W3C home > Mailing lists > Public > public-html@w3.org > August 2007

Likert scale for survey?

From: Laura Carlson <laura.lee.carlson@gmail.com>
Date: Sat, 4 Aug 2007 07:03:59 -0500
Message-ID: <1c8dbcaa0708040503s6d4423d5j7549dd1e15cbedb5@mail.gmail.com>
To: public-html@w3.org, "Dan Connolly" <connolly@w3.org>

In the minutes, of the HTML WG 2007-08-02 phone conference [1] a
survey on the design principles is mentioned. [2]

Something to consider to help prevent bias, might be to have response
categories evened out over a Likert type of scale [3].

Instead of the current  1 neutral, 3 positive, and 1 negative scale of:

> I don't mind if it's there, but it doesn't appeal to me much.
> Yes, as stated.
> Yes, though I have suggestions...
> Yes, but only if changes are made
> No, because...
>
> Comments (or a URI pointing to your comments):

Perhaps use a five point scale indicating level of agreement like:

- Strongly Agree
- Agree
- Neutral
- Disagree
- Strongly Disagree

Then have the text box to include comments, suggestions, proposed changes  like:
"Comments, suggestions, proposed changes (or a URI pointing to your ideas)":

Like it says at wikipedia a lot of things can cause bias in surveys.
People tend to avoid using extreme response categories (central
tendency bias), agree with statements as presented (acquiescence
bias), or try to portray themselves or their organization in a more
favorable light (social desirability bias). Even the ordering of the
questions and ordering or the responses can insert bias.

Best Regards,
Laura

[1] http://www.w3.org/2007/08/02-html-wg-minutes.html
[2] http://www.w3.org/2002/09/wbs/40318/dprv/
[3] http://en.wikipedia.org/wiki/Likert_scale

-- 
Laura L. Carlson
http://www.d.umn.edu/goto/webdesign/
Received on Saturday, 4 August 2007 12:04:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:03 GMT