SPSSX Discussion

May Day Conundrum

Classic

List

Threaded

13 messages Options

John F Hall

May Day Conundrum

I've been at this nearly all day, going right back to scratch and working one step at a time.

I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ "respondent earnings quartiles (dv):Q1377"

ExPrtFul "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable

FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Skip,not in paid work	1558	45.5	45.5	45.5
	1 less than 11999	395	11.5	11.5	57.1
	2 12000- 19999	414	12.1	12.1	69.2
	3 20000- 31999	467	13.7	13.7	82.8
	4 32000 or more	413	12.1	12.1	94.9
	7 Refused information	147	4.3	4.3	99.2
	8 Don't know	27	0.8	0.8	100.0
	Total	3421	100.0	100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Not employee	1919	56.1	56.1	56.1
	2 full-time 30+ hours/week	1080	31.6	31.6	87.7
	3 part-time	409	12.0	12.0	99.6
	8 Don't know	1	0.0	0.0	99.6
	9 Refusal	12	0.4	0.4	100.0
	Total	3421	100.0	100.0

CROSSTABS exprtful by rearnq.

Warnings

The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

*ExPrtFul Respondent: Is job full or part-time? :Q1007 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
Count
		REarnQ respondent earnings quartiles (dv):Q1377							Total
		-1 Skip,not in paid work	1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	7 Refused information	8 Don't know	Total
ExPrtFul Respondent: Is job full or part-time? :Q1007	Not employee	64	395	414	467	413	139	27	1919
	full-time 30+ hours/week	1080	0	0	0	0	0	0	1080
	part-time	409	0	0	0	0	0	0	409
	Don't know	1	0	0	0	0	0	0	1
	Refusal	4	0	0	0	0	8	0	12
Total		1558	395	414	467	413	147	27	3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Win7 Pro, SPSS 24

Email: [hidden email]

Website: http://surveyresearch.weebly.com/

SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

===================== 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

Rick Oliver

Re: May Day Conundrum

Well, if the two -1 categories were user-missing, that would explain the results, since Crosstabs automatically removes any completely blank rows or columns -- but they don't appear to be user-missing.

On Wed, May 1, 2019 at 12:32 PM John F Hall <[hidden email]> wrote:

I've been at this nearly all day, going right back to scratch and working one step at a time.
I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ            "respondent earnings quartiles (dv):Q1377"
ExPrtFul           "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable
FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377

Frequency
Percent
Valid Percent
Cumulative Percent
Valid
-1 Skip,not in paid work
1558
45.5
45.5
45.5
1 less than 11999
395
11.5
11.5
57.1
2 12000- 19999
414
12.1
12.1
69.2
3 20000- 31999
467
13.7
13.7
82.8
4 32000 or more
413
12.1
12.1
94.9
7 Refused information
147
4.3
4.3
99.2
8 Don't know
27
0.8
0.8
100.0
Total
3421
100.0
100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007

Frequency
Percent
Valid Percent
Cumulative Percent
Valid
-1 Not employee
1919
56.1
56.1
56.1
2 full-time 30+ hours/week
1080
31.6
31.6
87.7
3 part-time
409
12.0
12.0
99.6
8 Don't know
1
0.0
0.0
99.6
9 Refusal
12
0.4
0.4
100.0
Total
3421
100.0
100.0

CROSSTABS exprtful by rearnq.

Warnings
The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation
Count

REarnQ respondent earnings quartiles (dv):Q1377
Total
-1

Skip,not in paid work
1

less than 11999
2

12000- 19999
3

20000- 31999
4

32000 or more
7

Refused information
8

Don't know
ExPrtFul Respondent: Is job full or part-time? :Q1007
Not employee
64
395
414
467
413
139
27
1919
full-time 30+ hours/week
1080
0
0
0
0
0
0
1080
part-time
409
0
0
0
0
0
0
409
Don't know
1
0
0
0
0
0
0
1
Refusal
4
0
0
0
0
8
0
12
Total
1558
395
414
467
413
147
27
3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall
[Retired academic survey researcher]
IBM-SPSS Academic Author 9900074
Win7 Pro, SPSS 24

Email:             [hidden email]
Website:          http://surveyresearch.weebly.com/
SPSS course:   http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html
Research:        http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

===================== 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

Rich Ulrich

Re: May Day Conundrum

In reply to this post by John F Hall

I bet that you are wrong when you state that no Missing Values

are "declared" for either variable. Check that with Display Dictionary....

I also note the table shows results that I consider highly suspicious.

Namely, in the table produced when you include Missing, all the

income categories greater than 0 occur for the -1 group, "Not employed."

Rich Ulrich

From: SPSSX(r) Discussion <[hidden email]> on behalf of John F Hall <[hidden email]>
Sent: Wednesday, May 1, 2019 1:31 PM
To: [hidden email]
Subject: May Day Conundrum

I've been at this nearly all day, going right back to scratch and working one step at a time.

I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ "respondent earnings quartiles (dv):Q1377"

ExPrtFul "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable

FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Skip,not in paid work	1558	45.5	45.5	45.5
	1 less than 11999	395	11.5	11.5	57.1
	2 12000- 19999	414	12.1	12.1	69.2
	3 20000- 31999	467	13.7	13.7	82.8
	4 32000 or more	413	12.1	12.1	94.9
	7 Refused information	147	4.3	4.3	99.2
	8 Don't know	27	0.8	0.8	100.0
	Total	3421	100.0	100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Not employee	1919	56.1	56.1	56.1
	2 full-time 30+ hours/week	1080	31.6	31.6	87.7
	3 part-time	409	12.0	12.0	99.6
	8 Don't know	1	0.0	0.0	99.6
	9 Refusal	12	0.4	0.4	100.0
	Total	3421	100.0	100.0

CROSSTABS exprtful by rearnq.

Warnings

The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

*ExPrtFul Respondent: Is job full or part-time? :Q1007 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
Count
		REarnQ respondent earnings quartiles (dv):Q1377							Total
		-1 Skip,not in paid work	1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	7 Refused information	8 Don't know	Total
ExPrtFul Respondent: Is job full or part-time? :Q1007	Not employee	64	395	414	467	413	139	27	1919
	full-time 30+ hours/week	1080	0	0	0	0	0	0	1080
	part-time	409	0	0	0	0	0	0	409
	Don't know	1	0	0	0	0	0	0	1
	Refusal	4	0	0	0	0	8	0	12
Total		1558	395	414	467	413	147	27	3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Win7 Pro, SPSS 24

Email: [hidden email]

Website: http://surveyresearch.weebly.com/

SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

Rick Oliver

Re: May Day Conundrum

The Frequencies results indicate the categories are valid, but running Display Dictionary to confirm makes sense.

On Wed, May 1, 2019, 10:24 PM Rich Ulrich <[hidden email]> wrote:

I bet that you are wrong when you state that no Missing Values

are "declared" for either variable. Check that with Display Dictionary....

I also note the table shows results that I consider highly suspicious.

Namely, in the table produced when you include Missing, all the

income categories greater than 0 occur for the -1 group, "Not employed."

--

Rich Ulrich

From: SPSSX(r) Discussion <[hidden email]> on behalf of John F Hall <[hidden email]>
Sent: Wednesday, May 1, 2019 1:31 PM
To: [hidden email]
Subject: May Day Conundrum

I've been at this nearly all day, going right back to scratch and working one step at a time.

I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ            "respondent earnings quartiles (dv):Q1377"

ExPrtFul           "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable

FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

-1 Skip,not in paid work

1558

45.5

45.5

45.5

1 less than 11999

395

11.5

11.5

57.1

2 12000- 19999

414

12.1

12.1

69.2

3 20000- 31999

467

13.7

13.7

82.8

4 32000 or more

413

12.1

12.1

94.9

7 Refused information

147

4.3

4.3

99.2

8 Don't know

27

0.8

0.8

100.0

Total

3421

100.0

100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

-1 Not employee

1919

56.1

56.1

56.1

2 full-time 30+ hours/week

1080

31.6

31.6

87.7

3 part-time

409

12.0

12.0

99.6

8 Don't know

1

0.0

0.0

99.6

9 Refusal

12

0.4

0.4

100.0

Total

3421

100.0

100.0

CROSSTABS exprtful by rearnq.

Warnings

The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation

Count

REarnQ respondent earnings quartiles (dv):Q1377

Total

-1

Skip,not in paid work

1

less than 11999

2

12000- 19999

3

20000- 31999

4

32000 or more

7

Refused information

8

Don't know

ExPrtFul Respondent: Is job full or part-time? :Q1007

Not employee

64

395

414

467

413

139

27

1919

full-time 30+ hours/week

1080

0

0

0

0

0

0

1080

part-time

409

0

0

0

0

0

0

409

Don't know

1

0

0

0

0

0

0

1

Refusal

4

0

0

0

0

8

0

12

Total

1558

395

414

467

413

147

27

3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Win7 Pro, SPSS 24

Email:             [hidden email]

Website:          http://surveyresearch.weebly.com/

SPSS course:   http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research:        http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

===================== 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

spss.giesel@yahoo.de

Re: May Day Conundrum

I have seen such things happen when variables turned out to be string formatted.
Maybe you check that first.

Mario Giesel
Munich, Germany

Am Donnerstag, 2. Mai 2019, 05:28:08 MESZ hat Rick Oliver <[hidden email]> Folgendes geschrieben:

The Frequencies results indicate the categories are valid, but running Display Dictionary to confirm makes sense.

On Wed, May 1, 2019, 10:24 PM Rich Ulrich <[hidden email]> wrote:

I bet that you are wrong when you state that no Missing Values

are "declared" for either variable. Check that with Display Dictionary....

I also note the table shows results that I consider highly suspicious.

Namely, in the table produced when you include Missing, all the

income categories greater than 0 occur for the -1 group, "Not employed."

--

Rich Ulrich

From: SPSSX(r) Discussion <[hidden email]> on behalf of John F Hall <[hidden email]>
Sent: Wednesday, May 1, 2019 1:31 PM
To: [hidden email]
Subject: May Day Conundrum

I've been at this nearly all day, going right back to scratch and working one step at a time.

I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ            "respondent earnings quartiles (dv):Q1377"

ExPrtFul           "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable

FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

-1 Skip,not in paid work

1558

45.5

45.5

45.5

1 less than 11999

395

11.5

11.5

57.1

2 12000- 19999

414

12.1

12.1

69.2

3 20000- 31999

467

13.7

13.7

82.8

4 32000 or more

413

12.1

12.1

94.9

7 Refused information

147

4.3

4.3

99.2

8 Don't know

27

0.8

0.8

100.0

Total

3421

100.0

100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

-1 Not employee

1919

56.1

56.1

56.1

2 full-time 30+ hours/week

1080

31.6

31.6

87.7

3 part-time

409

12.0

12.0

99.6

8 Don't know

1

0.0

0.0

99.6

9 Refusal

12

0.4

0.4

100.0

Total

3421

100.0

100.0

CROSSTABS exprtful by rearnq.

Warnings

The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation

Count

REarnQ respondent earnings quartiles (dv):Q1377

Total

-1

Skip,not in paid work

1

less than 11999

2

12000- 19999

3

20000- 31999

4

32000 or more

7

Refused information

8

Don't know

ExPrtFul Respondent: Is job full or part-time? :Q1007

Not employee

64

395

414

467

413

139

27

1919

full-time 30+ hours/week

1080

0

0

0

0

0

0

1080

part-time

409

0

0

0

0

0

0

409

Don't know

1

0

0

0

0

0

0

1

Refusal

4

0

0

0

0

8

0

12

Total

1558

395

414

467

413

147

27

3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Win7 Pro, SPSS 24

Email:             [hidden email]

Website:          http://surveyresearch.weebly.com/

SPSS course:   http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research:        http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

===================== 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

David Marso

Re: May Day Conundrum

Administrator

In reply to this post by John F Hall

Appears that your data are munged.
Inspect carefully and believe nothing.

-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
--
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

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

Bruce Weaver

Re: May Day Conundrum

Administrator

In reply to this post by John F Hall

> I've been at this nearly all day, going right back to scratch and working
> one step at a time.
>
> I'm total baffled and it's doing my head in.
>
>
>
> In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as
> downloaded from the UKData Service) contains two variables:
>
>
> REarnQ "respondent earnings quartiles (dv):Q1377"
>
> ExPrtFul "Respondent: Is job full or part-time? :Q1007"
>
>
>
> No missing values are declared for either variable
>
> --- snip ---

-----
--
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/).

bdates

Re: May Day Conundrum

Yes I could.

BTW, Jane Morgan's group, JFM consulting, is applying. I had this thought last night about other potential competitors. I'm sure MPHI will apply since they do so much with the state. Also perhaps schools of public health or maybe even med schools.

Brian

From: SPSSX(r) Discussion <[hidden email]> on behalf of Bruce Weaver <[hidden email]>
Sent: Thursday, May 2, 2019 10:44:22 AM
To: [hidden email]
Subject: Re: May Day Conundrum

John, can anyone download the data file? If so, please provide the link. Or
failing that, could you save a version of the file that has variables rearnq
and exprtful, and upload it via the Nabble archive for the mailing list?
Then others could try to replicate what you're seeing.

Link to this thread in Nabble:
http://spssx-discussion.1045642.n5.nabble.com/May-Day-Conundrum-td5737754.html

Syntax to (try to) replicate what John is seeing:

FREQUENCIES rearnq exprtful.
CROSSTABS exprtful by rearnq.
CROSSTABS exprtful by rearnq / missing include.

HTH.

John F Hall wrote
> I've been at this nearly all day, going right back to scratch and working
> one step at a time.
>
> I'm total baffled and it's doing my head in.
>
>
>
> In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as
> downloaded from the UKData Service) contains two variables:
>
>
> REarnQ "respondent earnings quartiles (dv):Q1377"
>
> ExPrtFul "Respondent: Is job full or part-time? :Q1007"
>
>
>
> No missing values are declared for either variable
>
> --- snip ---

-----
--
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: May Day Conundrum

Thanks to everyone for their comments.

Yes, you can access the data, but you have to sign conditions.

New European General data Protection Regulations (GDPR) mean that, from October 2018, users wishing to download and use data from the British Social Attitudes Survey (BSAS) must

1: have (or open) an account at UKDS

2: be registered for use of BSAS data

3: be logged in to their UKDS account

It can be quite complicated to access the data as the new UKDS beta version of instructions on downloading data sets are confusing.

Once logged in, users are asked to specify a new or existing usage, identify the data set(s) they want and add it/them to the usage list.

UKDS SN Project Purpose

167695 Earnings from paid work Teaching Δ Request access

The data set(s) then need Depositor Approval (usually automatic by email).

167695 Earnings from paid work Teaching Active

You can only download one set at a time. To get the entire series I had to go this process 32 times!

It was a bit of a palaver even for an experienced user like me, but I have written a detailed guide in:

https://surveyresearch.weebly.com/uploads/2/9/9/8/2998485/downloading_bsas_data_from_the_u_k_data_service.pdf

It covers:

1: Locating a previously existing project

2: Assigning data sets to the project from the British Social Attitudes Survey

3: Downloading *.zip folders containing data and documentation deposited by SCPR/Natcen

4: Extracting the SPSS saved files

It is intended as a supplement to, not a replacement for, the UKDS guide. Hopefully it will help users navigate the new system.

BSAS 2009 link: https://beta.ukdataservice.ac.uk/datacatalogue/studies/study?id=6695

DOI: 10.5255/UKDA-SN-6695-1

BSAS series link: https://beta.ukdataservice.ac.uk/datacatalogue/series/series?id=200006#!/access

. . but it's much quicker to read my guide.

This morning I downloaded the original data again from UKDS, made a copy to work on, and am systematically going through it, searching for incompatibilities and inconsistencies which may underlie the problem. I'm even trying a complicated workaround involving multiple conditional transformations to derive a (full-time/part-time) work variable which will create fully populated earnings tables.

The original data set was generated using SPSS 12 and the final version was uploaded to UKDS on 8^th Feb 2011.

British Social Attitudes Survey, 2009. [data collection]. National Centre for Social Research. (2011).

UK Data Service. SN: 6695

An extensive reply to David's "mudged" comment has taken all afternoon to compose. It will be sent when it makes sense and has enough detail for you gurus.

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Email: [hidden email]

Website: http://surveyresearch.weebly.com/

SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Dates, Brian
Sent: 02 May 2019 16:52
To: [hidden email]
Subject: Re: May Day Conundrum

Yes I could.

Brian

From: SPSSX(r) Discussion <[hidden email]> on behalf of Bruce Weaver <[hidden email]>
Sent: Thursday, May 2, 2019 10:44:22 AM
To: [hidden email]
Subject: Re: May Day Conundrum

bdates

Re: May Day Conundrum

I want to apologize to everyone for the message sent in error earlier. I was managing two emails, and I sent my response to the wrong one. So much for my multi-tasking abilities. Take care.

Brian

From: John F Hall <[hidden email]>
Sent: Thursday, May 2, 2019 12:18:48 PM
To: Dates, Brian; [hidden email]
Cc: 'bweaver'
Subject: RE: May Day Conundrum

Thanks to everyone for their comments.

Yes, you can access the data, but you have to sign conditions.

New European General data Protection Regulations (GDPR) mean that, from October 2018, users wishing to download and use data from the British Social Attitudes Survey (BSAS) must

1: have (or open) an account at UKDS

2: be registered for use of BSAS data

3: be logged in to their UKDS account

It can be quite complicated to access the data as the new UKDS beta version of instructions on downloading data sets are confusing.

Once logged in, users are asked to specify a new or existing usage, identify the data set(s) they want and add it/them to the usage list.

UKDS SN Project Purpose

167695 Earnings from paid work Teaching Δ Request access

The data set(s) then need Depositor Approval (usually automatic by email).

167695 Earnings from paid work Teaching Active

You can only download one set at a time. To get the entire series I had to go this process 32 times!

It was a bit of a palaver even for an experienced user like me, but I have written a detailed guide in:

https://surveyresearch.weebly.com/uploads/2/9/9/8/2998485/downloading_bsas_data_from_the_u_k_data_service.pdf

It covers:

1: Locating a previously existing project

2: Assigning data sets to the project from the British Social Attitudes Survey

3: Downloading *.zip folders containing data and documentation deposited by SCPR/Natcen

4: Extracting the SPSS saved files

It is intended as a supplement to, not a replacement for, the UKDS guide. Hopefully it will help users navigate the new system.

BSAS 2009 link: https://beta.ukdataservice.ac.uk/datacatalogue/studies/study?id=6695

DOI: 10.5255/UKDA-SN-6695-1

BSAS series link: https://beta.ukdataservice.ac.uk/datacatalogue/series/series?id=200006#!/access

. . but it's much quicker to read my guide.

The original data set was generated using SPSS 12 and the final version was uploaded to UKDS on 8^th Feb 2011.

British Social Attitudes Survey, 2009. [data collection]. National Centre for Social Research. (2011).

UK Data Service. SN: 6695

An extensive reply to David's "mudged" comment has taken all afternoon to compose. It will be sent when it makes sense and has enough detail for you gurus.

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Email: [hidden email]

Website: http://surveyresearch.weebly.com/

SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

From: SPSSX(r) Discussion <[hidden email]> On Behalf Of Dates, Brian
Sent: 02 May 2019 16:52
To: [hidden email]
Subject: Re: May Day Conundrum

Yes I could.

Brian

From: SPSSX(r) Discussion <[hidden email]> on behalf of Bruce Weaver <[hidden email]>
Sent: Thursday, May 2, 2019 10:44:22 AM
To: [hidden email]
Subject: Re: May Day Conundrum

John F Hall

Re: May Day Conundrum

In reply to this post by Rick Oliver

I agree David's "mudged" but could think of stronger language.

The actual data are safeguarded and I am not authorised to share: I would get banned from UKDS.

I have downloaded the original data again from UKDS, made a copy to work on, and am systematically going through it, searching for incompatibilities and inconsistencies which may underlie the problem.

The original data set was generated using SPSS 12 and the final version was uploaded to UKDS on 8^th Feb 2011.

British Social Attitudes Survey, 2009. [data collection]. National Centre for Social Research. (2011).

UK Data Service. SN: 6695

There's one questionnaire for the main survey with complex overlapping sub-samples using three different versions of an additional self-completion questionnaire. There are 847 variables and 3421 cases. Specification of metadata (variable and value labels, levels of measurement, missing values etc.) is sometimes confusing and often incorrect or absent. This may be due to repeated use of automatic software at UKDS or of SPSS "heuristics". Fully specified SPSS *.sav files are often stripped of their metadata. I also suspect that metadata from earlier waves are used on subsequent waves as the same anomalies appear again.

Part of the problem with the 2009 data is the accompanying documentation which clearly served Natcen's working requirements for internal data management, data analysis for clients and authors of chapters in the main report:

British Social Attitudes Survey, 2009. [data collection]. National Centre for Social Research. (2011).

UK Data Service. SN: 6695

The documentation is not necessarily aimed at users wishing to do secondary analysis.

Some explanations of derived variables (e.g. names changed from source variables which are not in the file) can only be found in footnotes to items in the questionnaire (201 pages). It seems that, in the work section, variables beginning R*** relate to the Respondent, those beginning E*** relate to Employees, those beginning S*** relate to Self-employed eg:

REmploye " Is R an employee or self-employed currently? dv:Q987" [If working: Respondent]

REarnQ " respondent earnings quartiles (dv):Q1377" [If has earnings: Respondent]

ExPrtFul " Respondent: Is job full or part-time? :Q1007" [If working: Employees only]

REarnQ is a derived variable grouping earnings into 4 categories based on quartiles.

ExPrtFul is asked of employees .

DISPLAY DICTIONARY variables exprtful rearnq.

Variable Information
Variable	Position	Label	Missing Values
ExPrtFul	438	Respondent: Is job full or part-time? :Q1007	Lowest through -1
REarnQ	619	respondent earnings quartiles (dv):Q1377
Variables in the working file

Variable Values
Value		Label
ExPrtFul	-1^a	Not employee
	1	full-time 30+ hours/week
	2	part-time
	8	Don't know
	9	Refusal
REarnQ	-1	Skip,not in paid work
	1	less than 11999
	2	12000- 19999
	3	20000- 31999
	4	32000 or more
	7	Refused information
	8	Don't know
a. Missing value

Here's what I am working on:

Elaboration model

X → Y . T (the effect of X on Y controlling for T) where:

Y = Dependent variable

X = Independent variable

T = Test variable(s)

Y (Dependent) X (Independent) T_n (Test or control)

REarnQ "Gross earnings from paid work" Rsex ExPrtFul "Working full time or part time"

Initial frequency counts

FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377
		Frequency	Percent		Valid Percent		Cumulative Percent
Valid	Skip,not in paid work	1558	45.5		45.5		45.5
	less than 11999	395	11.5		11.5		57.1
	12000- 19999	414	12.1		12.1		69.2
	20000- 31999	467	13.7		13.7		82.8
	32000 or more	413	12.1		12.1		94.9
	Refused information	147	4.3		4.3		99.2
	Don't know	27	0.8		0.8		100.0
	Total	3421	100.0		100.0
Values -1 " Skip,not in paid work" 7 " Refused information" and 8 " Don't know" need to be treated as missing. ExPrtFul Respondent: Is job full or part-time? :Q1007
		Frequency		Percent		Valid Percent		Cumulative Percent
Valid	full-time 30+ hours/week	1080		31.6		71.9		71.9
	part-time	409		12.0		27.2		99.1
	Don't know	1		0.0		0.1		99.2
	Refusal	12		0.4		0.8		100.0
	Total	1502		43.9		100.0
Missing	Not employee	1919		56.1
Total		3421		100.0

In addition to -1 "Not employee" values 8 " Refused information" and 9 " Don't know" need to be treated as missing.

MISSING VALUES REarnQ (-1, 7, 8) ExPrtFul (-1 8 9).

FREQUENCIES REarnQ ExPrtFul.

REarnQ Respondent earnings quartiles
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1 less than 11999	395	11.5	23.4	23.4
	2 12000- 19999	414	12.1	24.5	47.9
	3 20000- 31999	467	13.7	27.6	75.5
	4 32000 or more	413	12.1	24.5	100.0
	Total	1689	49.4	100.0
Missing	-1 Skip,not in paid work	1558	45.5
	7 Refused information	147	4.3
	8 Don't know	27	0.8
	Total	1732	50.6
Total		3421	100.0

[NB: The value labels in the above tables date from before the £ sign was allowed in SPSS]

ExPrtFul Respondent: Is job full or part-time? :Q1007
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	full-time 30+ hours/week	1080	31.6	72.5	72.5
	part-time	409	12.0	27.5	100.0
	Total	1489	43.5	100.0
Missing	Not employee	1919	56.1
	Don't know	1	0.0
	Refusal	12	0.4
	Total	1932	56.5
Total		3421	100.0

Zero order tables[1] X →Y

T → Y

CROSSTABS rsex by rearnq /cells row.

*RSex SEX OF respondent? :Q356 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
% within RSex SEX OF respondent? :Q356
		REarnQ respondent earnings quartiles (dv):Q1377				Total
		less than 11999	12000- 19999	20000- 31999	32000 or more	Total
RSex SEX OF respondent? :Q356	Male	12.3%	22.4%	29.7%	35.6%	100.0%
RSex SEX OF respondent? :Q356	Female	34.1%	26.6%	25.6%	13.6%	100.0%
Total		23.4%	24.5%	27.6%	24.5%	100.0%

CROSSTABS exprtful by rearnq /cells row.

*ExPrtFul Respondent: Is job full or part-time? :Q1007 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
% within ExPrtFul Respondent: Is job full or part-time? :Q1007
		REarnQ respondent earnings quartiles (dv):Q1377		Total
		Skip,not in paid work	Refused information	Total
ExPrtFul Respondent: Is job full or part-time? :Q1007	full-time 30+ hours/week	100.0%		100.0%
	part-time	100.0%		100.0%
	Don't know	100.0%		100.0%
	Refusal	33.3%	66.7%	100.0%
Total		99.5%	0.5%	100.0%

CROSSTABS exprtful by rearnq /cells row /missing include.

*ExPrtFul Respondent: Is job full or part-time? :Q1007 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
% within ExPrtFul Respondent: Is job full or part-time? :Q1007
		REarnQ respondent earnings quartiles (dv):Q1377							Total
		Skip,not in paid work	less than 11999	12000- 19999	20000- 31999	32000 or more	Refused information	Don't know	Total
ExPrtFul Respondent: Is job full or part-time? :Q1007	Not employee	3.3%	20.6%	21.6%	24.3%	21.5%	7.2%	1.4%	100.0%
	full-time 30+ hours/week	100.0%							100.0%
	part-time	100.0%							100.0%
	Don't know	100.0%							100.0%
	Refusal	33.3%					66.7%		100.0%
Total		45.5%	11.5%	12.1%	13.7%	12.1%	4.3%	0.8%	100.0%

Because of the tabulation problem I am now searching through other variables to use as test variables:

Remploye (Employee or Self-employed)

FREQUENCIES remploye.

REmploye Is R an employee or self-employed currently? dv:Q987
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-3 Skip, not now in paid work	1494	43.7	44.5	44.5
	1 Emp	1594	46.6	47.5	92.0
	2 SEmp	260	7.6	7.7	99.7
	8 Don't know	1	0.0	0.0	99.8
	9 Refusal	8	0.2	0.2	100.0
	Total	3357	98.1	100.0
Missing	-1 Skip, never had a job	64	1.9
Total		3421	100.0

There are more than three values to be treated as missing, so:

RECODE remploye (8=-8)(9=-9).

MISSING VALUES remploye (-9 thru -1).

FREQUENCIES remploye.

*REmploye Is R an employee or self-employed currently? dv:Q987 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
Count
		REarnQ respondent earnings quartiles (dv):Q1377				Total
		less than 11999	12000- 19999	20000- 31999	32000 or more	Total
REmploye Is R an employee or self-employed currently? dv:Q987	Emp	327	362	428	350	1467
	SEmp	68	52	39	63	222
Total		395	414	467	413	1689

CROSSTABS remploye by rearnq /cells row.

*REmploye Is R an employee or self-employed currently? dv:Q987 REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
% within REmploye Is R an employee or self-employed currently? dv:Q987
		REarnQ respondent earnings quartiles (dv):Q1377				Total
		less than 11999	12000- 19999	20000- 31999	32000 or more	Total
REmploye Is R an employee or self-employed currently? dv:Q987	Emp	22.3%	24.7%	29.2%	23.9%	100.0%
	SEmp	30.6%	23.4%	17.6%	28.4%	100.0%
Total		23.4%	24.5%	27.6%	24.5%	100.0%

There are problems using [ExPrtFul] combined with [REarnQ] which produce a message saying the table is empty.

Warnings

The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

There are two separate variables for hours worked.

[EJbHrCal] " Hours R works per week, including overtime [employee]. DV:Q1008"

[SJbHrCaI] " Hours R works per week, including overtime [self-employed]. DV:Q1010"

These two variables are mutually exclusive and this may be the cause of the problem.

FREQUENCIES EJbHrCaI SJbHrCaI.

EJbHrCaI Hours R works per week, including overtime [employee]. DV:Q1008
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1 10-15 hours a week	82	2.4	5.1	5.1
	2 16-23 hours a week	183	5.3	11.5	16.6
	3 24-29 hours a week	90	2.6	5.6	22.3
	4 30 or more hours a week	1228	35.9	77.0	99.2
	5 Varies too much to say	7	0.2	0.4	99.7
	8 Don't know	4	0.1	0.3	99.9
	9 Refusal	1	0.0	0.1	100.0
	Total	1595	46.6	100.0
Missing	-1 Not employee	1826	53.4
Total		3421	100.0

SJbHrCaI Hours R works per week, including overtime [self-employed]. DV:Q1010
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1 10-15 hours a week	16	0.5	6.2	6.2
	2 16-23 hours a week	38	1.1	14.6	20.8
	3 24-29 hours a week	16	0.5	6.2	26.9
	4 30 or more hours a week	183	5.3	70.4	97.3
	5 Varies too much to say	5	0.1	1.9	99.2
	9 Refusal	2	0.1	0.8	100.0
	Total	260	7.6	100.0
Missing	-1 Not self-employed	3161	92.4
Total		3421	100.0

MISSING VALUES EJbHrCaI (5 8 9) SJbHrCaI (5 9).

FREQUENCIES EJbHrCaI SJbHrCaI.

EJbHrCaI Hours R works per week, including overtime [employee]. DV:Q1008
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Not employee	1826	53.4	53.6	53.6
	1 10-15 hours a week	82	2.4	2.4	56.0
	2 16-23 hours a week	183	5.3	5.4	61.3
	3 24-29 hours a week	90	2.6	2.6	64.0
	4 30 or more hours a week	1228	35.9	36.0	100.0
	Total	3409	99.6	100.0
Missing	5 Varies too much to say	7	0.2
	8 Don't know	4	0.1
	9 Refusal	1	0.0
	Total	12	0.4
Total		3421	100.0

SJbHrCaI Hours R works per week, including overtime [self-employed]. DV:Q1010
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	-1 Not self-employed	3161	92.4	92.6	92.6
	1 10-15 hours a week	16	0.5	0.5	93.1
	2 16-23 hours a week	38	1.1	1.1	94.2
	3 24-29 hours a week	16	0.5	0.5	94.6
	4 30 or more hours a week	183	5.3	5.4	100.0
	Total	3414	99.8	100.0
Missing	5 Varies too much to say	5	0.1
	9 Refusal	2	0.1
	Total	7	0.2
Total		3421	100.0

An alternative variable [workmode] was created using the following syntax:.

if (not (missing (rearnq))) hoursemp = EJbHrCaI.

if (not (missing (rearnq))) hoursself = SJbHrCaI.

frequencies var hoursemp hoursself.

do if (missing (hoursself)).

compute workmode = hoursemp.

else if (missing (hoursemp)) .

compute workmode = hoursself.

end if.

recode workmode (2 3=1)( 4 =2).

variable labels workmode (R works full- or part- time).

value labels workmode

1 'Part-time' 2 'Full-time' 5 ' Varies too much to say' 8 " Don't know" 9 ' Refusal'.

frequencies workmode.

workmode (R works full- or part- time)
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1 Part-time	386	11.3	23.0	23.0
	2 Full-time	1292	37.8	77.0	100.0
	Total	1678	49.0	100.0
Missing	5 Varies too much to say	8	0.2
	8 Don't know	2	0.1
	9 Refusal	1	0.0
	System	1732	50.6
	Total	1743	51.0
Total		3421	100.0

[NB: If SPSS finds missing values for variables in the intermediate steps it automatically sets sysmis as a missing value for [rearngrp]

CROSSTABS workmode by rearnq .

*workmode (R works full- or part- time) REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
Count
		REarnQ respondent earnings quartiles (dv):Q1377				Total
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
workmode (R works full- or part- time)	1 Part-time	241	86	41	18	386
workmode (R works full- or part- time)	2 Full-time	150	328	423	391	1292
Total		391	414	464	409	1678

CROSSTABS workmode by rearnq /cells row.

*workmode (R works full- or part- time) REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation**
% within workmode (R works full- or part- time)
		REarnQ respondent earnings quartiles (dv):Q1377				Total
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
workmode (R works full- or part- time)	1 Part-time	62.4%	22.3%	10.6%	4.7%	100.0%
workmode (R works full- or part- time)	2 Full-time	11.6%	25.4%	32.7%	30.3%	100.0%
Total		23.3%	24.7%	27.7%	24.4%	100.0%

crosstabs rsex by rearnq by workmode /cells row.

*RSex SEX OF respondent? :Q356 REarnQ respondent earnings quartiles (dv):Q1377 * workmode (R works full- or part- time) Crosstabulation**
% within RSex SEX OF respondent? :Q356
workmode (R works full- or part- time)			REarnQ respondent earnings quartiles (dv):Q1377				Total
workmode (R works full- or part- time)			1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
1 Part-time	RSex SEX OF respondent? :Q356	1 Male	49.4%	25.8%	10.1%	14.6%	100.0%
	RSex SEX OF respondent? :Q356	2 Female	66.3%	21.2%	10.8%	1.7%	100.0%
	Total		62.4%	22.3%	10.6%	4.7%	100.0%
2 Full-time	RSex SEX OF respondent? :Q356	1 Male	7.6%	22.2%	32.2%	38.0%	100.0%
	RSex SEX OF respondent? :Q356	2 Female	16.9%	29.6%	33.4%	20.1%	100.0%
	Total		11.6%	25.4%	32.7%	30.3%	100.0%
Total	RSex SEX OF respondent? :Q356	1 Male	12.1%	22.6%	29.9%	35.4%	100.0%
	RSex SEX OF respondent? :Q356	2 Female	34.1%	26.7%	25.5%	13.7%	100.0%
	Total		23.3%	24.7%	27.7%	24.4%	100.0%

Sorted, but what a clutter! CTABLES is better:

CTABLES

/TABLE rsex BY rearnq [ROWPCT.COUNT TOTALS [COUNT]]

/CATEGORIES VARIABLES= rsex rearnq TOTAL=YES

/TABLE workmode BY rearnq [ROWPCT.COUNT TOTALS [COUNT]]

/CATEGORIES VARIABLES= rsex rearnq TOTAL=YES .

		REarnQ respondent earnings quartiles (dv):Q1377
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
		Row N %	Row N %	Row N %	Row N %	Count
RSex SEX OF respondent? :Q356	1 Male	12.3%	22.4%	29.7%	35.6%	831
	2 Female	34.1%	26.6%	25.6%	13.6%	858
	Total	23.4%	24.5%	27.6%	24.5%	1689

		REarnQ respondent earnings quartiles (dv):Q1377
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
		Row N %	Row N %	Row N %	Row N %	Count
workmode (R works full- or part- time)	1 Part-time	62.4%	22.3%	10.6%	4.7%	386
	2 Full-time	11.6%	25.4%	32.7%	30.3%	1292
	Total	23.3%	24.7%	27.7%	24.4%	1678

Just what the doctor ordered (almost). Reduce clutter even further with:

CTABLES

/TABLE rsex BY rearnq [ROWPCT.COUNT f5.1 "%" totals [count "n= 100%"]]

/CATEGORIES VARIABLES= rearnq TOTAL=YES

/TABLE workmode BY rearnq [ROWPCT.COUNT f5.1 "%" totals [count "n= 100%"]]

/CATEGORIES VARIABLES= rearnq TOTAL=YES .

(modified from syntax supplied some years ago by the saintly Jignesh Sutar)

		REarnQ respondent earnings quartiles (dv):Q1377
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
		%	%	%	%	n= 100%
RSex SEX OF respondent? :Q356	1 Male	12.3	22.4	29.7	35.6	831
RSex SEX OF respondent? :Q356	2 Female	34.1	26.6	25.6	13.6	858

		REarnQ respondent earnings quartiles (dv):Q1377
		1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
		%	%	%	%	n= 100%
workmode (R works full- or part- time)	1 Part-time	62.4	22.3	10.6	4.7	386
workmode (R works full- or part- time)	2 Full-time	11.6	25.4	32.7	30.3	1292

Coup de résistance:

*To nest sex within workmode.

ctables /TABLE workmode > rsex by rearnq

[ROWPCT.COUNT f5.1 "%" totals [count "n= 100%"]]

/CATEGORIES VARIABLES= rearnq TOTAL=YES.

				REarnQ respondent earnings quartiles (dv):Q1377
				1 less than 11999	2 12000- 19999	3 20000- 31999	4 32000 or more	Total
				%	%	%	%	n= 100%
workmode (R works full- or part- time)	1 Part-time	RSex SEX OF respondent? :Q356	1 Male	49.4	25.8	10.1	14.6	89
	1 Part-time	RSex SEX OF respondent? :Q356	2 Female	66.3	21.2	10.8	1.7	297
	2 Full-time	RSex SEX OF respondent? :Q356	1 Male	7.6	22.2	32.2	38.0	735
	2 Full-time	RSex SEX OF respondent? :Q356	2 Female	16.9	29.6	33.4	20.1	557

With this table it's easy to compare men and women and how the global distribution:

23.4%

24.5%

27.6%

24.5%

1689

. . has been partitioned.

By copying cells to Excel the epsilons (percentage point differences) between men and women can be calculated.

+49.4

-8.4

-22.6

-18.4

The gradient isn't as clear in this table as it is for other years, but it's a technique which students can understand and serves as a platform for more advanced statistical analysis.

Mindful of David's preference for writing commans in UPPER CASE, I'm afraid I got lazy and lapsed into lower case, but resisted the temptation to use abbreviated syntax.

All in all, a good day's work.

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: 02 May 2019 05:28
To: [hidden email]
Subject: Re: May Day Conundrum

The Frequencies results indicate the categories are valid, but running Display Dictionary to confirm makes sense.

On Wed, May 1, 2019, 10:24 PM Rich Ulrich <[hidden email]> wrote:

I bet that you are wrong when you state that no Missing Values
are "declared" for either variable. Check that with Display Dictionary....

I also note the table shows results that I consider highly suspicious.
Namely, in the table produced when you include Missing, all the
income categories greater than 0 occur for the -1 group, "Not employed."

--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of John F Hall <[hidden email]>
Sent: Wednesday, May 1, 2019 1:31 PM
To: [hidden email]
Subject: May Day Conundrum

I've been at this nearly all day, going right back to scratch and working one step at a time.
I'm total baffled and it's doing my head in.

In the 2009 British Social Attitudes Survey (BSAS) file bsa09.sav (as downloaded from the UKData Service) contains two variables:

REarnQ            "respondent earnings quartiles (dv):Q1377"
ExPrtFul           "Respondent: Is job full or part-time? :Q1007"

No missing values are declared for either variable
FREQUENCIES rearnq exprtful.

REarnQ respondent earnings quartiles (dv):Q1377

Frequency
Percent
Valid Percent
Cumulative Percent
Valid
-1 Skip,not in paid work
1558
45.5
45.5
45.5
1 less than 11999
395
11.5
11.5
57.1
2 12000- 19999
414
12.1
12.1
69.2
3 20000- 31999
467
13.7
13.7
82.8
4 32000 or more
413
12.1
12.1
94.9
7 Refused information
147
4.3
4.3
99.2
8 Don't know
27
0.8
0.8
100.0
Total
3421
100.0
100.0

ExPrtFul Respondent: Is job full or part-time? :Q1007

Frequency
Percent
Valid Percent
Cumulative Percent
Valid
-1 Not employee
1919
56.1
56.1
56.1
2 full-time 30+ hours/week
1080
31.6
31.6
87.7
3 part-time
409
12.0
12.0
99.6
8 Don't know
1
0.0
0.0
99.6
9 Refusal
12
0.4
0.4
100.0
Total
3421
100.0
100.0

CROSSTABS exprtful by rearnq.

Warnings
The crosstabulation of ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 is empty.

CROSSTABS exprtful by rearnq / missing include.

ExPrtFul Respondent: Is job full or part-time? :Q1007 * REarnQ respondent earnings quartiles (dv):Q1377 Crosstabulation
Count

REarnQ respondent earnings quartiles (dv):Q1377
Total
-1

Skip,not in paid work
1

less than 11999
2

12000- 19999
3

20000- 31999
4

32000 or more
7

Refused information
8

Don't know
ExPrtFul Respondent: Is job full or part-time? :Q1007
Not employee
64
395
414
467
413
139
27
1919
full-time 30+ hours/week
1080
0
0
0
0
0
0
1080
part-time
409
0
0
0
0
0
0
409
Don't know
1
0
0
0
0
0
0
1
Refusal
4
0
0
0
0
8
0
12
Total
1558
395
414
467
413
147
27
3421

Does SPSS have secret hidden codes? Can someone please explain what is going on?

John F Hall
[Retired academic survey researcher]
IBM-SPSS Academic Author 9900074
Win7 Pro, SPSS 24

Email:             [hidden email]
Website:          http://surveyresearch.weebly.com/
SPSS course:   http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html
Research:        http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

===================== 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

[1] ⁴ SPSS command CROSSTABS produces very cluttered output: CTABLES is more complex to use, but the output is far less cluttered. (See: 3.2.1.4 Elaboration 4 (Income differences 2009 – 2014_CTABLES)

Rich Ulrich

Re: May Day Conundrum

I see a minor ambiguity. What if there are hours for both

Employed and Self-employed? You end up with workmode

as undefined if you use what you wrote -

do if (missing (hoursself)).

compute workmode = hoursemp.

else if (missing (hoursemp)) .

compute workmode = hoursself.

end if.

I don't see how you can usefully add two non-missing

responses, but you can take the maximum of the two

(if there aren't any high-valued missing).

COMPUTE workmode= max(hoursemp, hoursself).

Rich Ulrich

John F Hall

Re: May Day Conundrum

Rich

You left the first bit of syntax out.

if (not (missing (rearnq))) hoursemp = EJbHrCaI.

if (not (missing (rearnq))) hoursself = SJbHrCaI.

frequencies var hoursemp hoursself.

My syntax then works:

do if (missing (hoursself)).

compute workmode = hoursemp.

else if (missing (hoursemp)) .

compute workmode = hoursself.

end if.

hoursemp and hoursself don't exist in the original file: they are used in an intermediate stage.

I modified your syntax to:

COMPUTE workmode2 = max (EJbHrCaI, SJbHrCaI).

Neat trick: 1 line to replace 8! Must remember that.

frequencies workmode.

workmode
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1	98	2.9	5.3	5.3
	2	221	6.5	11.9	17.2
	3	106	3.1	5.7	22.9
	4	1411	41.2	76.1	99.0
	5	12	0.4	0.6	99.6
	8	4	0.1	0.2	99.8
	9	3	0.1	0.2	100.0
	Total	1855	54.2	100.0
Missing	System	1566	45.8
Total		3421	100.0

*workmode workmode2 Crosstabulation**
Count
		workmode2							Total
		1	2	3	4	5	8	9	Total
workmode	1	98	0	0	0	0	0	0	98
	2	0	221	0	0	0	0	0	221
	3	0	0	106	0	0	0	0	106
	4	0	0	0	1411	0	0	0	1411
	5	0	0	0	0	12	0	0	12
	8	0	0	0	0	0	4	0	4
	9	0	0	0	0	0	0	3	3
Total		98	221	106	1411	12	4	3	1855

I was hoping the labels would be copied as well. I tried variations of APPLY DICTIONARY but got nowhere, so did:

recode workmode (2 3=1)( 4 =2)(else = copy) into workmode3.

variable labels workmode3 (R works full- or part- time).

value labels workmode3

1 'Part-time' 2 'Full-time' 5 "Varies too much to say"

8 "Don't know" 9 "Refusal".

missing values workmode3 (-1, 5 thru 9).

frequencies workmode3.

workmode3 (R works full- or part- time)
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1 Part-time	425	12.4	23.1	23.1
	2 Full-time	1411	41.2	76.9	100.0
	Total	1836	53.7	100.0
Missing	5 Varies too much to say	12	0.4
	8 Don't know	4	0.1
	9 Refusal	3	0.1
	System	1566	45.8
	Total	1585	46.3
Total		3421	100.0

Sorry about the type size. Supposed to be 10 but looks more like 18. I think Microsoft have been messing about with 365 as their Outlook, Word, Powerpoint and Excel icons in the taskbar have all changed since yesterday.

John

John F Hall

[Retired academic survey researcher]

IBM-SPSS Academic Author 9900074

Email: [hidden email]

Website: http://surveyresearch.weebly.com/

SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html

Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html

From: Rich Ulrich <[hidden email]>
Sent: 02 May 2019 22:42
To: [hidden email]; John F Hall <[hidden email]>
Subject: Re: May Day Conundrum

I see a minor ambiguity. What if there are hours for both

Employed and Self-employed? You end up with workmode

as undefined if you use what you wrote -

do if (missing (hoursself)).

compute workmode = hoursemp.

else if (missing (hoursemp)) .

compute workmode = hoursself.

end if.

I don't see how you can usefully add two non-missing

responses, but you can take the maximum of the two

(if there aren't any high-valued missing).

COMPUTE workmode= max(hoursemp, hoursself).

Rich Ulrich