Full Terms & Conditions of access and use can be found at
Download by: [22.214.171.124] Date: 17 January 2017, At: 10:48
Journal of Advertising
ISSN: 0091-3367 (Print) 1557-7805 (Online) Journal homepage: http://www.tandfonline.com/loi/ujoa20
Why Narrative Ads Work: An Integrated Process
Eunjin (Anna) Kim, S. Ratneshwar & Esther Thorson
To cite this article: Eunjin (Anna) Kim, S. Ratneshwar & Esther Thorson (2017): Why Narrative
Ads Work: An Integrated Process Explanation, Journal of Advertising
To link to this article: http://dx.doi.org/10.1080/00913367.2016.1268984
Published online: 17 Jan 2017.
Submit your article to this journal
View related articles
View Crossmark data
Why Narrative Ads Work: An Integrated Process Explanation
Eunjin (Anna) Kim
Southern Methodist University, Dallas, Texas, USA
University of Missouri, Columbia, Missouri, USA
Michigan State University, Lansing, Michigan, USA
We propose and test an integrated process explanation for why
narrative ads in general are more persuasive than non-narrative
ads. A study involving a random sample of 25 narrative and 25
non-narrative TV commercials and 484 nonstudent research
participants conﬁrmed that a set of four process variables—
emotive response, ad hedonic value, ad credibility, and perceived
goal facilitation—collectively mediate the positive effects of
narrative (versus non-narrative) ads on attitude toward the ad
) and brand attitudes. The results also supported moderated
mediation processes, with type of featured product and product
involvement as important moderators. Implications for theory as
well as practice are discussed.
Scholars have argued that the human brain is hardwired to
process narratives because they are a fundamental way of
organizing the events of our worlds (Fisher 1989). Accord-
ingly, researchers have demonstrated that narrative ads are
generally more persuasive than non-narrative ads, with the for-
mer generating more positive feelings, more positive cognitive
responses, fewer negative cognitive responses, and more
favorable ad and brand attitudes (e.g., Chang 2009a, 2009b;
Escalas 2004a, 2004b; Escalas, Moore, and Britton 2004;
McQueen et al. 2011).
Many different explanations have
been offered for why narrative ads outperform non-narrative
ads. Deighton, Romer, and McQueen (1989) suggested that
dramatic narrative ads, which draw the viewer into an ad,
enhance persuasion through empathetic processing. Green and
Brock (2000) proposed a different mechanism, called
“transportation,” which refers to a pleasurable, ﬂowlike state
in the viewer. Green and Brock posited that transportation
enhances narrative persuasion by fostering fewer negative
thoughts and more positive affective responses. De Graaf et al.
(2012) emphasized viewer identiﬁcation with the characters in
an ad, and Escalas (2004b) suggested that identiﬁcation can
promote a “self-brand connection,” that is, assimilation of an
advertised brand into a viewer’s self-concept. Escalas (2004b)
also proposed that mental simulation plays an important role
in narratives by enabling viewers to rehearse the events in an
ad and visualize using the advertised product/brand.
Notwithstanding these theoretical advances, signiﬁcant
gaps exist in the literature. First, we currently have a frag-
mented state of knowledge for why narrative ads work, with
different conceptual explanations that involve overlapping the-
oretical constructs. Second, as we discuss in detail later, the lit-
erature displays substantial overlaps and redundancies among
key theoretical constructs. Third, some of the proffered
explanations may hold only in the context of the speciﬁc stim-
uli used in the authors’ studies. For example, in Green and
Brock’s (2000) work, narratives were presented in a textual
format, whereas Escalas (2004b) and Deighton, Romer, and
McQueen (1989) used imagery-rich stimuli involving ﬁlms
and storyboards. But imagery is often processed differently
from text (Childers and Houston 1984), and Lien and Chen
(2013) have shown that transportation mediates the effects of
verbal narrative ads but not visual narrative ads (but see
Murphy et al. 2013). Finally, prior empirical studies have
mainly relied on laboratory experiments (an exception is Dei-
ghton, Romer, and McQueen 1989). Although experimental
methods are important for demonstrating causality between
Address correspondence to Eunjin (Anna) Kim, Temerlin Advertis-
ing Institute, Southern Methodist University, 3300 Dyer Street, 239B
Umphrey Lee Center, Dallas, TX 75275. E-mail:firstname.lastname@example.org
Eunjin (Anna) Kim, (PhD, University of Missouri) is an assistant
professor of advertising, Temerlin Advertising Institute, Meadows
School of Arts, Southern Methodist University.
S. Ratneshwar (PhD, Vanderbilt University) is a professor emeri-
tus of marketing, Robert J. Trulaske Sr. College of Business, Univer-
sity of Missouri.
Esther Thorson (PhD, University of Minnesota) is a professor,
College of Communication Art and Sciences, Michigan State University.
Color versions of one or more of the ﬁgures in the article can be
found online at www.tandfonline.com/ujoa.
Journal of Advertising, 0(0), 1–14
Copyright !2017, American Academy of Advertising
ISSN: 0091-3367 print / 1557-7805 online
variables, the stimuli used in experiments are usually impover-
ished representations of real ads, lacking considerably in eco-
logical validity. It is not clear which of the insights generated
from laboratory experiments would actually apply to real-
world narrative ads, such as those in actual TV commercials.
In sum, the literature on narrative advertising has resulted
in important insights. But when taken as a whole, it does not
provide a compelling theoretical explanation for the persua-
sive power of narrative ads. Given the current state of knowl-
edge, the present research takes an important ﬁrst step toward
an overarching theory of narrative advertising. Speciﬁcally,
we posit and test an integrated process explanation, one which
is empirically veriﬁed in the context of real-world ads. Our
process framework builds on prior literature, with four key
variables, namely, emotive response, ad hedonic value, ad
credibility, and perceived goal facilitation. The framework
also proposes that the type of featured product in an ad and
product involvement are important moderating variables,
either attenuating or enhancing the effects of the four process
variables on attitude toward the ad (A
) and brand attitudes.
We test and validate the process framework with a quasi-
experiment involving a large sample of actual TV
THE CONCEPT OF NARRATIVE ADVERTISING
Narrative ads—also called story ads (Padgett and Allen
1997), transformational ads (Puto and Wells 1984), drama ads
(Deighton, Romer, and McQueen 1989), and slice-of-life ads
(Mick 1987)—communicate through a storylike format
(Escalas 1998), while non-narrative ads typically communi-
cate via argument and explanation. A story is characterized by
central themes, protagonists’ goals, actions taken to fulﬁll
those goals, and the outcomes of those actions (Schank and
Berman 2002). Padgett and Allen (1997) deﬁned a story ad as
one involving “actors with motives, an event sequence, a
setting that has physical, social, and temporal components” (p.
53). Similarly, Escalas (1998) suggested that a narrative ad is
characterized by two structural components: (1) chronology,
in other words, a particular sequence of events, and (2)
causality, or deﬁned relationships between story elements and
what causes things to happen.
Synthesizing these ideas from prior literature, we deﬁne a
narrative ad in structural terms as an ad that tells a story with
the following necessary elements: who, what, when, where,
why, how, and chronology. “Who” refers to the presence of
main actors, characters, or protagonists in an ad. “What” refers
to actions taken by the actors, the outcomes of those actions,
and what ultimately happens to the actors. “When” and
“where” refer to the situation, setting, or context for the
actions. “Why” refers to the goals or motives (explicitly stated
or not) underlying the actors’ actions. “How” is a portrayal of
the actions involving the actors. Finally, chronology refers to
the sequence of events in the ad and the time frame over which
they occur. The elements of who, what, when, where, why, and
how (“5 Ws and an H”) parallel the rhetorical elements of the
“dramatistic pentad” proposed by Burke (1945). Chronology
is included because of its centrality in the work of scholars
such as Escalas (1998).
MAJOR THEMES IN THE PRIOR LITERATURE ON
Narrative Ads and Empathetic Processing
Prior research has shown that narratives with characters, plots,
and dramatic structures can engage viewers in ads and evoke
empathetic processing in the form of positive affect (Deighton,
Romer, and McQueen 1989). It has also been shown that narrative
(versus non-narrative) ads are likely to produce more positive cog-
nitive responses and fewer negative cognitive responses (e.g.,
Chang 2009a, 2009b; Escalas 2004b). Empathetic processing is
characterized by factors such as whether a viewer feels drawn into
an ad, whether the actions depicted in an ad seem authentic, and
whether an ad portrays feelings and events to which a viewer
could relate. Empathetic processing is enhanced when a viewer
feels so immersed in a story that he or she is “transported” into it
(Green and Brock 2000). Empathetic processing of narrative ads
is also boosted when the viewer is able to mentally simulate the
events portrayed in a story (Escalas 2012). Both of these themes
are discussed in more detail in the material that follows.
Narrative Ads and Transportation
Transportation can be thought of as a synthesis of attention,
imagery, and a pleasurable, ﬂowlikestate,whereonelosestrack
of time and awareness of surrounding events (Green and Brock
2000; Tellegen and Atkinson 1974; for other perspectives, see
Laer et al. 2014). The presence of main characters and story plot
elements in narrative ads makes it more likely that such ads will
cause transportation in the viewer, relative to non-narrative ads
(Murphy et al. 2013). Transportation enhances narrative persua-
sion by fostering fewer negative thoughts, stronger and more posi-
tive affective responses, and a greater sense of verisimilitude of
the events portrayed in the ad (Green and Brock 2000). For exam-
ple, in the context of health communications, researchers have
shown that messages conducive to transportation are likely to
enhance processing of new and/or difﬁcult information (Kreuter
et al. 2007), reduce resistance to change (Green 2006), and pro-
duce desired shifts in attitudes and behaviors (McQueen et al.
2011; Murphy et al. 2013). Further, the imagery aspects of trans-
portation are linked to the fact that narratives can engender mental
simulation of the portrayed events (Escalas 2007).
Narrative Ads and Mental Simulation
Mental simulation is a constructive cognitive act, involving
imitative mental representations of the events in the original
2E.(A.) KIM ET AL.
stimulus ad as well as the generationofhypotheticalscenariosthat
resemble those events (Chang 2013; Escalas 2004a). Narrative
ads that involve vivid images and details are the ones that are most
likely to prompt mental simulation in the viewer (e.g., Escalas
2004a; Petrova and Cialdini 2008). Because mental simulation
enables the viewer to rehearse the events in an ad, it makes it easier
for viewers to visualize themselves using the advertised product/
brand and beneﬁting from it (Escalas 2004a, 2004b; Phillips,
Olson, and Baumgartner 1995). Further, because the imagery pro-
cesses implicit in mental simulation often require substantial cog-
nitive resources, fewer cognitiveresourcesareavailablefor
counterarguments to ad claims (Adaval and Wyer 1988; Escalas
2004a). Mental simulation promoted by narratives also has affec-
tive consequences because of ﬂuency effects (Chang 2013). Spe-
ciﬁcally, mental simulation makes it relatively easy for the viewer
to process narratives, resulting inpositiveaffectthatislikelyto
transfer over to ad and product/brand evaluations (Chang 2013).
Narrative Ads and Identiﬁcation With Ad Characters
Identiﬁcation is a process wherein a viewer takes on the
perspective of the main character in a narrative ad and per-
ceives the events in the narrative from that character’s vantage
point (Busselle and Bilandzic 2008). In so doing, the viewer is
likely to experience vicariously the main character’s thoughts,
feelings, attitudes, and actions (Slater and Rouner 2002).
Because narrative (versus non-narrative) ads are more likely
to have central characters, situations, and story lines to which
viewers can relate, they are more conducive to identiﬁcation
(de Graaf et al. 2012). Identiﬁcation fosters self-brand connec-
tions; and because self-brand connections involve a person’s
ego, they make it difﬁcult for a viewer to counterargue in
response to an ad’s claims (Escalas 2004b). Also, because nar-
rative ads usually portray the main character in a positive light,
the vicarious transfer of that character’s experiences, thoughts,
and feelings, which are integral to the process of identiﬁcation,
are likely to engender positive affect in the viewer (Deighton,
Romer, and McQueen 1989; Puto and Wells 1984).
CONCEPTUAL FRAMEWORK AND HYPOTHESES
As discussed in the previous section, the prior literature has
yielded signiﬁcant insights on the topic of narrative advertis-
ing. Nonetheless, the themes and theoretical constructs in this
literature indicate signiﬁcant overlaps and redundancies. For
example, empathetic processing is highly related to, and possi-
bly a consequence of, both transportation and mental simula-
tion. Similarly, the imagery aspects of transportation
correspond fairly directly to the construct of mental simula-
tion. Furthermore, mental simulation of the story lines and
events in a narrative ad likely fosters viewer identiﬁcation
with the main characters in the ad. And, as discussed earlier,
the process of identiﬁcation, in turn, bears on the issue of
Accordingly, in developing our conceptual framework, we
build on the insights of the prior literature but deliberately
eschew some of the preexisting theoretical constructs and ter-
minology. Our framework includes four key explanatory vari-
ables (see Figure 1). These variables were carefully chosen so
as to avoid conceptual redundancy; in the empirical section of
this article, we provide solid evidence of their independent sta-
tus. Speciﬁcally, we propose that narrative ads with greater
emotive response, ad hedonic value, ad credibility, and per-
ceived goal facilitation will result in more favorable A
brand attitudes. Each construct is discussed in the sections that
follow and then linked to the effectiveness of narrative ads.
Emotive response refers to the extent to which a viewer
feels emotionally involved in an ad. Prior research has shown
that narrative ads produce strong affective responses by engag-
ing viewers with stories involving characters and situations
(e.g., Aaker, Stayman, and Hagerty 1986; Escalas, Moore, and
Britton 2004; Deighton, Romer, and McQueen 1989). Further,
Laer et al. (2014) have argued that identiﬁable characters and
plots in narrative ads enable viewers to vicariously experience
FIG. 1. Hypothesized model.
WHY NARRATIVE ADS WORK 3
the main characters’ thoughts and feelings, thereby transport-
ing viewers into the ad stories. Therefore, we propose that nar-
rative ads in general will produce greater emotive response
than non-narrative ads and that greater emotive response in
turn will engender more positive A
Ad Hedonic Value
Ad hedonic value refers to the extent to which a viewer
ﬁnds the content of an ad to be pleasurable and entertaining.
Building hedonic value into ads has become vital for capturing
viewer attention (Olney, Holbrook, and Batra 1991; Teixeira,
Picard, and Kaliouby 2014). Further, Batra and Ahtola (1991)
have shown that the viewing pleasure elicited by an ad’s
hedonic elements is an important determinant of A
. In this
context, prior studies suggest that ads with stories can be
inherently engaging, entertaining, and enjoyable, and dramatic
narrative ads typically score better on these aspects than ads
with rational arguments (Deighton, Romer, and McQueen
1989; Escalas 2012).
From a process standpoint, we suggest the viewing plea-
sure engendered by hedonic elements of narrative ads
should have beneﬁcial effects on A
, both directly and
indirectly (Ayleworth and MacKenzie 1998; Escalas 2004a;
Friestad and Thorson 1993). First, the pleasure experienced
by the viewer can transfer directly in terms of a positive
evaluation of the ad, in other words, a positivity bias in
terms of a classic halo effect (Dean 1999). Second, Bless
et al. (1990) have argued that, compared to people in a
neutral or negative state of mind, people who are in a plea-
surable state of mind are more likely to focus on the posi-
tive aspects of an ad, such as the positive consequences of
product consumption, as people are inherently motivated to
maintain a pleasurable state(seealsoPham,Geuens,and
De Pelsmacker 2013). Note that an important difference
between ad hedonic value and emotive response is that, by
deﬁnition, the former is always positively valenced,
whereas the latter refers to the strength and not the valence
of the affective responses induced by an ad in a viewer.
Ad credibility refers to the extent to which a viewer
believes that the information in an ad is truthful.
It has been
shown that ad credibility has a positive impact on consumers’
cognitive and affective responses and their behavioral inten-
tions (Drake and Ritchie 2007; Obermiller, Spangenberg, and
MacLachlan 2005; Yoo and MacInnis 2005). In the context of
narrative advertising, Dal Cin, Zanna, and Fong (2004) and
Slater and Rouner (2002) have shown that viewers are less
likely to engage in the type of critical thinking needed for gen-
erating counterarguments to narrative ads when compared to
non-narrative ads. Fewer counterarguments in turn should lead
to perceptions of greater ad credibility in the case of narrative
(versus non-narrative) ads.
Researchers have advanced several reasons for why narra-
tive (versus non-narrative) ads may prompt fewer counterargu-
ments and decreased skepticism toward the ad. First, narrative
ads typically depict the life experiences of other individuals. It
is relatively difﬁcult to discount the lived experiences of
others—even when such experiences are portrayed in an ad
(Dal Cin, Zanna, and Fong 2004; Slater 2002). Consequently,
viewers may be less likely to challenge the content of narrative
ads. Put differently, the inherent structure of narrative ads,
involving protagonists and stories but not explicit message
arguments, lends itself well for representations that have great
verisimilitude (Avery and Ferraro 2000; Dal Cin, Zanna, and
Fong 2004). Second, it likely that identiﬁcation with the pro-
tagonists in a narrative ad blocks one’s ability to criticize the
ad’s content (Dal Cin, Zanna, and Fong 2004; de Graaf et al.
2012; Green 2006). Third, Deighton, Romer, and McQueen
(1989) have argued that the dramatization typical of many nar-
rative ads leads to greater viewer immersion in an ad, which in
turn results in fewer counterarguments. Finally, Escalas
FIG. 2. Final model.
4E.(A.) KIM ET AL.
(2004a) has argued that narrative transportation depletes the
cognitive resources that are necessary for critical thinking on
the part of the viewer. Consistent with Escalas’s (2004a) view,
Green and Brock (2000) have proposed that comprehending
narratives demands both cognitive and emotional resources.
Devoting the resources necessary for processing and compre-
hending narratives is likely to impair viewers’ ability and/or
motivation to question the veracity of the stories contained in
All in all, based on the aforementioned arguments, we pro-
pose that, in general, narrative ads will be perceived as more
credible than non-narrative ads. Further, in view of the strong
relationship between ad credibility and ad evaluations (Petty
and Cacioppo 1986; Yoo and MacInnis 2005), we predict that
narrative (versus non-narrative) ads will lead to more favor-
. Formally, we propose the following hypotheses for
the three variables discussed thus far:
H1: Narrative (versus non-narrative) ads will result in greater emo-
tive response, ad hedonic value, and ad credibility, and more posi-
H2: Emotive response, ad hedonic value, and ad credibility will
mediate the positive effects of narrative (versus non-narrative) ads
Relationship Between A
and Brand Attitudes
It is well established that A
is a precursor of brand atti-
tudes and mediates the effects of ad content on brand attitudes
(e.g., Batra 1986; Mackenzie and Lutz 1989). Therefore, we
expect emotive response, ad hedonic value, and ad credibility
to have positive effects on brand attitudes and these effects to
be mediated by A
H3: Narrative (versus non-narrative) ads will result in more posi-
tive brand attitudes.
will mediate the positive effects of emotive response, ad
hedonic value, and ad credibility on brand attitudes.
Perceived Goal Facilitation
Perceived goal facilitation refers to the extent to which a
viewer of an ad believes that the advertised product/brand will
facilitate various consumption goals. For present purposes,
goals can be broadly classiﬁed as “being” goals and “doing”
goals (Belk 1988). “Being” goals refer to a person’s identity
projects and can be accomplished via a product/brand’s self-
enriching beneﬁts, which validate or signal a sense of identity
(Park, Eisingerich, and Park 2013). “Doing” goals refer to a
person’s intentions of engaging in speciﬁc consumption activi-
ties and can be accomplished through a product/brand’s self-
enabling beneﬁts, in other words, the instrumentality of a
product/brand for carrying out the intended consumption activ-
ity (Park, Eisingerich, and Park 2013; Richins 2013).
Narrative ads can be an effective way of communicating a
product/brand’s self-enriching and self-enabling beneﬁts for
two reasons. First, ads that tell stories can grab viewers’ atten-
tion and thereby effectively communicate the advertised prod-
uct/brand’s beneﬁts (Escalas 1998). Second, narrative ads
typically portray consumption experiences with vivid imagery
and concrete details (Escalas 2004a, 2007). Vivid images and
concrete details are known to trigger mental simulation in the
viewer, thereby making the ad more personally relevant and
memorable (Escalas 2007). Mental simulation also helps the
viewer appreciate the manner in which the advertised product/
brand can serve as a means to achieving his or her goals. In
contrast, because of their expository nature, non-narrative ads
tend to be more abstract in the way they link the advertised
brand to goal fulﬁllment (Green 2006; Padgett and Allen
1997). In other words, the structure of the narrative ads is
inherently more conducive than that of non-narrative ads for
consumers to perceive how their goals might be fulﬁlled via
consumption of the advertised product/brand. The positive
effects of narrative (versus non-narrative) ads on perceived
goal facilitation in turn should lead to more positive brand atti-
tudes. Note that unlike the effects of the other three explana-
tory variables (emotive response, ad hedonic value, and ad
credibility), we posit a direct path from perceived goal facilita-
tion to brand attitudes and not a path going through A
is because the former variables pertain to elements of an ad,
whereas perceived goal facilitation relates to the manner in
which the advertised product/brand can help viewers achieve
various consumption goals.
H5: Narrative (versus non-narrative) ads will result in greater per-
ceived goal facilitation.
H6: Perceived goal facilitation will mediate the positive effect of
narrative (versus non-narrative) ads on brand attitudes.
Moderating Role of Type of Featured Product
We now consider important distinctions in the types of
products that can be featured in advertising. We draw here on
Holbrook and Hirschman’s (1982) classiﬁcation scheme,
where products can be of two types: (1) primarily utilitarian
products, which focus on functional beneﬁts (e.g., baking
soda, washing machines, and over-the-counter [OTC] medica-
tions), and (2) primarily hedonic products, whose consumption
produces pleasure and enjoyment (e.g., chocolates, perfumes,
and video games). Utilitarian products are typically consumed
for cognitively oriented or “rational” beneﬁts, whereas
hedonic products are consumed mainly for sensory gratiﬁca-
tion. Narrative ads that feature hedonic rather than utilitarian
products are more likely to (1) enhance emotional involvement
via imagery processing and mental simulation of consumption
WHY NARRATIVE ADS WORK 5
situations (Escalas 2004a) and (2) evoke affective responses
among viewers (Holbrook and Hirschman 1982; Pham,
Geuens, and De Pelsmacker 2013). Therefore, although narra-
tive ads in general should produce greater emotive response
than non-narrative ads (hypothesis 1), this effect should be
enhanced when the featured product in an ad is hedonic (ver-
sus utilitarian). We do not predict a similar moderating effect
for type of featured product on the effects of narrative ads on
ad credibility and perceived goal facilitation, because these
two process variables are primarily cognitive in nature. We
also do not hypothesize this moderating effect in the case of
ad hedonic value, because the enhancing effects of narrative
(versus non-narrative) ads on this variable are due to the story-
telling that is characteristic of narrative ads; the type of fea-
tured product (e.g., candy versus OTC medicines) is unlikely
to affect the beneﬁcial effects of storytelling per se.
H7: The positive effect of narrative (versus non-narrative) ads on
emotive response will be enhanced when the type of featured prod-
uct is hedonic rather than utilitarian.
Moderating Role of Product Involvement
We now consider the moderating role of product involve-
ment. Product involvement can be conceptualized as the per-
ceived relevance of a product category to an individual
consumer, based on his or her inherent values, needs, and
interests (Zaichkowsky 1985). We expect product involvement
to moderate the relationships between (1) the mediator’s emo-
tive response and ad credibility and A
and (2) the mediator-
perceived goal facilitation and brand attitudes. It is only when
product involvement is high that viewers will care sufﬁciently
about an ad for their emotive responses and perceptions of ad
credibility to impact their A
. Conversely, when product
involvement is low, A
is likely to be driven by superﬁcial
processing rather than deep emotional responses or thoughtful
cognitive responses (Petty and Cacioppo 1986). Analogously,
we expect that perceived goal facilitation will make a differ-
ence to brand attitudes only in high product involvement situa-
tions. When product involvement is low, the viewer is
unlikely to care sufﬁciently about goal fulﬁllment for a narra-
tive ad to actually impact brand attitudes. Note that we do not
posit any moderators for the explanatory role of ad hedonic
value, because hedonic enjoyment of an ad and its consequent
impact on A
are unlikely to require much involvement or
effort (Waterman 1993).
H8: The mediating effect of emotive response for the positive
effect of narrative (versus non-narrative) ads on A
enhanced for high product involvement.
H9: The mediating effect of ad credibility for the positive effect of
narrative (versus non-narrative) ads on A
will be enhanced for
high product involvement.
H10: The mediating effect of perceived goal facilitation for the
positive effect of narrative (versus non-narrative) ads on brand atti-
tudes will be enhanced for high product involvement.
Overview. The study employed 50 randomly selected TV
commercials (25 narrative and 25 non-narrative), which had
aired in the United States in 2015. Each study participant
watched one commercial out of the 50 and then responded to
the dependent measures. Participants (ND530) were recruited
through Amazon Mechanical Turk and received monetary
compensation. All participants were at least 18 years old, U.S.
residents, and native speakers of English; 63% were female.
Participants were widely distributed in age, with 15% in the
18 to 24 age group, 58%, 25 to 44, and 27%, 45 to 64. Almost
all had completed high school, and 49.5% had college degrees.
Stimuli. The 50 TV commercials were sampled from all the
commercials aired by CBS between 5:30 p.m. and 11:30 p.m.
over a two-week period (February 20 to March 6, 2015). Note
that we recorded all TV programs including commercials that
aired during those two weeks to sample the 50 TV commercials.
CBS was selected because it was the top-rated network among
the four major broadcast TV networks, both for total viewers and
adults ages 18 to 49, in January–February 2015 (Nielson). A total
of 312 unique commercials aired during the aforementioned
two weeks on CBS, and two judges classiﬁed each of them as
either narrative or non-narrative, using the criteria discussed ear-
lier. Commercials were classiﬁed as narrative if they told a story
with the necessary elements of who, what, when, where, why,
how, and chronology. The remaining commercials were
classiﬁed as non-narrative. The two judges agreed 100% on the
classiﬁcation of the ﬁnal set of 25 narrative and 25 non-narrative
The classiﬁcation process resulted in 64 narrative and 248
non-narrative commercials. The 25 narrative and 25 non-nar-
rative commercials were then randomly selected from the two
groups of commercials. Next, these commercials were scruti-
nized to detect any that were highly targeted, for example,
because of gender or products such as prescription drugs. Such
commercials were eliminated from the sample because of their
highly targeted nature. This process resulted in the elimination
of three narrative commercials and six non-narrative commer-
cials. The eliminated commercials were replaced with new
commercials that were selected at random from the same
groups of commercials as the original sets. The commercials
in the ﬁnal sample ranged in length from 15 seconds to two
minutes. Breaking this down, 18 commercials of each type
were 30 seconds long. Two narrative and four non-narrative
commercials were 15 seconds long; ﬁve narrative and two
non-narrative commercials were one minute long; and one non
narrative commercial was two minutes long. The 25 narrative
6E.(A.) KIM ET AL.
commercials included a variety of brands for consumer prod-
ucts and services, such as Ally Bank, Cesar dog food, Delta,
Farmers Insurance, Kleenex, Kohler, Lexus, M&Ms, U.S.
Cellular, Wendy’s, and Eggo. Similarly, the 25 non-narrative
commercials included brands such as Advil PM, Applebee’s,
Boeing, Cadillac, Coke, H&R Block, Ofﬁce Depot, Purina
Go-Cat, T-Mobile, and Twizzlers.
Procedure. The study was conducted online with Qual-
trics software. The stimulus commercials were ﬁrst down-
loaded from YouTube and saved into a private YouTube
account. Next, the commercials were cleaned using HTML
programming so as to disable embedded ads and additional
video suggestions. Links to the commercials were then placed
on the appropriate page of the Qualtrics survey. Participants
were randomly assigned to either the narrative or the non-nar-
rative condition, and then one of 25 commercials within that
condition was randomly assigned to the participants, that is,
each person saw only one commercial. Each commercial was
seen by 10 to 13 participants. Participants were informed that
they were participating in an “ad response study,” and that the
purpose of the study was to better understand how people
respond to certain ads. They were also told that there were no
right or wrong answers and we would appreciate frank
answers. Participants were instructed to complete the study
without taking a break and to do the study only on a laptop or
desktop computer. Participants were asked to watch the com-
mercial as they would watch any other ad and to watch it on
After watching the commercial, participants responded to
scale items for various dependent measures in the following
, brand attitudes, perceived goal facilitation, emotive
response, ad credibility, ad hedonic value, manipulation check
for narrative (versus non-narrative) ads, type of featured prod-
uct, and product involvement. Note that A
and brand atti-
tudes were measured prior to the other variables to minimize
the possibility of support for the hypotheses due to halo effects
in answering the scales. Note also that our procedure of obtain-
ing attitudinal measures before the process variables in the
framework is consistent with classic papers in persuasion,
such as Petty and Cacioppo (1984). Order of presentation of
items for each measure was randomized. Further, to ensure
that participants were reading the scale items for the measures
carefully, an attention-check question was included among the
dependent measures. A total of 24 participants answered this
attention-check question incorrectly and hence were elimi-
nated from the data. In addition, 22 other participants were
eliminated due to insincere responses, such as a uniform
response pattern on all measures. Elimination of participants
in this manner left 484 participants in the ﬁnal data set.
Toward the end of the survey, participants were administered
a suspicion probe to check on hypotheses guessing with the
question, “In your opinion, what was the objective of this
research study? If you have an idea of the research hypothesis,
please describe.” Scrutiny of the data for this suspicion probe
question conﬁrmed that no participants were able to guess any
of the hypotheses.
and brand attitudes were assessed with
9-point semantic differential scales. Emotive response, ad
credibility, and perceived goal facilitation were measured with
9-point Likert scales. Ad hedonic value was indexed by
summing the responses to 16 Yes/No dichotomous items, with
Yes responses scored as 1 (scale range 0 to 16).
(aD.98) and brand attitudes (aD.98) were assessed
with ﬁve items each (e.g., Negative/Positive, Unfavorable/
Favorable), taken from Mackenzie and Lutz (1989). For emo-
tive response (aD.75), one of the ﬁve items, “The ad affected
me emotionally,” was taken from Green and Brock (2000).
The other four items (e.g., “The ad hooked me in terms of my
feelings”) were created by the authors. The three items (aD
.87) for ad credibility (e.g., “This ad is generally truthful”)
were taken from Obermiller, Spangenberg, and MacLachlan
(2005). The six items (aD.90) for perceived goal facilitation
(e.g., “I would be able to express who I wish to be”; “I would
be able to accomplish some of my personal projects”) were
adapted from Park, Eisingerich, and Park (2013) and Richins
(2013). Regarding ad hedonic value, previous studies suggest
that a variety of elements, such as humor, fantasy, arousal, and
fun, acting individually or in conjunction, can contribute to the
hedonic enjoyment of an ad (Holbrook and Hirschman 1982;
Hung 2014; Olney, Holbrook, and Batra 1991). Therefore, to
effectively assess ad hedonic value, it is important to account
for the many elements that can contribute to it. Accordingly,
ad hedonic value was measured by an index consisting of 16
items. The items (e.g., “This ad is not fun/fun; not entertain-
ing/entertaining; not fantasy-ﬁlled/fantasy-ﬁlled”) were taken
from sources such as Holbrook and Hirschman (1982), Olney,
Holbrook, and Batra (1991), and Voss, Spangenberg, and
Grohmann (2003). For assessing whether the featured product
in the ad was perceived as hedonic (versus utilitarian), 3-item
semantic bipolar (1 to 9) scales were used for each construct.
The items were taken from Voss, Spangenberg, and Grohmann
(2003; e.g., hedonic: “The product/brand shown in the com-
mercial is not enjoyable/enjoyable”; utilitarian: “The product/
brand shown in the commercial is necessary/not necessary”).
An index measure was created for hedonic (versus utilitarian)
type of product by dividing the summed score of the hedonic
items (aD.84) by the summed score for the utilitarian items
(aD.85). The four items (aD.91) for product involvement
(e.g., “I am very interested in anything related to this product”)
were taken from Stokburger-Sauer, Ratneshwar, and Sen
Regarding the manipulation check for narrative (versus
non-narrative) commercials, ﬁve items (9-point Likert scales)
were used, from sources such as Escalas (1998), Deighton,
Romer, and McQueen (1989), and Padgett and Allen (1997).
Speciﬁcally, these items (aD.90) measured the degree to
which participants perceived the commercial in terms of “the
commercial tells a story,” “the commercial shows the main
WHY NARRATIVE ADS WORK 7
actors or characters in a story,” “the commercial shows how a
series of events unfolded in a story format,” “the commercial
shows when and where things happened in a story,” and “the
commercial shows why things happened in a story.”
Overview of data analyses. All data analyses were done
after aggregating participants’ responses across the 50 com-
mercials. Table 1 shows descriptive statistics and correlations
for the study measures. Psychometric properties of the mea-
surement model were assessed with conﬁrmatory factory anal-
yses (CFA), using EQS 6.2 software. Subsequently, the data
were subjected to mixed-model GLM analysis and tests of
mediation using bootstrapping procedures (Hayes 2013). Ad
type was coded as follows: non-narrative ad (1) and narrative
Manipulation check for narrative (versus non-narrative)
ads. Participants who watched narrative (versus non-narra-
tive) ads scored signiﬁcantly higher on the measure for narra-
tive ads (M
D6.00 versus M
482) D191.9, p<.001).
Measurement model. To assess convergent and discrimi-
nant validity of the variables, CFA was conducted. The CFA
model was estimated by the elliptical reweighted least square
(ERLS) procedure of the EQS program (see Bentler 1995).
The adequacy of model ﬁt was evaluated by examining
chi-square statistics (x
), the comparative ﬁt index (CFI), the
Bentler-Bonett normed ﬁt index (NFI), and root mean square
error of approximation (RMSEA). The results met the criteria
suggested by Bentler (1995), with a good ﬁt of the model to
the data: x
(252) D1155.7, CFI D.95, NFI D.95, RMSEA D
.076 (90% conﬁdence interval [CI] of .071 to .081). All factor
loadings were signiﬁcant (ps<.01), and the standardized
factor loadings for all items exceeded the minimum level of
.50. Construct reliabilities (CR) and average variance
extracted (AVE) of each construct were calculated and found
to exceed the minimum criteria of 0.70 and 0.50, respectively.
These results hence provide evidence of adequate convergent
validity. Discriminant validity was examined with Fornell and
Larcker’s (1981) criterion of checking whether the AVE for
each construct exceeds the square of the pairwise correlations
between constructs. All AVE estimates were greater than the
squared interconstruct correlations, thus conﬁrming discrimi-
Tests of hypotheses. We ﬁrst conducted one-way analyses
of variance (ANOVAs) to check whether narrative (versus
non-narrative) ads had effects on the two moderating varia-
bles, type of featured product (hedonic versus utilitarian) and
product involvement. The ANOVA results showed that narra-
tive (versus non-narrative) ads did not have an effect on type
of featured product (MD1.14 versus 1.09; F(1, 482) <1).
However, narrative (versus non-narrative) ads did have a sig-
niﬁcant effect on product involvement (MD4.16 versus 5.24;
F(1, 482) D31.60, p<.001, h
For testing the main effects of type of ad on the dependent
variables in the hypothesized model (hypothesis 1, hypothesis
3, and hypothesis 5), the data were analyzed by a mixed-model
multivariate analysis of variance (MANOVA), with narrative
(versus non-narrative) ads as a between-subjects variable and
emotive response, ad hedonic value, ad credibility, perceived
goal facilitation, A
, and brand attitudes as six within-subject,
dependent variables. In addition, in view of the signiﬁcant
effect of narrative (versus non-narrative) ads on product
involvement, we controlled for the effects of product involve-
ment by including it as a covariate in the MANOVA.
The MANOVA results showed a large main effect of narra-
tive (versus non-narrative) ads (F(1, 481) D533.18, p<.001,
D.53). Inspection of the means for the dependent variables
conﬁrmed, as predicted, that narrative (versus non-narrative)
commercials yielded signiﬁcantly higher scores on all six vari-
ables (see Table 2 for covariate-adjusted means and standard
errors). Speciﬁcally, narrative (versus non-narrative) ads
resulted in more emotive response (hypothesis 1: M
5.41 versus M
D3.00; p<.001), more ad hedonic
Descriptive Statistics and Correlations Between Measures
Variable M SD 12345
1. Emotive response 4.28 1.97
2. Ad hedonic value 10.09 3.94 .42
3. Ad credibility 5.73 1.63 .42 .36
4. Perceived goal facilitation 4.71 1.95 .23 .36 .27
6.51 1.58 .46 .56 .60 .29
6. Brand attitudes 6.28 1.54 .26 .45 .51 .26 .65
Note. The theoretical scale range is 1 to 9 for all variables, except
ad hedonic value (0 to 16); for more details of the measures, see text.
Dependent Measures as a Function of Narrative versus
Emotive response 3.00 (.101) 5.41 (.102)
Ad hedonic value 7.92 (.205) 11.93 (.206)
Ad credibility 4.86 (.090) 6.48 (.090)
Perceived goal facilitation 4.32 (.115) 5.13 (.116)
5.32 (.070) 7.60 (.070)
Brand attitudes 5.44 (.082) 7.08 (.083)
Notes. The numbers in the columns represent covariate adjusted
means after controlling for the effects of product involvement, with
standard errors in parentheses.
***Indicates that the means in the two columns are signiﬁcantly
8E.(A.) KIM ET AL.
value (hypothesis 1: M
D11.93 versus M
7.92; p<.001), more ad credibility (hypothesis 1: M
6.48 versus M
D4.86; p<.001), more perceived
goal facilitation (hypothesis 5: M
D4.32; p<.001), more positive A
D7.60 versus M
D5.32; p<.001), and
more positive brand attitudes (hypothesis 3: M
D5.44; p<.001). This main effect for
narrative (versus non-narrative) ads was qualiﬁed by an inter-
action of this independent variable with the within-subject fac-
tor (Wilks’ λD.76, F(5, 477) D30.99, p<.001, h
However, this interaction does not pose a threat to the results
supporting the hypotheses, since the mean values for all six
dependent measures were consistently greater for narrative ads
than non-narrative ads. The statistically signiﬁcant interaction
merely indicates that narrative (versus non-narrative) ads
resulted in different effect sizes for the various dependent vari-
ables. Note also that the product involvement covariate was
signiﬁcant in the MANOVA (F(1, 481) D44.29, p<.001,
Tests of mediation effects using bootstrapping procedures. A
parallel multiple-mediator model was estimated for the effects of
narrative (versus non-narrative) ads on A
using the bootstrap-
ping procedure (5,000 samples) of the “PROCESS” macro
(model 4, Hayes 2013). The results failed to support a mediating
role for emotive response in the effect of narrative (versus non-
narrative) ads on A
(hypothesis 2: BD¡.05, 95% CI D¡.19
to .06). But the bootstrap results conﬁrmed that the effect of nar-
rative (versus non-narrative) ads on A
was mediated both by ad
hedonic value (hypothesis 2: BD.41, 95% CI D.28 to .56) and
ad credibility (hypothesis 2: BD.49, 95% CI D.37 to .65). Also,
even with all three mediators in the model, the direct effect of
narrative (versus non-narrative) ads on A
was signiﬁcant (p<
.05), indicating this set of three mediators accounted for partial
rather than full mediation.
Our prediction (hypothesis 4) of A
mediating the effects
of emotive response, ad hedonic value, and ad credibility on
brand attitudes was tested individually for each of these three
explanatory variables, while controlling for the effects of the
other two explanatory variables on brand attitudes (Hayes
2013). The result of this analysis conﬁrmed that A
the effects of emotive response (BD.05, 95% CI D.02 to
.08), ad hedonic value (BD.08, 95% CI D.06 to .10), and ad
credibility (BD.21, 95% CI D.16 to .26). Note also that the
direct effects of all three variables on brand attitudes were still
signiﬁcant in the mediation results (ps<.05), indicating that
accounted for partial rather than full mediation. Finally,
the results also conﬁrmed that the effect of narrative (versus
non-narrative) ads on brand attitudes was additionally medi-
ated by perceived goal facilitation (hypothesis 6: BD.08,
95% CI D.03 to .16). In this case also the direct effect of nar-
rative (versus non-narrative) ads on brand attitudes was signiﬁ-
cant (p<.01), suggesting that perceived goal facilitation
accounted for partial rather than full mediation.
Tests of moderated mediation using bootstrapping
procedures. Hypothesis 7 and hypothesis 8 were tested
simultaneously with the PROCESS macro (model 21, Hayes
2013). As hypothesized, the positive effect of narrative (versus
non-narrative) ads on emotive response was greater when the
type of featured product was hedonic rather than utilitarian
(hypothesis 7: BD.62, p<.01), and the mediating effect of
emotive response for the positive effect of narrative (versus
non-narrative) on A
was greater for high product involve-
ment (hypothesis 8: BD.02, p<.05; see Table 3 for details).
Further, when product involvement was high, emotive
response played a signiﬁcant mediating role in the positive
effect of narrative (versus non-narrative) ads on A
hedonic featured products (BD.31, 95% CI D.13 to .51) and
utilitarian featured products (BD.20, 95% CI D.09 to .36).
However, when product involvement was low, the mediating
effect of emotive response on A
was not signiﬁcant for
hedonic featured products as well as utilitarian featured
products (CI values included zero in both cases).
Hypothesis 9, regarding the moderating role of product
involvement on the ad credibility mediator, was tested with
the PROCESS macro (model 7, Hayes 2013). However, the
hypothesis was not supported. Hypothesis 10, regarding the
moderating role of product involvement on the mediating
effects of perceived goal facilitation on brand attitudes, was
tested with the PROCESS macro (model 14, Hayes 2013). As
shown in Table 3, this hypothesis was supported (hypothesis
10: BD.04, p<.01). As predicted, perceived goal facilitation
played a signiﬁcant mediating role in the positive effect of nar-
rative (versus non-narrative) ads on brand attitudes when prod-
uct involvement was high (BD.15, 95% CI D.07 to .26), but
not when product involvement was low (BD.01, 95% CI D
¡.06 to .08).
Prior studies in advertising have demonstrated the superior-
ity of narrative ads over non-narrative ads, with researchers
offering many different explanations for this phenomenon
(e.g., Chang 2009a, 2009b; Deighton, Romer, and McQueen
1989; Escalas 2004a, 2004b; Murphy et al. 2013). These
explanations have invoked constructs such as empathetic proc-
essing (e.g., Deighton, Romer, and McQueen 1989), transpor-
tation (e.g., Green and Brock 2000), mental simulation
(Escalas 2004a), and identiﬁcation (e.g., de Graaf et al. 2012).
Although many valuable insights have resulted from this liter-
ature, the current state of theory in this area involves substan-
tial conceptual overlaps and redundancies, thereby precluding
a coherent understanding of the phenomenon. Also, most of
the prior studies used laboratory stimuli that are impoverished
representations of real-world commercials, thus reducing the
ecological validity of the ﬁndings. Given this context of prior
research, our research objectives were (1) to develop and test
an integrative process explanation for why and how narrative
WHY NARRATIVE ADS WORK 9
ads, in general, are more persuasive than non-narrative ads and
(2) to test and validate our process explanation with actual TV
The conceptual framework for our research posited four
explanatory process variables, namely, emotive response, ad
hedonic value, ad credibility, and perceived goal facilitation.
Tests of Moderated Mediation Hypotheses
A. Mediator: Emotive Response
Mediator Variable Model (DV DEmotive Response)
Predictors B SE t- Value p- Value
Ad Type (Narrative vs. Non-Narrative) 2.97 0.29 10.33 0.000
Product Type 0.83 0.29 2.85 0.005
Ad Type X Product Type 0.62 0.23 2.75 0.006
D.35, F(3, 480) D85. 74, p<.001
Dependent Variable Model (DV DA
Predictors B SE t- Value p- Value
Emotive Response -0.03 0.06 -0.61 0.539
Ad Type (Narrative vs. Non-Narrative) 2.15 0.12 17.96 0.000
Product Involvement -0.05 0.05 -1.10 0.272
Emotive Response X Product
0.02 0.01 2.10 0.036
D.55, F(4, 479) D147.86, p<.001
Conditional Indirect Effects at Product Type and Product Involvement
Mediator Condition 1 Condition 2 BBoot SE Boot 95% CI
Emotive response Hedonic Product Low Involvement 0.05 0.13 -.25 to .27
High Involvement 0.31 0.10 .13 to .51
Utilitarian Product Low Involvement 0.03 0.09 -.16 to .18
High Involvement 0.20 0.07 .09 to .36
B. Mediator: Perceived Goal Facilitation
Mediator Variable Model (DV DPerceived Goal Facilitation)
Predictors B SE t- Value p- Value
Ad Type (Narrative vs. Non-Narrative) 0.84 0.17 4.84 0.000
D.04, F(1, 482) D23.38, p<.001
Dependent Variable Model (DV DBrand Attitude)
Predictors B SE t- Value p- Value
Perceived Goal Facilitation 0.10 0.03 3.36 0.001
Ad Type (Narrative vs. Non-Narrative) 1.60 0.12 13.65 0.000
Product Involvement 0.08 0.03 2.62 0.009
Perceived Goal Facilitation X Product
0.04 0.01 3.06 0.002
D.36, F(4, 479) D77.61, p<.001
Conditional Indirect Effects at Product Involvement
Mediator Condition BBoot SE Boot 95% CI
Perceived Goal Facilitation Low Involvement 0.01 0.04 -.06 to .08
High Involvement 0.15 0.05 .07 to .26
Notes.Brefers to unstandardized regression coefﬁcient. Bootstrap resamples D5000.
10 E.(A.) KIM ET AL.
Further, two moderating variables, type of featured product
and product involvement, were brought into the framework. It
was argued that narrative (versus non-narrative ads), in gen-
eral, produce more emotive response, more ad hedonic value,
more ad credibility, and more perceived goal facilitation. Fur-
ther, we posited that the effects of emotive response, ad
hedonic value, and ad credibility on brand attitudes would be
mediated by A
, whereas perceived goal facilitation would
directly impact brand attitudes. Hypotheses were also offered
for the two moderating variables in terms of moderated media-
The empirical study, conducted with 25 non-narrative and
25 narrative commercials and a large sample of nonstudent
research participants, provided solid support for the indepen-
dent status of our four explanatory variables (see Results,
Measurement Model). The study also yielded support for
almost all of the ten speciﬁc research hypotheses. The study
conﬁrmed that narrative (versus non-narrative) commercials
result in more emotive response, more ad hedonic value, more
ad credibility, more perceived goal facilitation, more positive
, and more positive brand attitudes. Further, the results sup-
ported the mediating role of ad hedonic value and ad credibil-
ity in terms of the effects of narrative (versus non-narrative)
ads on A
. However, the mediating role of emotive response
in the effects of narrative (versus non-narrative) ads on A
was not supported in the simultaneous mediation test. It is
likely that this failure to ﬁnd support for the mediating role of
emotive response was due to some multicollinearity between
this variable and the other two mediators, ad hedonic value
and ad credibility (Pearson rD.46 with both variables). We
conducted a follow-up analysis, where we examined the medi-
ating role of emotive response as a single mediator. The boot-
strap result conﬁrmed that the effect of narrative (versus non-
narrative) ads on A
was mediated by emotive response (BD
.16, SE D.07; 95% CI D.01 to .31).
The results also supported the hypotheses that the effects of
emotive response, ad hedonic value, and ad credibility on
brand attitudes would be mediated by A
. The results further
conﬁrmed that the effects of narrative (versus non-narrative)
ads on brand attitudes were additionally mediated by per-
ceived goal facilitation. Regarding moderated mediation pro-
cesses, we conﬁrmed that the mediating role of emotive
response on A
was greater when the featured product in the
ad was hedonic rather than utilitarian and product involvement
was high rather than low. We were also able to conﬁrm that
the mediating role of perceived goal facilitation on brand atti-
tudes was enhanced when product involvement was high.
Theoretical and Practical Implications
Researchers previously have proposed individual process
mechanisms, involving potentially redundant theoretical con-
structs, for explaining the superiority of narrative (versus non-
narrative) ads. In contrast, we posited and empirically veriﬁed
that multiple yet conceptually different process variables work
in conjunction to enhance narrative persuasion. Speciﬁcally,
we showed that four process variables, emotive response, ad
hedonic value, ad credibility, and perceived goal facilitation,
play important roles in explaining why narrative ads are more
persuasive than non-narrative ads. Two of these variables,
emotive response and ad hedonic value, are primarily affective
in nature; the other two, ad credibility and perceived goal
facilitation, are primarily cognitive in nature. Prior research
has touched on both the affective and cognitive effects pro-
duced by narrative ads (e.g., Chang 2013; Deighton, Romer,
and McQueen 1989; Escalas 2004a) but has not provided a
systematic framework for organizing the explanatory varia-
bles. Our research provides an important step toward future
theorizing in this area by identifying speciﬁc and conceptually
distinct processes that lead to more narrative persuasion.
Of the four explanatory variables in our framework, emo-
tive response is closely related to the construct of empathetic
processing proposed by Deighton, Romer, and McQueen
(1989), Green and Brock (2000), and others. The construct of
ad hedonic value, which refers to the extent to which the ad
provides pleasure and enjoyment, is related to the ideas
implied in discussions by Deighton, Romer, and McQueen
(1989), Escalas (2012), and others. However, these previous
researchers did not clearly deﬁne the construct and also did
not investigate it empirically. As such, the present research
provides both an explication of this construct as well as empir-
ical veriﬁcation of its value in explaining the persuasiveness
of narrative advertising. Regarding the construct of ad credi-
bility, previously researchers have suggested that one of the
advantages of narrative ads is that they prompt fewer counter-
arguments from viewers (e.g., Chang 2009b; Dal Cin, Zanna,
and Fong 2004; Green and Brock 2000; McQueen et al. 2011).
Researchers have also suggested that narrative ﬁdelity is
important for narrative ads to be effective (Fisher 1989). Our
research builds on these ideas by specifying the construct of
ad credibility and demonstrating its signiﬁcant explanatory
value. Finally, the construct of perceived goal facilitation is
loosely related to the idea of self-brand connections posited by
Escalas (2004b) but is broader in scope. Our conceptualization
of this construct draws on research on branding as well as con-
sumer goals and motivation (e.g., Park, Eisingerich, and Park
2013; Richins 2013). The results conﬁrm the important role of
perceived goal facilitation in explaining the effects of narrative
advertising on brand attitudes.
Our research also provides straightforward yet important
implications for advertising practitioners. In addition to docu-
menting the superiority of narrative ads over non-narrative
ads, our study indicates that in creating narrative ads, adver-
tisers in general should create narrative structures that (1) pro-
duce emotive response in terms of high emotional
involvement while viewing an ad, (2) create a positive hedonic
experience for viewers, (3) appear credible, and (4) communi-
cate how consumers in the target audience can achieve their
WHY NARRATIVE ADS WORK 11
consumption goals by buying and using the advertised prod-
uct/brand. Undoubtedly, it may be difﬁcult to achieve all four
of these objectives with the same narrative ad. Still, our ﬁnd-
ings provide useful guidelines to advertisers for creating effec-
tive narrative ads.
Limitations and Future Research
We believe our research provides important insights on the
topic of narrative advertising. Nonetheless, some limitations
need to be acknowledged; we present them in this section,
along with several suggestions for future research. First, to
emphasize ecological validity, we designed the empirical
study as a quasi-experiment, with a random sample of narra-
tive and non-narrative TV commercials. Given the inclusion
of 25 commercials of each type, spanning many different prod-
uct categories, it is unlikely that alternative explanations exist
for the present ﬁndings. Nonetheless, one limitation of using
actual TV commercials is that one cannot exclude variations
extraneous to the independent variable. Future research may
be able to further validate the present hypotheses with care-
fully controlled laboratory stimuli. Second, we focused on the
medium of TV commercials and did not examine other media.
Future research could examine the applicability of the present
ﬁndings to media such as print or radio. For example, it may
turn out that for narrative ads in print media, when compared
to TV commercials, ad credibility and perceived goal facilita-
tion are more important, whereas emotive response and ad
hedonic value are less important. Third, as in many other aca-
demic studies of advertising, we asked our research partici-
pants to watch the stimulus ads such that ad exposure was
intentional rather than incidental. We did not of course man-
date that participants pay attention to the stimuli. Nonetheless,
future research could conﬁrm the generalizability of the pres-
ent ﬁndings to incidental viewing situations where ads are
embedded within program content. In so doing, researchers
could also examine potential interactions between speciﬁc
types of program content (e.g., sports, sitcoms, reality shows)
and narrative ad characteristics (see also Zheng and Phelps
2012). Fourth, our ﬁndings are based on responses from
Mechanical Turk participants in the United States. Future
investigators may be able to conﬁrm the generalizability of
our ﬁndings to other segments of the U.S. population and also
to people in other countries.
Finally, we focused on scaled variables to examine the
hypothesized process mechanisms. But additional insights
may be possible for these process mechanisms with the use of
thought protocols. Such thought protocols may also be able to
provide insights on issues related to the process mechanisms
by inquiring into topics such as the level of integration of the
advertised product/brand into the narrative, the activation and
use of persuasion knowledge by viewers, and the role of prior
brand attitudes. These types of research problems are just a
few examples of the many interesting issues that surround the
important topic of building a theory of narrative advertising.
1. Note that in some other research areas it has been shown that statis-
tical evidence can be more persuasive than narrative evidence
(e.g., Allen and Preiss 1997; Zebregs et al. 2015). Nonetheless, a
preponderance of advertising studies has shown that a narrative
format is more effective than a non-narrative format.
2. Note that the construct of ad credibility is different from the con-
struct of source credibility (e.g., Chaiken and Maheswaran 1994).
3. A separate set of data analyses on the 36 30-second commercials
yielded results very similar to those reported here for the full set of
50 commercials. More details of the results of the additional analy-
sis are available on request from the authors.
The authors thank Margaret Duffy, Shelly Rodgers, and
Glen Leshner for their suggestions.
Financial support for the research was provided by
American Academy of Advertising and the Robert J. Trulaske
Sr. College of Business.
Aaker, David A., Douglas M. Stayman, and Michael R. Hagerty (1986),
“Warmth in Advertising: Measurement, Impact and Sequence Effects,”
Journal of Consumer Research, 12 (4), 365–81.
Allen, Mike, and Raymond W. Preiss (1997), “Comparing the Persuasiveness
of Narrative and Statistical Evidence Using Meta-Analysis,” Communica-
tion Research Reports, 14 (2), 125–31.
Avery, Rosemary Jane and Rosellina Ferraro (2000), “Verisimilitude or
Advertising? Brand Appearances on Primetime Television,” Journal of
Consumer Affairs, 34 (2), 217–44.
Aylesworth, Andrew B., and Scott B. MacKenzie (1998), “Context Is Key:
The Effect of Program-Induced Mood on Thoughts About the Ad,” Jour-
nal of Advertising, 27 (2), 17–31.
Batra, Rajeev (1986), “Affective Advertising: Role, Processes, and Meas-
urement,” in The Role of Affect in Consumer Behavior: Emerging Theories
and Applications, Robert A. Peterson, Wayne D. Hoyer, and William R.
Wilson, eds., Lexington, MA: Lexington Books, pp. 53–85.
———, and Olli T. Ahtola (1991), “Measuring the Hedonic and Utilitarian
Sources of Consumer Attitudes,” Marketing Letters, 2 (2), 159–70.
Belk, Russel W. (1988), “Possessions and the Extended Self,” Journal of Con-
sumer Research, 15 (2), 139–68.
Bentler, Peter M. (1995), EQS Structural Equations Program Manual, Los
Angeles: Multivariate Software.
Bless, Herbert, Gerd Bohner, Norbert Schwarz, and Fritz Stack (1990), “Mood
and Persuasion: A Cognitive Response Analysis,” Personality and Social
Psychology Bulletin, 16 (2), 331–45.
Burke, Kenneth (1945), A Grammar of Motives, Berkeley: University of Cali-
Busselle, Rick, and Helena Bilandzic (2008), “Fictionality and Perceived Real-
ism in Experiencing Stories: A Model of Narrative Comprehension and
Engagement,” Communication Theory, 18 (2), 255–80.
12 E.(A.) KIM ET AL.
Chaiken, Shelly, and Durairaj Maheswaran (1994), “Heuristic Processing Can
Bias Systematic Processing: Effects of Source Credibility, Argument
Ambiguity, and Task Importance on Attitude Judgment,” Journal of Per-
sonality and Social Psychology, 66 (3), 460–73.
Chang, Chingching (2009a), “Being Hookedby Editorial Content: TheImplications
for Processing Narrative Advertising,” Journal of Advertising,38(1),21–34.
_____ (2009b), “Repetition Variation Strategies for Narrative Advertising,”
Journal of Advertising, 38 (3), 51–65.
_____ (2013), “Imagery Fluency and Narrative Advertising Effects, “Journal
of Advertising, 42 (1), 54–68.
Childers, Terry L., and Michael J. Houston (1984), “Conditions for a Picture-
Superiority Effect on Consumer Memory,” Journal of Consumer Research,
11 (2), 643–54.
Dal Cin, Sonya, Mark P. Zanna, and Geoffrey T. Fong (2004), “Narrative Per-
suasion and Overcoming Resistance,” in Resistance and Persuasion, E.S.
Knowles and J.A. Linn, eds., Mahwah, NJ: Erlbaum, 175–91.
Dean, Dwane H. (1999), “Brand Endorsement, Popularity, and Event Sponsor-
ship as Advertising Cues Affecting Consumer Pre-Purchase Attitudes,”
Journal of Advertising, 28 (3), 1–12.
de Graaf, Anneke, Hans Hoeken, Jos!
e Sanders, and Johannes W.J. Beentjes
(2012), “Identiﬁcation as a Mechanism of Narrative Persuasion,” Commu-
nication Research, 39 (6), 802–23.
Deighton, John, Daniel Romer, and Josh McQueen (1989), “Using Drama to
Persuade,” Journal of Consumer Research, 16 (3), 335–43.
Drake, Peter R., and Robin J.B. Ritchie (2007), “The Defensive Consumer:
Advertising Deception, Defensive Processing, and Distrust,” Journal of
Marketing Research, 44 (1), 114–27.
Escalas, Jennifer Edson (1998), “Advertising Narratives: What Are They and
How Do They Work?,” in Representing Consumers: Voices, Views, and
Visions, B. Stern, ed., New York: Routledge & Kegan Paul, 267–89.
_____ (2004a), “Imagine Yourself in the Product,” Journal of Advertising, 33
_____ (2004b), “Narrative Processing: Building Consumer Connections to
Brands,” Journal of Consumer Psychology, 14 (1/2), 168–80.
_____ (2007), “Self-Referencing and Persuasion: Narrative Transportation versus
Analytical Elaboration,” Journal of Consumer Research,33(4),421–29.
_____ (2012), “Success Stories: How Marketing Managers Can Leverage the
Psychology of Narratives,” in Cracking the Code: How Managers Can
Drive Proﬁts by Leveraging Principles of Consumer Psychology, Steve
Posavac, ed., Armonk, NY: M.E. Sharpe, 181–99.
———, Marian C. Moore, and Julie Edell Britton (2004), “Fishing for Feel-
ings? Hooking Viewers Helps!,” Journal of Consumer Psychology, 14 (1/
Fisher, Walter R. (1989), Human Communication as Narration: Toward a Phi-
losophy of Reason, Value, and Action, Columbia: University of South Car-
Fornell, Claes, and David F. Larcker (1981), “Evaluating Structural Equation
Models with Unobservable Variables and Measurement Error,” Journal of
Marketing Research, 18 (1), 39–50.
Friestad, Marian and Esther Thorson (1993), “Remembering Ads: The Effects
of Encoding Strategies, Retrieval Cues, and Emotional Response,” Jour-
nal of Consumer Psychology, 2 (1), 1–23.
Green, Melanie C. (2006), “Narratives and Cancer Communication,” Journal
of Communication, 56 (S1), S163–83.
———, and Timothy C. Brock (2000), “The Role of Transportation in the Per-
suasiveness of Public Narratives,” Journal of Personality and Social Psy-
chology, 79 (5), 701–21.
Hayes, A.F. (2013), An Introduction to Mediation, Moderation, and Condi-
tional Process Analysis, New York: Guilford Press.
Holbrook, Morris B., and Elizabeth C. Hirschman (1982), “The Experiential
Aspect of Consumption: Consumer Fantasy, Feelings, and Fun,” Journal
of Consumer Research, 9 (2), 132–40.
Hung, Kineta (2014), “Why Celebrity Sells: A Dual Entertainment Path Model
of Brand Endorsement,” Journal of Advertising, 43 (2), 155–66.
Kreuter, Matthew W., Melanie Green, Joseph N. Capella, Michael D. Slater,
Meg E. Wise, Doug Storey, Eddie M. Clark, Daniel J. O’Keefe, Deborah
O. Erwin, Kathleen Holmes, Thomas Houston, and Sabra Woolley (2007),
“Narrative Communication in Cancer Prevention and Control: A Frame-
work to Guide Research and Application,” Annals of Behavioral Medicine,
33 (3), 221–35.
Laer, Tom Van, Ko de Ruyter, Luca M. Visconti, and Martin Wetzels (2014),
“The Extended Transportation-Imagery Model: A Meta-Analysis of the
Antecedents and Consequences of Consumers’ Narrative Transportation,”
Journal of Consumer Research, 40 (5), 797–817.
Lien, Nai Hwa, and Yi Ling Chen (2013), “Narrative Ads: The Effect of Argument
Strength and Story Format,” Journal of Business Research,66(4),516–22.
MacKenzie, Scott B., and Richard J. Lutz (1989), “An Empirical Examination
of the Structural Antecedents of Attitude toward the Ad in an Advertising
Pretesting Context,” Journal of Marketing, 53 (2), 48–65.
McQueen, Amy, Matthew W. Kreuter, Bindu Kalesan, and Kassandra I.
Alcaraz (2011), “Understanding Narrative Effects: The Impact of Breast
Cancer Survivor Stories on Message Processing, Attitudes, and Beliefs
among African American Woman,” Health Psychology, 30 (6), 674–82.
Mick, David Glen (1987), “Toward a Semiotic of Advertising Story
Grammars,” in Marketing and Semiotics: New Directions in the Study of
Signs for Sale, J. Umiker-Sebeok, ed., Berlin: Walter de Gruyter, 249–78.
Murphy, Sheila T., Laurren B. Frank, Joyee S. Chatterjee, and Lourdes Baez-
conde-Garbanti (2013), “Narrative versus Non-Narrative: The Role of
Identiﬁcation, Transportation, and Emotion in Reducing Health Dispar-
ities,” Journal of Communication, 63 (1), 116–37.
Nielson (2015), Nielsen TV Ratings, available at http://www.nielsen.com/us/
Obermiller, Carl, Eric Spangenberg, and Douglas L. MacLachlan (2005), “The
Consequence of Disbelief,” Journal of Advertising, 34 (3), 7–77.
Olney, Thomas J., Morris B. Holbrook, and Rajeev Batra (1991), “Consumer
Responses to Advertising: The Effects of Ad Content, Emotions, and Atti-
tude toward the Ad on Viewing Time,” Journal of Consumer Research, 17
Padgett, Dan, and Douglas Allen (1997), “Communicating Experiences: A
Narrative Approach to Creating Service Brand Image,” Journal of Adver-
tising, 26 (4), 49–62.
Park, Choong W., Andreas B. Eisingerich, and Jason Whan Park (2013),
“Attachment-Aversion (AA) Model of Customer–Brand Relationships,”
Journal of Consumer Psychology, 23 (2), 229–48.
Petrova, Petia K., and Robert B. Cialdini (2008), “Evoking the Imagination as
a Strategy of Inﬂuence,” in Handbook of Consumer Psychology, C.P.
Haugtvedt, P.M. Herr, and F.R. Kardes, eds., New York: Psychology Press
Petty, Richard E., and John T. Cacioppo (1984), “The Effects of Involvement on
Responses to Argument Quantity and Quality: Central and Peripheral Routs to
Persuasion,” Journal of Personality and Social Psychology,46(1),69–81.
———, and ——— (1986), Communication and Persuasion: Central and
Peripheral Routes to Attitude Change, New York: Springer-Verlag.
Pham, Michel T., Maggie Geuens, and Patrick De Pelsmacker (2013), “The
Inﬂuence of Ad-evoked Feelings on Brand Evaluations: Empirical General-
izations from Consumer Responses to More Than 1000 TV Commercials,”
International Journal of Research in Marketing,30(4),383–94.
Phillips, Diane M., Jerry C. Olson, and Hans Baumgartner (1995),
“Consumption Visions in Consumer Decision Making,” Advances in Con-
sumer Research, 22 (1), 280–84.
Puto, Christopher P., and William D. Wells (1984), “Informational and Trans-
formational Advertising: The Differential Effects of Time,” Advances in
Consumer Research, 11 (1), 638–43.
Richins, Marsha L. (2013), “When Wanting Is Better Than Having: Material-
ism, Transformation Expectation, and Product-Evoked Emotions in the
Purchase Process,” Journal of Consumer Research, 40 (1), 1–18.
Schank, Roger C., and Tamara R. Berman (2002), “The Perceived Role of Sto-
ries in Knowledge and Action,” in Narrative Impact: Social and Cognitive
WHY NARRATIVE ADS WORK 13
Foundations, Melanie C. Green, Jeffrey J. Strange, and Timothy C. Brock,
eds., Mahwah, NJ: Erlbaum, 287–313.
Slater, Michael D. (2002), “Entertainment Education and the Persuasion
Impact of Narratives,” in Narrative Impact: Social and Cognitive Founda-
tions, Melanie C. Green, Jeffrey J. Strange, and Timothy C. Brock, eds.,
Mahwah, NJ: Erlbaum, 157–81.
———, and Donna Rouner (2002), “Entertainment Education and Elaboration
Likelihood: Understanding the Processing of Narrative Persuasion,” Com-
munication Theory, 12 (2), 173–91.
Stokburger-Sauer, Nicola, S. Ratneshwar, and Sankar Sen (2012), “Drivers of
Consumer-Brand Identiﬁcation,” International Journal of Research in
Marketing, 29 (4), 406–18.
Teixeira, Thales, Rosalind Picard, and Rana el Kaliouby (2014), “Why, When,
and How Much to Entertain Consumers in Advertisements? A Web-Based
Facial Tracking Field Study,” Marketing Science, 33 (6), 809–27.
Tellegen, Auke, and Gilbert Atkinson (1974), “Openness to Absorbing and
Self-Altering Experiences (‘Absorption’), a Trait Related to Hypnotic Sus-
ceptibility,” Journal of Abnormal Psychology, 83 (3), 268–77.
Voss, Kevin E., Eric R. Spangenberg, and Bianca Grohmann (2003),
“Measuring the Hedonic and Utilitarian Dimensions of Consumer
Attitude,” Journal of Marketing Research, 40 (3), 310–20.
Waterman, Alan S. (1993), “Two Conceptions of Happiness: Contrasts of Per-
sonal Expressiveness (Eudaimonia) and Hedonic Enjoyment,” Journal of
Personality and Social Psychology, 64 (4), 678–91.
Yoo, Changjo, and Deborah MacInnis (2005), “The Brand Attitude Formation
Process of Emotional and Informational Ads,” Journal of Business
Research, 58, 1397–1406.
Zaichkowsky, Judith L. (1985), “Measuring the Involvement Construct,” Jour-
nal of Consumer Research, 12 (3), 341–52.
Zebregs, Simon, Bas van denPutte, Peter Neijens, and Anneke de Graaf (2015),
“The Differential Impact of StatisticalandNarrativeEvidenceonBeliefs,Atti-
tude, and Intention: A Meta-Analysis,” Health Communication,30(3),282–89.
Zheng, Lu, and Joseph E. Phelps (2012), “Working toward an Understanding
of Persuasion via Engaging Narrative Advertising: Reﬁning the Transpor-
tation-Imagery Model,” in Advertising Theory, S. Rodgers and Esther
Thorson, eds., New York: Routledge, 255–68.
14 E.(A.) KIM ET AL.