designing surveys

[Eric Hansen <ehansen -at- COMWARE -dot- COM>]

Tony's point about even versus odd number scales is not quite accurate.
Whether your survey should have and even or odd number of choices is
really answered by that old reliable:  "It depends."
An odd numbered scale does allow people to take the middle ground, but
sometimes that is appropriate.  An even numbered scale forces people to
make a choice, and sometimes that is appropriate.   Which kind of scale
is appropriate in what situations is a judgment call.  I suggest, if at
all possible, that you and the client together decide.
BTW, this type of scale is called Likert scale.  I assume it's named
after someone named Likert, but I do not know that for sure.
Eric Hansen
Instructional Designer
Comware Incorporated

        -----Original Message-----
                From:   Anthony Mohr [mailto:amohr -at- CASTEK -dot- COM]
<mailto:[mailto:amohr -at- CASTEK -dot- COM]>
                Sent:   Thursday, October 08, 1998 9:16 AM
                Subject:        Re: Designing a survey

Suzette Seveny wrote: "I have a couple of surveys I use on a regular
basis;
one for documentation and
one for training.  I always use a scale of one to five
with an explanation
as
follows:

        1 -     Very dissatisfied
        2 -     Somewhat dissatisfied
        3 -     Neither satisfied nor dissatisfied
        4 -     Somewhat satisfied
        5 -     Very satisfied"

I read somewhere that when you design a survey with a
scale, the scale
should always be made up of an even number of options.
If the scale has an
uneven number of options, then people who are unsure
about their response
tend to choose the option in the middle.  However, if
the scale is made up
of an even number of options, people responding to the
survey are forced to
choose an option that most closely aligns with their opinion.
Tony Mohr
Technical Writer
Castek Software Factory
Toronto, ON


 From ??? -at- ??? Sun Jan 00 00:00:00 0000=