Hi Ben, all,
Accidentally, I’m the head of the team responsible for – among other things - Eurobarometer archiving at GESIS.
Let me just confirm that using the posted EB file for PSPP purposes is entirely fine, or let’s even say very welcome, and even redistributing the shortened file for instruction purposes
is okay.
We would be grateful, though, if both the citations for the data producer, that is the European Commission as given in the questionnaire that ‘ftr’ attached, and the actual data file
were reproduced. The recommended citation covers both, like this:
European Commission, Brussels (2014). Flash Eurobarometer 383 (Firearms in the European Union).
GESIS Data Archive, Cologne. ZA5897 Data file Version 1.0.0, https://doi.org/10.4232/1.11958.
Thanks for considering this, and for your continued work on PSPP!
Let me know if there are more questions on data access and re-use at GESIS.
Best,
Markus
Von: Pspp-users <pspp-users-bounces+markus.quandt=gesis.org@gnu.org>
Im Auftrag von ft gmail
Gesendet: Montag, 17. Januar 2022 15:09
An: pspp-users@gnu.org; Ben Pfaff <blp@cs.stanford.edu>
Betreff: Re: seeking a good data set for CTABLES examples
Hi Ben,
This is a shortened Flash Eurobarometer 383 survey from September 2013 which includes a multi-response variable
Q9 Which of the following are reasons why you own or used to own a firearm?
and some accompanying socio-demographic nominal, ordinal and scale variables,
with n = 26,555 in 28 countries.
I shortened the labels for the multi-response question Q9 so that you can produce a readable table.
See attached save file and the questionnaire.
I am looking very much forward to your implementation of CTABLES.
Regards,
ftr
On 17/01/2022 07:37, Ben Pfaff wrote:
Here's an example of what I can do currently with this dataset and CTABLES. Syntax:
CTABLES /TABLE QN105BA[c] + QN105BB[c] + QN105BC[c] + QN105BD[c]
/CLABELS ROWLABELS=OPPOSITE.
Output:
Custom Tables
╭──────────────────────────────────────────────────────────────┬───────────┬────────┬───────────┬─────────────┬──────────╮
│ │ Almost │ Very │ Somewhat │ Somewhat │ Very │
│ │ certain │ likely │ likely │ unlikely │ unlikely │
│ ├───────────┼────────┼───────────┼─────────────┼──────────┤
│ │ Count │ Count │ Count │ Count │ Count │
├──────────────────────────────────────────────────────────────┼───────────┼────────┼───────────┼─────────────┼──────────┤
│105b. How likely is it that drivers who have had too much to │ 700│ 1502│ 2763│ 1307│ 609│
│drink to drive safely will A. Get stopped by the police? │ │ │ │ │ │
│105b. How likely is it that drivers who have had too much to │ 1100│ 2819│ 2417│ 430│ 140│
│drink to drive safely will B. Have an accident? │ │ │ │ │ │
│105b. How likely is it that drivers who have had too much to │ 1149│ 2037│ 2032│ 994│ 622│
│drink to drive safely will C. Be convicted for drunk driving? │ │ │ │ │ │
│105b. How likely is it that drivers who have had too much to │ 1101│ 1834│ 2307│ 1095│ 549│
│drink to drive safely will D. Be arrested for drunk driving? │ │ │ │ │ │
╰──────────────────────────────────────────────────────────────┴───────────┴────────┴───────────┴─────────────┴──────────╯
Here is one data set that seems to suit the purpose:
https://catalog.data.gov/dataset/2008-national-survey-of-drinking-and-driving-attitudes-and-behaviors
It's not perfect because the variable names are poor (they are simply
named for question numbers) and because a lot of the variables have a
wrong measurement level, but I'm going to start from it.
Please feel free to send me more data sets.
On Sat, Jan 15, 2022 at 10:40 AM Ben Pfaff <blp@cs.stanford.edu> wrote:
>
> Hi! I'm getting to the point with work on CTABLES that I need a good
> data set for use in examples. A good data set would need to be:
> * Publicly available and freely redistributable.
> * Medium size (at least hundreds of cases).
> * Have a mix of categorical and scale variables.
> * Contain some variables suitable for multiple response sets.
>
> I can't use the data sets that come with SPSS because it's not clear
> that they are freely redistributable.
>
> I'd appreciate advice and pointers.
>
> Thanks,
>
> Ben.