8. Survey Question Considerations
To frame questions for these surveys these considerations were made.
clear English and non technical terms
Some respondents may not have English as their first language
and most will likely not have medical background. The words
used were chosen on purpose to be down to earth and based
on experience of the day to day.
People have limited time to fill out surveys. An attempt was
made to ask question the shortest way possible.
To respect people’s time, questions were often asked in
checklist form or multiple choice so that many aspects
of a question could be gone through at the same time.
Questions were made optional within a survey, just as surveys were themselves
optional. This was done so that if a person wanted to not
divulge an answer or did not find a question relevant they
did not then have to stop the survey but could just skip
that question. Some surveys require an answer to every question.
This was one designed to be optional..
This is one of the most comprehensive surveys every done
of dystonia. It asks the very detailed questions people
mention in books, anecdotally or on social media that
clinical studies either may not have thought to ask
or did not have time to ask. There were i
over 2000 questions in these surveys. That number enables
a very wide study of the nature of this very poorly understood
condition. The questions may seem small specific about
time of day when pain is worse, or whether a bra strap
hurts but these features may matter.
no labels of types of dystonia
The names for types of dystonia have varied widely over
the years. People with a diagnosis 15 years ago may
nowadays be under a different category. In addition there
are often two ore more labels for the same dystonia. A dystonia may be
both focal and cervical for instance, given categories and
sub categories. To be the easy to understand, questions
focused on the body part involved instead.
A real effort was made to avoid words that have stigma. The surveys
do not ask ‘Are you depressed?” which is a medical label some may not
like. To determine mood, questions were more specific asking about trouble sleeping, interest in eating.etc.
no bias in the question
It is easy to inadvertently hint at answers in questions
if the questioner expects a certain answer. Great effort was made
to avoid such bias. For instance instead of asking
“Is your pain worse at night or in the daytime?” which
assumes there is pain, the question was framed to ask
if there is pain, and only later when that pain occurs.
clarity of question
Many studies seek a numerical answer for statistical analysis.
People are asked if they have trouble sleeping some or most
or all of the time for instance. From the respondent perspective
though, the differences between ‘some’ and ‘most ‘ may be
unclear. The survey tries to use scales only when the options are
clear though that is sometimes a challenge.
An effort was made to avoid confusion. Sometimes in surveys, the question
asked is so precise that it has become contorted. For instance
if a person asks if the respondent agrees or disagrees with
the statement “I do not feel much energy each day” there
may be confusion. The respondent is trying to figure out
if he disagrees that he does not, so that would mean he does.
An attempt was made to ask questions clearly though this
is sometimes easiest to know if a focus group does the
questions and gives feedback.
An effort was made to get feedback on the right topic and
to frame questions that could not mislead. To ask if dystonia
has made a person have to eat only soft food, if the question
asked “Do you eat applesauce or yogurt daily?” the person
who responds yes may only do so because they have always
eaten applesauce daily because they like them not due to
number of questions
The Survey Monkey platform can handle 5000 questions.
This series did eventuallyget responses to several thousand questions
often in checklist format. To keep the task of responding less daunting,
a short short survey of 25 questions was created for those who only wanted a quick way to describe their experience.
adjustments as project progressed
errors of design
It was noticed once the questions were online and a few
people were doing them, that there were some typographical
errors and a few other missteps such as listing something
as multiple choice (one answer possible only) when it should
have been checklist (as many answers as one wants).
When errors were noticed they were corrected except when
there had already been people answering the question in such
numbers that data of their responses would be lost if the
question was rewritten. In that case, the question was left as
it was, in order to keep the survey results accurate. However
ideally such errors would have been caught sooner in
the design stage.
new input half way through the project
Over the course of a year, research about dystonia
continues and social media and patients continued
having insights. Six months into the study it became clear
that a few other very good questions could have been asked. At
the half way point therefore, a second series of questions was put
up, called “one more go”. These questions attempted
a follow up to the first set, and provided the study with the luxury
of being its own follow up. However this set was not as well
subscribed to, was not as well advertised and it is likely that
most respondents were tired of the process and felt they
had done enough. It is not clear if the “one more go’ series
then is going to be as useful, though some of those surveys
did reach fairly high respondent numbers. A few did not.
What seemed useful in the end was to have the second set question
on paper so that researchers considered it as a question
useful to ask, were they to design a follow up survey.