Ratings Scales - Odd vs Even or JUSTER scale...
Ismail Parsa
...
|> I've done some research myself and it seems that the difference between
|> say, a four point scale and a five point scale is not much at all. Using
|> an Alpha test on a Likert scale shows the difference between a four and a
|> five is next to nothing.
|>
|> So I think it depends on the survey, what you are asking, what
|> restrictions you have to work with etc.
|> Basically use whateevr you feel comfortable with.
|>
|> Anyone else got an opinion on this ?
My comments are not really related to the subject but...
I would like to remind the market researchers and polsters about
the JUSTER scale as an alternative to likert scale in mail surveys
or one on one interviews.
Juster scale is not really an alternative to Likert scale in telephone
surveys unless you could find a shorter way of describing the scale
over the phone.
Juster scale goes as follows:
10 Certain, practically certain (99 in 100 chance)
9 Almost sure ( 9 in 10 chance)
8 Very probable ( 8 in 10 chance)
7 Probable ( 7 in 10 chance)
6 Good possibility ( 6 in 10 chance)
5 Fairly good possibility ( 5 in 10 chance)
4 Fair possibility ( 4 in 10 chance)
3 Some possibility ( 3 in 10 chance)
2 Slight possibility ( 2 in 10 chance)
1 Very slight possibility ( 1 in 10 chance)
0 No chance, almost no chance ( 1 in 100 chance)
Juster developed this probability scale which bears his name in
response to the poor predictive performance of buying intention scales.
He observed that many respondents who stated no buying intention
accounted for a large proportion of purchases, while only a proportion
of those who said they intended to buy actually did so. The scale really
combines the verbal probability descriptions and numeric probabilities.
Multiplying the number of responses for each probability by that
probability and dividing the results by the total number of responses
gives an estimate of the mean population rate (purchase or other.)
For more on Juster scale see Juster, F (1966.) _Consumer Buying Intentions
and Purchase Probability_. Occasional paper 99, National Bureau of
Economic Research, Columbia University Press. Also see Hoek and
Gendall in Journal of the Market Research Society, Volume 35
number 4, _A new Method of Predicting Voting Behavior_.
I have used Juster scale in conjoint studies and obtained very
consistent and reliable results.
Regards,
*-----------------------------*
| Ismail Parsa |
| Epsilon Data Management |
| 50 Cambridge Street |
| Burlington MA 01803 USA |
| |
| E-MAIL: s...@epsilon.com |
| V-MAIL: (617) 273-0250*6734 |
| FAX: (617) 272-8604 |
| |
| The Usual Caveat Applies |
*-----------------------------*