The Marketing Book 5th Edition

236 The Marketing Book

research and international marketing. For a
detailed example of an application of LISREL,
see Moutinho and Meidan (1989). Table 9.13
summarizes the applications of the two causal
models in marketing.

Hybrid models

Under this category there are three different
types of methods used in solving marketing
problems: dynamic programming; heuristic
programming; and stock control models.

Dynamic programming

As stated by Budnick et al. (1977), dynamic

programming is a recursive approach to opti-

mization problems which works on a step-

by-step basis utilizing information from the

preceding steps. The model has been used to

aid decision making in areas such as distribu-

tion (i.e. the minimization of transportation

costs), the distribution of salespeople to various

territories (in such a way that maximum profits

will be obtained), and determining the best

combination of advertising media and fre-

quency under a budgetary constraint.

Table 9.13 Applications of causal models in marketing (the techniques,

advantages, limitations and when recommended to use)

Model Content Marketing

applications

Advantages Limitations

Path analysis Provides the means

for studying the

direct and indirect

effects of variables,

by offering

quantitative

information on the

basis of qualitative

data on causal

relations

Tourism marketing

(McDonaghet al.,

1992), other

applications

(Wasserman, 1989)

Displays graphically

the pattern of

causal relationships

Assumes

relationships among

variables are linear

LISREL (linear

structural

relations)

A structural

equation modelling,

that enables one to

decompose

relations among

variables and test

causal models that

involve both

observable and

unobservable

variables

Consumer

behaviour, personal

selling, marketing

strategy,

international

marketing

(Moutinho and

Meidan, 1989)

Provides an integral

approach to data

analysis andtheory

construction. The

method easily

handles errors in

measurement.

Ability to bring

together

psychometric and

econometric

analyses

Requires an a priori

theory for

structure analysis