Hi everyone! I am not sure if this is the right newsgroup but can't find a
better one. I hope you are willing to take a look at it.
For a Likert type kind of question (something like "does this product
satisfy your needs?"), I have to choose between two lists that I can offer
respondents as potential answers. Both are possible but which would you say
is preferable?

1. Not at all
2. A little
3. Somewhat
4. Quite a lot
5. To a great extent

or

1. Not at all
2. To a small extent
3. To a moderate extent
4. To a large extent
5. To a great extent

Hi everyone! I am not sure if this is the right newsgroup but can't find a
better one. I hope you are willing to take a look at it.
For a Likert type kind of question (something like "does this product
satisfy your needs?"), I have to choose between two lists that I can offer
respondents as potential answers. Both are possible but which would you say
is preferable?

1. Not at all
2. A little
3. Somewhat
4. Quite a lot
5. To a great extent

or

1. Not at all
2. To a small extent
3. To a moderate extent
4. To a large extent
5. To a great extent

On Sat, 7 Jun 2008 22:14:55 +0200, "Burger" <burger1... at (no spam) hotpop.com
wrote:



Hi everyone! I am not sure if this is the right newsgroup but can't find a
better one. I hope you are willing to take a look at it.
For a Likert type kind of question (something like "does this product
satisfy your needs?"), I have to choose between two lists that I can offer
respondents as potential answers. Both are possible but which would you say
is preferable?

1. Not at all
2. A little
3. Somewhat
4. Quite a lot
5. To a great extent

or

1. Not at all
2. To a small extent
3. To a moderate extent
4. To a large extent
5. To a great extent

Like Stan says, the gradation is not completely clear
on 4 and 5.

British subjects are apt to read "quite" as either what
Americans mean as "quite" or else as "not quite".

You might find any textbook with the word Psychometrics
in the title. Google-scholar can help.

Also, what you have are items on a 1-5 scale, intended
to specify equal intervals.
I know that the term "Likert" gets used loosely, but the
original Likert scales specified "agreement-disagreement"
using symmetrical terms and where the expected average
was in the middle. I don't think you need to use
"Likert" here unless you at least achieve the symmetry.

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html

...Very Satisfied;
Satisfied; Neither Satisfied or Dissatisfied; Dissatisfied; Very
Dissatisfied.

...Very Satisfied;
Satisfied; Neither Satisfied or Dissatisfied; Dissatisfied; Very
Dissatisfied.

You will find a comprehensive white paper on survey best practices at
this link.

http://www.zoomerang.com/whitepapers/customersat.pdf

Thanks,
Kiran

Interesting paper -- thanks for posting.

They listed the response levels as going from Very satistfied to Very
dissatisfied. Potentially that may create a positive response bias:
seeing the positive levels first (assunming one reads from left to
right), people may be more inclined to check the initial response
levels.

Further, since this layout is, in a sense, mildly counter-intuitive
(i.e., the usual convention, as with a simple number line or x-axis is
to proceed from smaller/negative values on the left to larger/positive
values on the right). Thus the resopondent may see that the
researcher has reversed this convention, and take it as a subtle cue
to make positive responses. (And, of course, such a bias is often
already implicit in the task, as frequently such surveys are used by
organizations to look good, and many people are happy to oblige.)

I'm not agruging strongly for any of these points -- just want to
mention them as things to consider in designing a survey.

John Uebersax PhD


On Jun 10, 7:25 pm, Kiran <kiranga... at (no spam) gmail.com> wrote:
You will find a comprehensive white paper on survey best practices at
this link.

http://www.zoomerang.com/whitepapers/customersat.pdf

Thanks,
Kiran

How satisfied are you with the service you received?

1 2 3 4 5
+---------+---------+---------+---------+
Very Very
Dis. Dis. Neither Satisf. Satisf.

... people do not always look as closely at the labels
as a data analysist would want them to and tend to answer such

questions with respect to the outer labels. For this reason, I'd like
to put a minor variation of the above up for discussion:

Were you rather satisfied or dissatisfied ...

  1         2         3         4         5
  +---------+---------+---------+---------+
Dis.                                   Satisf.

It depends on how you want to analyse and present the data, but IMHO
this often has some advantages.

if it's better to omit the intervening labels, why didn't Likert
[...]
I don't see any advantages here. Unless perhaps you have a true
continuous 'thermometer' scale.