Information Technology Reference
In-Depth Information
considered to be a true reflection of a particular phenomenon or issue
being studied. Similarly, researchers often speak of
reliability
. What they
mean is that a research instrument is reliable if it will provide the same
results each time it is administered to the same population. Since we are
not attempting to aid you in the actual development and design of your
research instrument, beyond this brief discussion, we encourage you to
seek other resources on survey design when you begin to develop your
instrument.
Web-based surveys
Web technologies make it possible to publish surveys with a potentially very
large target audience. Other advantages of web-based surveys are reduced
time and costs for data collection, reduction in transcription errors, and the
possibility of more sophisticated interactions. As a way of better understand-
ing the specifics of online surveys, you might want to keep in mind the pre-
ceding information as you read the following pages. One of the major
benefits of going online to conduct your survey research is access to a very
broad population. Given the need for large samples, researchers often look to
the internet with its possibility of access to a very large population. Be
warned, however, that this promise of accessing a large population can be
illusive. Many researchers find that response rates can be lower than
expected (Witmer, Colman & Katzman, 1999), even though the ease of tak-
ing the survey may lead to an expectation of higher return rates. Some sus-
pect that question and instrument length may contribute to low return rates,
mindful that attention spans online have a tendency to be short due to the
nature of the environment. Witmer et al. (1999) found that there was no sig-
nificant difference in return rates based on length, suggesting that there may
be other issues of concern such as ease of answering questions, interest in the
topic, and online interaction with the researcher.
There are some general rules of thumb that researchers can follow to
improve response rates. In general, the easier it is to complete a survey, the
more likely people will do so. Witmer et al. note that response rates can be
improved by avoiding fragmented and varied questionnaires (ibid.: 156). It is
also useful to contact potential participants with some type of pre-notification
message, explaining the nature of the research, a little about the researchers,
any technical details regarding security, how the data will be used, and so forth
(Krishnamurthy, 2004). This initial notification could be outlined in such a way
as to constitute an informed consent, with the participants' reply of a willing-
ness to participate in the survey as an implied agreement (though researchers
are still encouraged to consider including an explanation and formal informed
consent as part of the survey instrument).
