Information Technology Reference
In-Depth Information
looking at the site will consider taking the survey. A link to a web-based survey
can also be included in the body of a message and emailed to a group of poten-
tial participants.
Emailed surveys tend to look and feel like any other instrument you might
receive in the mail or be handed by a researcher. It takes no programming skill
to paste a survey into the body of an email or attach a word processing docu-
ment to an email. However, doing so severely limits the benefits of conducting
survey research online. One of the issues with emailed surveys is that there can
be conflicts between versions of word processing programs, making it difficult
or impossible for a participant to open and/or complete the survey. If the sur-
vey is included in the body of an email, the format can be altered, causing ques-
tions to appear strange to the person completing the survey and adding
extraneous symbols (e.g. ?,
>
, and *). All of the data gathered still need to be
cleaned and prepared for analysis much like data collected offline.
Creating a web-based survey instrument mitigates many of the downsides to
emailing a survey. The instrument, for example, will appear the same to all par-
ticipants in its formatting, images, coloring and so forth (Mann & Stewart,
2000). It's also possible to format responses with check boxes
, radio buttons
, and text boxes for short-answer questions, making it significantly
easier for people to provide answers. Analysis is much easier as well, given that
the instruments are formatted consistently and, typically, there is a back-end
database collecting all of the response data.
The limitations of choosing a web-based survey model for your research are
relatively obvious. The first is the need for technical skill. Survey sites are built
as a progression of HTML web pages containing a number of form items for the
user to respond to questions via text fields, check boxes, multiple choice buttons
and perhaps sliders. The interaction is via CGI and thus survey sites are usually
built with the usual dynamic web scripting tools, very often with PHP. The
returned data is stored on the server, usually in a database system. The technol-
ogy required is quite similar to that which enables blogs and wikis. Formatting
a survey takes time and skill. It also requires some type of hosting environment
such as an institutional server or an ISP (Internet Service Provider). In addition
to the web programming (in PHP for instance) and database skills necessary for
constructing the survey instrument, it is necessary to be comfortable using FTP
(File Transfer Protocol), or other methods, for uploading your files on to your
host server and downloading data as they are acquired. In some instances,
though, the need for manual (FTP) downloads can be avoided by the use of a
script to automate data downloads. If the technical skills noted need to be
acquired, keep in mind that you will require a budget for that support.
If you are debating which way to go with your research, we hope that budget
expenses for technical skills will not be the primary deciding factor. Keep in mind
that other types of surveys would involve mailing costs, copying services, paper
expense, and the staff to input data. Even if surveys are emailed, someone has to
clean up the data and prepare them for analysis. Therefore, we encourage you to
