Please enable JavaScript.
Coggle requires JavaScript to display documents.
Theory: issues and aspects (Deterministic vs. stochastic theory (Much…
Theory: issues and aspects
Classificational
schemata
For different kinds of goods, stores, wholesalers, pricing policies etc
organizing phenomena into classes or groups that are amenable to systematic investigation and theory development
Are not themselves theoretical because they lack the requisite lawlike generalizations! (theoretical constructions will contain these as components)
Classification systems involve a portioning of some universe of objects, events, or other phenomena into classes or sets that are homogenous with respect to some categorical properties
Logical
portioning
Deductive
classification,
a priori
classification
Classificational schema is always developed
before the researcher analyzes data
-> researcher imposes a classificational system on the data
Usually results in
monothetic classifications
: all members of a category possess all of the characteristics or properties used to identify the category
Can result in either
single-level or multilevel schemata
There may exist
empty classes
: a proper application of the categorical terms may generate a class to which no phenomenon belongs
Presupposes understanding of the phenomena being investigated
or else the classifications involved may be totally unrealistic -> any universe of phenomena can be classified in an infinite variety of ways
Grouping
Inductive
classification,
ex post
classification
Researchers generate their schemata
only after they analyze data
--> researcher lets the data suggest the classificational system
Grouping procedures are designed to conveniently accommodate
larger number of propertie
s than in case with LP
All grouping procedures share the common characteristics that they determine categories by an analysis of a specific data set
Correspondence analysis, factor analysis, cluster analysis… ->
separate phenomena into groups that maximize both the degree of likeness within each group and the degree of differences between groups according to some objective function
Polythetic classes:
the phenomena in any given class may share many characteristics in common, but no individual phenomenon need to possess all of the characteristics of the class
Does not generate empty classes
because those can be formed only from existing observations
Requires less a priori knowledge
concerning which specific properties are likely to be powerful for classifying phenomena
Criteria for evaluating
classificational schemata
Does the schema adequately
specify
the phenomenon
to be classified?
Does the schema adequately
specify the properties/characteristics
that will be doing the classifying?
Does the schema have
categories that are
mutually exclusive?
if an item fits one class, it will not fit any other class -> no single item may fit 2 different classes at the same level
Does the schema have
categories that are
collectively exhaustive?
every item that is to be classified should have a “home” -> “other”-category
Is the schema
useful
? -> does it serve its purposes?
Positive vs
normative theory
Positive theories:
systematically related sets of statements, including some lawlike generalizations,
that are empirically testable and that increase
scientific understanding through the explanation
and prediction of the phenomena
In marketing,
normative (rational theories)
Structure
: positive theories must contain lawlike generalizations, normative theories prescriptive statements (can’t explain the phenomena)
Purpose
Positive:
to increase our understanding of
marketing phenomena by providing
systematized structures capable of
explaining and predicting phenomena
Normative:
is to assist marketers in making better decisions
Validation
criteria
Positive:
checking the
internal logic and mathematics
of the theory and then exposing
it to empirical tests in the real world
Normative:
the internal logic and mathematics can be checked and verified, further, the usefulness of the theory can be evaluated -> cannot be empirically tested
Deterministic vs.
stochastic theory
Much marketing theory uses tendency laws/relationships -> these relationships are stochastic? -> Bass: all marketing behavior may well be fundamentally stochastic
A theory is deterministic if and only if,
given the
values of its state variables for some initial period,
the theory logically determines a unique set
of values for those variables for any other period
If the state description of a theory is defined by the values of a set of statistical parameters, and if those parameters can be uniquely specified through theoretical laws, given their values at some initial time, the
theory would be deterministic
Three sources of
uncertainty in scientific explanations
result from errors in measurement, the logical relationship between statements in the explanans and explanandum, and the inability to predict the occurrence of individual phenomena
D-N expl.:
are deterministic, because measurement error is the only form of uncertainty
D-S expl.:
have both measurement error and uncertainty resulting from their inability to predict the occurrence of individual phenomena -> deterministic because their explananda (statistical parameter)
are logically subsumed under the explanans
I-S expl.
: not deterministic, because they contain all three sources of uncertainty
Deterministic theory of some type is a
legitimate goal of research in marketing
Both the deterministic and stochastic schools of research in marketing are developing theory whose
ultimate objective is deterministic