messages Options

John F Hall

Finding variables with the GUI

I have a file with 847 variables, one of which is [REarn] on row 618.

When using Analyze >> Descriptive Statistics >> Frequencies from the GUI it will take forever scrolling up and down to find [REarn]. However, in Frequencies, if you type "rearn" quickly in the left pane, it immediately finds the variable.

That's a new one, even for me! Is it documented anywhere?

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

spss.giesel@yahoo.de

Re: Finding variables with the GUI

Hi, John,

I've been using this for years and it's probably not documented.
This kind of literal search is quite common for a lot of programs.

Personally, I would like if you could write the variable name SLOWLY and find it more quickly that way. To write the name quickly appears a bit annoying to me, especially if the variable name is long and you have many similar named variables.

GL,
  Mario

Mario Giesel
Munich, Germany

Am Donnerstag, 11. Oktober 2018, 17:07:34 MESZ hat John F Hall <[hidden email]> Folgendes geschrieben:

I have a file with 847 variables, one of which is [REarn] on row 618.

That's a new one, even for me! Is it documented anywhere?

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

Bruce Weaver

Re: Finding variables with the GUI

Administrator

A search option for Utilities > Variables would be useful sometimes too.

Mario Giesel-2 wrote

> Hi, John,
> I've been using this for years and it's probably not documented.This kind
> of literal search is quite common for a lot of programs.
> Personally, I would like if you could write the variable name SLOWLY and
> find it more quickly that way. To write the name quickly appears a bit
> annoying to me, especially if the variable name is long and you have many
> similar named variables.
> GL, Mario
>
> Mario GieselMunich, Germany
>
> Am Donnerstag, 11. Oktober 2018, 17:07:34 MESZ hat John F Hall <

> johnfhall@

> > Folgendes geschrieben:
>
>

> I have a file with 847 variables, one of which is [REarn] on row 618.
>
>
>
> When using Analyze >> Descriptive Statistics >> Frequencies from the GUI
> it will take forever scrolling up and down to find [REarn]. However, in
> Frequencies, if you type "rearn" quickly in the left pane, it immediately
> finds the variable.
>
>
>
> That's a new one, even for me! Is it documented anywhere?
>
>
>
> John F Hall MA (Cantab) Dip Ed (Dunelm)
>
> [Retired academic survey researcher]
>
>
>
> Email:

> johnfhall@

>
>
> Website:     Journeys in Survey Research
>
> Course:       Survey Analysis Workshop (SPSS)
>
> Research:   Subjective Social Indicators (Quality of Life)
>
>
> =====================To manage your subscription to SPSSX-L, send a
> message

> toLISTSERV@.UGA

> (not to SPSSX-L), with no body text except thecommand. To leave the list,
> send the commandSIGNOFF SPSSX-LFor a list of commands to manage
> subscriptions, send the commandINFO REFCARD
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

> (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

PLEASE NOTE THE FOLLOWING:
1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above.
2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/).

John F Hall

Re: Finding variables with the GUI

In reply to this post by spss.giesel@yahoo.de

Mario

Yep. Most of the variable names in the British Social Attitudes Survey are unique after 4 or 5 characters so it's relatively quick to find a variable. I'm the laziest key-tapper in the world, so I dread to think of variables where the first 15 characters were all the same! Myself, I usually find variables by highlighting the Name column and using Ctrl F, but unfortunately these days everyone wants to be pointy-clicky and swipey.

Relevant to the above, at the beginning of October 2018, the UK Data Service (UKDS) implemented a new beta site which experienced a few technical teething problems (understatement of the year!). All previous accounts and projects are still there, but, because of new Data Protection Regulations, each data set previously assigned to a project now needs individual Depositor Approval before it can be accessed and downloaded. This process is time-consuming and occasionally confusing (not to say extremely irritating for a veteran user like me).

I'm writing a step-by-step demonstration work-through showing how to identify and download the original SPSS saved files, which I'll upload to my site and post a link to the list.

My major problem is that the metadata in the original files are mutually incompatible: they cannot be combined to enable analysis across time. Attributes are not the same for each survey. Variables with the same name have different data types, different widths (for string variables), different levels of measurement, different variable labels, different coding categories, different value labels and different missing values. They are also generated by different releases of SPSS.

They demonstrate changes from SPSS 12 in 1983 (UPPER CASE only, 40 character limits on variable labels, 16 character limits on value labels, printed in two blocks of 8 characters, no £ sign ( $ the only available currency symbol) to SPSS 19 in 2015 (Mixed case labels, 256 character limit for both variable and value labels, £ symbol allowed).

At least they're not still using SPSS-X, but the setup files for the early waves are clearly based on 80-column Hollerith card formats and the data prep info is printed on the facsimile questionnaires. Nowadays that's all gone: questionnaires only available in CAPI/BLAISE format.

What's a late septuagenarian insomniac pedantic purist Luddite to do?

John

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Mario Giesel
Sent: 11 October 2018 17:28
To: [hidden email]
Subject: Re: Finding variables with the GUI

Hi, John,

I've been using this for years and it's probably not documented.

This kind of literal search is quite common for a lot of programs.

Personally, I would like if you could write the variable name SLOWLY and find it more quickly that way. To write the name quickly appears a bit annoying to me, especially if the variable name is long and you have many similar named variables.

GL,

Mario

Mario Giesel

Munich, Germany

Am Donnerstag, 11. Oktober 2018, 17:07:34 MESZ hat John F Hall <[hidden email]> Folgendes geschrieben:

I have a file with 847 variables, one of which is [REarn] on row 618.

That's a new one, even for me! Is it documented anywhere?

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

Anthony Babinec

FW: Finding variables with the GUI

In reply to this post by Bruce Weaver

Edit menu has a Go to Variable selection. It lists variables in
alphabetic order. When you select a variable, it takes you to
that column in the Data Editor window. This of course is not the
same as the "typomatic" facility being discussed.

Tony Babinec

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Bruce Weaver
Sent: Thursday, October 11, 2018 10:47 AM
To: [hidden email]
Subject: Re: Finding variables with the GUI

A search option for Utilities > Variables would be useful sometimes too.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

John F Hall

Re: Finding variables with the GUI

Tony
Just tried that. Never done it before, but great to know. I'll use it in the new work-through demonstration. Is all this stuff documented for idiot users like me?
John

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email: [hidden email]
Website: Journeys in Survey Research
Course: Survey Analysis Workshop (SPSS)
Research: Subjective Social Indicators (Quality of Life)

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Anthony Babinec
Sent: 11 October 2018 18:07
To: [hidden email]
Subject: FW: Finding variables with the GUI

Edit menu has a Go to Variable selection. It lists variables in alphabetic order. When you select a variable, it takes you to that column in the Data Editor window. This of course is not the same as the "typomatic" facility being discussed.

Tony Babinec

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Bruce Weaver
Sent: Thursday, October 11, 2018 10:47 AM
To: [hidden email]
Subject: Re: Finding variables with the GUI

A search option for Utilities > Variables would be useful sometimes too.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon Peck

Re: Finding variables with the GUI

First letter search in dialog boxes is not documented by SPSS, because it is a standard Windows action. The multi-letter navigation in a dialog box is news to me, but it seems to work.

In the Data Editor, notice that there is a toolbar icon that looks like three bars attacking a grid and the corresponding menu item View > Variables or ctrl-T that brings up a dialog showing variable information for whatever variable you select in the dialog, and it has a Go To button. It supports multiple selection and has a Paste button to copy the selected names into the Syntax Editor. Also available from Utilities > Variables. Tony's suggestion of Edit > Go to Variable is quicker if you already know the name and just want to see it in the Data Editor.

In a similar vein, variable sets can be defined via Utilities > Define Variable Sets, and one or more used via Utilities > Use Variable Sets. This limits the variables displayed in a dialog or the Data Editor to those appearing in at least one of the sets in use while not affecting the data file contents. So John could, for example, define a set for each year's variables and operate on one or a few years at a time more easily. The set definitions are saved with the data.

Once you have defined variable sets, you can use the SETSMACRO extension command to create a macro listing the variables in the selected variable sets for use in syntax.

Finally, the SPSSINC SELECT VARIABLES extension command (Utilities > Define Variable Macro) allows you to define a macro listing variables based on variable metadata such as patterns in the names, variable types, measurement level, custom attributes, etc.

On Thu, Oct 11, 2018 at 10:32 AM John F Hall <[hidden email]> wrote:

Tony
Just tried that. Never done it before, but great to know. I'll use it in the new work-through demonstration. Is all this stuff documented for idiot users like me?
John

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email: [hidden email]
Website: Journeys in Survey Research
Course: Survey Analysis Workshop (SPSS)
Research: Subjective Social Indicators (Quality of Life)

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Anthony Babinec
Sent: 11 October 2018 18:07
To: [hidden email]
Subject: FW: Finding variables with the GUI

Edit menu has a Go to Variable selection. It lists variables in alphabetic order. When you select a variable, it takes you to that column in the Data Editor window. This of course is not the same as the "typomatic" facility being discussed.

Tony Babinec

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Bruce Weaver
Sent: Thursday, October 11, 2018 10:47 AM
To: [hidden email]
Subject: Re: Finding variables with the GUI

A search option for Utilities > Variables would be useful sometimes too.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon K Peck
[hidden email]

spss.giesel@yahoo.de

Re: Finding variables with the GUI

Hi, John F.,

looks like a lot of work! To me the most critical is unequal value labels.
But metadata information would be needed to be sure.
Obviously, you have to homogenize this meta information for each data set.

A possible approach would be: For each file and variable 
a) generate a new homogenized variable with favored attributes (variable name, variable labels, value labels, type, measurement level, missing values, width etc.)
b) generate its values from source variables (RECODE, IF, DO IF)
c) save file as newest sav version keeping only the homogenized variables

SPSS macros can probably make the job less cumbersome if you have a lot of variables.

GL,
Mario

Mario Giesel
Munich, Germany

Am Donnerstag, 11. Oktober 2018, 19:48:49 MESZ hat Jon Peck <[hidden email]> Folgendes geschrieben:

First letter search in dialog boxes is not documented by SPSS, because it is a standard Windows action. The multi-letter navigation in a dialog box is news to me, but it seems to work.

Once you have defined variable sets, you can use the SETSMACRO extension command to create a macro listing the variables in the selected variable sets for use in syntax.

On Thu, Oct 11, 2018 at 10:32 AM John F Hall <[hidden email]> wrote:

Tony
Just tried that. Never done it before, but great to know. I'll use it in the new work-through demonstration. Is all this stuff documented for idiot users like me?
John

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email: [hidden email]
Website: Journeys in Survey Research
Course: Survey Analysis Workshop (SPSS)
Research: Subjective Social Indicators (Quality of Life)

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Anthony Babinec
Sent: 11 October 2018 18:07
To: [hidden email]
Subject: FW: Finding variables with the GUI

Edit menu has a Go to Variable selection. It lists variables in alphabetic order. When you select a variable, it takes you to that column in the Data Editor window. This of course is not the same as the "typomatic" facility being discussed.

Tony Babinec

-----Original Message-----
From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Bruce Weaver
Sent: Thursday, October 11, 2018 10:47 AM
To: [hidden email]
Subject: Re: Finding variables with the GUI

A search option for Utilities > Variables would be useful sometimes too.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon K Peck
[hidden email]

Bruce Weaver

Re: Finding variables with the GUI

Administrator

In reply to this post by Jon Peck

Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon Peck

Re: Finding variables with the GUI

Variable Sets were designed as a feature to help manage the gui when there are a lot of variables. The backend knows nothing about them (hence the extension command). So it was expected that users would create these interactively. There is a sneaky way to retrieve the definitions via a Python api, but there is no way to set them programmatically.

On Fri, Oct 12, 2018 at 8:40 AM Bruce Weaver <[hidden email]> wrote:

Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon K Peck
[hidden email]

Rick Oliver

Re: Finding variables with the GUI

In reply to this post by Bruce Weaver

It's GUI-specific functionality, although there are other GUI-specific features that can be controlled through syntax, like column width in the Data Editor.

On Fri, Oct 12, 2018 at 9:40 AM Bruce Weaver <[hidden email]> wrote:

Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

John F Hall

Re: Finding variables with the GUI

The information about Utilities > Define Variable Sets and Utilities > Use Variable Sets looked useful, but it was just as complex to use.

I'm using data from the British Social Attitudes Survey to explore the following research questions:

1: Is there a difference between the earnings (from paid work) of men and women?

2: What other variables might account for differences in earnings?

3: What effect do they have by themselves?

4: What happens to any differences in earnings between men and women when controlling for these other variables?

(See page: https://surveyresearch.weebly.com/32-three-or-more-variables.html)

I now wish to extend the tutorials to cover all 32 waves of the survey 1983 to 2015 (no survey 1988 or 1992). The analysis uses earnings groups based on quartiles (separately derived for each year). CROSSTABS and CTABLES are then used to produce zero-, first-, second- and n-order contingency tables from which epsilon (percentage point difference) is calculated for the difference between men and women in earnings from paid work. (Grateful acknowledgments to Jon Peck for Python, Jignesh Sutar for CTABLES syntax and Bruce Weaver for Elaboration macro).

An intermediate step is to generate a cumulative SPSS "mother" file containing a reduced data set as above, from which I can generate 32 mutually compatible data sets, any two or more of which can then be combined to enable analysis across time.

To date I have been using clunky syntax SAVE OUT <pathway\filename>/KEEP <varlist> . If one or more vars are not in the file, SPSS reports an error and the culprit names have to be edited out. This means that the varlist is different for each year, depending on which varnames are present, and results in a multiplicity of *.sps files. It doesn't help that the coding of earnings intervals changes halfway through the series and then changes from earning per year to earnings per month, but that's a different problem.

Utilities > Define Variable Sets

After highlighting the first variable, you can find the variable you want by typing the first (few) letters of the varname (very quickly). If you try to find a second variable which is above that one in the list, it can't find it, so you have to go back up the list to a point above the next variable.

Use Variable Sets

When you add the variable set, the Data Editor appears full size displaying row numbers and selected variables, but with all other variables pale blue-d out. I can get the same info from DISPLAY LABELS, but that didn't work on the blanked out matrix: the output was for all 1080 variables. I tried File > save as >, but that didn't work either.

Thanks for all the comments, but my time is too precious: think I'll stick to clunky for now.

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Rick Oliver
Sent: 12 October 2018 18:53
To: [hidden email]
Subject: Re: Finding variables with the GUI

It's GUI-specific functionality, although there are other GUI-specific features that can be controlled through syntax, like column width in the Data Editor.

On Fri, Oct 12, 2018 at 9:40 AM Bruce Weaver <[hidden email]> wrote:

Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Jon Peck

Re: Finding variables with the GUI

As I said, the backend does not know about variable sets, although the SETSMACRO extension command could generate a variable list from one.

However if the variable names by year have a pattern in the names such as containing the year at the end, you could use the SPSSINC SELECT VARIABLES to define the list for a year and use that in your KEEP.

On Fri, Oct 12, 2018 at 11:55 PM John F Hall <[hidden email]> wrote:

The information about Utilities > Define Variable Sets and Utilities > Use Variable Sets looked useful, but it was just as complex to use.

I'm using data from the British Social Attitudes Survey to explore the following research questions:

1:    Is there a difference between the earnings (from paid work) of men and women?
2:    What other variables might account for differences in earnings?
3:    What effect do they have by themselves?

4:   What happens to any differences in earnings between men and women when controlling for these other variables?

(See page: https://surveyresearch.weebly.com/32-three-or-more-variables.html)

I now wish to extend the tutorials to cover all 32 waves of the survey 1983 to 2015 (no survey 1988 or 1992).   The analysis uses earnings groups based on quartiles (separately derived for each year). CROSSTABS and CTABLES are then used to produce zero-, first-, second- and n-order contingency tables from which epsilon (percentage point difference) is calculated for the difference between men and women in earnings from paid work. (Grateful acknowledgments to Jon Peck for Python, Jignesh Sutar for CTABLES syntax and Bruce Weaver for Elaboration macro).

An intermediate step is to generate a cumulative SPSS "mother" file containing a reduced data set as above, from which I can generate 32 mutually compatible data sets, any two or more of which can then be combined to enable analysis across time.

To date I have been using clunky syntax SAVE OUT <pathway\filename>/KEEP <varlist> . If one or more vars are not in the file, SPSS reports an error and the culprit names have to be edited out. This means that the varlist is different for each year, depending on which varnames are present, and results in a multiplicity of *.sps files. It doesn't help that the coding of earnings intervals changes halfway through the series and then changes from earning per year to earnings per month, but that's a different problem.

Utilities > Define Variable Sets

After highlighting the first variable, you can find the variable you want by typing the first (few) letters of the varname (very quickly). If you try to find a second variable which is above that one in the list, it can't find it, so you have to go back up the list to a point above the next variable.

Use Variable Sets

When you add the variable set, the Data Editor appears full size displaying row numbers and selected variables, but with all other variables pale blue-d out. I can get the same info from DISPLAY LABELS, but that didn't work on the blanked out matrix: the output was for all 1080 variables. I tried File > save as >, but that didn't work either.

Thanks for all the comments, but my time is too precious: think I'll stick to clunky for now.

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Rick Oliver
Sent: 12 October 2018 18:53

To: [hidden email]
Subject: Re: Finding variables with the GUI

It's GUI-specific functionality, although there are other GUI-specific features that can be controlled through syntax, like column width in the Data Editor.

On Fri, Oct 12, 2018 at 9:40 AM Bruce Weaver <[hidden email]> wrote:
Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

Jon K Peck
[hidden email]

Jon Peck

Re: Finding variables with the GUI

Here’s another thought, albeit with more initial work. If you create a custom attribute (or attribute array) in your master file listing all the years for which a variable exists, you could then generate an exact list of all the applicable variables using SPSSINC SELECT VARIABLES as a macro for use in your SAVE or other syntax. Variable attributes can be create with VARIABLE ATTRIBUTE syntax, so TO or even other macro definitions could save some manual effort.

On Sat, Oct 13, 2018 at 7:46 AM Jon Peck <[hidden email]> wrote:

As I said, the backend does not know about variable sets, although the SETSMACRO extension command could generate a variable list from one.

However if the variable names by year have a pattern in the names such as containing the year at the end, you could use the SPSSINC SELECT VARIABLES to define the list for a year and use that in your KEEP.

On Fri, Oct 12, 2018 at 11:55 PM John F Hall <[hidden email]> wrote:
The information about Utilities > Define Variable Sets and Utilities > Use Variable Sets looked useful, but it was just as complex to use.

I'm using data from the British Social Attitudes Survey to explore the following research questions:

1:    Is there a difference between the earnings (from paid work) of men and women?
2:    What other variables might account for differences in earnings?
3:    What effect do they have by themselves?

4:   What happens to any differences in earnings between men and women when controlling for these other variables?

(See page: https://surveyresearch.weebly.com/32-three-or-more-variables.html)

I now wish to extend the tutorials to cover all 32 waves of the survey 1983 to 2015 (no survey 1988 or 1992).   The analysis uses earnings groups based on quartiles (separately derived for each year). CROSSTABS and CTABLES are then used to produce zero-, first-, second- and n-order contingency tables from which epsilon (percentage point difference) is calculated for the difference between men and women in earnings from paid work. (Grateful acknowledgments to Jon Peck for Python, Jignesh Sutar for CTABLES syntax and Bruce Weaver for Elaboration macro).

An intermediate step is to generate a cumulative SPSS "mother" file containing a reduced data set as above, from which I can generate 32 mutually compatible data sets, any two or more of which can then be combined to enable analysis across time.

To date I have been using clunky syntax SAVE OUT <pathway\filename>/KEEP <varlist> . If one or more vars are not in the file, SPSS reports an error and the culprit names have to be edited out. This means that the varlist is different for each year, depending on which varnames are present, and results in a multiplicity of *.sps files. It doesn't help that the coding of earnings intervals changes halfway through the series and then changes from earning per year to earnings per month, but that's a different problem.

Utilities > Define Variable Sets

After highlighting the first variable, you can find the variable you want by typing the first (few) letters of the varname (very quickly). If you try to find a second variable which is above that one in the list, it can't find it, so you have to go back up the list to a point above the next variable.

Use Variable Sets

When you add the variable set, the Data Editor appears full size displaying row numbers and selected variables, but with all other variables pale blue-d out. I can get the same info from DISPLAY LABELS, but that didn't work on the blanked out matrix: the output was for all 1080 variables. I tried File > save as >, but that didn't work either.

Thanks for all the comments, but my time is too precious: think I'll stick to clunky for now.

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Rick Oliver
Sent: 12 October 2018 18:53

To: [hidden email]
Subject: Re: Finding variables with the GUI

It's GUI-specific functionality, although there are other GUI-specific features that can be controlled through syntax, like column width in the Data Editor.

On Fri, Oct 12, 2018 at 9:40 AM Bruce Weaver <[hidden email]> wrote:
Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Jon K Peck
[hidden email]

Jon K Peck
[hidden email]

John F Hall

Re: Finding variables with the GUI

Jon

Sounds complicated, but I've solved the problem using Rich's suggestion of RANK, with a modification by me which saves having to run separate jobs on each year file. See subject: [RE: Estimating actual earnings from earnings groups]

John

John F Hall MA (Cantab) Dip Ed (Dunelm)

[Retired academic survey researcher]

Email: [hidden email]

Website: Journeys in Survey Research

Course: Survey Analysis Workshop (SPSS)

Research: Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Jon Peck
Sent: 13 October 2018 21:03
To: [hidden email]
Subject: Re: Finding variables with the GUI

On Sat, Oct 13, 2018 at 7:46 AM Jon Peck <[hidden email]> wrote:

As I said, the backend does not know about variable sets, although the SETSMACRO extension command could generate a variable list from one.

However if the variable names by year have a pattern in the names such as containing the year at the end, you could use the SPSSINC SELECT VARIABLES to define the list for a year and use that in your KEEP.

On Fri, Oct 12, 2018 at 11:55 PM John F Hall <[hidden email]> wrote:
The information about Utilities > Define Variable Sets and Utilities > Use Variable Sets looked useful, but it was just as complex to use.

I'm using data from the British Social Attitudes Survey to explore the following research questions:

1:    Is there a difference between the earnings (from paid work) of men and women?
2:    What other variables might account for differences in earnings?
3:    What effect do they have by themselves?

4:   What happens to any differences in earnings between men and women when controlling for these other variables?

(See page: https://surveyresearch.weebly.com/32-three-or-more-variables.html)

I now wish to extend the tutorials to cover all 32 waves of the survey 1983 to 2015 (no survey 1988 or 1992).   The analysis uses earnings groups based on quartiles (separately derived for each year). CROSSTABS and CTABLES are then used to produce zero-, first-, second- and n-order contingency tables from which epsilon (percentage point difference) is calculated for the difference between men and women in earnings from paid work. (Grateful acknowledgments to Jon Peck for Python, Jignesh Sutar for CTABLES syntax and Bruce Weaver for Elaboration macro).

An intermediate step is to generate a cumulative SPSS "mother" file containing a reduced data set as above, from which I can generate 32 mutually compatible data sets, any two or more of which can then be combined to enable analysis across time.

To date I have been using clunky syntax SAVE OUT <pathway\filename>/KEEP <varlist> . If one or more vars are not in the file, SPSS reports an error and the culprit names have to be edited out. This means that the varlist is different for each year, depending on which varnames are present, and results in a multiplicity of *.sps files. It doesn't help that the coding of earnings intervals changes halfway through the series and then changes from earning per year to earnings per month, but that's a different problem.

Utilities > Define Variable Sets

After highlighting the first variable, you can find the variable you want by typing the first (few) letters of the varname (very quickly). If you try to find a second variable which is above that one in the list, it can't find it, so you have to go back up the list to a point above the next variable.

Use Variable Sets

When you add the variable set, the Data Editor appears full size displaying row numbers and selected variables, but with all other variables pale blue-d out. I can get the same info from DISPLAY LABELS, but that didn't work on the blanked out matrix: the output was for all 1080 variables. I tried File > save as >, but that didn't work either.

Thanks for all the comments, but my time is too precious: think I'll stick to clunky for now.

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]

Email:          [hidden email]
Website:     Journeys in Survey Research
Course:       Survey Analysis Workshop (SPSS)
Research:   Subjective Social Indicators (Quality of Life)

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Rick Oliver
Sent: 12 October 2018 18:53

To: [hidden email]
Subject: Re: Finding variables with the GUI

It's GUI-specific functionality, although there are other GUI-specific features that can be controlled through syntax, like column width in the Data Editor.

On Fri, Oct 12, 2018 at 9:40 AM Bruce Weaver <[hidden email]> wrote:
Jon, there appears to be no command syntax for defining or using variable
sets. Why would that be?

Jon Peck wrote
> --- snip ---
> In a similar vein, variable sets can be defined via Utilities > Define
> Variable Sets, and one or more used via Utilities > Use Variable Sets.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Jon K Peck
[hidden email]

Jon K Peck
[hidden email]