From: kb on
Hi All,

I am trying to recode a likert scale question (1-9 responses) and must
change response 7 to response 8.
But the number of responses that must be changed should be 10% of the
sum of the frequency of 8,9 for each variable.(Top two box)

I am dealing with 40 variables at a time and must do this moving of
10% of response 7 to response 8 for each one of them. I am using Proc
Surveyselect method, then recoding the data and trying to merge the
data thereafter. But this messes up the metadata.

Any suggestion on how this problem can be dealt?
Ranuni function would not work as the frequency is varying for each
variable.

Please help!
From: "Data _null_;" on
Maybe you could post some sample data and the code you are using and
someone would be able to help. I don't really follow your description
of the process.

On 9/18/09, kb <bhattkavita(a)gmail.com> wrote:
> Hi All,
>
> I am trying to recode a likert scale question (1-9 responses) and must
> change response 7 to response 8.
> But the number of responses that must be changed should be 10% of the
> sum of the frequency of 8,9 for each variable.(Top two box)
>
> I am dealing with 40 variables at a time and must do this moving of
> 10% of response 7 to response 8 for each one of them. I am using Proc
> Surveyselect method, then recoding the data and trying to merge the
> data thereafter. But this messes up the metadata.
>
> Any suggestion on how this problem can be dealt?
> Ranuni function would not work as the frequency is varying for each
> variable.
>
> Please help!
>
From: kb on
On Sep 18, 6:28 pm, iebup...(a)GMAIL.COM ("Data _null_;") wrote:
> Maybe you could post some sample data and the code you are using and
> someone would be able to help. I don't really follow your description
> of the process.
>
> On 9/18/09, kb <bhattkav...(a)gmail.com> wrote:
>
>
>
> > Hi All,
>
> > I am trying to recode a likert scale question (1-9 responses) and must
> > change response 7 to response 8.
> > But the number of responses that must be changed should be 10% of the
> > sum of the frequency of 8,9 for each variable.(Top two box)
>
> > I am dealing with 40 variables at a time and must do this moving of
> > 10% of response 7 to response 8 for each one of them. I am using Proc
> >Surveyselectmethod, then recoding the data and trying to merge the
> > data thereafter. But this messes up the metadata.
>
> > Any suggestion on how this problem can be dealt?
> > Ranuni function would not work as the frequency is varying for each
> > variable.
>
> > Please help!- Hide quoted text -
>
> - Show quoted text -

Hi All,

Here is the sample data.
I must change 10% of the total of frequency of 8 and 9 i.e. the number
7 response will be changed from 7 to 8. (will be recoded)

Suppose 10% of freq for var b1 is 20 and only 1 response in B1has
responded 7 then the other 1 will be taken from 6 i.e. in this case
the 6 response will be changed to 8.

Could anyone please throw some light on how the syntax should be?I am
having trouble with this part.
I am trying to work this out by using the proc surveyselect.

Please help. Attached is the link to the sample data.

http://spreadsheets.google.com/ccc?key=0AvGEZYn1kYTudFFOR3pCZW8xdUlVYzBZTkNsMHpkSUE&hl=en

Thanks.