Paul Beck wrote:
> Dear all
>
> We are looking at requirements for polls and questionnaires for our
website. This will be everything from quick and easy polls to more
in-depth surveys of our student population's activities and interests etc.
>
> Does anyone have any experience of good software packages (there seem
> to be an awful ot of them) or have you built your own polls?
We have been looking at a number of packages as part of our
e-consultation research project (http://econsult.mgt.qub.ac.uk/). We
have not, however, thoroughly reviewed all the products, so all I can
give are some initial pointers.
1. It is important, for academic use, to support quite sophisticated
question types. These include matrices, with a scale across the top
and items down the left, with restrictions on where you can click.
One form of answer that is particularly difficult to implement, and
so gets left out of many products, is the ability to rank a dozen
items. For political science questions, ranking reveals potential
compromises that a Likert scale will not (see http://www.deborda.org/).
This requirement for sophisticated questions eliminates a lot of the
packages (both commercial and free software).
2. The usability of the packages is crucial. At the first difficulty,
respondents give up. Some packages show their origin in paper
questionnaire design tools. The web interfaces have been fitted as an
afterthought - and it shows. The SPSS online questionnaire tool is an
example of this. You design the questionnaire as if it is on paper, and
the users have to fill in long forms with oddly spaced items. When a
tool is developed by people experienced in web design, you get fewer
usability problems.
3. A poll is not the same as a survey. In formal surveys you want people
to carefully consider answers to a lot of questions, and not be affected
by other people's answers. Web polls ask one question, and immediately
take you to a graph of results. Don't like the results? Vote again (like
in Irish elections).
4. In some survey situations, you want to be able to invite a particular
sample of people to vote (e.g. via e-mail invitations), and then track
who has responded. Not all packages support e-mail invitations and
response tracking.
These criteria reduce the choices somewhat. Having worked through those,
the products we are looking at in more detail are:
phpsurveyor (http://phpsurveyor.sourceforge.net/),
Open Survey Pilot (http://osp.serie-a.de/english/),
mod_survey (http://gathering.itm.mh.se/modsurvey/), and
Responsian Consultation Suite (http://www.responsian.com/).
No doubt there are others.
--
Dr. David R. Newman, Queen's University Belfast, School of
Management and Economics, Belfast BT7 1NN, Northern Ireland (UK)
Tel. +44 28 9097 3643 FAX: +44 28 9097 5156
mailto:[log in to unmask]
http://www.qub.ac.uk/mgt/
|