180 The Marketing Book
The following three conditions (Tull and Haw-
kins, 1993) should be met if observation is to be
used successfully:
1 The action must be accessible and overt; thus,
the measurement of feelings, motivation,
attitude etc. is ruled out.
2 Actions should be frequent, repetitive and
predictable.
3 Actions should encompass a reasonably short
time span.
Modes of observation are classed according to
four main factors: naturalness, openness, struc-
ture and directness.
Questionnaires and their design
A questionnaire is an ordered set of questions
which may be employed in a variety of research
situations. They may vary in structure, i.e. the
amount of freedom which is allowed to the
respondent in answering the questions. Highly
structured questionnaires, with set answer for-
mats, are usually easier to administer, answer
and analyse; unstructured questionnaires are
usually harder to administer, need more
thought on the respondent’s part and require
considerable interpretative skills in their
analysis. The situation-specifics of the research
context will condition, largely, the type of
questionnaire to be employed.
The format of the questions may be dichot-
omous (a yes/no type answer), multiple-choice
(where respondents are invited to select one or
a number of responses from a predetermined
list) or open-ended (where the respondents
reply using their own words). Though these
questions are more difficult to interpret, they go
some way in eliminating interviewer bias.
When deciding on the questions, Webb
(2000) suggests that the following questions be
asked:
(i) Is the question necessary?
(ii) Will the respondent comprehend the
question?
(iii) Is the question sufficient to elicit the
required data?
(iv) Does the respondent have the necessary
data to answer the question?
(v) Is the respondent willing/able to answer the
question?When phrasing the questions, the questionnaire
design should ensure that the vocabulary used
is appropriate for the respondent being ques-
tioned, and that only the clearest and simplest
words are used. Also, vague/ambiguous ques-
tions should be avoided, as should biased words
or questions which might ‘lead’ the respondent.
Those questions which contain estimates or
which rest on implicit assumptions may be
difficult to analyse, and should only be asked if
absolutely necessary. All questionnaires should
undergo rigorous pre-testing on a sub-sample of
potential respondents before use.Qualitative research methods
Not all research objectives may be met by the
use of a question and answer format, good
though these methods are at gathering data
concerning knowledge of facts, incidents of
past/present behaviour patterns etc.; other
areas of human activity do not fall into such
convenient and relatively easily accessed cate-
gories. Such areas include respondents’ atti-
tudes, motivations, opinions, feelings etc., as
well as other types of question which might
cause respondents to experience heightened
levels of anxiety or embarrassment, or where
they might feel a difficulty in putting their
answers into words.
Qualitative research methods are
employed to uncover other ways of gaining
access to such types of data; they seek to
answer the ‘why’ and ‘how’ questions, rather
than the ‘what happened’ or ‘how many’ types
of enquiry.
A comparison between quantitative and
qualitative methods has already been given by