Date: Fri, 10 Sep 2004 09:41:02 +0800
Reply-To: "Little, Jonathon" <Jonathon.Little@health.wa.gov.au>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: "Little, Jonathon" <Jonathon.Little@health.wa.gov.au>
Subject: Re: Ordinal scale and using mean
Content-Type: text/plain; charset="us-ascii"
I am not entirely sure that was the original question, although yes
kendall tau-b is tickety-boo from my point of view.
I fail to see how the problem differs from any other in psychology where
Likert type scales are used. Yes, I agree with Hector which argued that
vales assigned are arbitrary, although I wonder if in changing anchor
values when presenting the instrument to respondents you would create an
instrument with different psychometric properties. If I were a
respondent and were presented with a question with anchors such as
-1000 (Strongly Agree) to 1500000 (Strongly Disagree)
I would get very confused, and the psychometric properties of the scales
may be affected. In most instances such as these, instruments items are
treated as interval variables, although true, they fall somewhere
between an ordinal scale and an interval one. The difficulty in using a
measure of central tendency, such as a mean, is that a mean is a
parametric statistic which assumes equal intervals between successive
responses. In this sense, a mean, or variance, or standard deviation
etc, will be in error to the extent that respondents do not interpret
intervals within the scale as equal. Strictly speaking the only
appropriate measure of central tendency for ordinal data is a median.
Ideally the values accompanying each anchor would have been presented to
the respondents when they answered and not assigned after the fact.
Senior Research Analyst
Office of Mental Health
Department of Health
2nd Floor C-Block, 189 Royal Street, East Perth, WA 6004
Tel:08 9222 4089
Fax: 08 9222 2351
Visit our website at : www.mental.health.wa.gov.au
The contents of this e-mail transmission are confidential and may be
protected by professional privilege. The contents are intended only for
the named recipients of the e-mail. If you are not the intended
recipient, you are hereby notified that any use, reproduction,
disclosure or distribution of the information contained in this e-mail
is prohibited. Please notify the sender immediately.