Questionnaire Design
- The design of a SAMT questionnaire depends
on two main factors:
- the type of information you want to elicit
- how you intend to process and present the results.
- There are two main types of questionnaire,
one which generates discursive responses and one which produces quantitative
data. In general, a discursive questionnaire is suitable for a module with
small numbers and a quantitative questionnaire more suitable for large numbers.
For example, an individual member of staff wishing to gain an impression of
student opinion on a module with relatively small numbers is likely to find a
questionnaire inviting discursive responses appropriate. However, if a
department wishes to use a single format for assessing all its modules, in order
to provide comparative information, it will be better to design a questionnaire
which generates quantitative data. Each approach is equally valid.
Questionnaires can also be designed to generate quantitative data and discursive
responses, e.g. as a minimum, a quantitative questionnaire could invite students
to add any written comments in a designated space or on the reverse.
- On a discursive SAMT questionnaire, it is
recommended that a minimum of five or six questions is used, e.g.:
- What was the
best (or most useful) thing about the module?
- What was the worst
(or least useful) thing about the module?
- Was the module enjoyable
(or interesting)? If not, why not?
- Overall, how good is the
module? (For this question it is appropriate to include a scale by which
students can respond, e.g. excellent, good, satisfactory, poor, very poor.)
- Any other comments?
- On quantitative questionnaires, the key
factor is the scale which is provided for students’ responses. It is
strongly advised that an odd number of options is used for responses, e.g. three
or five, in order to provide more useful data on aggregation (the average
relates to the middle point on the scale rather than falling between two
points). A three point response scale will provide less detailed
information and may not be appropriate when the range of responses might be:
good, satisfactory, poor. However, it may be useful for other types of
question, For example, Was the length of the module appropriate for the material
covered? Too long, About Right, Too Short.
- Variation in the number of points on the
scale on a single questionnaire is acceptable but may cause difficulties when
the results are processed, particularly if a computerised method is to be used.
However, the range of responses on the five point scale can vary, e.g. from
‘very appropriate to very inappropriate’ to ‘very good to very poor’. In
fact, on a questionnaire with a large number of questions and tick boxes in
columns, variation in the types of response requested will mitigate against the
possibility of the respondent simply filling in the questionnaire randomly or
giving the same response for each question.
- It is important to clarify at the top of a
quantitative questionnaire which end of the scale relates to positive responses
and which to negative responses, e.g.
5-very good, 4-
good, 3-
satisfactory, 2 -
poor, 1- very poor
- It is also important to be consistent in the use of this scale throughout the
department. If there is variation students will become confused about
whether 1 or 5 means very good and so on and the results from some
questionnaires could be confused or invalid.
- On any questionnaire, it is generally good
practice to limit the number of questions in order to ensure the highest
possible response rate. Where a large number of questions is asked, it is
best to break the questionnaire up into different sections, each containing a
small number of questions.
Format of Questions
- On a discursive questionnaire, open
questions which require a detailed response (rather than a Yes or No answer) are
most suitable. On a quantitative questionnaire, the range of responses
must relate to the scale provided for the answers, e.g. How interesting is the module? Very much so
to Not at all. or Overall, how good is the module?
Very good to Very poor.
- It is also important to guard against mixing
‘negative’ and ‘positive’ questions. For example, on a questionnaire with
a five point scale (Agree to Disagree), the following two questions: 'The
classes were helpful' or 'The set work did not link
well with the material covered in the classes' should
generate responses at opposite ends of the spectrum.
Although it may be tempting to include a negative question in the middle of a
set of positive questions because its provocative nature may generate useful
feedback, it is likely to confuse students and will make processing the results
more difficult.
Timing of the Survey
- Where possible, the same periods of
assessment should be used throughout a department to minimise disruption for
both staff and students. This will also allow departments using a single
questionnaire for all modules to make any efficiency gains.
- It is recommended that the assessment takes
place in classes or other supporting tuition, rather than in lectures, since the
response rate is likely to be greater.
Anonymity
In accordance with Senate policy, students
should not be asked to write their names on survey questionnaires or to provide
any other means of identification, e.g. registration number.
Page last updated:
02 May 2013
