Marketing research 185
research projects are not intended to measure
such tangible factors; some measurements are
far more complicated because of the ambiguous
nature of the answer to the question: ‘In order
to meet the objectives of the research, what is to
be measured?’
Torgerson (1958) said measurement is ‘...
the assignment of numbers to objects to repre-
sent amounts of degrees of a property pos-
sessed by all of the objects’. Now while it is a
relatively simple process to see how such a
definition applies, for example, to age (number
of years since birth), weight (number of units of
gravitational attraction) etc., it is not such a
clear-cut process when it is necessary to meas-
ure those factors, important in a social research
setting, which are far more covert. For example,
there is no universally agreed system for ‘the
assignment of numbers’ to a respondent’s
attitude towards a certain brand of coffee, or
their motivation in purchasing one brand of
motor car in preference to another.
In measuring abstract constructs such as
beliefs, motivations, feelings, attitudes etc.,
marketing research may have to express them
‘in terms of still other concepts whose meaning
is assumed to be more familiar to the inquirer’
(Green et al., 1988). To evaluate a research
situation, there is a requirement to measure
factors/variables, overt and covert, which are
relevant, but there is also a need to know what
to measure. Attitude, for example, can be
defined in many ways, some of which have
more and some less relevance to a specific
situation, because attitude is a multidimen-
sional concept. In making a definition of atti-
tude, some of these components are excluded,
thus researchers must note that measurement is
never fully able to translate reality into sets of
numbers – representation can only ever be
incomplete.
Variables are factors relevant to a research
situation which vary and in doing so affect the
state of that situation. For example, many
research studies are concerned with consumer
responses to proposed changes in a product’s
price: thus, the dependent variable is consumer
response; the independent variables might
include packaging, the price of competing
brands/products and brand loyalty. The
research objectives will have stipulated what
the outcome of interest is – the dependent
variable (here, consumer response). There may
also have been implicit assumptions as to what
the independent variables are, but the implicit
needs to be made explicit. One way to accom-
plish this is to construct a model of the research
situation. If there is insufficient information to
do this, then some additional research may be
required – exploratory, for example. It may also
be possible to determine the significant vari-
ables by means of a thorough analysis of the
literature – secondary data search.
Southern (1988) writes that there are three
important components of the measurement
process:1 Measurement is a process; it is controlled and
open, not arbitrary or intuitive.
2 Measurement translates qualities into
quantities; the numbers may then be
manipulated. However, numbers themselves
have no meaning, and those who manipulate
them must exercise care if the validity of the
relationship between number and
characteristic is to be preserved.
3 Measurement has formal rules which may vary
depending upon the manipulation, but once
set, they must be followed consistently if
reliability of data is to be guaranteed.Green et al. (1988) say that number systems
have:1 Order.
2 Distance– differences between numbers are
ordered.
3 Origin– number systems will have a unique
origin indicated by zero.Scales in marketing research
There are four main levels of measurement –
nominal, ordinal, interval and ratio – and each