summary distribution with multiple observations per respondent

[Amy Corning]

I have SPSS data containing variables that identify each respondent's age
at the time of 8 different events. I would like to find a way in SPSS to
obtain a single summary distribution of ages over the 8 variables, but
measures of central tendency are not appropriate. For the moment I am
generating frequency distributions for each of the 8 variables, then
averaging them by hand, but I'm wondering whether anyone can suggest a more
efficient approach? (Essentially, I think I may need a way of treating each
respondent's 8 responses as 8 separate cases.)

Thank you! -Amy

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

[Maguin, Eugene]

Amy,

I think you need to restructure your file from wide to long. Read up on the
varstocases command.

Gene Maguin



-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Amy
Corning
Sent: Saturday, August 13, 2011 4:29 PM
To: [hidden email]
Subject: summary distribution with multiple observations per respondent

I have SPSS data containing variables that identify each respondent's age
at the time of 8 different events. I would like to find a way in SPSS to
obtain a single summary distribution of ages over the 8 variables, but
measures of central tendency are not appropriate. For the moment I am
generating frequency distributions for each of the 8 variables, then
averaging them by hand, but I'm wondering whether anyone can suggest a more
efficient approach? (Essentially, I think I may need a way of treating each
respondent's 8 responses as 8 separate cases.)

Thank you! -Amy

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

[Art Kendall]

see if the syntax below does the kind of thing you want.
Open a new instance of SPSS.  Copy the syntax from the email.  Paste it to a syntax window. Run it.

Art Kendall
Social Research Consultants

data list list /id (f4) age1 to age3 (3f2) gender (f1).
begin data
    001 17 18 19 1
    002 18 19 17 1
    003 19 17 18 1
    004 17 18 19 2
    005 18 19 17 2
    006 19 17 18 2
    007 17  0  0 1
    008 0  17 19 1
    009 17 18  0 1
    010 19  0  0 2
    011  0  0  0 2
    012  0 17  0 2
end data.
dataset name 'original'
    var labels
    age1 'age event type 1'
    age2 'age event type 2'
    age3 'age event type 3'.
value labels gender
    1 'male'
    2 'female'.
mult response
    /groups = ages 'ages for events' (age1 to age3 (1, 104))
    /variables = gender (1,2)
    /frequencies = ages
    /tables = ages by ages by ages /ages by gender
    /cells = all.



On 8/13/2011 4:28 PM, Amy Corning wrote:

I have SPSS data containing variables that identify each respondent's age
at the time of 8 different events. I would like to find a way in SPSS to
obtain a single summary distribution of ages over the 8 variables, but
measures of central tendency are not appropriate. For the moment I am
generating frequency distributions for each of the 8 variables, then
averaging them by hand, but I'm wondering whether anyone can suggest a more
efficient approach? (Essentially, I think I may need a way of treating each
respondent's 8 responses as 8 separate cases.)

Thank you! -Amy

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

[Maguin, Eugene]

Ok, glad it worked for you. On to the next question. Although you don't say
so, I'm going to assume that you are thinking of t-test or Anovas. I'd
suggest you look at the mixed command because you have what looks like a two
level analysis. However, more experienced list-folks may have better advice.

Gene Maguin

-----Original Message-----
From: Amy Corning [mailto:[hidden email]]
Sent: Sunday, August 14, 2011 11:23 AM
To: Gene Maguin
Subject: Re: summary distribution with multiple observations per respondent


Thank you, Gene! I was not aware that such a command even existed, but it
does just what I need and is simple to use. One further question: does SPSS
provide any way to adjust for the clustering that now occurs (e.g., if I
were to use any significance tests on the summary variable in the
restructured data)? Not essential, but I am wondering if you might know.
Again, thanks so much for your help! Best, Amy

On Sat, 13 Aug 2011 16:46:16 -0400, Gene Maguin <[hidden email]>
wrote:
> Amy,
>
> I think you need to restructure your file from wide to long. Read up on
the

> varstocases command.
>
> Gene Maguin
>
>
>
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
> Amy
> Corning
> Sent: Saturday, August 13, 2011 4:29 PM
> To: [hidden email]
> Subject: summary distribution with multiple observations per respondent
>
> I have SPSS data containing variables that identify each respondent's age
> at the time of 8 different events. I would like to find a way in SPSS to
> obtain a single summary distribution of ages over the 8 variables, but
> measures of central tendency are not appropriate. For the moment I am
> generating frequency distributions for each of the 8 variables, then
> averaging them by hand, but I'm wondering whether anyone can suggest a

more
> efficient approach? (Essentially, I think I may need a way of treating
each

> respondent's 8 responses as 8 separate cases.)
>
> Thank you! -Amy
>
> =====================
> 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

[Richard Ristow]

At 04:28 PM 8/13/2011, Amy Corning wrote:

>I have SPSS data containing variables that identify each
>respondent's age at the time of 8 different events. I would like to
>find a way in SPSS to obtain a single summary distribution of ages
>over the 8 variables,

The first question is, what do you want to do, conceptually?

Among other things, do the events occur in sequence, so each person
is always older when the second event occurs than when the first one
does, etc.? Or can a person's earliest event be the first, third, or
whatever on your list of 8?

>For the moment I am generating frequency distributions for each of
>the 8 variables, then averaging them by hand, but I'm wondering
>whether anyone can suggest a more efficient approach? (Essentially,
>I think I may need a way of treating each respondent's 8 responses
>as 8 separate cases.)

It's often helpful to do precisely that. Use VARSTOCASES; if you're
using the menus, DATA -> RESTRUCTURE.

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