Quantitative methods in marketing 235
use of resources, with the exception that it
requires separate computational techniques not
normally used in other models.
The transportation model seems to have
limited application, and the area of marketing
where it has been used is in making distribu-
tion decisions. This limited usage is the result of
the way in which the model is formulated.
The major deterministic operational research
techniques and their marketing applications,
limitations and advantages are summarized in
Table 9.12.
Causal models
Under this heading we have two main tech-
niques: path analysis and linear structure
relations (LISREL). Both these methods are
relatively new to marketing. Of the two, LISREL
is more widely applied because of its versatility.
Path analysis
Path analysis (PA) is a method for studying
patterns of causation among a set of variables,
which was popularized in the sociological
literature (Hise, 1975). Though path diagrams
are not essential for numerical analysis, they
are useful for displaying graphically the pat-
tern of causal relationships among sets of
observable and unobservable variables. Path
analysis provides means for studying the direct
and indirect effects of variables. PA is intended
not to accomplish the impossible task of deduc-
ing causal relations from the values of the
correlation coefficients, but to combine the
quantitative information given by the correla-
tions with such qualitative information as may
be at hand on causal relations, to give quantita-
tive information.
Path analytic models assume that the
relationships among the variables are linear
and additive. Path coefficients are equivalent to
regression weights. Direct effects are indicated
by path coefficients. Indirect effects refer to the
situation where an independent variable affects
a dependent variable through a third variable,
which itself directly or indirectly affects the
dependent variables. The indirect effect is given
by the product of the respective path
coefficients.
In a recent example, developed for a
tourism marketing study (McDonagh et al.,
1992), PA was used to measure the effects of
three major environmental factors (exogenous
variables) – preservation of local landscape,
preservation of architectural values, and over-
crowding – as a direct causal impact on two
critical endogenous variables: (1) concern
towards a policy of global conservation and (2)
preservation of cultural values.LISREL
LISREL (linear structural relations) is a method
of structural equation modelling that allows the
researcher to decompose relations among vari-
ables and to test causal models that involve both
observable (manifest) and unobservable (latent)
variables. Path analysis and LISREL models are
two important analytical approaches for testing
causal hypotheses. Essentially, the analyst wants
the reproduced correlations to be close to the
original correlations. The LISREL model allows
the researcher to simultaneously evaluate both
the measurement and causal (i.e. structural)
components of a system.
LISREL allows for a holistic, more realistic
conception of social and behavioural phenom-
ena. It recognizes that measures are imperfect,
errors of measurement may be correlated, resid-
uals may be correlated and that reciprocal causa-
tion is a possibility. A prioritheory is absolutely
necessary for covariance structure analysis. An
important strength of structural equation mod-
elling is its ability to bring together psychomet-
ric and econometric analyses.
Many applications of LISREL modelling
can be seen in such areas as consumer behav-
iour, personal selling, new product adoption,
marketing strategy, organizational decision
making, distribution channels, advertising