finding missing values

[Talino Bruno]

I have  a number of measures with 113 questions. I have set up my SPSS data
window  so that the first column is the subject Id and the rest are  the
individual quetsion numbers (e.g Q1...Q113) with applicable scores. When
the subject did not answer a particular question it was left blank in the
data window. How can I determine  the amount of questions that were left
unanswered by each subject? I need to know for each subject which questions
they may have missed.

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

[Albert-Jan Roskam]

Hi!

Try this untested code:

compute miss_counter = 0.
do repeat #q = q1 to q113.
+if ( missing(#q) ) miss_counter = miss_counter + 1.
end repeat print.

It assumes that q1 to q113 are contiguously ordered in the data file.

Cheers!!
Albert-Jan

--- On Mon, 7/13/09, Talino Bruno <[hidden email]> wrote:

> From: Talino Bruno <[hidden email]>
> Subject: finding missing values
> To: [hidden email]
> Date: Monday, July 13, 2009, 1:39 AM
> I have  a number of measures
> with 113 questions. I have set up my SPSS data
> window  so that the first column is the subject Id and
> the rest are  the
> individual quetsion numbers (e.g Q1...Q113) with applicable
> scores. When
> the subject did not answer a particular question it was
> left blank in the
> data window. How can I determine  the amount of
> questions that were left
> unanswered by each subject? I need to know for each subject
> which questions
> they may have missed.
>
> =====================
> 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

[John McConnell-2]

Hello Talino

SPSS has a module which allows you to report/visualise missing patterns
and also potentially impute into them

http://www.spss.com/statistics/missing_values/

If you don't have that you might try to use the COUNT procedure on the
following menu path...

Transform|Count Values within Cases...

HTH

John

John McConnell
analytical people
500 Chiswick High Road
London
W4 5RG, United Kingdom

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Talino Bruno
Sent: 13 July 2009 00:40
To: [hidden email]
Subject: finding missing values

I have  a number of measures with 113 questions. I have set up my SPSS
data
window  so that the first column is the subject Id and the rest are  the
individual quetsion numbers (e.g Q1...Q113) with applicable scores. When
the subject did not answer a particular question it was left blank in
the
data window. How can I determine  the amount of questions that were left
unanswered by each subject? I need to know for each subject which
questions
they may have missed.

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

[Melissa Ives]

I've seen the suggestions to use DO REPEAT or COUNT.
I might use the NVALID function like this:

Compute nmiss=113-(NVALID(Q1 to Q113)).

But the other half of Talino's question is which ones are missing.
It seems to me that something like this might give what you want (a list of missing items per record).
The format for QMISS may need to be changed depending on how much missing data you have.

STRING QMISS (A50).
VECTOR Q=Q1 TO Q113.
LOOP #I=1 TO 113.
+  DO IF MISSING (Q(#I)).
+  COMPUTE QMISS=CONCAT(RTRIM(QMISS),",",STRING(#I,F3)).
+  END IF.
END LOOP.
EXECUTE.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Talino Bruno
Sent: Sunday, July 12, 2009 6:40 PM
To: [hidden email]
Subject: [SPSSX-L] finding missing values

I have  a number of measures with 113 questions. I have set up my SPSS data window  so that the first column is the subject Id and the rest are  the individual quetsion numbers (e.g Q1...Q113) with applicable scores. When the subject did not answer a particular question it was left blank in the data window. How can I determine  the amount of questions that were left unanswered by each subject? I need to know for each subject which questions they may have missed.

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

PRIVILEGED AND CONFIDENTIAL INFORMATION
This transmittal and any attachments may contain PRIVILEGED AND
CONFIDENTIAL information and is intended only for the use of the
addressee. If you are not the designated recipient, or an employee
or agent authorized to deliver such transmittals to the designated
recipient, you are hereby notified that any dissemination,
copying or publication of this transmittal is strictly prohibited. If
you have received this transmittal in error, please notify us
immediately by replying to the sender and delete this copy from your
system. You may also call us at [phone] for assistance.

=====================
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]

Melissa Ives wrote

I've seen the suggestions to use DO REPEAT or COUNT.
I might use the NVALID function like this:

Compute nmiss=113-(NVALID(Q1 to Q113)).

--- snip ---

How about this?

Compute nmiss= NMISS(Q1 to Q113).

[parisec]

Hi all,

i have data in 24-hr time where 12:00 AM is = 00:00:00.

how can i convert this so that 12:00 AM is 24:00:00?

i tried this:

        recode mytime (TIME.HMS(0,0,0)=TIME.HMS(24,0,0)).

with no success.
any ideas?

thanks
Carol

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

[Oliver, Richard]

This may be somewhat inelegant, but...

 

dataset close all.
output close all.
new file.
data list free /timevar (time5).
begin data
0:00 0:15 1:30 3:45 9:59 12:00 12:15 13:30 15:45 21:59
end data.
string ampm (a2).
do if xdate.hours(timevar)=0.
compute ampm='AM'.
compute timevar=timevar+(86400/2).
else if xdate.hours(timevar)=12.
compute ampm='PM'.
else if xdate.hours(timevar)<11.
compute ampm='AM'.
else if xdate.hours(timevar)>12.
compute ampm='PM'.
compute timevar=timevar-(86400/2).
end if.
list.

 

*86400/2 is the number of seconds in 12 hours.

---

From: SPSSX(r) Discussion on behalf of Parise, Carol A.
Sent: Mon 9/28/2009 2:00 PM
To: [hidden email]
Subject: recoding 24 hr time

Hi all,

i have data in 24-hr time where 12:00 AM is = 00:00:00.

how can i convert this so that 12:00 AM is 24:00:00?

i tried this:

        recode mytime (TIME.HMS(0,0,0)=TIME.HMS(24,0,0)).

with no success.
any ideas?

thanks
Carol

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

[Dale Glaser]

good afternoon all..I'm revisting an issue that mildly flummoxed me years ago, and as I was revisiting the issue (poring through Maxwell and Delaney and related texts/articles) I still don't have an unequivocal answer....so I would be most appreciative to get your take on the reporting of the proper effect size for a fully within (9 x 5) repeated measures design.

 

I know how the partial eta squared is calculated for the multivariate test (1 - Wilk's Lamda) and univariate (SSH/ SSH + SSE), but when I get such disparate effect sizes (e.g., .153 for multivariate vs. .069 for univariate for the first within subject factor..............687 for multivariate vs. .086 for univariate for the interaction) I honestly don't know which one is the proper one to report.  And this is not an anomolous finding for many of the GLModels I run.

 

So, when all is said and done, keeping in mind the pros and cons of interpreting multivariate vs. univariate effects (and the attendant assumptions), as well as the oft-cited preference for effect sizes such as omega-squared, I would be most appreciative to hear what many of you report for the partial eta squared in a repeated measures (t > 2) context.

 

Thank you!!....Dale Glaser

Dale Glaser, Ph.D.
Principal--Glaser Consulting
Lecturer/Adjunct Faculty--SDSU/USD/AIU
Past-President, San Diego Chapter of
American Statistical Association
3115 4th Avenue
San Diego, CA 92103
phone: [phone]
fax: [phone]
email: [hidden email]
website: www.glaserconsult.com

[Spousta Jan]

Hi Carol,

Suppose your variable is in an appropriate date format. Then try:

if (timevar = TIME.HMS(0,0,0) ) timevar = TIME.HMS(24,0,0) .
EXECUTE .

Hope this helps,

Jan


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Parise, Carol A.
Sent: Monday, September 28, 2009 9:00 PM
To: [hidden email]
Subject: recoding 24 hr time

Hi all,

i have data in 24-hr time where 12:00 AM is = 00:00:00.

how can i convert this so that 12:00 AM is 24:00:00?

i tried this:

        recode mytime (TIME.HMS(0,0,0)=TIME.HMS(24,0,0)).

with no success.
any ideas?

thanks
Carol

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



_____________
Tato zpráva a všechny připojené soubory jsou důvěrné a určené výlučně adresátovi(-ům). Jestliže nejste oprávněným adresátem, je zakázáno jakékoliv zveřejňování, zprostředkování nebo jiné použití těchto informací. Jestliže jste tento mail dostali neoprávněně, prosím, uvědomte odesilatele a smažte zprávu i přiložené soubory. Odesilatel nezodpovídá za jakékoliv chyby nebo opomenutí způsobené tímto přenosem.

Jste si jisti, že opravdu potřebujete vytisknout tuto zprávu a/nebo její přílohy? Myslete na přírodu.


This message and any attached files are confidential and intended solely for the addressee(s). Any publication, transmission or other use of the information by a person or entity other than the intended addressee is prohibited. If you receive this in error please contact the sender and delete the message as well as all attached documents. The sender does not accept liability for any errors or omissions as a result of the transmission.

Are you sure that you really need a print version of this message and/or its attachments? Think about nature.

-.- --

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

[Oliver, Richard]

First, my apologies for my earlier "solution" for a completely different problem. I should have read the post more carefully.

Second, the solution below is correct, but note that there is a slightly simpler (but arguably more cryptic) solution:

if timevar=0 timevar=86400.

This could also be expressed in a Recode:

recode timevar (0=86400).

Internally, times are stored as a number of seconds, and there are 86400 seconds in a day.


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Spousta Jan
Sent: Tuesday, September 29, 2009 4:01 AM
To: [hidden email]
Subject: Re: recoding 24 hr time

Hi Carol,

Suppose your variable is in an appropriate date format. Then try:

if (timevar = TIME.HMS(0,0,0) ) timevar = TIME.HMS(24,0,0) .
EXECUTE .

Hope this helps,

Jan


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Parise, Carol A.
Sent: Monday, September 28, 2009 9:00 PM
To: [hidden email]
Subject: recoding 24 hr time

Hi all,

i have data in 24-hr time where 12:00 AM is = 00:00:00.

how can i convert this so that 12:00 AM is 24:00:00?

i tried this:

        recode mytime (TIME.HMS(0,0,0)=TIME.HMS(24,0,0)).

with no success.
any ideas?

thanks
Carol

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



_____________
Tato zpráva a všechny připojené soubory jsou důvěrné a určené výlučně adresátovi(-ům). Jestliže nejste oprávněným adresátem, je zakázáno jakékoliv zveřejňování, zprostředkování nebo jiné použití těchto informací. Jestliže jste tento mail dostali neoprávněně, prosím, uvědomte odesilatele a smažte zprávu i přiložené soubory. Odesilatel nezodpovídá za jakékoliv chyby nebo opomenutí způsobené tímto přenosem.

Jste si jisti, že opravdu potřebujete vytisknout tuto zprávu a/nebo její přílohy? Myslete na přírodu.


This message and any attached files are confidential and intended solely for the addressee(s). Any publication, transmission or other use of the information by a person or entity other than the intended addressee is prohibited. If you receive this in error please contact the sender and delete the message as well as all attached documents. The sender does not accept liability for any errors or omissions as a result of the transmission.

Are you sure that you really need a print version of this message and/or its attachments? Think about nature.

-.- --

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

[parisec]

Hi all,

I received a solution offline which worked:

COMPUTE test1 = TIME.HMS(00,00,00) .
FORMAT  test1 (TIME10).
COMPUTE test2 = TIME.HMS(00,00,00) .
FORMAT  test2 (TIME10).
IF (test1=TIME.HMS(00,00,00)) test2=TIME.HMS(24,00,00).

But both of are great succint solutions.

Thanks so much for the replies.
Carol

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Oliver, Richard
Sent: Tuesday, September 29, 2009 7:57 AM
To: [hidden email]
Subject: Re: recoding 24 hr time

First, my apologies for my earlier "solution" for a completely different problem. I should have read the post more carefully.

Second, the solution below is correct, but note that there is a slightly simpler (but arguably more cryptic) solution:

if timevar=0 timevar=86400.

This could also be expressed in a Recode:

recode timevar (0=86400).

Internally, times are stored as a number of seconds, and there are 86400 seconds in a day.


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Spousta Jan
Sent: Tuesday, September 29, 2009 4:01 AM
To: [hidden email]
Subject: Re: recoding 24 hr time

Hi Carol,

Suppose your variable is in an appropriate date format. Then try:

if (timevar = TIME.HMS(0,0,0) ) timevar = TIME.HMS(24,0,0) .
EXECUTE .

Hope this helps,

Jan


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Parise, Carol A.
Sent: Monday, September 28, 2009 9:00 PM
To: [hidden email]
Subject: recoding 24 hr time

Hi all,

i have data in 24-hr time where 12:00 AM is = 00:00:00.

how can i convert this so that 12:00 AM is 24:00:00?

i tried this:

        recode mytime (TIME.HMS(0,0,0)=TIME.HMS(24,0,0)).

with no success.
any ideas?

thanks
Carol

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



_____________
Tato zpráva a všechny připojené soubory jsou důvěrné a určené výlučně adresátovi(-ům). Jestliže nejste oprávněným adresátem, je zakázáno jakékoliv zveřejňování, zprostředkování nebo jiné použití těchto informací. Jestliže jste tento mail dostali neoprávněně, prosím, uvědomte odesilatele a smažte zprávu i přiložené soubory. Odesilatel nezodpovídá za jakékoliv chyby nebo opomenutí způsobené tímto přenosem.

Jste si jisti, že opravdu potřebujete vytisknout tuto zprávu a/nebo její přílohy? Myslete na přírodu.


This message and any attached files are confidential and intended solely for the addressee(s). Any publication, transmission or other use of the information by a person or entity other than the intended addressee is prohibited. If you receive this in error please contact the sender and delete the message as well as all attached documents. The sender does not accept liability for any errors or omissions as a result of the transmission.

Are you sure that you really need a print version of this message and/or its attachments? Think about nature.

-.- --

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

[parisec]

Hi All,

 

I have more of a methodology question than an SPSS question but i suspect that many have you may have come across this before.

 

I have data that are something like this: 10 tasks, 2 conditions (hard, easy) and 3 performance groups (low, med, high)

 

Assumptions:

 

1) Task - Repeated measures factor 1:  It is not important how people scored on one task versus another. the tasks are considered independent

2) Condition - Repeated measures factor 2: It's already been determined that one condition is hard versus easy so we know that all people are going to perform better in the hard vs easy condition

3) Performance group - Between subjects factor: The performance groups were determined by the investigators based on scores.

 

My question of interest is "Within each performace group, how do people perform in the hard condition versus the easy condition on each of the tasks?"

 

While i could run a mixed design GLM with a doubly repeated measures, i am thinking of a simple approach that gets straight to the question:

 

Within each performance group, run a repeated measures t-test (or anova) comparing scores in the hard vs easy condition on a task. Now do this for each of the 10 tasks.

 

Argument against the t-tests: 1. It's a moral sin due to increased familywise error! 2. the same people performed on the tasks and this isn't being taken into account

 

Argument for the t-tests: 1.Too many conditions (10 tasks, 2 conditions, 3 groups) for the number of cases (35) for the mixed design,

   2. If i only cared about 1 task and was writing up the results, but decided that next month i wanted to analyze the same data for another task, etc. etc, etc, there would be no concer regarding running multiple tests so it's really not as big a concern as we make it out to be (i'm plagerizing this from a lecture i went to)

 

I'm curious as to your thoughts on this. I would also appreciate any references.

 

Thanks

Carol

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[Steve Simon, P.Mean Consulting]

Parise, Carol A. wrote:

> I have more of a methodology question than an SPSS question but i
> suspect that many have you may have come across this before.
>
> I have data that are something like this: 10 tasks, 2 conditions (hard,
> easy) and 3 performance groups (low, med, high)
>
> Assumptions:
>
> 1) Task - Repeated measures factor 1:  It is not important how people
> scored on one task versus another. the tasks are considered independent
> 2) Condition - Repeated measures factor 2: It's already been determined
> that one condition is hard versus easy so we know that all people are
> going to perform better in the hard vs easy condition
> 3) Performance group - Between subjects factor: The performance groups
> were determined by the investigators based on scores.
>
> My question of interest is "Within each performace group, how do people
> perform in the hard condition versus the easy condition on each of the
> tasks?"
>
> While i could run a mixed design GLM with a doubly repeated measures, i
> am thinking of a simple approach that gets straight to the question:
>
> Within each performance group, run a repeated measures t-test (or
> anova) comparing scores in the hard vs easy condition on a task. Now do
> this for each of the 10 tasks.
>
> Argument against the t-tests: 1. It's a moral sin due to increased
> familywise error! 2. the same people performed on the tasks and this
> isn't being taken into account
>
> Argument for the t-tests: 1.Too many conditions (10 tasks, 2 conditions,
> 3 groups) for the number of cases (35) for the mixed design,
>    2. If i only cared about 1 task and was writing up the results, but
> decided that next month i wanted to analyze the same data for another
> task, etc. etc, etc, there would be no concer regarding running multiple
> tests so it's really not as big a concern as we make it out to be (i'm
> plagerizing this from a lecture i went to)
>
> I'm curious as to your thoughts on this. I would also appreciate any
> references.

You will find outside resources about the Bonferroni adjustment and a
somewhat haphazard list of my thoughts on my category page on multiple
comparisons:

http://www.pmean.com/category/MultipleComparisons.html

As far as your particular application goes, you'll be swimming against
the tide if you run a series of multiple t-tests. Although a few journal
referees might accept your approach, most will swat your attempt down.

Here is an alternative suggestion. Any time a repeated factor in an
analysis is of no direct interest, you can simplify things greatly by
summing or averaging across this repeated factor. You said that the
distinction between hard and easy tasks is not important. So why not
create a new outcome measure, which is the total amount of time to
handle two tasks, one easy and one hard. Now you have a singly repeated
measures design, which is easier to run and easier to explain.

Now if you really wanted to simplify things, create an outcome measure
which is the total time across the 20 repeated conditions (10 task by 2
levels). Then it's a one factor ANOVA. In an earlier part of your email
you state that distinctions among tasks are not important, but later you
seem to contradict this.

If you are interested in things like an interaction involving difficulty
level and performance group or an interaction involving task and
performance group, then this would be a mistake.

Whatever choice you make, good luck!
--
Steve Simon, Standard Disclaimer
The Monthly Mean is celebrating its first anniversary.
Find out more about the newsletter that dares
to call itself "average" at www.pmean.com/news

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