Inert Set or Inferred Set? A Comment

Marketing Notes and Communications
sion has no research procedures for dealing with
deception in advertising.
I certainly must agree with Preston's call for a
different strategy on the part of social scientists—the call to learn how and why things hap-
59
pen. We suggested a very similar plea in 1974.**
Fortunately, I believe the current literature and
Preston's own major work demonstrate that many
are giving heed to this need.
4. Same reference as footnote 2, p. 46.
Inert Set or Inferred Set?
A Comment
Pradeep Kakkar
Kakkar questions the actual categorization procedures used to identify
evoked and inert set products.
N
ARAYANA and Markin's "alternative conceptualization " of consumer response in a multibrand choice situation was presented in terms of
a model of percepttial and/or attitudinal positioning of brands.' The authors make an important
point in suggesting that the concept of evoked set
should be supplemented by other "sets" based on
consumer evaluations of known brands. Such a
categorization process would be analogous not
only to the Sherifs' proposal^ cited by Narayana
and Markin, but also to Emerv's mapping model
with respect to price and quality perceptions.^
The existence of such zones of acceptance, neutrality, and rejection would have several implications for marketing strategy\ as the authors point
out, as well as for such fertile areas in consumer
research as attitude measurement, information
processing, and brand loyalty. For instance, it
may be hypothesized that new information about
a brand in the consumer's evoked set would be
processed differently from new information about
a brand categorized as inept. It is important,
therefore, that tests of the model should closely
examine the categorization process. It is the purpose of this note to point out methodological issues in the Narayana-Markin study that render
their test of the model invalid. Specifically, objection is raised with respect to three issues: (1)
1. Chem L. Narayana and Rom J. Markin, "Consumer Behavior and Product Performance: An Alternative Conceptualization," JOURNAL OF MARKETING, Vol. 39 (October 1975),
pp. 1-6.
2. Caroline W. Sherif, Muzafer Sherif, and Robert E.
Nebergall, Attitude and Attitude Change (Philadelphia: W. B.
Saunders Co., 1965).
3. F. Emery, "Some Psychological Aspects of Price," in
Pricing Strategy. B. Taylor and G. Wills, eds. (Philadelphia:
Brandon/Systems Press, 1970).
inconsistency between concept definition and
concept measurement, (2) measurement bias, and
(3) interpretation of the data.
With regard to the first issue, the NarayanaMarkin definitions of evoked, inert, and inept sets
are given in terms of consumer evaluation of
brands. "The brands in the evoked set are
evaluated positively. . . . The inert set consists of
those brands in the product category for which
the consumer has neither a positive nor a negative evaluation. . . . The brands in the inept (reject) set are negatively evaluated. . . ." However,
measurement of these concepts is in terms of intentions to btiy. This discrepancy illustrates the
conceptual argument repeatedly raised in research on the attitude-behavior relationship, most
notably by Fishbein: namely, that construct measurement should be consistent with construct
spjecification."* Thus, either the response categories
in the model should be defined in terms of intentions to buy or the measures obtained should be
directly those of e\aluation.
Second, the asymmetrical wording of the questions used by Narayana and Markin could be a
potential source of response bias. In particular, the
context suggested in measuring the evoked set is
very different from that implied in measuring the
inept set. Specifically, the former asked respondents to list brands that "you consider buying."
Instead of measuring the inept set similarly, for
example, by asking for a list of brands that respondents would not consider buying, Narayana and
Markin asked for names of brands that "you do not
4. Martin Fishbein, "Attitudes and the Prediction of Behavior," in Readings in Attitude Theory and Measurement.
Martin Fishbein, ed. (New York: John Wiley & Sons, 1975).
60
consider buying anytime at all." One possible result
of this differential polarity may have been to inflate
the size of the inert set if respondents were unwilling to adopt the strong position suggested by the
inept set question.
Finally, the method used to "measure" the inert
set leaves the interpretation of the data open to
question. The authors obtained the names of brands
in the awareness set, the evoked set, and the inept
set. Without a similar measure of the inert set, they
concluded by inference that brands in the awareness set not included in the evoked or inept sets
• ABOUT THE AUTHOR.
Pradeep Kakkar is assistant professor of marketing in
the Wharton School, University of Pennsylvania,
Philadelphia.
An Explanation
Journal of Marketing, July 1976
automatically constituted the inert set. In other
words, three states of nature were defined and
elements absent in two states were presumed to be
present in the third category. While this inferential
approach may be acceptable where the states of
nature are known to exist, it cannot be held up as
verification of a hypothesized model. The implication that a consumer who does not consider buying
a brand, and at the same time does not wish to
reject it absolutely, lacks information with respect
to that brand is not justified; the existence of the
inert set will have to be established by direct measurement for a valid test of the hypothesis.
Attention to these methodological issues in further tests of the Narayana-Markin proposal should
enhance its potential usefulness for marketing
practitioners and researchers.
Chem L. Narayana and Rom J. Markin
The comments of Professor Kakkar raise some
questions concerning the methodological issues in
our article.' It should be stated at the outset that
his comments are well thought out and strongly
support the conceptualization developed in the
original article. The theoretical model and its
conceptualization comprise the major contribution of the original paper, while the empirical
part was "strictly exploratory in nature" in order
to verify the model. The methodological issues
raised by Professor Kakkar relate to this empirical
part of the article.
With regard to the specific objections raised by
Professor Kakkar, there is no inconsistency between concept definition and concept measurement. The concept of evoked set is defined in the
article as "the set of brands of a product which
the buyer actually considers when making a
specific brand choice." Similarly, the inept set is
defined as "those brands that the consumer has
decided not to consider for purchase at all in their
present form." The positive and negative evaluation process is a precondition to this choice behavior. Hence, the concepts define intentions, not
the preceding evaluations. Thus, there is absolutely no inconsistency between concept definition and concept measurement.
It follows from the above explanation that there
is no measurement bias due to the asymmetrical
wording of the questions. The definitions of the
concepts necessitate the asymmetrical wording of
the questions.
Professor Kakkar's final objection concems the
interpretation of the inert set. It is true that the
conclusion on the inert set is inferred once the
evoked and inept sets are identified. We do not see
what difference it makes whether the inert set is
inferred or obtained directly. The implications
remain the same as far as the brands in the inert
set are concemed. It should also be pointed out
that there may be several resisons for any brand
to be in the inert set,^ not just the consumer's lack
of information as Professor Kakkar mentioned.
We are in the process of developing direct measurement scales to test some specific hypotheses.
We believe that the model is pragmatic and that
any further development of methodology and testing would enhance its potential usefulness for
marketing practitioners and researchers, as Professor Kakkar pointed out.
1. Pradeep Kakkar, "Inert Set or Inept Set? A Comment,"
JOURNAL OF MARKETING, Vol. 40 (Julv 1976), pp. 59-60; and
Chem L. Narayana and Rom J. Markin, "Consumer Behavior and Product Performance: An Alternative Conceptualization," JOURNAL OF MARKETING, Vol. 39 (October 1975),
pp. 1-6.
• ABOUT THE AUTHORS.
2. Narayana and Markin, same reference as footnote 1,
p. 2.
Chem L. Narayana is associate professor of marketing and Rom J. Markin is professor of marketing in
the College of Economics and Business, Washington
State University, Pullman.