Over the past decade, the use of
online methods for market
research has skyrocketed. Due to ever-increasing
technological advances, it has become possible for do-it-yourself
researchers to design, conduct and analyze their own surveys for
literally a
fraction of the cost and time it would
have taken in the past.
But are there any drawbacks compared to traditional methods (such
as mail, telephone and personal interviewing)? Today I'll
provide a list of several main
advantages and...
Continue Reading »
You have your research
objectives and survey
sample clearly defined. You have your questions and
response options written. So now,
how are you going to
convince as many respondents as possible to complete the
survey? Well, one of the many things you can do to
increase response rates is to
make sure you've included
a solid introduction to your survey.
A useful introduction not only provides the rationale for
completion; it also
answers the common questionsrespondents inevitably have about...
Continue Reading »
Some survey creators want to automatically include a “don’t know”
option on every opinion-based survey question they pair with a
scale. But this shouldn’t always be the case, particularly
because
respondents differ as to whether they will even
admit they don’t know enough to offer a legitimate answer
to a question (online surveys included).
"Don’t know" can obviously be a useful category if it truly
deserves its own response option (i.e., respondents really do not
know enough to have formed an...
Continue Reading »
The title of this post is a question asked early and often by
marketers, managers and researchers conducting survey research
projects. Unfortunately, there is no single answer to the question;
there isn't a "correct" size or minimum number of people to sample
that is appropriate for all survey designs. The closest thing I can
come to a simple answer is actually another question:
how
confident do you want to be in your findings?
The following points, however, can help us better understand all
the...
Continue Reading »
Smaller, targeted groups of people that are of interest to survey
and market researchers are called "rare populations." These groups
do not necessarily have to be small in terms of absolute size; it
is more a function of their relative size and composition in
comparison to the larger population (or frame).
Rare populations for our purposes might include examples such
as:
• Small demographic
segments – consisting of less than 3% of an overall
population (ethnicity, age, education, geography, etc.)
•...
Continue Reading »
Possibly the biggest advantage of using surveys as a primary
research method is that you can accurately describe the opinions
held within a very large and diverse population without having to
interview every member of that population. This advantage goes for
both traditional and online survey methods.
By implementing the ideas put forth by probability theory
(especially random sample selection),
we can accurately
estimate the views of many people by surveying a subset of a given
population......
Continue Reading »
Here at Cvent we encourage you to use random sampling for your
survey projects if you have an exhaustive population list and
accurate contact information. Random sampling can reduce many
types of error and bias. Coverage error is reduced by random
sampling because each unit of the population has an equal chance of
selection. Likewise, random sampling gets rid of
self-selection bias because you control who receives your
surveys and can ensure that your sample is representative of your
larger...
Continue Reading »
In most online survey research projects, respondents are either
sent e-mails with invitations to complete a survey, or a link is
posted on a website and people click and complete. I
recommend going the e-mail route, because it reduces a sampling
error called self-selection bias.
Imagine the scenario of a talk radio show: the host says he’s
conducting a poll, and listeners can call-in to express their
opinion. At the end of the show, the host tallies up the
"results," and concludes that the...
Continue Reading »
In a recent post I outlined three types of attitude
questions (those pertaining to affect, cognition and action), and
why it is important to create separate survey items pertaining to
each concept. Action-based questions attempt to
measure potential behavior, or the likelihood that an individual
will take certain actions, and thus could also be grouped
together with questions of behavior. I’d like to go into a
little more detail about the relationship between attitudes and
behavior - the...
Continue Reading »
Have you ever created frequency distributions from particular
survey questions, only to realize that some response options were
chosen only a few times, and others weren’t chosen at all?
This can occur for many types of questions with ordinal or interval
scales, and is especially common when likert scales are used. When survey
respondents or raters are hesitant to select choices at either end
of a scale, a phenomenon called central tendency
bias may be to blame.
Central tendency bias occurs...
Continue Reading »
Subscribe to this blog by RSS
Powered by Compendium
blog comments powered by