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 (September 2005)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Sat, 17 Sep 2005 11:10:43 -1000
Reply-To:     Bob Schacht <bobschacht@INFOMAGIC.NET>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Bob Schacht <bobschacht@INFOMAGIC.NET>
Subject:      Re: Likert scales
Comments: To: "Claus D. Hansen" <>
In-Reply-To:  <>
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 10:10 AM 9/17/2005, Claus D. Hansen wrote: >Dear list, > >...I have a group of books who say that Likert scales is the very common type >of scale where a group of respondents answer a set of items with 4 or 5 >categories ranging from strong disagree to strong agree (or something >similar). Their total scale score is then computed as the sum of the answers >to each of the items.

This makes sense only if the wording of the questions is constructed so that all of the responses point in the same direction. For example, if you word all the questions in the second person ("you"), always avoid negatives (or: always use negatives), and all of the questions have a common theme. Otherwise, the total score makes no sense.

>However, I also have another book who says that Rensis Likert in fact >envisaged a more complex method of attributing scale scores to the >respondents than the one just mentioned. In this book it says that a group >of test persons is to answer the questions. Each of the test persons is >given a scale score equal to the number of items they have agreed with. For >each item an average index score is computed based on only those persons >that has answered that they agree with this statement. The scale is then >distributed to the real respondents and their final scale score is computed >by summating the average index scores of those items they agree with. (this >explanation is taken from Babbie & Mouton: The Practice of Social Research, >1998, Oxford University Press) > > > >Does anyone know which of the two above mentioned choices are right? And >perhaps where I might be able to read more about it?

Again, it depends on how the questions are constructed. This cannot be over-emphasized. Consider this: Is it intuitively clear what a high score means? If not, I wouldn't trust the scoring.


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