Está en la página 1de 12

Monroe Friedman

Word-of-Author Advertising and the


Consumer: An Empirical Analysis of the
Quality of Product Brands Noted by
Authors of Popular Cultural Works

ABSTRACT. The term "word-of-author advertising" has been used to refer to the
fact that writers of screenplays, television dramas, novels, and other popular cultural
products are using brand names in the texts of their works. The study draws upon
data from 265 Consumer Reports tests conducted from 1950--1979 to determine
the quality of the products associated with brands which are frequently used in this
way as compared to those which are not. The results indicate that in a majority
of the tests, product quality was higher for the frequently used brand names in
word-of-author advertising, but a substantial minority of test reports proved to be
exceptions to this general finding. Implications of the study findings are drawn for
consumer policy, especially as it relates to educational programs.

In a recent series of articles, Friedman (1985a, b, 1986a, b) identified


a relatively new publishing practice which he called "word-of-author
advertising." The term was used to refer to the fact that writers of
screenplays, television dramas, novels, and other popular cultural
products are using brand names in the texts of their works. Illustrations of the practice are Don McLean's taking his "Chevy to the
levee" in the hit song American Pie and the steady diet of Valium
consumed by the frenzied urban characters in Neil Simon's hit plays.
Friedman distinguished between two types of word-of-author advertising practices: those that are commercially inspired ("sponsored
word-of-author advertising") and those that are not ("unsponsored
word-of-author advertising"). While the former practice has received
much attention of late in the media, especially the commercial
purchase of "plugs" for branded products in major films (e.g.,
Friendly, 1983), it seems likely that the latter practice is by far the
more common. And the series of studies noted earlier found that
unsponsored word-of-author advertising has increased markedly
since World War II in bestselling American novels, popular American songs, and both American and British hit plays.
Whether sponsored or unsponsored, word-of-author advertising
Journal of Consumer Policy ! 0 (1987) 307--318.
1987 by D. Reidel Publishing Company.

308

MonroeFriedman

should be of interest to marketers since, like the related phenomenon


of word-of-mouth advertising, it places commercial messages in a
nonpromotional context that both theory and practice would suggest
is less likely to evoke critical responses on the part of consumer
audiences. As Williams has noted (1982, p. 382), drawing upon a
principal marketing implication of attribution theory, "Advertisers
cannot fully escape the attribution that a product is recommended
for monetary gain." One way for advertisers to deal with this
problem, according to Schiffman and Kanuk (1983), is through the
citation of products in a neutral context (i.e., a context other than a
paid advertisement). Schiffman and Kanuk believe that this practice
gives the consumer more confidence in a product's value. The
popular cultural works which are the milieu of concern of the
present study would appear to supply an ideal neutral context for
such citations. And, indeed, the fact that business firms have engaged
in the practice of sponsored word-of-mouth advertising since the
early years of the postwar era (Tie-in advertising, 1951) suggests that
sales may have benefitted from this marketing application of attribution theory.
Word-of-author advertising can be expected to prompt concern
on the part of consumer advocates, many of whom would no doubt
see consumer recipients as unduly vulnerable to the influence of
such messages. One likely source of concern is the potentially
harmful effects of cigarette and liquor word-of-author advertising,
especially on children, since ads for such products are not allowed
on television. Some may object to word-of-author advertising for any
product, simply because, as Schudson has noted (see Iggers, 1985),
consumers are likely to "give it more credibility than it deserves."
Nevertheless, advocates are more likely to be concerned if the brand
names used in word-of-author advertising are those associated with
products of low quality. For in such instances, the suspension of
critical faculties on the part of consumer recipients of the messages
could lead to undue influence in the direction of poor choices. Of
course, even products of high quality may be considered to have
shortcomings (e.g., to be too expensive, too wasteful of energy, or
too damaging to the environment). In general, however, if the brand
names used in word-of-author advertising are those associated with
high product quality, it would seem that there is likely to be less
cause for consumer concern.
This study attempts to determine the level of product quality

Word-of-Author Advertising

309

associated with brands varying in frequency of use in word-of-author


advertising. The study addresses this question by first identifying
those brands that are high on this dimension (relative to their
competitors), and then determining how these brands rate on quality
(again, relative to their competitors). In what follows, the study's
methodology is described in detail; subsequent sections present the
study findings and discuss their implications for American consumers.

PROCEDURES

The first procedural step was the identification of a set of frequently


used brand names in word-of-author advertising. Since the earlier
cited studies of unsponsored brand name usage in popular cultural
works in the postwar era had generated frequency of use data for
many different brand names, these studies became the focus of our
attention.
Why these brand name usages are believed to be unsponsored
requires a word of explanation. The criteria for selecting these
popular books, plays, and songs included one which gave high
priority to the early works of the represented authors. This was done
in part to avoid the formulaic (e.g., the thirtieth Perry Mason thriller
from the pen of Erle Stanley Gardner), and to emphasize works
thought to be more likely to reflect contemporary cultural developments. One consequence of implementing this criterion is that almost
all of the selected works represented the efforts of authors who were
relatively unknown prior to the publication of these works. Thus
these early works would have constituted a poor choice for sponsors
seeking to pay for "plugs" of their brand names.
In the author's research, brand name frequencies were examined
for a collection of 315 popular American novels, plays, and songs, all
of which were published in the years 1950--1979. Since literally
hundreds of different brand names were found in this collection a
method had to be devised to select a smaller sample for intensive
study. It was decided to limit attention to brand names associated with
products for which objective quality information was available over
the 30-year period of study. Consumer Reports magazine, which was
published continuously over this 30-year period, was selected as our
arbiter of product quality, and brand names associated with products
not tested in Consumer Reports were eliminated from our pool

310

Monroe Friedman

of candidates. Among those brands dropped at this point were many


popular magazines (e.g., Time) and soft drinks (e.g., Coca Cola).
Next we looked for brands in the remaining pool (we call them
"tested brands") which had scored high on two measures of frequency of usage. The first (Measure 1) was total number of productions (be it novels, plays, or songs) in which a tested brand appeared
and the second (Measure 2) was total number of appearances in
these productions. Thus a tested brand which appeared twice in one
novel and three times in one play would achieve a score of two on
Measure 1 (a total of two productions) and a score of five on
Measure 2 (a total of five appearances in all).
Once we found tested brands which had scored high on the two
measures we looked among these candidates for instances in which a
tested brand's competitors had not scored high on either measure
over the 30-year period of study. The objective here was to compare
the product quality of brands high on the two measures of word-ofauthor advertising with the product quality of competing brands
which were low on these measures.
Six of the tested brand names met this criterion in that their
scores on each of the two measures were at least 100% higher than
the scores of competing brand names in the same product categories.
In what follows we refer to them as "Word-of-Author Brands." Their
names, along with their scores on Measures 1 and 2, respectively, are
as follows: A & P (6, 10), Buick (11, 28), Cadillac (18, 61), Kleenex
(10, 19), Levis (6, 19), and Sears (6, 12).
Close examination of the product tests for the Word-of-Author
Brands appearing in Consumer Reports revealed the repeated presence of certain competitors of these brands over the course of the
1950--1979 period. These "Leading Competitor Brands" for the
Word-of-Author Brands and their respective scores on Measures 1
and 2 are as follows:
A&P:
Buick:
Cadillac:
Kleenex:
Levis:
Sears:

IGA (0, 0), Kroger (0, 0), and Safeway (1, I)


Chrysler (5, 14) and Mercury (1, 2)
Chrysler (5, 14) and Lincoln (1, 3)
Northern (0, 0), Scotties (0, 0), and Vanity Fair (0, 0)
Blue Bell (0, 0), Farah (0, 0), and Lee (0, 0)
J. C. Penney (0, 0) and Wards (0, 0)

To determine how the Word-of-Author Brands compared on product

Word-of-Author Advertising

311

quality with all of their competitors ("All Competing Brands") and


the Leading Competitor Brands, the total set of Consumer Reports
tests was identified for the 1950--1979 period which provided
overall evaluations of the Word-of-Author Brands and all of their
Leading Competitor Brands.
The individual product brands and models evaluated in each test
report were then assigned rank values according to their overall
evaluations in Consumer Reports. Since some brands were represented by more than one model in a single test report, a way had to
be found, for purposes of data analysis, to identify a single rank
value for such brands. After considering the pros and cons of several
alternative plans, it was decided to opt for the rank value of the
highest quality model of each multiple-model brand (as determined
by Consumer Reports) in a test report. In effect, then, the study
findings compare the highest quality representatives of the brands
included in a test report, e.g., how the highest ranked Sears model
compares with the highest ranked model of J. C. Penney, and the
highest ranked model of Wards, etc.

RESULTS
Tables I and II present the major study results. Each table reveals
how many test reports found the rank of the highest quality model of
a Word-of-Author Brand scoring higher than, the same as, and lower
than the average rank of the highest quality models of the brap.ds
representing a comparison group. The comparison groups are All
Competing Brands in Table I and Leading Competitor Brands in
Table II.
Before examining the data appearing in these tables it is of
interest to note that tests associated with one Word-of-Author Brand
appear with far greater frequency than any other. In particular, of the
265 test reports in Consumer Reports which served as our data
source, 80% were generated by Sears products and their competitors.
On the other hand, Cadillac, Kleenex and Levis, in combination with
their competitors, account for only 3% of the 265 tests. The reason
for the large number of Sears-related tests is that this brand, more
than any other in the set of six, is associated with a wide range of
consumer products which were evaluated by Consumer Reports.
Included here are major and minor appliances, television sets,

312

Monroe Friedman

TABLE I
Number and Percentage of Consumer Reports Test Reports in Which the Best
Model of a Word-of-Author Brand Ranked Higher Than, the Same As, and Lower
Than the Average of the Best of the Models of the Remaining Brands in the Test
Report (All Competing Brands)
Word-of-Author Brands
Relative Rank of
Word-of-Author
Brands

A &P
N(%)

Buick Cadillac Kleenex Levis


N(%) N(%)
N(%) N(%)

Sears
N(%)

Total
N(%)

Higher than
Comparison Group 11(61) 16(73) 5(100)

3(100)

3(75)

Same as
Comparison Group

0(0)

0(0)

0(0)

0(0)

0(0)

16(8)

Lower than
Comparison Group

7(39)

6(27) 0(0)

1(0)

1(25)

93(44) 108(41)

Total

18

22

103(49) 141(53)

212

16(6)

265

TABLE II
Number and Percentage of Consumer Reports Test Reports in Which the Best
Model of a Word-of-Author Brand Ranked Higher Than, the Same As, and Lower
Than the Average of the Best of the Models of the Leading Competitor Brands in
the Test Report
Word-of-Author Brands
Relative Rank of
Word-of-Author
Brands

A&P
N(%)

Buick
N(%)

Cadillac Kleenex Levis Sears


N(%)
N(%) N(%) N(%)

Total
N(%)

Higher than
Comparison Group

9(50) 14(64) 5(100)

2(50)

3(75)

98(46) 131(49)

Same as
Comparison Group

3(17)

1(25)

0(0)

28(13)

Lower than
Comparison Group
Total

6(33)
18

1 ( 5 ) 0(0)
7(32) 0(0)
22

1(25)

1(25)

33(12)

86(41) 101(38)
212

265

clothing, gardening equipment, automobile replacement parts, h o m e


i m p r o v e m e n t materials, and household furnishings.
Turning next to an examination of the data we find in Table I that
the highest ranked models of the W o r d - o f - A u t h o r Brands tended to
o u t p e r f o r m the highest ranked models of All C o m p e t i n g Brands.

Word-of-AuthorAdvertising

313

Overall, 53% of the 265 tests yielded results in favor of the Wordof-Author Brands while the reverse was true in only 41% of the
tests. And the same pattern emerged consistently across all six
Word-of-Author Brands, a result which is statistically significant at
the 0.05 level (Binomial Test).
Looking next at the data of Table II, we find a similar dominance
of the Word-of-Author Brands over the Leading Competitor Brands.
In particular, the highest ranked models of the Word-of-Author
Brands tended to outperform the highest ranked models of the
Leading Competitor Brands. This tendency was evident for 49% of
the 265 tests while the opposite tendency appeared in only 38% of
the tests. And once again, the tendency was evident in all the
comparisons involving the six Word-of-Author Brands, a result
which is statistically significant at the 0.05 level (Binomial Test).
To sum up, the study findings for unsponsored Word-of-Author
Brands, both individually and in aggregate, reveal a consistent
pattern of relatively high overall quality. And the pattern is manifested when these brands are compared with All Competing Brands
as well as the Leading Competing Brands.

DISCUSSION

Some Methodological Concerns


Before examining the study findings, a comment is called for regarding two methodological concerns. The first considers potential
problems with product evaluations conducted by Consumer Reports.
Although widely used, the magazine's tests are not without faults. To
illustrate, the tests exclude many regional brands, arbitrarily combine
scores on evaluative dimensions for various brands and models into
a single overall rating, and typically employ extremely small samples.
Nonetheless, as Morris (197 l) has noted, the test ratings are "simply
the best widely-available source of specific information we have on
commodity quality in our country" (p. 121).
A second methodological concern is the process by which we
selected, for purposes of analysis, a particular model to represent a
given brand for the case of a Consumer Reports test which examined
multiple models of the brand. It will be recalled that under such
circumstances we decided to select the highest ranking model of the

314

MonroeFriedman

brand (in short, the one with the highest overall evaluation in the
Consumer Reports test) as our data point. As a result, the study
findings do not always compare "typical" models of Word-of-Author
Brands with their competitors; instead, the study findings compare
the best available model for each Word-of-Author Brand with the
best available model for each of its competitors. This is a limitation
of the study, reflecting the fact that "typical" models are impossible
to identify for many brands, especially if there are, as often occurs
for the multiple-model case, only two models of a brand from which
to pick (Morris & Block, 1968).
Looking next at the actual study findings we see that when the
best model of each Word-of-Author Brand is compared with its
counterparts for All Competing Brands, the Word-of-Author Brands
consistently come out ahead. Since some evidence (Morris, 1971)
suggests that "name brand" models of products may be superior to
others, we decided to include a second comparison in the data
analysis to see if the Table I finding might be more parsimoniously
viewed as a "name brand" effect. As indicated in Table II, this
control comparison found no evidence of such an effect in that the
Word-of-Author Brands outperformed the Leading Competitor
Brands despite the fact that all of the brands in this comparison are
well-known "name-brand" types.

Interpreting the Study Findings: Some Cautions


It is well to ask what we are to make of the study findings. Do they
really suggest, as was indicated earlier, that consumers have less
reason to worry about word-of-author advertising since many of
the brands being cited over the years are represented by models
with relatively high ratings on Consumer Reports' tests? Practically
speaking, this would appear to be the case but we should be quick to
add that three considerations suggest the exercise of caution. First,
the study findings are historical in nature and what held for the
1950--1979 period may not continue to hold in the present and
future. Second, the data analysis revealed effects which are quite
modest in magnitude even though they reach customary levels of
statistical significance. In particular, 53% of the 265 tests reveal the
superiority of the Word-of-Author Brands over All Competing
Brands, but 41% (a very sizable minority) reveal the opposite effect.

Word-of-Author Advertising

315

This means that in two of five cases the brand name being "plugged"
in word-of-author advertising is likely to be relatively low in overall
quality.
A third consideration suggesting caution concerns the nature of
the brand names being studied. Earlier we noted that they emanated
from studies of unsponsored word-of-author advertising in popular
American books, plays, and songs. Our findings hold then for
unsponsored word-of-author advert!sing and not necessarily for their
sponsored counterparts. Since business firms are usually not eager to
disclose their sponsorship of word-of-author advertising, little is
known about what branded products are sponsored, although popular
news articles (e.g., Igger, 1985) have stressed the frequency of
packaged food and beverages. Examples given are Reese's Pieces in
"E. T.," Bertolli olive oil in "North Dallas Forty," and Pepsi Cola and
Popov vodka in "Back to the Future." In unsponsored word-ofauthor advertising, on the other hand, the brands most frequently
used have been those of automobiles and magazines (Friedman,
1985b, 1986a, 1986b).
Another difference between the sponsored and unsponsored
varieties of word-of-author advertising is suggested by the discussion
of food and beverage products. In particular, while the unsponsored
variety can and does occur in many popular communication channels
(novels, plays, songs, etc.), discussions of sponsored word-of-author
advertising suggest that the focus is exclusively on films. The reasons
for this restricted focus no doubt stem from the enormous popularity
of hit films, as compared, say, to bestselling novels, as well as the
opportunity for product association with a major film star in a highly
dramatic context.
As one reviewer of this paper has noted, the argument that
unsponsored word-of-author advertising c a n affect consumer purchase behavior seems sound, based as it is on the attribution theory
literature and the practice of sponsored word-of-author advertising.
However, this reviewer cautions, should we not first attempt to
demonstrate actual effects before worrying about the quality of the
branded products being cited in word-of-author communications?
The point appears to be reasonable until one considers the enormous
methodological problems in designing and executing a behavioral
study to determine what influences on buying decisions over the
years can be attributed to word-of-author advertising above and
beyond all other forms of advertising. As consumer research on a

316

Monroe Friedman

similar topic has shown (Zajonc's mere exposure effect), the question
is far easier to pose than to answer.

Policy Implications of the Study Findings


The policy implications of the study findings would seem most direct
for the field of consumer education. It is common practice for
textbooks in this field to have chapters or sections of chapters
dealing with advertising and the consumer (e.g., Miller, 1984;
Troelstrup & Hall, 1978). Typically they discuss the role of advertising in the economy and alert readers to the biases which may exist in
advertising messages, such as omission, puffery, and exaggeration.
Since, however, the definitions of advertising employed in these
textbooks are usually limited to producer-generated information
about consumer products and services, the focus of this study -unsponsored word-of-author advertising -- is not considered.
The study findings suggest that this may be a mistake. As indicated earlier, both psychological theory and marketing practice
suggest that unsponsored word-of-author advertising can affect
consumer perceptions of frequently cited brands. And the accumulated results of 30 years of Consumer Reports tests indicate that
in two of five cases the brand names being cited frequently are
relatively low in quality.
It would seem then that educators should alert consumers to the
possibility of commercial influences in popular forms of communication which have long been considered non-commercial in character.
Since the study findings show that such influences may not infrequently be in the direction of promoting relatively low quality
products, it is especially important that consumers be adequately
forewarned.
Before concluding, the reader should be reminded that this study
looks at only one question of concern relating to the effects of
word-of-author advertising. Other questions with policy implications
for consumers concern the nature of the branded products cited, the
popular literature audiences for the word-of-author advertising, and
the actual process by which the advertising affects consumer behavior. Of special interest in this regard is the volatile combination of
harmful or potentially harmful products, such as cigarettes and
alcoholic beverages, with citations in books popular with youth (e.g.,

Word-of-Author Advertising

317

those a u t h o r e d by the bestselling A m e r i c a n writers J u d y B l u m e and


Stephen King).

CONCLUSIONS
T h e study finds that, with regard to the quality of p r o d u c t s
p u r c h a s e d by consumers, u n s p o n s o r e d w o r d - o f - a u t h o r advertising
does not a p p e a r to have constituted a c o u n t e r p r o d u c t i v e source of
c o n s u m e r influence, in that m o s t b r a n d s represented in p o p u l a r
literature are associated with p r o d u c t s with relatively high overall
ratings in tests u n d e r t a k e n by C o n s u m e r R e p o r t s in the 1 9 5 0 - - 1 9 7 9
period. H o w e v e r , m a n y exceptions were f o u n d to this s u m m a r y
finding suggesting that educators should alert c o n s u m e r s to the
possibility of u n t o w a r d influence.

REFERENCES
Friedman, M. P. (1985a). Are Americans becoming more materialistic? Advances
in Consumer Research, 12, 385--386. Provo, UT: Association for Consumer
Research.
Friedman, M. P. (1985b). The changing language of a consumer society: Brand
name usage in popular American novels in the postwar era. Journal of Consumer
Research, 11,927--938.
Friedman, M. P. (1986a). Commercial influences in popular literature: An empirical
study of brand name usage in American and British hit plays in the postwar era.
Empirical Studies of the Art, 4, 63--77.
Friedman, M. P. (1986b). Commercial influences in the lyrics of popular music of
the postwar era. Journal of Consumer Affairs, 20, 193--214.
Friendly, D. T. (1983). Selling it at the movies. Newsweek, July 4, p. 46.
Iggers, J. (1985). Today's movies contain subtle ads. Detroit Free Press, October 23,
p. C1.
Miller, R. T. (1984). Economic issues for consumers. St. Paul, MN: West Publishing
Co.
Morris, R. T. (1971). Major firms comparatively evaluated. Journal of Consumer
Affairs, 5, 119--139.
Morris, R. T., & Block, B. (1968). The instability of quality. Journal of Consumer
Affairs, 2, 39--60.
Schiffman, L. G., & Kanuk; L. L. (1983). Consumer behavior. Englewood Cliffs, NJ:
Prentice-Hall.
Tie-in advertising (1951). Consumer Reports, January, pp. 43--44.
Troelstrup, A. W., & Hall, E. C. (1978). The consumer in American society. New
York: McGraw-Hill.
Williams, T. G. (1982). Consumer behavior: Fundamentals and strategies. St. Paul,
MN: West Publishing Co.

318

Monroe Friedman

ZUSAMMENFASSUNG

Werbung in populiiren Kunstwerken und der Verbraucher: eine empirische Analyse


der Qualitiit solcher Produktmarken, die in populiiren Kunstwerken genannt werden.
Der Begriff "Word-of-Author'-Werbung bezieht sich auf die F/ille, wo Autoren yon
Biihnen- und Fernsehstficken, von Romanen und anderen popul~iren Kunstprodukten in ihren Texten bekannte Markennamen verwenden. Der Beitrag berichtet fiber
eine empirische Studie anhand von 265 Tests der amerikanischen Zeitschrift Consumer Reports aus der Zeit von 1950 bis 1979, um Marken, deren Name in dem
genannten Sinne h~iufig verwendet werden, zu vergleichen mit solchen Marken, wo
dies nicht der Fall ist.
Die Ergebnisse zeigen, dab die Mehrzahl der h~iufig verwendeten Marken zwar
tatsfichlich ein h6heres Qualit~itsurteil erzielte als andere Produkte. Dies gilt aber
nicht fiir alle: Eine beachtfiche Minderheit von Marken macht eine Ausnahme.
Abschliel3end werden einige verbraucherpolitisehe Schlul3folgerungen gezogen
und insbesondere gefordert, die untersuchte Art von Konsumgfiterwerbung st~irker
als bisher in der werbekundlichen Verbrauchererziehung zu berficksichtigen.

THEAUTHOR
Monroe Friedman is Professor of Psychology at Eastern Michigan University,
Ypsilanti, M148197, USA.

También podría gustarte