LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (May 2009)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:   Tue, 26 May 2009 16:45:46 -0600
Reply-To:   ViAnn Beadle <vab88011@gmail.com>
Sender:   "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:   ViAnn Beadle <vab88011@gmail.com>
Subject:   Re: converting 1-5 rating into 1-6 rating scale
Comments:   To: king@kingdouglas.com
Content-Type:   multipart/alternative;

Also I'd look at the frequencies for both samples. Odd scales tend to show larger hits for the middle point--which is frequently a proxy answer for "I don't know" or "I don't care". Even scales force the respondent to a make a positive or negative choice. ----- Original Message ----- From: King Douglas To: SPSSX-L@LISTSERV.UGA.EDU Sent: Tuesday, May 26, 2009 3:05 PM Subject: Re: converting 1-5 rating into 1-6 rating scale

Ajay,

Is there any reason you need the new values to be integers? If you are going to aggregate the data and report means, then the converted scale should maintain the z-score for each case. The following formula does the trick for converting a 5-pt scale to a 6-pt scale. Scale end-points are maintained and the converted z-scores will match the original z-scores.

COMPUTE NEWSCORE = (OLDSCORE-3)*(5/4) + 3.5.

What the preceding does is subtract from each original score the original scale midpoint, then multiplies the result by (the range of the new scale divided by the range of the old scale), then adds to that result the midpoint of the new scale.

To check the accuracy of the conversions, run descriptives of old and new variables. Include the subcommand /SAVE in order save the z-scores of both the original and the converted scores. Any score that is an endpoint on the original scale should be an endpoint on the new scale.

DESCRIPTIVES VARIABLES = OLDSCORE NEWSCORE /SAVE /SORT.

Next, run descriptives of the saved z-score variables to check the high and low z-scores, which should be the same for the original and the converted scales.

DESCRIPTIVES VARIABLES = ZOLDSCORE ZNEWSCORE /SORT.

--- On Tue, 5/26/09, ajay atluri <atluriak@gmail.com> wrote:

From: ajay atluri <atluriak@gmail.com> Subject: converting 1-5 rating into 1-6 rating scale To: SPSSX-L@LISTSERV.UGA.EDU Date: Tuesday, May 26, 2009, 12:39 PM

Hi Team,

I am working on SPSS 16 Base. I have 500 participants and 70 questions in the data. All these questions are of 1-5 point scale in 2007 data. But, 2009 data we have same set of participants and questions but rating was changed from 1-5 scale rating to 1-6 rating scale.

At this juncture, I would like to convert all 2007 data into 6 point scale. To do this i have defined some formulae.

We need to convert 75% who has given the answer as 5 and 12% who have given the answer as 4 into 6. And for getting number of responses for 5, I will be converting 5-5, 4-5 and 3-5 by considering the following percentage of respondents 25%, 70% and 15% into 5. This list will go on.

Does any one in the forum has done in a similar way? If so, please share the syntax with me.

Thanks, Atluri

===================== To manage your subscription to SPSSX-L, send a message to LISTSERV@LISTSERV.UGA.EDU (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


[text/html]


Back to: Top of message | Previous page | Main SPSSX-L page