Sie sind auf Seite 1von 11

Journal of Consumer Research, Inc.

Detecting and Correcting Deceptive Advertising


Author(s): Gary M. Armstrong, Metin N. Gurol and Frederick A. Russ
Source: Journal of Consumer Research, Vol. 6, No. 3 (Dec., 1979), pp. 237-246
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/2488984 .
Accessed: 25/06/2014 06:55

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR to
digitize, preserve and extend access to Journal of Consumer Research.

http://www.jstor.org

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
Detecting and Correcting Deceptive
Advertising

GARY M. ARMSTRONG
METIN N. GUROL
FREDERICK A. RUSS*

This article investigates the use of the salient belief technique for measuring
deception and the effectiveness of corrective ads. The technique is applied to
Listerine advertising, which has already been found deceptive by the courts.
Comparisons between this technique and the normative belief technique
yielded similar results except for source effects over time.

The subject of deceptive advertisinghas been dis- occurred.' Similarly, most definitions of deception in
cussed in law journals for over half a century, but advertising developed during the past several years,
articles on deceptive and corrective advertising have and most of the procedures advanced for measuring
appeared in marketing publications within only the past deception are based on consumer research.
decade. Despite progress in defining and measuring In this article the approach is taken that: "Deception
deception, and similar progress in developing and im- occurs when a consumer perceives and believes an
plementing corrective advertising, few studies have advertising claim (explicit or implied) that is false, but
sought both to measure deception and to assess means that deception is of little concern unless the claim is
of correcting it. relevant" (Armstrong and Russ 1975, p. 24). This
definition suggests three important components of sa-
DECEPTIVE ADVERTISING lient deception: belief, falsity, and relevance.
Studies attempting to measure deception have used
In legal literature there is no general statutory defini- one of two major approaches. The first approach is to
tion of deceptive advertising. In marketing literature measure deception by asking subjects if certain adver-
only a few authors have tried to define deception in tisements, claims, or situations are deceptive (Ford,
advertising (Aaker 1974; Armstrong, Kendall, and Russ Kuehl and Reksten 1975; Haefner 1972; King and Wise
1975; Gardner 1975; Haefner 1972; Howard and Hul- 1974; Rao 1974). The second approach is to measure
bert 1973; Jacoby and Small 1975), and they too have advertising deception on the basis of brand attribute
not been able to develop a well accepted definition. beliefs of consumer respondents (Armstrong, Kendall,
The courts have generally proclaimed that the Fed- and Russ 1975; Kuehl and Dyer 1976, 1977).
eral Trade Commission (FTC) has expert judgment in Evaluation of the literature shows that measuring
deceptive advertising cases, and often determined the deception by asking subjects whether certain adver-
existence of deception based on FTC interpretations of tisements or advertising claims are deceptive is inap-
how consumers might construe an advertisement. Dur- propriate. Consumers who perceive an ad to be decep-
ing the past decade, however, this expert judgment has tive are not deceived by it. The most promising avenue
been questioned by a number of lawyers, policy mak- seems to be measurement of brand attribute beliefs of
ers, marketing researchers, and advertisers, as well as respondents.
some Commission members themselves, who have The use of brand attribute beliefs to detect deception
made strong arguments for the use of consumer re- appears to have followed two distinctly different ap-
search in the determination of whether deception has
'See, for example, Aaker 1974; Armstrong, Kendall, and Russ
1975; Armstrong and McLennan 1973; Armstrong and Russ 1975;
*Gary M. Armstrong and Frederick A. Russ are Associate Profes- Bemacchi 1974; Cohen 1969; 1972; 1974; 1975; Dillon 1973; Dyer
sors of Business Administration, both at the University of North 1972; Ford, Kuehl, and Reksten 1975; Gardner 1975; Gellhorn 1969;
Carolina, Chapel Hill, NC 27514. Metin N. Gurol is Associate Profes- Gerlach 1972; Haefner 1972; Haefner, Birchmore, and Permut 1974;
sor of Marketing at the University of Baltimore, Baltimore, MD Howard 1974; Howard and Hulbert 1973; Jacoby and Small 1975;
21201. Jones 1971; Pollay 1969; Preston 1976; Rosch 1975; Wilkie 1973;
Wilkie and Gardner 1974; Woodworth 1971.

237
C JOURNAL OF CONSUMER RESEARCH * Vol. 6 . December 1979

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
238 THE JOURNAL OF CONSUMER RESEARCH

proaches. One approach is the normative belief tech- 1. The measure used. NBT uses beliefs about brands.
nique, first advocated by Gardner (1975). It has been SBT uses perceptions and beliefs about claims made
employed in several studies of corrective advertising, by an ad. The former has the advantage of not requir-
and was used by Dyer and Kuehl (1978) and Kuehl and ing measurements of perception of the claims, which
Dyer (1976; 1977) to establish that deception had oc- could be time consuming or unreliable. The latter has
the advantage of identifying the claims perceived in an
curred in various advertising instances, after which cor- ad being tested, and thus can obviate the need for a
rective advertising was implemented and assessed. control group.
Conceptually, the normative belief technique
suggests that the standard against which allegedly de- 2. The standard used. NBT uses normative beliefs about
ceptive ads can be judged is the set of attribute beliefs the product category. SBT uses judgments of truth
made externally. NBT, therefore, has an advantage
about a product category held by a representative group when dealing with claims that are difficult to test sci-
of knowledgeable consumers. An advertisement that entifically, and a disadvantage in that inferences about
causes a relevant audience to hold a more favorable set deception cannot be made with certainty without ref-
of beliefs about a specific brand in the product category erence to some externally measured truth. For exam-
is considered to be potentially deceptive. Conceptual ple, before Listerine ads could be judged deceptive by
problems with this approach have been acknowledged Dyer and Kuehl, it was necessary for them to know or
by Kuehl and Dyer (1976, p. 375). assume that the normative beliefs held by consumers
Operationally, the normative belief technique in- about the other three brands tested (the product class)
volves identifying the major functional attributes asso- were an accurate representation of reality, and that
ciated with a product class, measuring the normative Listerine was not in any way differentiated from the
other brands on the attribute in question. Requiring
brand beliefs about these attributes for the product that beliefs about a specific brand match beliefs about
class, and measuring the specific brand beliefs held by the entire product class leads to difficulties, either
an audience after exposure to the ad being studied. when a brand is truly differentiated or when there are
Using this approach, Dyer and Kuehl (1978) dis- misperceptions about the product class as a whole.
covered that consumer beliefs about Listerine's ability
to (1) prevent colds and sore throats and (2) to kill germs 3. The issue of salience. NBT assumes that all beliefs are
equally relevant, or that their relevance is not a con-
effectively were significantly higher than the average of cern. SBT measures relevance, and can use the meas-
other brands in the product class. This result was taken ure to weight the belief score. Although the measure-
as corroboration of findings of deception by the FTC in ment of salience is not a requirement for the meas-
the Listerine case. urement of deception, it offers two advantages: (1) it
A second approach using beliefs in the measurement permits public policy makers to identify those decep-
of deception is the salient belief technique, first advo- tions that are adversely affecting either consumers or
cated by Armstrong and Russ (1975) and tested sub- competitive balance, and (2) it recognizes that attri-
sequently by Armstrong, Kendall, and Russ (1975). bute saliences can also be the target of persuasive
Conceptually, the salient belief technique suggests that communications. An ad that suggests the sound of a
deception occurs when consumers perceive and believe door slamming as an indication of automobile quality
false claims either made or implied by an ad. To be of may make a claim about a particularcar's sound more
salient than it should be in a consumer's decision
any concern, however, these false claims must be rele- process.
vant to consumers' decisions to purchase a brand in the
product category. Falsely held beliefs about an irrele- This article reports research that compares results of
vant attribute are technically a deception, but because using the salient belief technique with results found
they are unlikely to affect the decision process it seems using the normative belief technique in Dyer and Kuehl
useful to distinguish between them and falsely held (1978).
beliefs that are relevant.
Operationally, the salient belief technique involves CORRECTIVE MESSAGES
(1) identifying claims that relevant consumers (not
necessarily knowledgeable) perceive that the ad is mak- Many conceptual articles deal with the behavioral
ing, (2) determining externally which of the claims are aspects of corrective messages (Aaker 1974; Cohen
false, and (3) measuring the perceptions, beliefs, and 1975; Dyer and Kuehl 1974; 1976; 1977; Hunt 1972a;
saliences of a representative sample of consumers ex- Kassarjian, Carlson, and Rosin 1975; Mazis and Adkin-
posed to the ad. son 1976; Rosch 1975; Sawyer 1976; Wilkie and Gard-
Armstrong, Kendall, and Russ (1975) used the tech- ner 1974). Wilkie's treatment of the subject is probably
nique to measure and compare the deceptiveness of the most comprehensive (1973; 1974; 1975).
television ads for a food product, an insecticide, a Empirical research on corrective advertising can be
finance company, and a deodorant soap, which were divided into two groups according to the dependent
shown to adults in a convenience sample drawn from variables-attitudes toward the brand (Dyer and Kuehl
church and PTA groups. 1974; Hunt 1972a,b; Kassarjian, Carlson, and Rosin
The normative belief technique (NBT) and the salient 1975) and brand beliefs, on a multiattribute basis (Dyer
belief technique (SBT) differ in three important ways: and Kuehl 1978; Kuehl and Dyer 1976, 1977; Mazis and

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
DETECTINGAND CORRECTINGDECEPTIVEADVERTISING 239

Adkinson 1976). The four empirical studies on remedial ment are cross-sectional, and there is a lack of attention
messages that used attitudes toward the brand as their to the time dimension. Current laboratory evidence
main dependent variable consistently found that reme- may overestimate the effectiveness of corrective adver-
dial messages are effective in undoing the damage of a tising because the persistence of its impact has not been
deceptive advertisement by significantly changing the assessed.
attitudes in the negative direction.
Studies by Dyer and Kuehl (1978), Kuehl and Dyer PURPOSE
(1976; 1977), and Mazis and Adkinson (1976) represent
substantial progress, because they used brand attribute The original purposes of the research reported in this
beliefs rather that attitudes as the main dependent article were the following:
variables. These and other such studies, however, had 1. to investigate further the use of the salient belief tech-
several shortcomings. First, some of the brands used nique for measuring deception in an instance where
were unfamiliar or fictitious, eliminating the need to FTC and court decisions have already determined that
account for carry-over effects from previous advertising. an ad is deceptive;
For example, Hunt's (1972a,b) experiments were based
2. to assess the effectiveness of specially designed cor-
on a brand of gasoline that was never available locally. rective ads in reducing or eliminating deception in the
Kassarjian et al. (1975) used a not-very-popular short- and long-run;
motorcycle safety device with a regional brand that was
sold by a distant retailer. Kuehl and Dyer's (1976; 1977) 3. to provide additional information about the impact of
experiment involved two fictitious products. using different sources for corrective messages.
Second, only Kassarjian et al. (1975), Dyer and Additionally, the appearance of an article with a simi-
Kuehl (1978), and Kuehl and Dyer (1976; 1977) meas- lar purpose (Dyer and Kuehl 1978), after the research
ured the dependent variable after both the deceptive reported here was conducted, allows comparisons that
and corrective messages. The first measure is necessary strengthen previously drawn conclusions and provides
to find out if deception has taken place; otherwise, new insights that neither study could have generated
correction is unnecessary. Hunt (1972a,b) used a de- separately.
ceptive message, but did not measure the dependent
variable until the subjects were exposed to both decep-
tive and corrective messages. Thus, he administered METHODOLOGY
the corrective message without knowing whether sub-
jects were deceived. Dyer and Kuehl (1974) and Mazis
Experimental Procedure
and Adkinson (1976) did not use a deceptive message The experimental procedure used in this study is
before the remedial message. summarized in Table 1. The sample consisted of 134
Third, all the experiments on corrective messages subjects randomly assigned to four groups. For each
used either print or audio messages. Most of the FTC group, subjects were assembled in a laboratory setting
corrective advertising cases have involved television and shown various combinations of 60-second filmed
commercials. ads-deceptive and corrective ads for Listerine and
Fourth, although researchers have realized that irrelevant ads for other products. Sets of SBT measures
brand attribute beliefs should be the dependent vari- were taken at three points during the laboratory exper-
able, they have not concerned themselves with whether iment, and a fourth measure was mailed six weeks later.
these beliefs were caused by the specific ad being As Table I shows, subjects in all four groups completed
tested, or whether these beliefs were salient in indi- premeasurement questions (M) for the purpose of es-
viduals' purchase decisions. tablishing initial levels of claim beliefs and saliences,
Fifth, all but the last Dyer and Kuehl (1978) experi- and to obtain information on product usage and famil-

TABLE 1
EXPERIMENTAL PROCEDURE

First Post-deception Second Post-correction Six-week


Group Post-measure treatment measure treatment measure measure

1 Deceptive ad M21 Corrective ad M31 M41


(company
source)
2 M12 Deceptive ad M22 Corrective ad M32 M42
(FTC source)
3 M13 Deceptive ad M23 Irrelevant ad M33 M43
4 M14 Irrelevant ad M24 Irrelevant ad M34 M44

Mijrefers to the ith measure for the jth group.

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
240 THE JOURNAL OF CONSUMER RESEARCH

iarity. For the first treatment, Groups 1, 2, and 3 were Subjects


exposed to a deceptive Listerine television commercial
and to television commercials for irrelevent products (a The subjects were students enrolled at a large state
food product and a finance company). Group 4 was university, who responded to solicitations for volun-
shown ads for the food product, the finance company, teers and a small monetary inducement for participa-
and a third irrelevant product. A second set of questions tion. A sample consisting of student and nonstudent
(M2), similar to the first, was completed by individuals subjects might have been more representative of rele-
in each group to measure the changes in subjects' claim vant consumers. But students do represent a segment of
beliefs and saliences. Group 4 served as a control for relevant consumers (61 per cent were using Listerine or
measurement effects produced by the first set of ques- had used it previously), and there may not be that much
tions. difference between student consumers and nonstudent
For the second treatment, Group 1 saw a company- consumers. The SBT results for the food product and
source corrective ad for Listerine, along with ads for finance company ads in this study were similar to the
the other two products. Group 2 saw an FTC-source results of a previous study using the same ads with a
corrective ad along with the other two ads. Groups 3 sample consisting of nonstudent adults of a wide range
and 4 saw only irrelevant ads. Each subject then com- of ages (Armstrong, Kendall, and Russ 1975).
pleted a third set of questions (M3), the purpose of
which was to determine the effects of the corrective ads The Deceptive and Corrective Ads
for Listerine on claim beliefs and saliences. Group 3
served as a control for measuring the impact of the The deceptive Listerine television ad was chosen
corrective ads. Group 4 provided a way to examine full from a reel of 29 Listerine ads held by the FTC as
potential for reactivity in the experiment. evidence in the Warner-Lambert case. The particular
After six weeks, all subjects were mailed a fourth set film used was selected because it expressed or implied
of questions (M4), essentially identical to the first, to all three of the claims that the FTC had found to be
determine the extent to which the effects found in pre- deceptive: Listerine will (1) prevent, (2) ameliorate,
vious measures persisted over time. Of the original 134 and (3) cure colds and sore throats (Exhibit 1).
subjects, 106 (79 percent) returned this final question- The corrective ads used in this study were profes-
naire. sionally filmed for this experiment. Two ads were
Two types of effects were not directly controlled for produced-one with a company source and one with an
in this experiment: (1) the interaction between pre- FTC source. The scripts (read by an actor) were identi-
measures and the experimental treatments (test- cal in both filmed ads, except for the introductory sen-
treatment interaction effect) and (2) the reaction of sub- tences that identified the source of the message. To
jects to the experimenter and the experimental situation assure that subjects noted the source of the message,
(reactivity effect). relevant props were used in the ads to identify the
The test-treatment interaction effect was examined in
pretests with other, similar subjects. No test-treatment EXHIBIT1
interaction effect was found, so it was not controlled for TEXT OF THE DECEPTIVEADVERTISEMENT
in the experiment reported here.
Reactivity effects are practically impossible to elimi- [It is raining. Two mothers start talking. One mother has just
escorted her children to the school bus, the other (Muriel) is
nate or control in laboratory experiments. Reaction to checking the mailbox.].
the experimenter was reduced by using written, stan-
dardized instructions that were read by a neutral party. 1st Mother Muriel, where are Dave and Sue?
2nd Mother Oh, down with a cold again.
Reactions to the experimental situation might have 1st Mother Again?
been assessed by including a fifth group that was ex- 2nd Mother Oh, guess your family always seems fine.
posed to a neutral (nondeceptive or noncorrective) Lis- 1st Mother I got a theory.
terine ad. But, budget constraints plus the difficulty of 2nd Mother A theory? Nothing can prevent colds.
1st Mother You can help. Let's get out of the rain. [They go
identifying truly neutral Listerine ads (those with no inside the house.]
deceptive claims or no possibility of evoking memories 1st Mother Muriel, I make sure they have plenty of rest, and
of other Listerine ads that did have deceptive claims) I watch their diets.
made this strategy untenable. Thus, the possibility of 2nd Mother Uh-huh.
reactivity effects must be considered in evaluating the 1st Mother Then I have them gargle twice a day with Listerine.
2nd Mother Listerine antiseptic?
results of the study. 1st Mother Uh-huh. Ithink we've cut down on colds and those
Also, measures taken six weeks after exposure to we catch don't seem to last as long.
corrective advertisements are not the ultimate answer 2nd Mother Sure seems to work for your family.
to the question of long-term effects. Future research 1st Mother Yes, it does.
2nd Mother Well, I'll try it.
would do well to consider even longer-term effects of Male Voice During the cold-catching season, for fewer colds,
corrective ads-either single insertions or repeated milder colds, more people gargle with Listerine
ads. than any other oral antiseptic. Listerine. [Music]

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
DETECTINGAND CORRECTINGDECEPTIVEADVERTISING 241

EXHIBIT 2 The second section of the questionnaire contained a


TEXT OF THE CORRECTIVEMESSAGE set of questions for each of the ads subjects saw in the
first treatment. Sets of questions pertaining to a particu-
Hello, I am Walter Hughes (fictitious name), representing the lar ad were completed immediately after that ad was
(FTC or Warner-Lambert Company). viewed. The questions concerning Listerine were the
Contrary to prior advertising of Listerine, Listerine will not
prevent or cure colds or sore throats, and Listerine will not be same as those found in the premeasure except that (1)
beneficial in the treatment of cold symptoms or sore throats. the product familiarity question was deleted, and (2)
Listerine is an antiseptic that kills germs on contact. It is subjects were asked to indicate which of the listed
effective for general oral hygiene, bad breath, minor cuts, claims they perceived the ad to be making.
scratches, insect bites, and infectious dandruff. But it is not The third part of the questionnaire, completed after
effective against colds and cold symptoms, because colds are
caused by viruses and Listerine does not kill viruses. viewing corrective messages, was similar to the second
part except for two additions. Subjects were asked
about their purchase behavior for products shown in the
source visually throughout the message. The text of the ads, and a question was inserted to determine if they
corrective message is provided in Exhibit 2. had correctly perceived the source of the corrective
The first paragraph of the message identifies the message.
source. The second paragraph is from the final order of The final questionnaire, containing a fourth set of
the FTC, and corrects previous impressions about the measures similar to those in the first set of questions,
product.2 The first two sentences of the final paragraph was mailed to subjects six weeks later. During that time
state the benefits of Listerine as presented on the Lis- period there was no abnormal level of advertising or
terine label. The final sentence explains the reasoning publicity for Listerine.
behind the corrective message.
In both ads, the message is presented in a straightfor-
ward narrative. The message is objective-the FTC has ANALYSIS
medical evidence to support the negative claims and The main dependent variable in this experiment was a
Warner-Lambert has medical evidence to support the "Salient Deception Score" (SDS). An SDS was calcu-
positive claims. The message is two-sided-three nega- lated for every false claim by multiplying the belief
tive and six positive claims are made. The message was score of every subject by that subject' s salience score
designed to fit the spirit of Senator Magnuson's (1972, and summing over the three false claims. An SDS was
p. 4) suggestion that consumers have the right "to be calculated four times for each subject. Generally, such
given the facts needed to make an informed choice." data are analysed with a univariate analysis of variance,
which has provisions for correlated or repeated meas-
Questionnaires urements. But, in addition to the usual assumptions
concerning normality and homogeneity of variance
The first three sets of questions were combined in a within cells, Winer (1971) stresses that repeated-
single questionnaire. The questionnaire began with a measures analyses also require the assumption of
short cover letter that introduced the experiment and homogeneity of covariances (i.e., correlations) be-
explained that it was part of a study designed to meas- tween repeated assessments.
ure ad effectiveness. The first part of the questionnaire McCall and Appelbaum (1973, p. 401) state that "al-
contained the premeasures. In all cases, sets of ques- though researchers are accustomed to ignoring lack of
tions about Listerine were placed between sets of simi- homogeneity of within-cell variances in the analyses of
lar questions about the food product and the finance variance (ANOVA), a repeated-measures ANOVA is
company. This allowed subjects to become familiar not robust with respect to violations of the additional
with the format of the questionnaire while answering covariance (correlational) assumptions." Violations of
"irrelevant" questions. these assumptions produce artificially large F values
Regarding Listerine, subjects were first asked if they relative to the appropriate critical value (too many re-
were familiar with the brand. They were then given ten jections of the null hypotheses for the stated significance
claims about Listerine, and were asked to indicate the level).
extent to which they believed each claim was true For repeated-measures experiments such as this one,
(definitely true, probably true, probably false, definitely McCall and Appelbaum (1973, pp. 406-13) and Kirk
false). They were then asked to indicate how important
(1968, pp. 143-4) recommend the use of multivariate
(very, moderately, slightly, not at all) each claim would
be to their decision to buy the product, if the claim were techniques that make no covariance assumptions and
provide exact probability statements. Their recom-
true. Five of the claims used were mentioned in the
mendations were followed and multivariate analysis of
deceptive ad, and nine were stated in the corrective ads. variance was used to analyze SDSs. The belief and
salience scores were analyzed separately using the
2When this research was conducted, the courts had not yet elimi- same techniques. Multivariate tests of significance
nated the phrase "Contrary to prior advertising . . ." were calculated where appropriate.

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
242 THE JOURNAL OF CONSUMER RESEARCH

TABLE 2 FIGUREA
MEAN SALIENTDECEPTIONSCORES: SHORT-RUN MEAN SALIENTDECEPTIONSCORES: SHORT-RUN EFFECTS
EFFECTSa
SALIENT
Post- Post- DECEPTION
Pre- deception correction SCORE
Group measure (M1) measure (M2) measure (M3)
23
Company-source 17.60 20.51 11.80
corrective ad 22
FTC-source 18.10 22.35 11.41
corrective ad 21
Deceptive ad only 17.42 22.10 21.26
Control 18.19 20.09 20 <t 0;------

a134 subjects were tested.

18

17
RESULTS
16 -
Analysis of the data allows for tentative conclusions
on several issues concerning the impact of deceptive 15 -
and corrective advertising. The results are reported in
terms of the SDS and its two components-a belief 14
score and a salience score.3 The first three results re- 13-
ported here are based on the responses of the 134 sub- GROUPI (Company source corrective ad) \
jects who completed the first questionnaire (Tables 2 12 - GROUP2 (FTC source corrective ad)
and 3 and Figure A). The remaining results are based on GROUP3 (Deceptive ad only)
the responses of the 106 subjects who completed first - - GROUP4 (Control)
and second (after six weeks) questionnaires (Tables 4
and 5 and Figure B).
1. Changes in the level of deception produced by Ml M2 M3
exposure to the Listerine ad were determined by com- MEASURES
paring changes in deception (SDS) for Groups 1-3
(pooled) with changes in the unexposed control group
(Group 4). The means of the SDS measures are shown tributable to the corrective message (p < 0.001; Table 3,
in Table 2 and Figure A. Analysis of these changes row 2). This appears to be explained by a significant
shows a significant increase, when compared to the reduction in the belief component (p < 0.001) rather
control group, in the level of salient deception (p < 0.09; than in the salience component (p < 0.32). Again, this
Table 3, row 1 under Salient Deception Score). supports the Dyer and Kuehl (1978) findings. Despite
Although the means of the separate components are many design and measurement differences, both
not shown here, MANOVA performed on the SDS studies found that corrective ads had a significant effect
components shows a significant increase in the salience in reducing deception.
of false claims (p < 0.004; Table 3, row 1 under " Sali- 3. No significant short-run source effect was found.
ence") with little effect on claim belief (p < 0.98; Table There appears to be no significant difference (p < 0.28)
3, row 1 under "Belief'). between the changes in deception levels produced by
The overall result supports the findings of Dyer and the company source and the changes produced by the
Kuehl (1978) and the judgment of the FTC and the FTC source (Table 3, row 3). No significant differences
courts. were observed in either the belief component (p < 0. 18)
2. The impact of the corrective ads was measured by or the salience component (p < 0.76). These short-run
comparing changes in SDS for Groups 1 and 2 (pooled) findings appear to corroborate the results of Hunt
to changes in Group 3, which saw the deceptive ad but (1972a), Mazis and Adkinson (1976), and Dyer and
not a corrective ad. The MANOVA results suggest a Kuehl (1978).
significant short-run reduction in salient deception at- 4. The impact of the corrective messages appears to
have persisted over time-6 weeks. (The means of the
3TheSalient Deception Score for a single false claim could run from SDS measures are shown in Table 4 and Figure B.)
1 to 16. Because the Listerine ad contained three false claims, the When deception levels, after six weeks, were compared
SDS for the ad could run from 3 to 48. No single standard could be with initial levels, they had been significantly reduced
used as a cutoff between deception and its absence, so the analysis
was repeated using a variety of standards to identify the number of both in the group (Group 1) that received the company
deceptions occurring and the number of people deceived. None of source message (p < 0.040) and in the group (Group 2)
these analyses showed substantially different results. that received the FTC source message (p < 0.004).

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
DETECTING AND CORRECTING DECEPTIVE ADVERTISING 243

TABLE 3
MANOVA RESULTS: SHORT-RUN EFFECTSa

Salient
deception score Salience Belief
Group
comparisons d.f. F d.f. F d.f. F

Group 1-3 vs. Group 4, 1,129 2.95b 1,129 8.54c 1,130 .001
change from M1 to M2
Groups 1 and 2 vs. 1,95 32.18d 1,95 1.01 1,97 48.19d
Group 3, change from
M2 to M3
Group 1 vs. Group 2, 1,62 1.17 1,62 .10 1,64 1.88
change from M2 to M3

a134 subjects were tested.


bp < 0.1.

cp <0.01.
dp < 0.001.

TABLE 4 FIGURE B

MEAN SALIENT DECEPTION SCORES: LONG-RUN MEAN SALIENTDECEPTIONSCORES: LONG-RUN EFFECTS


EFFECTSa
SALIENT
Post- Post- DECEPTI ON
Pre- deception correction Six-week SCORE
measure measure measure measure
(M1) (M2) (M3) (M4) 23 -
Group

20.31 12.15 14.46 22 -


1 17.42
2 17.22 21.13 11.52 12.91
20.83 18.21
21 -
3 17.14 21.69
4 18.89 20.64 21.18
20
a106 subjects were tested.
19
TABLE 5 18
MANOVA RESULTS: LONG-RUN EFFECTSa 17
Group or Salient 16
comparison d.f. deception Salience Belief
15
Group 1, long-run 1,25 4.74C 49 6.47c
effect (M1 vs. M4) 14 _\
Group 2, long-run 1,22 10.60d .34 24.71e
effect (M, vs. M4) 13
Group 3, long-run 1,28 .46 3.30b .03
effect (M1 vs. M4) 12 GROUPI (Company source \
Group 1 vs. Group 2, 1,47 .75 .18 1.53 corrective ad)
GROUP2 (FTC source
difference in M4 II corrective ad)
Group 1 (M3 vs. M4) 1,25 4.0Qb 3.16b 11.21d ----GROUP 3 (Deceptive ad only)
Group 2 (M3 vs. M4) 1,22 1.18 .14 4.31c -GROUP4 (Control)

al06 subjects; scores reported are F ratios.


I ~~~I LI
bp < 0.1. Ml M2 M3 M4
cp < 0.05. MEASURES
dp < 0.01.
ep < 0.001.

Meanwhile, Group 3, which was exposed to the decep- (p < 0.02, p < 0.001) rather than salience-based (p <
tion ad, but did not receive the corrective message, 0. 49,p < 0.57; Table 5, columns 3 and 4). The results of
returned to a level of deception that was slightly above, the Dyer and Kuehl (1978) study also suggest that cor-
but not significantly different from, their original posi- rective ads are persistent, at least for three weeks.
tion (p < 0.50). (These results are shown in the first 5. There does not appear to be any long-run source
three rows of Table 5, column 1.) For both message effect (Table 5, row 4). The groups that received the
sources, the long-run effect appears to be belief-based corrective messages did not appear to differ

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
244 THE JOURNAL OF CONSUMER RESEARCH

significantly in their levels of deception, either at the EXHIBIT 3


beginning of the experiment or after six weeks (p < A COMPARISON OF THE APPROACHES USED IN THE
0.39). TWO STUDIES
6. There is no direct evidence that the long-run effect
is likely to continue, but differences in the short-run and Dyer-Kuehl study
long-run levels of deception for the two groups exposed Design factor (1978) This study
to corrective advertising can be examined. In this case, Approach NBT SBT
visual inspection (Figure B) suggests that, as might be Ad medium Print Television
expected, the levels of deception of those exposed to Subject task Study and interpret View advertisements
corrective messages are rising toward their original advertisements
Claims tested Prevents colds Prevents, ameliorates,
levels. However, the increase for the company-source and cures colds
group is significant (p < 0.06), whereas the increase for Measures Beliefs (about brand Perceptions, beliefs,
the FTC-source group is not (p < 0.29; Table 5, rows 5 and product class) and saliences (for
and 6). brand attribute
This result could be interpreted as suggesting that the claims)
Control Deceptive ad Irrelevant ad
impact of a company-source message (at least with a Subjects Adult-education College students
single forced exposure to it) is not likely to persist for as students
long as would the impact of an FTC-source message. Time Three weeks Six weeks
However, it must be remembered that the analysis
shows no difference in the effects of the two sources at sistent (at least for three to six weeks). The only impor-
the time of the final measure. tant difference in findings concerned source effects
This study and the Dyer and Kuehl (1978) study differ over time.
on this finding. The Dyer and Kuehl results show the In the Dyer and Kuehl study (1978) and in the one
company-source message gaining ground on the FTC- reported here the content of the corrective message is
source message in its impact on deception levels over held constant, whereas the source is varied. In reality,
time. This study, on the other hand, suggests that both however, the content of the message would depend
corrective messages lost ground over time, with the heavily on who sponsored it. This relationship must be
company-source losing its effectiveness faster. One considered in conclusions drawn about the effective-
possible explanation for the difference in results is the ness of message sources.
time span prior to the final measure (three and six Although this and other studies provide insights into
weeks). Another possible explanation is the difference a number of deceptive and corrective advertising is-
in the tasks that subjects were required to perform. sues, many gaps in our knowledge remain. For exam-
Dyer and Kuehl subjects were asked to study print ads, ple, which approach is most effective and efficient for
whereas subjects in this study were exposed to 60- detecting advertising deception and for assessing the
second television commercials. This suggests that, ig- effects of corrective campaigns? There are currently no
noring reach and frequency considerations, corrective studies that have compared the salient belief and nor-
ads might be more effective in print media, where they mative belief techniques.
can be studied, than in broadcast media, where they A question also remains concerning standards to
may flash by too quickly to have substantial or lasting which advertisements should be held. At least some
impact. individuals will misread or misinterpret even the most
conscientiously prepared advertisement. On the other
CONCLUSIONS AND IMPLICATIONS hand, requiring that beliefs about a brand match beliefs
about the entire product class might find deception in
These results provide for some interesting compari- cases where a brand is truly differentiated, or might
sons between the study reported in this article and the overlook deceptions where there are misperceptions
Dyer and Kuehl (1978) study. Both studies constitute about the entire product class.
efforts to examine the longitudinal aspects of the meas- Not enough is known about the side effects of correc-
urement of deception and its correction. Despite some tive advertising-about its impact on company images,
similarities in the two studies, there are many differ- beliefs about other advertising claims, and, ultimately,
ences (Exhibit 3). sales of products for which corrective ads have been
Given the number and types of differences, the two run.
studies might have been expected to produce dissimilar The experimental studies reviewed in this article in-
results. Such was not the case. Both studies suggest volved forced exposure to potentially deceptive ads,
that the Listerine ad produced significant increases in and forced exposure to a single corrective message.
levels of deception. Both suggest that corrective ads Subjects generally were students. The Listerine correc-
had a significant effect in reducing deception. The tive campaign, and others that may follow, afford good
studies agree that source effects were statistically in- opportunities to study "real" consumers in "real" ad-
significant. Both suggest that corrective effects are per- vertising environments. The FTC has a major study

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
DETECTINGAND CORRECTINGDECEPTIVEADVERTISING 245

underway to analyze the effects of the ten-million- Gellhorn, Ernest (1969), "Proof of Consumer Deception Be-
dollar corrective advertising campaign currently being fore the Federal Trade Commission," Kansas Law Re-
implemented for Listerine. This FTC study of the Lis- view, 17, 559-72.
terine campaign, along with similar studies being con- Gerlach, Gary G. (1972), The Consumer's Mind: A Prelimi-
ducted, should help fill gaps in our knowledge of decep- naiy Inquity into the Emerging Problems of Consumer
Evidence and the Law, Cambridge, MA: Marketing Sci-
tive and corrective advertising. ence Institute.
Gurol, Metin N. (1977), "Deception in Advertising: A Re-
[Received May 1978. Revised June 1979.] view of Past and Current FTC Practice and an Experi-
mental Evaluation of a New Approach for Detecting
Deception and a New Approach for Eradicating Its Ef-
REFERENCES fects," unpublished Ph.D. dissertation, Graduate School
of Business Administration, University of North
Aaker, David A. (1974), "Deceptive Advertising," in Con- Carolina at Chapel Hill.
sutnerism: Search for the Consulmer Interest, 2nd ed., Haefner, James Edward (1972), "The Perception of Decep-
eds. David A. Aaker and George S. Day, New York: The tion in Television Advertising, An Exploratory Investi-
Free Press, pp. 137-56. gation," unpublished Ph.D. dissertation, College of Bus-
Armstrong, Gary M., Kendall, CL, and Russ, Frederick A. iness Administration, University of Minnesota.
(1975), "Applications of Consumer Information Process- , Birchmore, Melinda R., and Permut, Steven E.
ing Research to Public Policy Issues," Comminications (1974), "Applications of Belief Measures to Specific and
Research, 2, 232-45. Overall Advertising Claims: A Program for Research,"
, and McLennan, James P. (1973), "The Federal Trade in Proceeding of the 1974 National Conference for Uni-
Commission and the Investigation and Regulation of De- versity Professors of Advertising, ed. S. K. Zeigler, New
ceptive Advertising," in Conceptual and Methodologi- York: American Academy of Advertising, pp. 311-7.
cal Foundations in Marketing, ed. Thomas V. Greer, Howard, John A. (1974), "The Dialogue That Might Have
Chicago: American Marketing Association, pp. 430-4. Happened," in Advertising and the Public Interest, ed.
, and Russ, Frederick A. (1975), "Detecting Deception S. F. Divita, Chicago: American Marketing Association,
in Advertising,"MSU Business Topics, 23, 21-32. pp. 27-33.
Bernacchi, Michael D. (1974), "Advertising and Its Discre- and Hulbert, James (1973), Advertising and the Pub-
tionary Control by the FTC: A Need for Empirically lic Interest, Chicago: Crain Communications, Inc.
Based Criteria," Joirnal of Urban Law, 2, 223-66. Hunt, H. Keith (1972a), "Deception, Inoculation, Attack:
Cohen, Dorothy (1969), "The Federal Trade Commission and Implications for Inoculation Theory, Public Policy and
the Regulation of Advertising in the Public Interest," Advertising Strategy," unpublished Ph.D. dissertation,
Journal of Marketing, 1, 40-4. Department of Marketing, Northwestern University.
(1972), "Surrogate Indicators of Deception in Adver- (1972b), "Source Effects, Message Effects and Gen-
tising," Journal of Marketing, 3, 10-5. eral Effects in Counteradvertising," in Proceedings of
(1974), "The Concept of Unfairness as it Relates to the Third Annuial Conference of the Association for Con-
Advertising Legislation," Journal of Marketing, 3, 8-13. sumer Research, ed. M. Venkatesan, Chicago: Associa-
(1975), "Remedies for Consumer Protection: Preven- tion for Consumer Research, pp. 370-81.
tion, Restitution, or Punishment," Journal of Marketing, Jacoby, Jacob, and Small, Constance (1975), "The FDA Ap-
39, 24-31. proach to Defining Misleading Advertising," Journal of
Dillon, Tom (1973), "How the FTC Stacks the Deck," paper Marketing, 39, 65-8.
presented at the annual meeting of the American Associ- Jones, Mary Gardiner (1971), "The FTC's Need for Science
ation of Advertising Agencies, White Sulphur Springs, Research," inProceedings, Second Annual Conference,
WV. Chicago: Association for Consumer Research, pp. 1-9.
Dyer, Robert Frederick (1972), "An Experimental Evalua- Kassarjian, Harold H., Carlson, Cynthia J., and Rosin, Paula
tion of the Federal Trade Commission's Corrective Ad- E. (1975), "A Corrective Advertising Study," in Ad-
vertising Remedy," unpublished Ph.D. dissertation, Col- vances in Consumer Research, Vol. 2, ed. M. J.
lege of Business and Management, University of Mary- Schlinger, Chicago: Association for Consumer Re-
land. search, pp. 631-42.
, and Kuehl, Philip G. (1974), "The 'Corrective Adver- King, Alan L., and Wise, Gordon L. (1974), "A Cross-
tising' Remedy of the FTC: An Experimental Evalua- Tabular Ranking of Advertisements Based on Extent of
tion," Journal of Mar-keting, 38, 48-54. Deceptiveness," in Proceedings: Southern Marketing
, and Kuehl, Philip G. (1978), "A Longitudinal Study Association 1974 Conference, ed. Barnett A. Greenberg,
of Corrective Advertising," Journal of Marketing Re- Atlanta, GA: Southern Marketing Association.
search, 15, 39-48. Kirk, Roger E. (1968), Experimental Design: Procedures for
Ford, Gary T., Kuehl, Philip G., and Reksten, Oscar (1975), the Behavioral Sciences, Belmont, CA: Brooks/Cole
"Classifying and Measuring Deceptive Advertising: An Publishing Co.
Experimental Approach," in American Marketing Asso- Kuehl, Philip G. and Dyer, Robert F. (1976), "Brand Belief
ciation 1975 Combined Proceedings, ed. Edward M. Measures in Deceptive Corrective Advertising: An Ex-
Mazze, Chicago: American Marketing Association, pp. perimental Assessment," in Proceedings: 1976 Educa-
493-7. tors' Conference, ed. Kenneth Bernhardt, Chicago:
Gardner, David M. (1975), "Deception in Advertising: A American Marketing Association, pp. 373-9.
Conceptual Approach," Journal of Marketing, 39, 40-6. , and Dyer, Robert F. (1977), "Applications of the

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions
246 THE JOURNAL OF CONSUMER RESEARCH

'Normative Belief Technique for Measuring the Effec- Rosch, Thomas J. (1975), "Marketing Research and the Legal
tiveness of Deceptive and Corrective Advertisements," Requirements of Advertising," Journal of Marketing,
in Advances in Consumer Research, Vol. 4, ed. William 39, 69-72.
Perreault, Jr., Chicago: Association for Consumer Re- Sawyer, Alan G. (1976), "The Need to Measure Attitudes and
search, pp. 204-12. Beliefs Over Time: The Case of Deceptive and Correc-
Magnuson, WarrenG. (1972), "Consumerism and the Emerg- tive Advertising," in Proceedings: 1976 Edutcators'
ing Goals of a New Society," in Consumerism: View- Conferenice of the American Marketing Association, ed.
points from Business, Government, and the Public Kenneth L. Bernhardt, Chticago: American Marketing
Interest, ed. Ralph M. Gaedeke and Warren W. Etche- Association, pp. 380-5.
son, San Francisco: Canfield Press. Wilkie, William L. (1973), Consum-er Research and Correc-
Mazis, Michael B., and Adkinson, Janice E. (1976), "An Ex- tive Advertising: A New Approach, Cambridge, MA:
perimental Evaluation of a Proposed Corrective Adver- Marketing Science Institute.
tising Remedy," Journal of Marketing Research, XIII, (1974), "Research on Counter and Corrective Adver-
178-83. tising," in Advertising and the Public Interest, ed. S. F.
McCall, Robert B., and Appelbaum, Mark (1973), "Bias in the Divita, Chicago: American Marketing Association, pp.
Analysis of Repeated-Measures Designs: Some Alterna- 189-202.
tive Approaches," Child Development, 44, 401-15. (1975), Applying Attitude Research in Public Policy,
Pollay, Richard W. (1969), "Deceptive Advertising and Con- Cambridge, MA: Marketing Science Institute.
sumer Behavior: A Case For Legislative and Judicial , and Gardner, David M. (1974), "The Role of Market-
Reform," Kansas Law Review, 17, 625-37. ing Research in Public Policy Decision Making," Journal
Preston, Ivan L. (1976), "A Comment on 'Defining Mislead- of Marketing, 38, 38-47.
ing Advertising' and 'Deception in Advertising'," Jour- Winer, B. J. (1971), Statistical Principles in Experimental
nal of Marketing, 3, 54-7. Design, New York: McGraw-Hill Book Co.
Rao, Chatrathi P. (1974), "Deceptive Advertising: Con- Woodworth, Fred L. (1971), "Policy Planning and Consumer
sumer Concern and Evaluative Beliefs," paper pre- Protection-The Need for Behavioral Perspective,"
sented at the Southern Marketing Association Confer- paper presented at Ohio State University, Columbus,
ence, Atlanta, GA. OH.

This content downloaded from 188.72.127.90 on Wed, 25 Jun 2014 06:55:08 AM


All use subject to JSTOR Terms and Conditions

Das könnte Ihnen auch gefallen