Sunteți pe pagina 1din 15

Session 9

Questionnaire Design
A Questionnaire is words and
communication
• What should be asked?
• How should it be phrased?
• In what sequence should questions be
arranged?
• What questionnaire layout is best?
• What level of rigidity / informality is
appropriate?
• Should the questionnaire be pre-tested?
Can Questionnaire design be
taught?
• Probably not !!!!!
• Each person has their own ways of
phrasing questions
• What can be taught are the many pitfalls in
designing questionnaires
• Be prepared to pre-test your questionnaire
several times until you get it right
• If there is a proven format – use it!
Close ended versus open ended
questions
• Close ended are very much easier to analyse
• They do however make the questionnaire very
rigid and formulaic
• Open ended are very tiresome to analyse, and
convey all shades of meaning which has to be
constrained into artificial categories
• Single choice versus multiple choice questions
• Depends on the way the survey is to be
administered
Pitfalls in design (1)
• Avoid complexity / use clear everyday
language
• Make sure the person has the answer. Is it
something they record (or is it a guess!)
• Avoid the subordinate sentence, which
leads to a leading question
• Avoid the “weasel words”. They often end
in ….ly.
Pitfalls in design (2)
• Be precise over sizes (big, bumper,
standard, etc)
• Do not overstretch peoples memory
• Avoid ambiguity (“Have you stopped
beating your wife yet?”)
• Avoid slang, derogatory words, or
localisms.
• Avoid assumptive questioning (“What did
you think of the football on TV last night?”)
Pitfalls in design (3)
• Non equal scaling (responses have an
inbuilt bias)
• Asking multiple questions in one question
• Embarrassing areas for questions
• Sequencing of questions in accordance
with peoples thought processes
• Put classificatory questions at the end of
the questionnaire
Coping with embarrassing areas
• Don’t ask for exactitude (range answer)
• Use “third person” technique
• Use proximity testing
• Use relativity (younger / older etc)
• Do you need to know!
• Use sympathetic terminology (make love /
have sex)
“Don’t Know”
• What does this response mean, should it
be a category?
• Does it mean “I genuinely do not know
which I prefer?”
• Does it mean “I don’t care?”
• Does it mean “I do not want to tell you my
opinion?”
• Does it mean “I don’t have any view on the
subject?”
Other issues in questionnaire
design
• Layout is everything. If it looks good it will
be completed
• Always remember to say “Thank you” at
the end of the questionnaire, together with
any instructions as to where to send the
results
• Allow sufficient space for free response
questions / ticks / crosses / circles / etc
Data typologies captured by
questionnaires
• Behavioural data: What the person does
or thinks, or has a view on

• Classificatory data: The demographic


variables by which the behavioural data is
classified.

• Data is both aggregate and segmental


Classificatory variables
• Demographic variables: Age, sex, social
class, marital status, employment status,
education, income, children, ethnicity,
religion, etc.
• Discreet variables: No problem
• Continuous variables: How many
classifications, at what points, and how are
they measured?
Some Issues in classification
• How big should the intervals be?
• Where should the intervals be placed?
• Should the intervals be equal in terms of
range / potential number of respondents?
(applies particularly to self selecting
samples)
• What is the classification a measure of,
the individual (as head of household) or
the household as a whole (social class)?
Social class
• There is no such thing as a “classless
society”
• Social class is based on three qualities a
person possesses; skill, power and wealth
• These three qualities are interchangeable
• Social class can be ascribed or achieved
• The social class structure of European
countries is moving from a pyramid to a
diamond structure
Social class measurement
• Social class is one of the most important
discriminators of behaviour in society
• It is not equal to income
• It is normally based on job responsibility of
the head of the household. If there are two
working it is that person with the more
senior job who defines social class
• There are SIX social classes in the UK: A,
B, C1, C2, D, and E

S-ar putea să vă placă și