Third Party Endorsement Influence on Perceptions
of Social Marketing Campaign
Attitudes, Credibility, Effectiveness, and Involvement
School of Journalism and Mass Communication
Dr. Terry Kinney
University of Minnesota
119 Murphy Hall
Minneapolis, MN 55455
[log in to unmask]
Manuscript submitted to the AEJMC
Public Relations Division
April 1, 2003
Third Party Endorsement Influence on Perceptions
of Social Marketing Campaign
Attitudes, Credibility, Effectiveness, and Involvement
This study examines the public relations concept of third party endorsement
in regards to a social marketing campaign. It was hypothesized that third
party endorsements would not influence the perceived legitimacy of the
campaign. In order to test this, a 2X3 (advertisement appeal type: fear,
humor by third party endorsement type: no endorsement, supporting,
opposing) experimental design was used. The findings supported the
assumption that third party endorsements do not significantly influence
Third Party Endorsement Influence on Perceptions
of Social Marketing Campaign
Attitudes, Credibility, Effectiveness, and Involvement
There is an overall need for research in the public relations field mainly
due to the fact that most public relations concepts are taught without
formal research to support them, leaving these concepts subject to debate
regarding importance, usefulness, or effectiveness (VanLeuven, 1990). In
addition, despite contrary viewpoints, public relations practitioners have
long been taught that getting "good publicity" in the form of independent
third party endorsements is an essential part of their jobs. Even the
"founder" of public relations, Edward Bernays posited that public relations
was manufacturing public compliance through propaganda, staged events, and
most of all - third party endorsements (as quoted in Levine, 1994).
Third party endorsements are a free (as advertising is paid for)
independent media message that the public assumes is credible and reliable,
as it does not originate from the organization being promoted (Wilcox et
al., 2000: Blyskal & Blyskal, 1985). It is because of this "independence"
from the promoting organization, that Bernays posited that companies obtain
legitimization through media attention to their cause. Consequently,
independent third party endorsements are portrayed as possibly the most
important and influential evaluation of the success of a campaign (Matera
et al., 2000). Not only is it assumed that third party endorsers such as
media legitimize certain ideas and behaviors, but also that they present
new ideas and behaviors to the public with credibility (Cutlip, 1994).
Thus, media attention is portrayed as extremely significant because a
majority of what society knows about campaigns reaches them through media
accounts (Olien, et al., 1989).
As a result, third party endorsements are depicted as essential to a public
communication campaign because, as McAlister (1989) and Salmon (1989)
noted, many social behaviors - presented in campaigns promoting social
change - can be learned from mass media discourse. In addition, because it
is speculated that the media have the ability to influence audiences
concerning social topics, it is important to study how the media discussion
of a social movement affects the movement.
Social marketing campaigns - often referred to as public service,
information, or public communication campaigns attempt to change an
aspect of society that is seen as detrimental in some way (Straubhaar &
LaRose, 2002) These campaigns try to catch the attention of the public and
the media gatekeepers by using public relations and advertising techniques
to "sell" people on pro-social (positive qualities that we want to
encourage in society) behaviors (Straubhaar & LaRose, 2002; Flay & Cook,
Social marketing campaigns merge marketing, advertising, and public
relations to disseminate information about an organization (or an
organization's cause) that the organization hopes will, inevitably create
changes in society (Solomon 1981). Thus, social marketing has an underlying
ethical and moral position inherent in the message. In addition, social
marketing, as an integrated communication approach to behavior change,
combines media, interpersonal influence, and carefully managed efforts to
influence public actions (Straubhaar & LaRose, 2002).
However, whether independent third party endorsements are or are not an
effective means to legitimize social movements, the field of public
relations is currently using the same tactics it originally utilized to
promote cigarettes to now attack them. Therefore, it is not only ironic
that the topic of this study is campaigning against the cigarette industry
(one of the first users of public relations). But, the particular
organization of study is especially controversial because it is attacking
an industry with its own funds.
This study uses a 2X3 (advertisement appeal type: fear, humor by third
party endorsement type: no endorsement, supporting, opposing) experimental
design to examine the effectiveness of the Minnesota based anti-tobacco
social marketing campaign called Target Market when paired with third party
endorsements implying either support or opposition to the organization's
message. In essence, the issue under study here is whether third party
endorsements influence the effectiveness of social marketing campaigns.
The present study differs from previous research by applying the concept
of third party endorsements to a social marketing campaign (as opposed to
advertising for a brand or product) and by investigating perceptions of
that campaign based on credibility, attitude, involvement, and perceived
effectiveness. Ironically, the study examines what some public relations
historians depict as the beginning of the "smoking movement" an
independent third party endorsement (Tate, 1999).
Although limited research exists on the influences of third party
endorsements, the literature that does is contradictory on the
effectiveness of this public relations concept. Third party endorsements of
products and brands, mainly in the form of media recommendations, have
largely dealt with attitude formation and purchase intention. Thus, this
review provides literature relevant to the theory of third party
endorsements as a tool to influence audiences/consumers.
In a study of product endorsements, Dean and Biswas (2001) used the third
party organization (e.g., Chrysler using its Motor Trend 'Car of the Year'
award as endorsement) endorsement in contrast to no endorsement, or a
celerity endorsement, and tested the variables of quality, attitude, risk,
and information value in regards to the endorsement types. Their 2X3
(products types: good, service; third party endorsement types: celebrity
endorsement, third party organization endorsement, no endorsement)
experimental design resulted in the designation of the third party
organization endorsement as a signal of quality of the product and
informational value. Despite the substantial findings that Dean and Biswas
work provided, their research had several limitations. First, as Dean and
Biswas noted, their conditions were "very artificial" (Dean & Biswas, 2001
p. 19) due to highly altered advertisements, confusion as to the third
party organization endorsers' identity, and the impossibility of the
celebrity endorser they used. Second, their results might not be
generalizable since they used a population (college students) with little
experience regarding the high-involvement and value items (desktop
computers and auto insurance) they used in their advertising stimuli.
Kamins and Marks (1991) used the concept of third party endorsement to
provide a "kosher claim" for familiar and unfamiliar brands of breakfast
cereal. Their study attempted to create "kosher" perceptions which they
suggested would in turn create differences in cognitions, attitude toward
the ad, attitude toward the brand, and purchase intentions. However, their
results were only significant for the familiar brand signaling that third
party endorsement of an unfamiliar product may not be effective. In
addition, they found that many participants perceived the "kosher"
Hallahan (1999) posited that the public does not necessarily notice
differences between advertising and news. In fact, Hallahan intimated that
third party endorsements can be discredited due to a number of factors:
preference for news, resistance to advertising, miscomprehension of
messages, or cognitive processing differences for persuasive messages.
Thus, Hallahan measured attitude toward advertising and news to determine
what perceptions people have of the two media in regards to their ability
to present products to the public based on trustworthiness, credibility,
reliability, and confidence in each media. In this study, it was determined
that the public's perceptions of each media tend to diminish the effects of
each, lending no support to the concept of third party endorsement
superiority as a persuasive message.
Though not called as such, Lazarsfeld and Merton (1948) broached the idea
of third party endorsement in their theory of status conferral. In it
Lazarsfeld and Merton posit, "[T]he mass media confer status on public
issues, persons, organizations and social movements" through media
coverage (p. 101). This coverage serves to legitimize the authority and
significance of organizations, except it was not applied to social
movements but to advertising testimonials. Through this research,
Lazarsfeld and Merton found that "celebrity" testimonials seemed to advance
the reputation of both the product and the presenter.
In summation, prior research on third party endorsements has been in
regards to effects on consumer perception of brands or products or
advertising campaigns. However, a systematic review of the literature
revealed no studies that examined the applications of this public relations
tool on a social marketing campaign.
Thus, this type of campaign research is important because it will help fill
a gap in the literature regarding effectiveness of third party endorsement
of social marketing campaigns. Most of those actually conducting PR
research would agree (Judd, 1990; Rosser, et al, 1990), that evaluating
campaign effectiveness is essential, not only in campaign evaluation, but
also in predicting whether campaign endorsements will affect public
attitudes and eventually, behaviors.
Theory and Hypothesis
It is apparent from this literature review that public relations research
will benefit greatly from the study of third party endorsement effects on
social marketing campaigns. Germane to this research is the assumption that
social marketing campaigns are persuasive messages which are often
perceived as advertisements, and that third party endorsements are seen as
independent media accounts. Thus, the general purpose of this study is to
examine third party endorsements as a public relations tool which - it is
suggested by some - should garner support for a social marketing campaign.
Consequently, the following research questions were posed: (1) Do third
party endorsements impact social marketing campaigns? (2) To what extent
does the type of third party endorsement influence the public's perception
of the campaign? and (3) Will the type of appeal used in the campaign
influence the effects of the third party endorsement?
In social marketing, an organization is trying to get certain themes or
messages out to the public. Ultimately, there are objectives the
organization wants to achieve with its message, not the least of which is
being perceived as legitimate by the public. Objectiveness of the press is
a quality that most of the public perceives as part of the news reporting
system. This idea is essential to the concept of third party endorsement
because it means that the press has "investigated" whatever cause they are
reporting about and are now giving the public the "facts" (Fairclough, 1995).
It is suggested that if media reports are favorable to an organization's
cause or overall appearance to the public, the organization's message then
becomes more credible and acceptable. Thus, it is speculated that how the
media portrays a social marketing campaign can often make or break it in
terms of effectiveness. As Fairclough (1995) posited, viewpoints are often
legitimized by newspaper accounts. In addition, van Dijk (1991) suggests
that most people's knowledge about the world comes from news
accounts. This should ultimately result in the media either endorsing or
legitimizing the organization's cause, or denouncing it.
However, recent research, most notably by Hallahan (1999), suggests that
third party endorsements may not actually persuade people in the way once
thought as most people are skeptical of all media. Consequently, the news
may not be any more persuasive than advertising. In order to test this
assumption we need to address the legitimacy of a campaign based on its
advertising and on third party endorsements. This study assumes that the
following concepts are part of a campaign's being deemed legitimate: 1) The
public believes the campaign is credible, 2) The public attitude toward the
campaign is favorable, 3) The public perceives that the campaign is
relevant, and 4) The public perceives the campaign as effective.
Notwithstanding, drawing from the work of Hallahan (1999), there should be
no difference between those who view advertising and those who read a news
article about the same item. For this study, this notion was applied to the
influence of third party endorsement on a social marketing campaign.
Subsequently, the following hypothesis was made:
Hypothesis: In regards to social marketing campaigns, there is no direct
relationship between a third party endorsement and the public's assessment
of the campaign's legitimacy in terms of attitude toward the campaign,
relevance of the campaign, credibility of the campaign, and perception of
the campaign's effectiveness.
A 2X3 (advertisement appeal type: fear, humor by third party endorsement
type: no endorsement, supporting, opposing) experimental design was used
wherein two different persuasive social marketing advertisements were
viewed by three different subject groups each. In the 2X3 (advertisement
appeal type: fear, humor by third party endorsement type: no endorsement,
supporting, opposing) design, one control group (C) watched only the
humorous PSA (H), one watched the humorous PSA and read an "opposing" third
party endorsement (OH), and one watched the humorous advertisement and read
a "supporting" third party endorsement (SH). The same happened for the
other three groups (CF, OF, SF) this time using the fear PSA. Following
each of these, participants filled out questionnaires designed to measure
their perceptions and attitudes regarding the PSA and social marketing
campaign organization, Target Market.
Estimation of Sample Size for Null Hypothesis
Since we hypothesized that there would be no relationship between the
third party endorsement and perceptions of social marketing campaign
legitimacy, a estimated power was required to establish how many
participants were desirable in each cell of the 2X3 (advertisement appeal
type: fear, humor by third party endorsement type: no endorsement,
supporting, opposing) experimental design to lead to statistically
significant results for the null hypothesis. Drawing from Cohen (1992) the
sample size necessary for statistical power was estimated prior to
recruiting participants as doing so this increases the probability that we
will get a significant result for our hypothesis.
Since the research is considered exploratory, the necessary significance
criterion needed for this study is between .10 and .05 of a two-sided test,
however we chose to use .05. Thus, using Cohen's effect size of .80 as the
objective power it was estimated that for a significance of .05 it was
necessary to use between 14 (large population effect size) and 35 (medium
population effect size) subjects in each cell of the 2X3 experimental design.
The participants included 183 undergraduate students from the University of
Minnesota Twin Cities. Participants were recruited from 10 different
communication studies or journalism and mass communication classes. They
were offered extra credit for class not less than one percent of their
total grade for participating in the study. Of these, N= 70 were male and
N= 112 were female. Moreover, N= 103 reported being either a past or
present smoker. In addition, most were Caucasian, with only 12 percent
minority students. The ages of the students were 17 to 29+, with the mode
age being 20-22 (N= 140) - making the majority of the population part of
the target market for the anti-tobacco messages presented. Demographic
characteristics of the sample population are shown in Table 1.
Insert Table 1 About Here
Participants were randomly assigned to one of the six cells of the 2X3
(advertisement appeal type: fear, humor by third party endorsement type: no
endorsement, supporting, opposing) design. In the humor control group (H)
there were 32 participants. The humor and "opposing" third party
endorsement (OH) cell consisted of 28 participants. There were 29
participants in the humor and "supporting" third party endorsement (SH)
cell. In the fear control group (F) there were 26 participants. The fear
and "opposing" third party endorsement (OF) cell consisted of 28
participants. And, there were 29 participants in the fear and "supporting"
third party endorsement (SF) cell.
Definition of the Variables
For this study, the independent variables were fear appeal, fear appeal
and opposing third party endorsement, fear appeal and supporting third
party endorsement, humor appeal, humor appeal and opposing third party
endorsement, and humor appeal and supporting third party endorsement.
The dependent variables were credibility, attitude, relevance, and
effectiveness. These dependent variables were operationalized as:
Credibility. Credibility was defined as "worthiness of being believed"
(West, 1994). In this study, "credibility" was measured in regards to the
PSA through two constructs, 1) quality and validity of the message claims,
and 2) credibility of the message. The perception of quality and validity
of the message claims were measured by seven items such as persuasive,
believable, and convincing. Perception of credibility of the message was
measured by seven statements such as: "The claims in the ad are true," and
"I think the information in this advertisement is credible."
Attitude. Attitude as defined by Petty and Caccioppo (1981) is a positive
or negative feeling regarding an item. In this study, attitude is divided
into two constructs, 1) attitude toward the organization, and 2) attitude
toward the message. Attitude toward the organization was measured using
nine items such as favorable, useful, and like a lot. Attitude toward the
message was measured using seven statements such as: "This advertisement is
offensive," and "This advertisement is informative."
Involvement. Involvement was defined as the individual's perception of the
relevance of the advertisement (Zaichowsky, 1985 & 1994), and was measured
by 12 items such as important, beneficial, relevant, and significant.
Effectiveness. Effectiveness was defined as individual perceptions of
significance, influence, powerfulness, and success of the campaign. In this
study," effectiveness" was measured by six statements such as: "I think
Target Market is significant in reaching youth smokers," and "I think
organizations like this are not influential to youth smokers."
Conceptually any difference in credibility, attitude, involvement, and
perceived effectiveness between the control groups (those simply viewing
the messages with no third party endorsement) and the experimental groups
(those viewing the messages with a third party endorsement) are likely the
result of the third party endorsement.
Fear and humor appeals were chosen as the replication variables for this
study as they were opposites in terms of effects on viewing audience
emotions. Although both types of appeals are seen as problematic (fear
appeals often make audiences uncomfortable, while humor appeals are
subjective as humor is seen differently by different ages, genders,
cultures, etc.), both however are known to stimulate emotional responses on
the part of the viewer (Schiffman & Kanuk, 1999; Cronkhite, 1969). In
addition, these are the two most common appeals used by Target Market in
their campaigns. This should lend validity to our results as any
differences between the two endorsements, without any appeals effects, will
mean that most likely only the endorsements affected perceptions.
Pre-Test of Stimuli. A small focus group of seven people (3 male and 4
female) aged 21-30 viewed the 24-advertisement archive of Target Market in
order to give their reactions to each public service announcement. Of these
seven, four were current or former smokers, and all were familiar with the
Target Market organization though none had ties to the organization. The
participants were shown each of the 24 public service announcements and
given 2 minutes in between each to rank the advertisement using a
five-point semantic differential scale based on potency (how strong or
weak) the fear, humor, sympathy, and attack characteristics of the
After viewing each of the public service announcements, the group was given
the opportunity to talk about the advertisements they had just seen. The
participants indicated different aspects of the advertisements (e.g., how
horrified or entertained they were) and identified four PSAs they thought
were most effective and creative. The investigator then summated the
semantic differential based on the four characteristics. The four the focus
group had spoken of (Driving to Work, Thank You, Secrets of a Tobacco
Executive, and Prisoner) were also the ones with the strongest opposing
ratings for fear and humor on the semantic differential scale. The two
strongest (and closest in attitude ranking) for fear and humor were thus
established as the public service announcements that would be used in the
study. These were "Thank You" with a humor attitude ranking of 60 out of 70
possible, and "Secrets of a Tobacco Executive" with a fear attitude ranking
of 65 out of 70 possible.
The fear appeal PSA (F) chosen for this study modeled itself after a
scary-movie preview about an average father who works at a company that is
studying the behavior of kids so it can "target" them for death. It is
entitled, "Secrets of a Tobacco Executive." The humor PSA (H) chosen for
this study depicted a booth where kids record messages to send to tobacco
companies. The title is, "Thank You," and each of the youth mockingly
asserts what it is emphysema, losing a loved one, male impotence, bad
breath, yellow skin and teeth that smoking has done in their lives. Both
are 60-second, full-color PSAs from the 2001 anti-tobacco archives of
Because it has been suggested that newspapers are the most influential
medium of opinion communication (Bremeck & Howell, 1976), two news articles
were used as third party endorsements of the Target Market campaign. These
were compilations of four actual anti-tobacco media articles printed in the
Minnesota Daily, Ad Age, North Country Times, and St. Paul Pioneer Press as
well as a video provided by Target Market of an anti-tobacco rally.
Moreover, both article manipulations were similar to the type of coverage
anti-tobacco organizations such as Target Market currently receive.
The first manipulation (opposing) contained descriptions of the
organization as "wrongfully attacking" tobacco companies and their
executives. The article depicts Target Market as an organization involved
in lawsuits for attacking tobacco companies -violating the 1998 Master
Settlement Agreement which prohibits personal attacks and "vilification" of
tobacco companies or their executives in anti-smoking advertising
campaigns. The article ends with Target Market denying any wrongdoing, and
trying to divert attention back to tobacco company advertising. (See
The second manipulation (supporting) described the success of the
organization in decreasing instances of youth smoking in Minnesota. The
article acts as the resulting coverage of a Target Market sponsored rally
called "Amnesty Day" - part of a special campaign that was supposed to give
tobacco companies a chance to apologize for all of the things they have
done. The article goes on to describe Target Market as successfully
"educating" youth about cigarettes by using the tobacco companies' own
money for events and advertisements to speak out against "Big Tobacco."
(See Appendix A).
Both manipulations were shown as being written by Stanley Wills (a
fictional journalist) from The New York Times and had a headline of,
"Anti-Tobacco Takes on Big Tobacco." In addition, the articles uniformly
contained approximately 400 words and had three quotes apiece. Intensity of
the articles as manipulations either supporting or opposing was moderate as
normal newspaper coverage of social marketing campaigns are generally not
In this study, the source of the persuasive communication complementing the
advertisement was a third party endorsement either supporting or opposing
the organization's tactics. Therefore, a traditional media credibility
scale was used as the manipulation check for the article. It consisted of
the same five-point scale used by Johnson and Kaye (1998) with
believability, accuracy, and fairness being subject rated from "not at all"
to "very much." However, the measure for the article used the additional
Flanagin and Metzger (2000) dimensions of bias as well as completeness or
depth of the coverage. As was the case with the Flanagin and Metzger's
credibility scales, in order to ensure that all measures indicate higher
levels of credibility, score reports for bias were reverse-coded.
In addition, a manipulation check for the stimuli was used to determine
that the independent variables of fear and humor were perceived correctly
by the audience. For this a 6-point version of the "amused" scale created
by Lacher and Mizerski (1994) and a 6-point version of the "fear" scale
created by Block and Keller (1995) ranging from "not at all" to "very much
so" were used. Block and Keller reported internal consistency reliability
coefficients for the "fear" scales ranging from .87 to .89 on prior
Credibility of the PSA was measured using a combination of the Putrevu and
Lord (1994) and Block and Keller (1995) credibility scales. These
statements were a 5-point version of the two ranging from "strongly agree"
to "strongly disagree" and were structured for participants to reflect on
the advertisement. The Putrevu and Lord scale reported an internal
consistency reliability coefficient of .81, while the Block and Keller
scale reported internal consistency reliability coefficients ranging from
.69 to .63 on previous experiments. In order to ensure that all measures
indicated higher levels of perceived credibility, score reports for
dishonesty, exaggeration and unbelievability were reverse-coded.
The quality and validity of the PSA claims were measured using a 5-point
semantic differential version of the "argument strength" scale developed by
Petty, Caccioppo, and Schumann (1983). The "argument strength" scale
reported internal consistency reliability coefficients ranging from .76 to
.94. In order for all measures to indicate higher levels of argument
strength, score reports were reverse-coded so that negative adjectives were
Attitude toward the PSA was measured using Neese and Taylor's (1996) scale
"attitude toward the ad." The measure consisted of seven 5-point items
which examine an individual's perception of an advertisement using
"strongly disagree" to "strongly agree." The "attitude toward the ad" scale
reported an internal consistency reliability coefficient of .83 on previous
experiments. In addition, attitude toward the organization was measured
using Homer's (1995) "attitude toward the business" scale which reported an
internal consistency reliability coefficient of .91 on previous
experiments. Both measures were reverse coded in order to ensure that
higher numbers indicated higher levels of agreement or favorability.
Following which, overall attitude in regards to the advertisement's
relevance questions were asked. This section was based on Zaichkowsky's
(1994) involvement inventory of opposing adjectives. Using a semantic
differential scale, participants could rank their overall attitude toward
the advertisement previously viewed. Adjective pairs such as
important/trivial, insignificant/relevant, worthless/valuable,
interesting/boring, appealing/disengaging were used to determine the
subject's perceived relevance of the advertisement and reported internal
consistency reliability coefficients ranging from .96 to .99 on previous
experiments. In order to ensure that all measures indicate higher levels of
relevance, score reports for positive adjectives were reverse-coded so that
higher numbers reflected higher relevancy.
Effectiveness of the organization was specifically designed for this study
and used a Likert Scale with five-point degrees of agreement to a series of
opinion statements expressing either a positive "strongly agree" or
negative "strongly disagree" viewpoint. The opinions ranged from "this
message is significant in reaching youth smokers" to "the message is
wrongfully slandering tobacco companies" to "I believe this message is
powerful." Statements were reverse-coded to ensure that higher numbers
indicated agreement with positive statements, while lower numbers indicated
agreement with negative statements. (See Appendix B for this measure).
Participants were randomly placed 10 at a time in one of two classrooms
(one room for the fear PSA, another for the humor PSA) containing tables
and chairs. At the front of each room was a 32" color television on a 4'
high stand. Packets containing the consent form, a manipulation article
(unless in the control group), and questionnaire were randomly distributed
and placed facedown on the tables and students were instructed to sit in a
chair in front of a packet. Students were then instructed that they would
be taking part in one of the following conditions: (1) watching a public
service announcement and filling out a questionnaire, or (2) watching a
public service announcement, reading an article, and filling out a
Participation in the study was conditional based on the student signing a
consent form that indicated the student was over 18 years old, that there
were no risks to participating, and that the student could refuse to answer
or withdraw from the study at any time. In addition, participants were
instructed that all answers would be confidential, as the consent form
would be detached from their questionnaire before turning it in to the
experimenter. Since the experimenter would not see the two forms together,
this also ensured that subjects would still receive credit for
participating even if they chose not to complete their questionnaire.
After filling out the consent form, participants were shown either the
60-second fear or humor PSA. They were then instructed to read the article
following the consent form (unless part of the control group which did not
receive articles) and when finished proceed to the questionnaire. Once they
were finished filling out the questionnaire, participants were asked to
remove the consent form from the front of their packet and place it in a
separate pile from the article/questionnaire section of the materials. The
entire process took approximately 7-15 minutes.
As mentioned earlier, this research was guided by three questions: (1) Do
third party endorsements impact social marketing campaigns? (2) To what
extent does the type of third party endorsement influence the public's
perception of the campaign? and (3) Will the type of appeal used in the
campaign influence the effects of the third party endorsement? Thus, the
following analyses were performed in order to establish inferences about
third party endorsements.
Manipulation checks. To make sure the independent variables were
effectively manipulated, the validity of the manipulation checks were first
examined. The first step was to conduct a factor analysis of the variables
that made up the article and stimuli measures. A Varimax rotation factor
analysis for the article manipulation check extracted two factors, however
one of the factors had only two variables (biased and dishonest) and was
dropped. The remaining factor was then examined for internal reliability.
Using coefficient alpha reliability outcomes (Cronbach, 1951), the article
manipulation check exceeded .70. A one-way analysis of variance (ANOVA)
showed that the manipulations were working correctly F=245.58; with d.f.=2,
The fear appeal manipulation was then factor analyzed using a Varimax
rotation factor analysis and resulted in loading on only one factor. The
coefficient alpha reliability outcomes for the fear appeal manipulation
exceeded .70. A t-test showed that the fear appeal worked correctly
t(180)=7.07, p<.001. In addition, the humor appeal manipulation also loaded
on only one factor and its coefficient alpha reliability outcome exceeded
.70, and the t-test showed that the humor appeal worked correctly
Credibility. Using a Varimax rotation factor analysis, one factor was
extracted for each of the two credibility measures. Quality and validity of
the message claims and credibility of the message both exceeded coefficient
alpha reliability outcomes of .70.
Attitude. Using a Varimax rotation factor analysis, one factor emerged for
the attitude toward the issue measure the coefficient alpha reliability of
the PSA measure exceeded .70 as did the reliability of the one factor that
emerged for the attitude toward the organization measure.
Relevance. Using a Varimax rotation factor analysis, two factors were
extracted from the relevance measure of the PSA. These were labeled as
Relevance and Involvement. Relevance was defined by the independent
variables of important, of concern to me, means a lot to me, valuable,
beneficial, relevant, significant to me, essential, and matters to me in
regards to the PSA. Involvement was defined by the independent variables of
interesting, exciting, and appealing in regards to characteristics of the
PSA. Coefficient alpha reliability outcomes for relevance and involvement
both exceeded .70.
Effectiveness. Using a Varimax rotation, two factors were extracted from
the effectiveness measure, however one of the factors had only two
variables (attacking and slandering) and was dropped. The remaining factor
was then examined for internal reliability. Using coefficient alpha
reliability outcomes, the effectiveness measure exceeded .70. Factor
loadings and alphas for all factors are reported in Table 2.
Insert Table 2 About Here
Research Questions and Hypothesis
Our first research question concerned the impact of third party
endorsements on public relations campaigns. To address this question,
control groups (those who only saw the public service announcements) were
compared to experimental groups (those who saw the public service
announcement and read the manipulated article) based on each of the
dependent variables using an ANOVA. Only credibility as a dependent measure
was significantly (p<.05) different among the three conditions. Consistent
with Cohen (1992) the power to detect differences in design at .05 medium
effects size is .80. Table 3 contains the means, F-values, and significance
for each dependent variable.
Insert Table 3 About Here
The second research question addressed the effects of third party
endorsements on public perceptions. It was hypothesized that there would be
no direct relationship between the type of third party endorsement and the
public's perception of the campaign's effectiveness based on attitude,
involvement, and credibility. To address the second research question, the
data were analyzed using multivariate analysis of variance (MANOVA).
A one-way MANOVA was conducted to determine the effect of third party
endorsements (no endorsement, supporting endorsement, opposing endorsement)
on two independent variables. The two independent variables were fear
appeal PSA and humor appeal PSA. In regards to the hypothesis that there is
no direct relationship between the third party endorsement and the public's
assessment of a social marketing campaign's legitimacy in terms of attitude
toward the campaign, relevance of the campaign, credibility of the
campaign, and perception of the campaign's effectiveness, only one factor
showed significant difference. Consistent with Cohen (1992) the power to
detect differences in design at .05 medium effects size is .80. Table 4
contains the means and standard deviations for the three endorsement groups.
Insert Table 4 About Here
Based on the MANOVA, only one factor showed significance - credibility of
the message, F=3.44; with d.f.=2, p<.05. The other factor that was reaching
the level of significance was validity and quality of message, F=2.69; with
The third research question concerned appeal type in regards to third party
endorsement. It was suggested that the relationship between third party
endorsement and public relations campaign would exist regardless of type of
advertising appeal used in the public relations campaign. To address this
question, those groups who saw the humor appeal PSA were compared to the
fear appeal PSA based on each of the dependent variables using ANOVA. For
the appeals, credibility of the message, and involvement were significantly
(p<.001) different between the two conditions. In addition, for exploratory
research such as this, attitude toward the message was reaching the level
of significance (p=.08). Table 5 contains the means, F-values, and
significance for each dependent variable.
Insert Table 5 About Here
By using statistical power analysis of .80 to detect differences in the
experimental design at .05 medium effects size, we were able to obtain
significant results that our null hypothesis for third party endorsements
was supported. Thus, there was substantiation for the prediction that there
would be no direct relationship between the third party endorsement and the
public's assessment of a social marketing campaign's legitimacy in terms of
attitude toward the campaign, relevance of the campaign, credibility of the
campaign, and perception of the campaign's effectiveness. Although
credibility was significant in the MANOVA, by isolating the effects of PSA
appeal and third party endorsement type on credibility (using ANOVAs) we
were able to see that the effects were due to perceptions of the PSA, not
the third party endorsement. In addition, none of the other measures of
legitimization were significant.
However, an unexpected finding in this study was that all participants were
significantly more involved with the fear appeal as compared to the humor
appeal. In addition, attitude toward the fear appeal was more favorable.
One possible explanation is that the majority of the participants reported
already seeing the humor appeal while the fear appeal was new to them, thus
more involving and more effective in attitude formation.
The results of this study confirm the literature which suggests that third
party endorsements do not affect perceptions. These findings are consistent
with those suggesting that third party endorsements do not aid in the
persuasion process. These data, along with those from other third party
endorsement studies imply that the concept of third party endorsements as a
public relations tool for legitimizing brands, products, organizations, or
even viewpoints through publicity needs to be rethought.
For Further Study
The findings of this study are curious and deserve further research.
Several items could be altered in supplementary studies that could further
examine the effects of third party endorsements on social marketing
campaigns. For example, to strengthen the manipulation, participants could
be shown a "package" of PSAs from a given campaign and read multiple
articles written about the campaign. This would give them more of a
foundation from which to base attitudes as well as establish the
credibility of both the organization and the issue.
Moreover, the strength of the manipulation could be changed to present
articles that more clearly support or oppose the organization or its
message. In this study, the support/opposition was a very weak manipulation
that could have simply been construed by participants as unbiased reporting.
Furthermore, while attitudes toward the advertisement were described in
these findings, we were in no way trying to predict behavioral intent.
Moreover, we were not trying to predict which type of third party
endorsement or type of advertising appeal would be more effective, we were
merely studying the effects that third party endorsements had on audience
perceptions. However, the linkages between attitudes and behaviors in
regards to social marketing messages would be of interest for further
study. In addition, a study examining the effectiveness of type of appeal
would be appealing.
This study is best described as a pilot study as there are several
limitations in its design. First, it should be noted that some scholars
disapprove of the questionnaire as a means of studying attitude (Dawes &
Smith, 1985; Scott, 1969) due to validity of attitude scales. In regard to
social marketing research this issue may be especially prevalent as
participants may not have wanted to give responses that they felt were not
within acceptable limits of societal norms. This was seen during the data
analysis as most participants refrained from opinions, attitudes, or
beliefs that were in strong agreement or strong disagreement with the
Failure to respond regarding immediate attitude measures such as these
could also be due to the familiarity of Minnesota participants in the study
with the Minnesota-based Target Market campaign. The stimuli and
manipulations may not have as much of an impact on Minnesota participants
(as many of them reported seeing the stimuli before the study). However, it
could also mean that the participants did not have strong opinions in
In regards to our population, the number of smokers (past or present) was
over half of the total population used for the study and is not
representative of the overall population of Minnesota youth. This could
have resulted in the lack of persuasion seen in this study instead of any
news favoritism or advertising bias as suggested by Hallahan (1999).
In addition, because a convenience sample was used, generalization to
populations should be made with caution as students who volunteered to
participate may not be representative of all 17-29 year olds. In addition,
use of only college students, when at least half of the target market for
this organization is under the age of 18 restricts generalization to the
target market. Thus, the results of this study should be considered
exploratory and in need of replication.
Atkin, Charles K. (1981), "Mass Media Information Campaign Effectiveness,"
from Public Communication Campaigns. London: Sage Publications, Inc.
Blyskal, Jeff and Marie (1985), How the Public Relations Industry Writes
the News. New York: William Morrow and Company, Inc.
Block and Keller (1995), "When to Accentuate the Negative: The effects of
Perceived Efficacy and Message Framing on Intentions to Perform a
Health-Related Behavior," Journal of Medical Research, 32 (May), 192-203.
Bremeck, Winston L. & Howell, William S. (1976), Persuasion: A Means of
Social Influence, Second Edition, Prentice-Hall, Inc., Englewood Cliffs,
Buffett, Warren (1995), "Smokers' Hacks," from Toxic Sludge is Good for
You. Edited by John Stauber and Sheldon Rampton. Common Courage Press.
Cohen, Jacob (1992), "A Power Primer," Psychological Bulletin, vol. 112
Cronkhite, Gary (1969), Persuasion: Speech and Behavioral Change, The
Bobbs-Merrill Company, Inc. Indianapolis and New York.
Cutlip, Scott M. (1994), The Unseen Power: Public Relations. A
History. Hillsdale, New Jersey: Lawrence Erlbaum Associates.
Dawes, R.M. & Smith, T.L. (1985), Attitude and Opinion Measurement in G.L.
Lindzey and E.A. Aronson (Eds.), Handbook of Social Psychology, Third
Edition, Random House, New York.
Dean, Dwane Hal and Biswas, Abhuijit (2001), "Third-Party Organization
endorsement of products: An advertising cue affecting consumer prepurchase
evaluation of goods and services," Journal of Advertising, (Winter).
Dillard, James Price and Peck, Eugenia (2000), "Affect and Persuasion:
Emotional Responses to Public Service Announcements," Communication
Research, vol. 27 (August).
Fairclough, Norman (1995), Media Discourse. London: Oxford University Press.
Flay, Brian R. and Cook, Thomas D. (1981), "Evaluation of Mass Media
Prevention Campaigns," from Public Communication Campaigns. London: Sage
Hallahan, Kirk (1999), "No, Virginia, It's not true what they say about
publicity's "Implied Third-Party Endorsement" effect," Public Relations
Homer, Pamela M. (1995), "Ad Size as an Indicator of Perceived Advertising
Costs and Effort: The Effects on Memory and Perceptions," Journal of
Advertising, 24 (Winter), 1-12.
Judd, Larry R. (1990), "Importance and use of formal research and
evaluation," Public Relations Review, vol. 16 (Winter).
Kamins, Michael A. and Marks, Lawrence J. (1991), "The perception of kosher
as a third party certification claim in advertising for familiar and
unfamiliar brands," Journal of the Academy of Marketing Science, (Summer).
Lacher, Kathleen T. and Mizerski, Richard (1994), "An Exploratory Study of
the Responses and Relationships Involved in the Evaluation of, and in the
Intention to Purchase New Rock Music," Journal of Communications Research,
21 (September), 366-380.
Lazarsfeld, Paul F. and Merton, Robert K. (1948), Mass Communication,
Popular Taste and Organized Social Action, in The Communication of Ideas,
Lyman Bryson ed., Harper and Brothers, New York and London.
Levine, Michael (1994), Guerrilla P.R.: How you can wage an effective
publicity campaign without going broke. New York: HarperCollins Publishers,
Matera, Fran R. and Artigue, Ray J. eds. (2000), Public Relations Campaigns
and Techniques: Building Bridges into the 21st Century. Boston: Allyn and
McAlister, Alfred (1981), "Antismoking Campaigns: Progress in Developing
Effective Communications," from Public Communication Campaigns. London:
Sage Publications, Inc.
McAlister, Alfred, Ramirez, Amelie G., Galavotti, Christine and Gallion,
Kipling J. (1989), "Antismoking Campaigns: Progress in Application of
Social Learning Theory," from Public Communication Campaigns, 2nd Edition.
London: Sage Publications, Inc.
Neese, William T. and Taylor, Ronald D. (1994), "Verbal Strategies for
Indirect Comparative Advertising," Journal of Advertising Research, 34
Olien, Clarice N., Tichenor, Phillip J. and Donohue, George A. (1989),
"Media Coverage and Social Movements," from Information Campaigns:
Balancing Social Values and Social Change. London: Sage Publications.
Perloff, Richard M. (1993), The Dynamics of Persuasion, Lawrence Erlbaum
Associates, Hillsdale, New Jersey.
Petty, Richard E. and Caccioppo, John T. (1981), Attitudes and Persuasion:
Classic and Contemporary Approaches, Wm. C. Brown Company Publishers,
Petty, Richard E., Caccioppo, John T. and Schumann, David (1983), "Central
and Peripheral Routes to Advertising Effectiveness: The Moderating Role of
Involvement," Journal of Communications Research, 10 (September), 135-146.
Putrevu, Sanjay and Lord, Kenneth R. (1994), "Comparative and
Noncomparative Advertising: Attitudinal Effects Under Cognitive and
Affective Involvement Conditions," Journal of Advertising, 23 (June), 77-90.
Rice, Ronald E. and Atkin, Charles K. eds.(1989), Public Communication
Campaigns, 2nd Edition. London: Sage Publications, Inc.
Rice, Ronald E. and Paisley, William J. eds. (1981), Public Communication
Campaigns. London: Sage Publications, Inc.
Rosser, Connie, Flora, June A., Chafee, Steven H., and Farquhar, John
W.(1990), "Using research to predict learning from a PR campaign," Public
Relations Review, vol. 16 (Summer).
Salmon, Charles T.(1989), "Campaigns for Social "Improvement": An Overview
of Values, Rationales, and Impacts," from Information Campaigns: Balancing
Social Values and Social Change. London: Sage Publications.
Salmon, Charles T. ed. (1989), Information Campaigns: Balancing Social
Values and Social Change. London: Sage Publications.
Schiffman, Leon G. & Kanuk, Leslie Lazar (1999), Consumer Behavior, Seventh
Edition, Prentice Hall, New Jersey.
Scott, W.A. (1969), Attitude Change Through Reward of Verbal Behavior,
Journal of Abnormal and Social Pscyhology, 55, 72-75.
Straubhaar, Joeseph and LaRose, Robert (2002), Media Now: Communications
Media in the Information Age, 3rd Edition. Belmont, CA: Wadsworth, Thomson
Solomon, Douglas S. (1981), "A Social Marketing Perspective on Campaigns,"
from Public Communication Campaigns. London: Sage Publications, Inc.
Tate, Cassandra (1999), Cigarette Wars: The Triumph of "The Little White
Slaver." New York: Oxford Free Press.
van Dijk, Teun A. (1991), "The interdisciplinary study of news as
discourse," A Handbook of Qualitative Methodologies for Mass Communications
Research. London and New York: Routledge.
van Dijk, Teun A. (1985), Handbook of Discourse Analysis, Volume 4. London:
VanLeuven, James K. (1990), "Body of knowledge: tomorrow's research
agenda," Public Relations Review, vol. 16 (Summer).
Wilcox, Dennis L., Ault, Phillip H., Agee, Warren K. and Cameron, Glen T.
eds. (2000), Public Relations Strategies and Tactics, 6th Edition. New
York: Addison Wesley Longman, Inc.
Zaichowsky, Judith Lynne (1994), "The Personal Involvement Inventory:
Reduction, Revision, and Application to Advertising," Journal of
Advertising, vol. 23, issue 4.
Zaichowsky, Judith L. (1985), "Measuring the Involvement Construct,"
Journal of Communication Research, 12 (December), 341-352.
Table 1: Demographic Characteristics of the Population Studied
Racial or Ethnic Background-
Table 2: Internal Consistency (Alpha) Coefficients and Factor Loadings for
Factors No. of Alpha Factor
Factor 1: Article Manipulation Check 9 .91
Factor 2: Manipulation Check for Fear 4 .85
Factor 3: Manipulation Check for Humor 3 .90
Factor 4: Attitude Toward PSA 7 .86
This advertisement is offensive. ( r ) .66
This advertisement is believable. .78
This advertisement is useful. .85
This advertisement is informative. .77
This advertisement is clear. .58
This advertisement is likable. .77
This advertisement is convincing. .82
Factors Continued No. of Alpha Factor
Factor 5: Relevance of PSA 9 .93
Of no concern to me. ( r ) .86
Means a lot to me. .81
Worthless ( r ) .77
Irrelevant ( r ) .76
Insignificant ( r ) .80
Doesn't matter to me. ( r ) .83
Factor 6: Involvement with PSA 3 .86
Unexciting ( r ) .87
Factor 7: Credibility of PSA 7 .85
The claims in the ad are true. .86
I trust the claims in the ad. .88
I think the ad is sincere. .73
I think the ad is dishonest. ( r ) .70
I think the information in this advertisement is credible. .82
I think the information in this advertisement is exaggerated. ( r ) .54
I think the information in this advertisement is unbelievable. ( r ) .63
Factor 8: Quality/Validity of PSA Message Claims 7 .86
Unpersuasive ( r ) .75
Unbelievable ( r ) .75
Unconvincing ( r ) .81
Bad argument ( r ) .80
Factors Continued No. of Alpha Factor
Factor 9: Attitude Toward TM 9 .96
Negative ( r ) .91
Unpleasant ( r ) .82
Disagreeable ( r ) .91
Worthless ( r ) .88
Dislike a lot ( r ) .89
Factor 10: Effectiveness of TM 4 .85
I think TM is significant in reaching youth smokers. .81
I think organizations like TM are not influential to youth smokers. ( r ) .84
I believe that TM is powerful. .79
I believe TM is ineffective. ( r ) .85
( r ) = reverse coded.
Table 3: One-way ANOVA in Influence Effects of Third Party Endorsement Types
Attitude toward Message
Table 4: Means and Standard Deviation for Third Party Endorsements of PSAs.
Attitude to PSA
H=humor appeal, HO= humor appeal and opposing third party endorsement, HS=
humor appeal and supporting third party endorsement. (Same for F=fear appeal).
Table 5: One-way ANOVA in Influence Effects of Social Marketing Appeal Types
Attitude toward Message
Attitude toward Organization
Articles Used as Third Party Endorsements
Anti-Tobacco Takes On Big Tobacco
By Stanley Wills
New York Times
New York, April 8 As part of "Corporate Tobacco Amnesty Day," four
Minnesota students made an overture to the New York-based tobacco company
Philip Morris while a live two-way satellite feed showed the over 300
Minnesota youth who were waiting for an answer.
Amnesty Day is part of a special campaign created by the Minnesota-based
anti-tobacco movement called Target Market (TM). The event was supposed to
give tobacco companies a chance to apologize for all of the things they
However, there was never actually a meeting with Philip Morris in New York.
The four Minnesota students, Julie Embry, Willie Dodd, Stacy McKenna and
Carrie Hiller, did make it to the lobby area though and spoke to a
receptionist who they said was polite, but unhelpful.
Meanwhile, the crowd of Minnesota youth who had gathered at the State
capitol in St. Paul, Minn. chanted and held signs mocking Marlboro, the
leading cigarette of Philip Morris. Student leaders Anna Lee and Darren
Miles rallied the crowd shouting, "We're waiting."
The live satellite feed broadcast the foursome going into the company's
headquarters and eventually being escorted outside by a security guard. The
youth then stood outside the corporate offices and offered tobacco
employees forgiveness if they admitted guilt and promised never to target
teens again. For employees who wanted to remain anonymous, a confessional
box was provided.
In addition, the crowd had sent with the New York group with over 10,000
postcards from students who were thanking Philip Morris for helping them
get hooked on smoking.
This was the latest in a series of activities sponsored by TM an
organization funded by the American Legacy Foundation (ALF) that was
created by part of a $6.1 billion settlement from the tobacco companies in
"We got $6.1 billion -- $435 million of that was set aside for tobacco
prevention," said Andy Berndt, a TM activist. "It was put into an
endowment, and we got $25 million of that."
The irony has not been lost on TM members.
"I think it's a great use of their [the tobacco companies'] money," said
Diana Harvey, the group's spokesperson. "To educate teens about the
manipulation with the end goal of stopping the increasing tide of the teen
Target Market hopes to drastically reduce teen smoking and that approaches
like Amnesty Day will get results. In a report released today, TM stated
that its campaign had lowered the teen smoking rate in Minnesota by 20
percent in the last year.
Anti-Tobacco Takes On Big Tobacco
By Stanley Wills
New York Times
New York, April 8 Making good on its month-old threat, Philip Morris
today sued the American Legacy Foundation (ALF) and Minnesota-based
anti-tobacco organization Target Market (TM) for "vilifying" tobacco
companies in its anti-smoking ads and is seeking court "guidelines" for
In today's suit filed in Superior Court in St. Paul, Minn., Philip Morris
claimed "many" of the organization's ads "have included personal attacks on
companies and individuals."
Philip Morris has alleged the organization's ads violate the 1998 Master
Settlement Agreement between tobacco makers and 46 state attorneys general.
The agreement prohibits personal attacks and "vilification" of tobacco
companies or their executives in anti-smoking advertising campaigns.
The agreement also required tobacco makers to pay $6.1 billion for the
creation of the ALF, using tobacco moneys to fund prevention campaigns. Of
that, $25 million was designated for use by TM.
Philip Morris executives however, said they aren't seeking punitive damages
and instead want the court to rule that the ads violate the agreement, and
that state attorneys general will use the decision as a guide to developing
In a statement today, TM spokesperson Diana Harvey called Philip Morris's
suit "meritless" and accused the tobacco maker of trying to crush its
advertising campaign because of its truthfulness. The ads "have not
vilified or personally attacked any person or any tobacco company," she said.
"We need to communicate positive messages about why youth should not
smoke," a spokesman for Philip Morris told a New York Times reporter last
week. "It's wrong to attack our company."
Philip Morris's "positive messages" however are also being questioned as
the American Legacy Foundation demanded that Philip Morris pull its youth
anti-smoking ads, citing that research shows the leading tobacco company's
$100 million campaign is not working and undermines the ALF's anti-tobacco
Philip Morris is currently running anti-smoking ads of their own aimed at
teens tagged "Think. Don't Smoke." However, a study in the American
Journal of Public Health found that exposure to the teen-targeted Philip
Morris ads decreased anti-smoking attitudes among 12- to 17-year olds and
increased the likelihood that non-smokers in this age group would say they
intended to smoke in the future after viewing the ads.
Philip Morris maintains that it does not air any ad unless 90 percent or
more of children and parents interviewed by the company clearly identify a
"don't smoke" message.
Effectiveness Measure Created
To what extent do the following statements characterize the organization
Target Market? Strongly Disagree Strongly Agree
I think Target Market is significant
in reaching youth smokers. 1 2 3 4 5
I think organizations like this are not
influential to youth smokers. 1 2 3 4 5
I feel that Target Market is rightfully
attacking tobacco companies. 1 2 3 4 5
I feel that Target Market is wrongfully
slandering tobacco companies. 1 2 3 4 5
I believe that Target Market is powerful. 1 2 3 4 5
I believe Target Market is ineffective. 1 2 3 4 5
 In the 1995 public relations expose Toxic Sludge is Good for You,
Warren Buffet, once the largest shareholder of R.J. Reynolds stock,
detailed the link between public relations and cigarette smoking:. The
affiliation began during the 1920's when cigarette companies turned to the
public relations industry to help "sell" cigarettes to society. During
this time, cigarettes were fashionable for men, but not for women. In
addition, the public relations profession was trying to solidify itself as
an important support-building business. Thus, public relations pioneer,
Edward Bernays, was hired by American Tobacco Company, the makers of Lucky
Strikes, to help "sell" women on cigarettes.
Throughout the 1929 New York City Easter Parade, Bernays had cunningly
placed many "prominent debutantes" holding cigarettes as "Torches of
Freedom," (Tate, 1999) to promote women as smokers. Although it was noted
that women had been smoking long before Bernays staged the event, it had
been taboo for women to do so in public. The unpopularity of female smoking
had also made it impossible for advertisers to target them, and so a large
untapped market existed. In essence, all Bernays did was get media coverage
of the event, which helped to alter or relax the social stigma attached to
women smoking. Thus, women had an "endorsement" to smoke.
 The term "Big Tobacco" was actually removed from both articles as it
was determined to be too strong of a manipulator and appeared as biased
coverage on the part of journalists.
 One participant did not answer regarding gender.
 One participant did not answer regarding racial or ethnic background.
 One participant did not answer regarding age.
 One participant did not answer whether they had smoker in the past or
were currently a smoker.