Descriptive design questionnaire

Questionnaire
• A questionnaire is a self-report instrument designed to gather ‘valid’
and ‘reliable’ information
• Questions may be structured (specified) or unstructured (open or
unspecified)
• The use of questionnaires as data-gathering research instruments is
based on 3 key assumptions:
1. Respondents can read and understand items
2. Respondents possess the necessary information to answer the
items
3. Respondents are willing to answer the items honestly
Principles of questionnaire design 1: The basics
• Select only those potential items (variables) that are consistent with:
(1) specific and theoretically-driven research questions, (2)
operationally-defined constructs, AND (3) the target sample or
population
• Limit questions/items to only those of major interest
• Item development is crucial and should be based on extensive:
(a) interviewing (b) piloting (c) pre-testing
12/26/2012 1

Principles of questionnaire design 2: Constraints
• Time:
- Not more than 30 minutes for adults (about 15 mins for mail
questionnaire)
- Not more than 10-15 minutes for children and adolescents
• Sensitivity or delicacy of content
• Design
• Avoid ‘non-response’ bias
- The attributes of a well-designed questionnaire do not arise naturally
out of the process of questionnaire construction, but are the result of
careful developmental work
Principles of questionnaire design 3
• Decide major research question(s) variables and their possible multiple
indicators
• Decide key demographic and classificatory variables [e.g., age, sex,
education, occupation, residential (or work) location, etc.]
• Design response categories to avoid potential ‘non-response bias’

12/26/2012 2

• Use simple nomenclature, direct statements/questions
• Avoid, at all costs, double- or multiple-barreled questions
• Wherever possible, avoid the use of negatively-worded
questions/items; i.e., state all items in the POSITIVE
• TRYOUT potential items with respondents drawn from the target
sample/population
• Revise wording and ordering of items
• Pay particular attention to LAYOUT
• Conduct an ‘empirical’ TRIAL of the instrument and be prepared to
revise the instrument drastically
• ADMINISTER the revised questionnaire

12/26/2012 3

Principles 2
SURVEY INSTRUMENT LENGTH
One outcome of a good pretest is to find
out how long it takes to complete a
survey instrument … The federal Office of
Management and Budget has set as a
guideline that surveys should take
less than half an hour unless there is a
compelling reason why more information
is needed. There are many academic
surveys, however, that last an hour or
longer. (Fowler, 1993, p. 103)

12/26/2012 4

Principles 4
Avoid also the use of negative statements
if possible (surprisingly confusing), and
double-barrelled questions (or two
questions in one). Making questions easy
to answer involves avoiding hypothetical
situations, jargon, technical language and
ambiguity. (Gorard, 2001, p. 95)

12/26/2012 5

Principles 5
Ordering of items
Before writing actual questions it is useful
to consider the overall design of your
questionnaire instrument. Perhaps the
most crucial here is the order in which
items will appear. This applies to the
order of the questions in each
section, and the order of each section
within the whole. (Gorard, 2001, pp.
87-88)

12/26/2012 6

Organisation of the questionnaire
Use of multiple questions
It is common interviewing technique to use several questions to meet a
single objective.
The sequence of questions within a topic
The funnel sequence of questions
The term refers to a procedure of asking the most general or unrestricted
questions in an area first, and following it with successively more
restricted questions. In this way the content is gradually narrowed to
the precise objectives:
1. How do you think this country is getting along in its relations with
other countries?
2. How do you think we are doing in our relations with Russia?
3. Do you think we ought to be dealing with Russia differently from
the way we are now?
4. (if yes) What should we be doing differently?
5. Some people say we should get tougher with Russia, and others
think we are too tough as it is. How do feel about it?
12/26/2012 7

Inverted funnel
There are times when it is desirable to invert the sequence and
start with the specific questions, concluding by asking the
respondent the most general question.
The inverted sequence is especially appropriate for topics in
which the respondent is without strong feelings or on which he
has not previously formulated a point of view.
Attitude scales
There are, of course, a number of other patterns by which the
ordering of questions may be prescribed. Among the various
methods is the attitude scale, in its many forms. Each of the
various methods of scaling specifies or implies its own criteria
for ordering questions. The techniques of Thurstone, Guttman,
Likert, and Coombs are relevant examples.
Organisation of topics
The sequence of topics is not haphazard or random; it is an
organised, systematic progression

12/26/2012 8

Labeled response categories
Sheatsley (1983, as cited in Maxim, 1999, p. 222) itemizes some of
the more popular. Among these are:
• excellent – good – fair – poor
• approve – disapprove; favour – oppose; good idea – bad idea
• agree – disagree
• too many – not enough – about right; too much – too little –
about right amount
• better – worse – about the same
• very – fairly – not al all
• regularly – often – seldom – never
• always – most of the time – some of the time – rarely or never
• more likely – less likely – no difference
12/26/2012 9

Likert scales
The origin of attitude scaling can be identified with the work of
Thurstone and Likert that appeared in the late 1920s and
early 1930s (Thurstone, 1928; Likert, 1932).
- Thurstone’s binary: generally yes/no
- Rensis Likert: continues to be one of the most commonly used
scale applications; the scale is simple to use; it seems to
work well in a wide range of circumstances; and it appears to
be fairly robust.
(Maxim, 1999, pp. 223-224)
1 = Strongly disagree
2 = Disagree to some extent
3 = Uncertain, neither agree nor disagree
4 = Agree to some extent
5 = Strongly agree

12/26/2012 10

 Likert’s primary concern: unidimensionality (all the
items measure the same thing)
 Probably 100 respondents: sufficient for most
purposes (but 250 or 300: not unusual)
 Decision: a high score to mean a favourable or an
unfavourable attitude  consistent from then on
 Adding up the items scores to obtain a total score:
- Pool: 132 items
- Possible range of total scores: from 132 to 660 (5
x 132) for each subject
 Table: a scale for mothers, dealing with acceptance
or rejection of children

12/26/2012 11

Descriptive design questionnaire

More Related Content

What's hot (20)

Similar to Descriptive design questionnaire (20)

More from tungnth (6)

Recently uploaded (20)

Descriptive design questionnaire