Time-use Research with SPSS / Sequence-analysis with SPSS

Time-use Research with SPSS / Sequence-analysis with SPSS

Post by Hans-Peter Schm » Fri, 04 Jul 2003 22:31:26



Dear Ladies and Gentlemen,

i am doing a project at the moment at the Transport Insitute of the
DLR (German Aerospace Centre, Institut fr Verkehrsforschung).

The focus of the project is to analyse a time use-study.
Various people had to report what they were doing on two subsequent
days (their time use from midnight to midnight).
For that, one day is splitted into 288 five-minute-segments. In each
of these 288 segments, one of 231 activities is available.

For every person, there are 288 variables in the SPSS-data set with
five-minute-segments (and with activities like sleep etc.), which
define that persons time use over one day.

My problem is related to that sort of data-structure:

I want to add the time-use for one person for all 288 segments.

With commands like count, recode etc. i am just able to recode/count
ONE variable and to get ONE new variable.
Is it possible to count / produce frequencies etc. for all of these
288 in one operation?
Eg to find out, that in 150 of the 288 five-minute segments that
person is watching TV (eg acitivity number 38)?

Does anyone has some experience with time-use research with SPSS

or

with such sequence-analysis with SPSS?

It would be great to get some hints.
Is there any literature concerning these problems?

Thank you very much!

Hans-Peter Schmid

 
 
 

Time-use Research with SPSS / Sequence-analysis with SPSS

Post by Griseknoe » Fri, 04 Jul 2003 23:04:47


Quote:> Eg to find out, that in 150 of the 288 five-minute segments that
> person is watching TV (eg acitivity number 38)?

S: Hmm... isn't COUNT working? A procedure lika below should work...
And if You want it percent, just divide by 288 and set the format to
percent. In syntax editor, like this:

count watch_tv=v1 to v288(1).

compute watch_tv=watch_tv/288*100 .

format watch_tv(pct5.2).

exe .

 
 
 

Time-use Research with SPSS / Sequence-analysis with SPSS

Post by Griseknoe » Fri, 04 Jul 2003 23:07:05


Quote:> count watch_tv=v1 to v288(1).

S: Sorry... for clarity, the above should be: "count watch_tv=v1 to
v288(38)."
 
 
 

Time-use Research with SPSS / Sequence-analysis with SPSS

Post by Rich Ulric » Sat, 05 Jul 2003 00:02:34



Quote:> ...
> The focus of the project is to analyse a time use-study.
> Various people had to report what they were doing on two subsequent
> days (their time use from midnight to midnight).
> For that, one day is splitted into 288 five-minute-segments. In each
> of these 288 segments, one of 231 activities is available.

> For every person, there are 288 variables in the SPSS-data set with
> five-minute-segments (and with activities like sleep etc.), which
> define that persons time use over one day.

> My problem is related to that sort of data-structure:

> I want to add the time-use for one person for all 288 segments.

> With commands like count, recode etc. i am just able to recode/count
> ONE variable and to get ONE new variable.

That last statement is wrong.  
A reasonable Count is always across several variables.  
So, what sort of output are you looking for?
Will you be happy with one variable for each Activity?

Quote:> Is it possible to count / produce frequencies etc. for all of these
> 288 in one operation?
> Eg to find out, that in 150 of the 288 five-minute segments that
> person is watching TV (eg acitivity number 38)?

You could use a Loop with a Vector-variable,
in order to have concise syntax to create 38 or so sums.

You could FLIP (transpose) the file, so that each Person
becomes a Variable; each of the 288 becomes a line.
Then the Frequencies on the Variables show what
Activities (sort in decreasing order?)  occurred.

The first, hard part of your problem is figuring out
what you want to produce, besides raw counts.
DO you need to inspect these results, or have them,
each one, in those 38 or more variables?

--

http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization."  Justice Holmes.

 
 
 

Time-use Research with SPSS / Sequence-analysis with SPSS

Post by Hans-Peter Schm » Wed, 09 Jul 2003 20:40:40




> > ...
> > The focus of the project is to analyse a time use-study.
> > Various people had to report what they were doing on two subsequent
> > days (their time use from midnight to midnight).
> > For that, one day is splitted into 288 five-minute-segments. In each
> > of these 288 segments, one of 231 activities is available.

> > For every person, there are 288 variables in the SPSS-data set with
> > five-minute-segments (and with activities like sleep etc.), which
> > define that persons time use over one day.

> > My problem is related to that sort of data-structure:

> > I want to add the time-use for one person for all 288 segments.

> > With commands like count, recode etc. i am just able to recode/count
> > ONE variable and to get ONE new variable.

> That last statement is wrong.  
> A reasonable Count is always across several variables.  
> So, what sort of output are you looking for?
> Will you be happy with one variable for each Activity?

> > Is it possible to count / produce frequencies etc. for all of these
> > 288 in one operation?
> > Eg to find out, that in 150 of the 288 five-minute segments that
> > person is watching TV (eg acitivity number 38)?

> You could use a Loop with a Vector-variable,
> in order to have concise syntax to create 38 or so sums.

> You could FLIP (transpose) the file, so that each Person
> becomes a Variable; each of the 288 becomes a line.
> Then the Frequencies on the Variables show what
> Activities (sort in decreasing order?)  occurred.

> The first, hard part of your problem is figuring out
> what you want to produce, besides raw counts.
> DO you need to inspect these results, or have them,
> each one, in those 38 or more variables?

Dear Group-members!

Thanks for the remarks!

I sorted out the "count-problem" :-)