Same variable, different files + spv copying

[John F Hall]

Dear listers

To date the SPSS tutorials on my website have used data from surveys conducted in the UK in the 1970s and 1980s, mainly because I used the data from these in the courses I taught from 1976 until I took early retirement in 1992.  These surveys were done either by myself or by close professional colleagues, and I had been involved in design, capture and analysis of the resulting data. 

However I now wish to start using data from more recent general population surveys in the public domain, especially if they are used in such textbooks as Babbie et al “Adventures in Social Research” (http://www.sagepub.com/books/Book237181?siteId=sage-us&prodTypes=any&q=Zaino&pageTitle=productsSearch)   which uses data from the NORC General Social Survey (http://www.norc.org/Research/Projects/Pages/general-social-survey.aspx) but I would be replicating their exercises and examples using SPSS syntax rather than the GUI.  More immediately I intend to use data from social science/social policy series such as the British Social Attitudes series (http://www.britsocat.com/Body.aspx?control=HomePage) and the European Social Survey (http://www.europeansocialsurvey.org/). 

I am currently working on some new SPSS tutorials for, and familiarising myself with, data from five waves of the biennial European Social Survey currently available (2002, 2004, 2006, 2008, 2010): the fieldwork for round 6 is due to start in September 2012.  The full European data set has around 50,000 cases per wave, but for the moment I am working with British data only (around 2,000 cases per wave).

The questionnaire is arranged in modules, not all of which are repeated in each wave.  The question numbers are always the same for each wave (although not always asked) and the corresponding  mnemonic variable names are always the same in the SPSS saved files.  Because the question numbers do not appear in the variable names or in the variable labels, this makes it difficult to work direct  from the questionnaire.  All the variable names in the files are limited to 8 characters ( a legacy from the old days). 

For instance, question B24 is always the same in each wave and the variable is always called stflife.

B24   All things considered, how satisfied are you with your life as a whole nowadays?

         Please answer using this card, where 0 means extremely dissatisfied and 10 means extremely satisfied.

 

Uses a showcard 10 with horizontal scale

 

Extremely dissatisfied [0]         ~ ~ ~ ~           Extremely dissatisfied [10]

 

It appears in all the files as:

---

Name              Label

 

stflife                How satisfied with life as a whole

---

There 662 variables in the file, so you can imagine the tortuous naming involved!

IMHO this is not best practice.  My preferred way of working would be to rename all the variables using the question numbers, or at least add the question numbers to the beginning of all the variable labels, then create a new set of saved files.  This way users can have open both the questionnaire pdf and the SPSS saved files, then simply scroll up and down to find what they need.  The early waves of the British Social Attitudes series had the same problem until I suggested they at least add question numbers to the variable labels: this they now do.

My question to the list concerns the best way of comparing responses across waves of the ESS.  For instance, I have produced summary statistics, frequency tables and barcharts for stflife from each of the waves 2002 to 2010, but in a cumbersome way, only one at a time by opening each file in turn and then running the same syntax:

freq stflife /sta mea med /bar .                   [to get five tables in the following format]

 

stflife
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	0	23	1.1	1.1	1.1
	1	15	.7	.7	1.9
	2	32	1.6	1.6	3.4
	3	84	4.1	4.1	7.5
	4	89	4.3	4.4	11.9
	5	223	10.9	10.9	22.8
	6	186	9.1	9.1	31.9
	7	393	19.2	19.2	51.1
	8	523	25.5	25.6	76.7
	9	262	12.8	12.8	89.5
	10	215	10.5	10.5	100.0
	Total	2045	99.7	100.0
Missing	77	1	.0
	88	6	.3
	Total	7	.3
Total		2052	100.0

 

 

Question 1:

When I created a 7 col x 12 row blank table in a Word doc and then tried to copy the Valid % column in the spv output into a column in  the Word table, it copied the whole column into a single cell.  Is there a way round this?

Question 2:

I can use MATCH FILES  ~ ~ ~/KEEP to create a file with stflife and essround (wave year) but is there a syntax way of doing the same operation successively on the five files, all of which will have been already dragged into SPSS, using dataset1 – dataset5 or file handles?

Thanks in advance: nothing too fancy, mind!

John F Hall (Mr)

 

Email:     [hidden email]

Website: www.surveyresearch.weebly.com

 

 

 

 

[Albert-Jan Roskam]

Hi John,

 

Ad 1.: Not sure which spss version you are using, but have you tried OUTPUT EXPORT to route the output item(s) to e.g. xls?

Ad 2. Perhaps you can use ADD FILES followed by CASESTOVARS:

add files / file = file 1 / file = file2 / file = file 3 / file = / file4 / file = file5 / keep = hid pid stflife essround.

sort cases by hid pid stflife.

casestovars / id = hid pid / index = stflife.

 (hid, pid = household, personal id)

 

Regarding the variable names: I must say variable names like "B24" make my eyes bleed ;-) The name "stflife" is (slightly) more descriptive. I agree with you that the the question number should be included in the variable label.

If the codebook is good enough (for copy/pasting), you could do that with some Python code.

 

Regards,
Albert-Jan

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a fresh

water system, and public health, what have the Romans ever done for us?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 

From: John F Hall <[hidden email]>
To: [hidden email]
Sent: Friday, August 24, 2012 12:54 PM
Subject: [SPSSX-L] Same variable, different files + spv copying

Dear listers

To date the SPSS tutorials on my website have used data from surveys conducted in the UK in the 1970s and 1980s, mainly because I used the data from these in the courses I taught from 1976 until I took early retirement in 1992.  These surveys were done either by myself or by close professional colleagues, and I had been involved in design, capture and analysis of the resulting data. 

However I now wish to start using data from more recent general population surveys in the public domain, especially if they are used in such textbooks as Babbie et al “Adventures in Social Research” (http://www.sagepub.com/books/Book237181?siteId=sage-us&prodTypes=any&q=Zaino&pageTitle=productsSearch)   which uses data from the NORC General Social Survey (http://www.norc.org/Research/Projects/Pages/general-social-survey.aspx) but I would be replicating their exercises and examples using SPSS syntax rather than the GUI.  More immediately I intend to use data from social science/social policy series such as the British Social Attitudes series (http://www.britsocat.com/Body.aspx?control=HomePage) and the European Social Survey (http://www.europeansocialsurvey.org/). 

I am currently working on some new SPSS tutorials for, and familiarising myself with, data from five waves of the biennial European Social Survey currently available (2002, 2004, 2006, 2008, 2010): the fieldwork for round 6 is due to start in September 2012.  The full European data set has around 50,000 cases per wave, but for the moment I am working with British data only (around 2,000 cases per wave).

The questionnaire is arranged in modules, not all of which are repeated in each wave.  The question numbers are always the same for each wave (although not always asked) and the corresponding  mnemonic variable names are always the same in the SPSS saved files.  Because the question numbers do not appear in the variable names or in the variable labels, this makes it difficult to work direct  from the questionnaire.  All the variable names in the files are limited to 8 characters ( a legacy from the old days). 

For instance, question B24 is always the same in each wave and the variable is always called stflife.

B24   All things considered, how satisfied are you with your life as a whole nowadays?

         Please answer using this card, where 0 means extremely dissatisfied and 10 means extremely satisfied.

 

Uses a showcard 10 with horizontal scale

 

Extremely dissatisfied [0]         ~ ~ ~ ~           Extremely dissatisfied [10]

 

It appears in all the files as:

Name              Label

 

stflife                How satisfied with life as a whole

There 662 variables in the file, so you can imagine the tortuous naming involved!

IMHO this is not best practice.  My preferred way of working would be to rename all the variables using the question numbers, or at least add the question numbers to the beginning of all the variable labels, then create a new set of saved files.  This way users can have open both the questionnaire pdf and the SPSS saved files, then simply scroll up and down to find what they need.  The early waves of the British Social Attitudes series had the same problem until I suggested they at least add question numbers to the variable labels: this they now do.

My question to the list concerns the best way of comparing responses across waves of the ESS.  For instance, I have produced summary statistics, frequency tables and barcharts for stflife from each of the waves 2002 to 2010, but in a cumbersome way, only one at a time by opening each file in turn and then running the same syntax:

freq stflife /sta mea med /bar .                   [to get five tables in the following format]

 

stflife
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	0	23	1.1	1.1	1.1
	1	15	.7	.7	1.9
	2	32	1.6	1.6	3.4
	3	84	4.1	4.1	7.5
	4	89	4.3	4.4	11.9
	5	223	10.9	10.9	22.8
	6	186	9.1	9.1	31.9
	7	393	19.2	19.2	51.1
	8	523	25.5	25.6	76.7
	9	262	12.8	12.8	89.5
	10	215	10.5	10.5	100.0
	Total	2045	99.7	100.0
Missing	77	1	.0
	88	6	.3
	Total	7	.3
Total		2052	100.0

 

 

Question 1:

When I created a 7 col x 12 row blank table in a Word doc and then tried to copy the Valid % column in the spv output into a column in  the Word table, it copied the whole column into a single cell.  Is there a way round this?

Question 2:

I can use MATCH FILES  ~ ~ ~/KEEP to create a file with stflife and essround (wave year) but is there a syntax way of doing the same operation successively on the five files, all of which will have been already dragged into SPSS, using dataset1 – dataset5 or file handles?

Thanks in advance: nothing too fancy, mind!

John F Hall (Mr)

 

Email:     [hidden email]

Website: www.surveyresearch.weebly.com

 

 

 

 

[David Marso]

Consider ADD FILES and a LAYERED SPLIT FILE?

John F Hall wrote

Dear listers

To date the SPSS tutorials on my website have used data from surveys
conducted in the UK in the 1970s and 1980s, mainly because I used the data
from these in the courses I taught from 1976 until I took early retirement
in 1992.  These surveys were done either by myself or by close professional
colleagues, and I had been involved in design, capture and analysis of the
resulting data.

However I now wish to start using data from more recent general population
surveys in the public domain, especially if they are used in such textbooks
as Babbie et al "Adventures in Social Research"
(http://www.sagepub.com/books/Book237181?siteId=sage-us
<http://www.sagepub.com/books/Book237181?siteId=sage-us&prodTypes=any&q=Zain
o&pageTitle=productsSearch> &prodTypes=any&q=Zaino&pageTitle=productsSearch)
which uses data from the NORC General Social Survey
(http://www.norc.org/Research/Projects/Pages/general-social-survey.aspx) but
I would be replicating their exercises and examples using SPSS syntax rather
than the GUI.  More immediately I intend to use data from social
science/social policy series such as the British Social Attitudes series
(http://www.britsocat.com/Body.aspx?control=HomePage) and the European
Social Survey (http://www.europeansocialsurvey.org/).

I am currently working on some new SPSS tutorials for, and familiarising
myself with, data from five waves of the biennial European Social Survey
currently available (2002, 2004, 2006, 2008, 2010): the fieldwork for round
6 is due to start in September 2012.  The full European data set has around
50,000 cases per wave, but for the moment I am working with British data
only (around 2,000 cases per wave).

The questionnaire is arranged in modules, not all of which are repeated in
each wave.  The question numbers are always the same for each wave (although
not always asked) and the corresponding  mnemonic variable names are always
the same in the SPSS saved files.  Because the question numbers do not
appear in the variable names or in the variable labels, this makes it
difficult to work direct  from the questionnaire.  All the variable names in
the files are limited to 8 characters ( a legacy from the old days).

For instance, question B24 is always the same in each wave and the variable
is always called stflife.

B24   All things considered, how satisfied are you with your life as a whole
nowadays?

         Please answer using this card, where 0 means extremely dissatisfied
and 10 means extremely satisfied.



Uses a showcard 10 with horizontal scale



Extremely dissatisfied [0]         ~ ~ ~ ~           Extremely dissatisfied
[10]



It appears in all the files as:

  _____

Name              Label



stflife                How satisfied with life as a whole

  _____

There 662 variables in the file, so you can imagine the tortuous naming
involved!

IMHO this is not best practice.  My preferred way of working would be to
rename all the variables using the question numbers, or at least add the
question numbers to the beginning of all the variable labels, then create a
new set of saved files.  This way users can have open both the questionnaire
pdf and the SPSS saved files, then simply scroll up and down to find what
they need.  The early waves of the British Social Attitudes series had the
same problem until I suggested they at least add question numbers to the
variable labels: this they now do.

My question to the list concerns the best way of comparing responses across
waves of the ESS.  For instance, I have produced summary statistics,
frequency tables and barcharts for stflife from each of the waves 2002 to
2010, but in a cumbersome way, only one at a time by opening each file in
turn and then running the same syntax:

freq stflife /sta mea med /bar .                   [to get five tables in
the following format]




stflife




Frequency

Percent

Valid Percent

Cumulative Percent


Valid

0

23

1.1

1.1

1.1


1

15

.7

.7

1.9


2

32

1.6

1.6

3.4


3

84

4.1

4.1

7.5


4

89

4.3

4.4

11.9


5

223

10.9

10.9

22.8


6

186

9.1

9.1

31.9


7

393

19.2

19.2

51.1


8

523

25.5

25.6

76.7


9

262

12.8

12.8

89.5


10

215

10.5

10.5

100.0


Total

2045

99.7

100.0




Missing

77

1

.0






88

6

.3






Total

7

.3






Total

2052

100.0









Question 1:

When I created a 7 col x 12 row blank table in a Word doc and then tried to
copy the Valid % column in the spv output into a column in  the Word table,
it copied the whole column into a single cell.  Is there a way round this?

Question 2:

I can use MATCH FILES  ~ ~ ~/KEEP to create a file with stflife and essround
(wave year) but is there a syntax way of doing the same operation
successively on the five files, all of which will have been already dragged
into SPSS, using dataset1 - dataset5 or file handles?

Thanks in advance: nothing too fancy, mind!

John F Hall (Mr)



Email:      <mailto:[hidden email]> [hidden email]

Website:  <http://surveyresearch.weebly.com/> www.surveyresearch.weebly.com

[John F Hall]

Albert-Jan, David

Thanks for pointing me to ADD FILES: I'd forgotten about that and it's a
long time since I did anything with files like these.  I can probably get
what I want with CROSSTABS and MEANS.

I'm sticking with varnames like B24 as, in the SPSS saved file, I can then
easily find it as all substantive questions and variables are in
alphabetical and questionnaire order and, in the questionnaire, I can
immediately find the question and see what the actual wording was.

This same question asking about "your life as a whole" was used in my
Quality of Life in Britain surveys in 1971 and 1975, also with 0-10 scale,
but with a vertical ladder rather than horizontal, and used "how satisfied
or dissatisfied are you" rather than the ESS "how satisfied are you"
(possible response bias there).

At least this question is comparable, but other questions in the ESS (eg
3-item Trust in Others scale) have 0-10 scales, whereas in my QoL survey,
with the same wording (co-ordinated with the ISR Quality of American Life
survey) had Yes/No.  The ISR study and my 1973 QoL survey used 1-7 scales
instead of 0-10 and some of the British Social Attitudes scales are 1-5, all
of which makes comparison difficult, but raises interesting methodological
questions for students of survey research.  In any case I've long suspected
that mean life-satisfaction on a 0-10 scale is a constant rather than a
variable.


John F Hall (Mr)

Email:    [hidden email]
Website: www.surveyresearch.weebly.com






-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
David Marso
Sent: 24 August 2012 15:36
To: [hidden email]
Subject: Re: Same variable, different files + spv copying

Consider ADD FILES and a LAYERED SPLIT FILE?


John F Hall wrote

>
> Dear listers
>
> To date the SPSS tutorials on my website have used data from surveys
> conducted in the UK in the 1970s and 1980s, mainly because I used the
> data from these in the courses I taught from 1976 until I took early
> retirement in 1992.  These surveys were done either by myself or by
> close professional colleagues, and I had been involved in design,
> capture and analysis of the resulting data.
>
> However I now wish to start using data from more recent general
> population surveys in the public domain, especially if they are used
> in such textbooks as Babbie et al "Adventures in Social Research"
> (http://www.sagepub.com/books/Book237181?siteId=sage-us
> &lt;http://www.sagepub.com/books/Book237181?siteId=sage-us&amp;prodTyp
> es=any&amp;q=Zain
> o&amp;pageTitle=productsSearch&gt;
> &prodTypes=any&q=Zaino&pageTitle=productsSearch)
> which uses data from the NORC General Social Survey
> (http://www.norc.org/Research/Projects/Pages/general-social-survey.asp
> x)
> but
> I would be replicating their exercises and examples using SPSS syntax
> rather than the GUI.  More immediately I intend to use data from
> social science/social policy series such as the British Social
> Attitudes series
> (http://www.britsocat.com/Body.aspx?control=HomePage) and the European
> Social Survey (http://www.europeansocialsurvey.org/).
>
> I am currently working on some new SPSS tutorials for, and
> familiarising myself with, data from five waves of the biennial
> European Social Survey currently available (2002, 2004, 2006, 2008,
> 2010): the fieldwork for round
> 6 is due to start in September 2012.  The full European data set has
> around
> 50,000 cases per wave, but for the moment I am working with British
> data only (around 2,000 cases per wave).
>
> The questionnaire is arranged in modules, not all of which are
> repeated in each wave.  The question numbers are always the same for
> each wave (although not always asked) and the corresponding  mnemonic
> variable names are always the same in the SPSS saved files.  Because
> the question numbers do not appear in the variable names or in the
> variable labels, this makes it difficult to work direct  from the
> questionnaire.  All the variable names in the files are limited to 8
> characters ( a legacy from the old days).
>
> For instance, question B24 is always the same in each wave and the
> variable is always called stflife.
>
> B24   All things considered, how satisfied are you with your life as a
> whole
> nowadays?
>
>          Please answer using this card, where 0 means extremely
> dissatisfied and 10 means extremely satisfied.
>
>
>
> Uses a showcard 10 with horizontal scale
>
>
>
> Extremely dissatisfied [0]         ~ ~ ~ ~           Extremely
> dissatisfied
> [10]
>
>
>
> It appears in all the files as:
>
>   _____
>
> Name              Label
>
>
>
> stflife                How satisfied with life as a whole
>
>   _____
>
> There 662 variables in the file, so you can imagine the tortuous
> naming involved!
>
> IMHO this is not best practice.  My preferred way of working would be
> to rename all the variables using the question numbers, or at least
> add the question numbers to the beginning of all the variable labels,
> then create a new set of saved files.  This way users can have open
> both the questionnaire pdf and the SPSS saved files, then simply
> scroll up and down to find what they need.  The early waves of the
> British Social Attitudes series had the same problem until I suggested
> they at least add question numbers to the variable labels: this they
> now do.
>
> My question to the list concerns the best way of comparing responses
> across waves of the ESS.  For instance, I have produced summary
> statistics, frequency tables and barcharts for stflife from each of
> the waves 2002 to 2010, but in a cumbersome way, only one at a time by
> opening each file in turn and then running the same syntax:
>
> freq stflife /sta mea med /bar .                   [to get five tables in
> the following format]
>
>
>
>
> stflife
>
>
>
>
> Frequency
>
> Percent
>
> Valid Percent
>
> Cumulative Percent
>
>
> Valid
>
> 0
>
> 23
>
> 1.1
>
> 1.1
>
> 1.1
>
>
> 1
>
> 15
>
> .7
>
> .7
>
> 1.9
>
>
> 2
>
> 32
>
> 1.6
>
> 1.6
>
> 3.4
>
>
> 3
>
> 84
>
> 4.1
>
> 4.1
>
> 7.5
>
>
> 4
>
> 89
>
> 4.3
>
> 4.4
>
> 11.9
>
>
> 5
>
> 223
>
> 10.9
>
> 10.9
>
> 22.8
>
>
> 6
>
> 186
>
> 9.1
>
> 9.1
>
> 31.9
>
>
> 7
>
> 393
>
> 19.2
>
> 19.2
>
> 51.1
>
>
> 8
>
> 523
>
> 25.5
>
> 25.6
>
> 76.7
>
>
> 9
>
> 262
>
> 12.8
>
> 12.8
>
> 89.5
>
>
> 10
>
> 215
>
> 10.5
>
> 10.5
>
> 100.0
>
>
> Total
>
> 2045
>
> 99.7
>
> 100.0
>
>
>
>
> Missing
>
> 77
>
> 1
>
> .0
>
>
>
>
>
>
> 88
>
> 6
>
> .3
>
>
>
>
>
>
> Total
>
> 7
>
> .3
>
>
>
>
>
>
> Total
>
> 2052
>
> 100.0
>
>
>
>
>
>
>
>
>
> Question 1:
>
> When I created a 7 col x 12 row blank table in a Word doc and then
> tried to copy the Valid % column in the spv output into a column in
> the Word table, it copied the whole column into a single cell.  Is
> there a way round this?
>
> Question 2:
>
> I can use MATCH FILES  ~ ~ ~/KEEP to create a file with stflife and
> essround (wave year) but is there a syntax way of doing the same
> operation successively on the five files, all of which will have been
> already dragged into SPSS, using dataset1 - dataset5 or file handles?
>
> Thanks in advance: nothing too fancy, mind!
>
> John F Hall (Mr)
>
>
>
> Email:      &lt;mailto:johnfhall@&gt; johnfhall@
>
> Website:  &lt;http://surveyresearch.weebly.com/&gt;
> www.surveyresearch.weebly.com
>

-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email
me.
--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Same-variable-different-files-
spv-copying-tp5714840p5714844.html
Sent from the SPSSX Discussion mailing list archive at 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