You are on page 1of 15

Advertising Engagement: A Driver of

Message Invoivement on Message Effects

ALEX WANG Engagement plays a contingent role in the effectiveness of advertising processing that
University of corresponds to the message effects created during the process. Such message
Connecticut-Stamford
effects are advertising recall, message involvement, message beiievability. attitude
alex.w3ng@uc0nn.edu
toward the message (AM), and attitude toward the advertisement {AAD)- This study's
objective is to examine whether higher engagement initiated by contextual relevance
increases advertising recall, message involvement, message beiievability, AM. and AAD-
The results have revealed that higher engagement increases advertising recall,
message invoivement, message beiievability, AM. and AAQ. Moreover, message
invoivement mediates the engagement effect on message beiievability, whereas AM
mediates message beiievability on AAD. implications based on the findings
demonstrate the importance of engagement as a driver of message involvement and
a metric for advertising effectiveness.

ON MARCH 21, 2006, the Advertising Research and Jaworski, 1989). Involvement, the motivation
Foundation (ARF) announced its new official def- to process information and an important factor of
inition of engagement. Delivered by ARF Chief advertising effectiveness (Greenwald and Leavitt,
Research Officer Joe Plummer at the annual con- 1984), may occur when a person's perception or
ference, the ARF unveiled a working definition of attention is directed toward an advertisement (An-
engagement: "engagement is turning on a pros- drews, 1988). While advertising researchers tradi-
pect to a brand idea enhanced by the surrounding tionally examine involvement in processing
context" (Elliott, 2006). Consumers increasingly advertisements, a primary antecedent of involve-
insist on being able to consume media when and ment in processing an advertisement is the per-
where they want, on any platform or device. Con- ceived need for relevant information (Bumkrant
sumers have a newfound control over their media and Sawyer, 1983). Thus, engagement initiated by
experiences because the technology and media contextual relevance may be an important driver
industries are obliging them. Consequently, adver- of involvement because engagement may be a
tisers are eager to increase the engagement of precondition to the level of involvement that in-
their advertising because advertisers are increas- fluences the consequence of message effects on
ingly demanding accountability for the money attitude formation (Ephron, 2006; Harvey, 1997).
they spend on their advertising. Specifically in the advertising setting, engage-
ment is defined here as a measure of the contex-
Terms often associated with engagement are
tual relevance in which a brand's messages are
involvement and relevance. A majority of adver-
framed and presented based on its surrounding
tisements do not receive any active processing
context. Tliis definition conforms to the ARF's
(Webb and Ray, 1979), due in part to advertising
working definition of engagement and supports
clutter and consumer involvement in tasks that
the concept of contextual targeting, which is
occupy attention and limit processing (Macliinis

DOt: 10.2501/S0021849906060429 December 2 0 0 6 JOUROflL OfflDUERTISlOGRESEflflGH 3 5 5


ADVERTISING ENGAGEMENT

Engagement is turning on a prospect to a brand idea the message (AM), and attitude toward
the advertisement (/I^D); tincl
enhanced by the surrounding context. 2. what relations among engagement, mes-
sage involvement, message beiievabil-
ity. AM, and A,\o are.
placing highly relevant advertisements ad- to engagement initiated by contextual
This study specifically examines whether
jacent to editorial contents expected to be relevance.
an online advertisement that invites con-
\'isited by target consumers (Harvey, While involvement related to process-
sumers to play a game while they are
2(JU6b). Por example, American Express' ing advertisements has been an impor-
playing an online game (engagement ini-
advertising campaign (Figure 1) features tant construct of advertising research,
tiated by contextual relevance) generates
a tennis star, Andy Roddick, playing ten- empirical in\'estigations of the engage-
better advertising recall, message involve-
nis with d fictitious character and oppo- ment effect and its relationship with in-
ment, message beiievability, A^, and AAD
nent. Pong. Tine advertisement was placed volvement have been limited, A question
than an online advertisement that does
within an online news article about Andy that has not been addressed in the liter-
not invite consumers to play a game. Be-
Roddick to invite readers to play tennis ature remains whether advertisements that
cause message involvement, message be-
with Pong. In this case, the advertisement engage consumers more could elicit
iievability. AM. and AAU are often necessary
highly related to the article (surrounding greater message effects. Thus, the present
conditions for persuasion and branding
context) aims to engage readers (pros- study aims at studying engagement by
(Laczniak, Kempf, and Muehling, 1999),
pects) to play a tennis game sponsored by examining:
showing whether or not higher engage-
American Express (a brand idea). While
ment enhances message effects and under-
readers are involved in reading the article 1. whether higher engagement initiated
standing their relations would be an
about Andy Roddick, the advertisement by contextual relevance increases ad-
important contribution to the understand-
may trigger consumer attention and acti- vertising recall, message involvement,
ing of engagement effects on online ad-
vate processing of the advertisement due message beiievability, attitude toward
vertising effectivL'ness that significantly
affects brand choice.

LITERATURE REVIEWS
Engagement and contextual relevance
As the ad\^ertihing industry grapples with
4: :,:t ' ^ •

ytUSf •lil. I t i l n t toil *!IJ>F«»mMl nlMJH &» I ^ " " -.aTrrr Urc
the profound changes in media such as the
Tennis internet, the concept oi engagement has
emerged as a demand creation paradigm
than the reach or awareness focused para-
digm (Ephron, 2005). Advertisers have iden-
•hi U e i'tniiflhiio^ir iiw l<ab !«'• urn;''Sac Sink la>
tified engagement as a crucial component
that underlies consumer brand choice in re-
sponse to communications. Engagement oc-
curs because of a brand idea or media the
consumer experiences (Barocci, 2006). It is
a critical measurement of when consumers
are strongly engaged in brands, brand mes-
sages, and their surrounding environments.
Recent studies have revealed that there
are various dimensions or drivers of en-
gagement. They range from an exposure-
Figure 1 Engagement Based on Contextual Relevancy based standpoint such as attentiveness to

3 5 6 JOUflflflL OFflOUEHTISIflGRESKH December 2 0 0 6


ADVERTISING ENGAGEMENT

qualitative aspects and relevance (Eph- their relevance-accessibility model and sures of engagement (Harvey, 2006a). Thf
ron, 2U05). Given the complexity of as- found that a message appeal is most likely initial results have shown that the edito-
sessing how consumers interact with to influence consumer choice goal when it rial environment of contextual relevance
advertising, these dimensions or drivers is both relevant and accessible. Their find- can increase attention to advertisements
of engagement are not necessarily sepa- ings have also suggested that when con- and advertising awareness. Because a con-
rate from one another. Measuring time sumers are satisfied wifh the relevant sumer's need for information is shaped
spent with a medium is a fimdamental information they receive for choice-goal de- by the degree of perceived relevance
component of looking at engagement—ti liberation, their choice goal is to buy the (Bumkrant and Sawyer, 1983), this study
consumer must spend time with a me- first acceptable brand they encounter. When argues that contextual relevance between
dium in order to experience its advertis- consumers believe that there are no signif- an advertisement and its surrounding con-
ing. Other possible drivers of engagement icant differences among the brands, they text can initiate higher engagement. An
are surprise, utility/relevancy, and emo- are likely to seek product features or prices advertisement that creates contextual rel-
tional bonding, identified by Joe Plum- that are most relevant to his or her choice evance to its surrounding context may
nier, ARF's Chief Research Officer (Hai"vey, goal. As a result, the degree of information produce a catalyst for increased amount
2006a). relevancy influences how consumers uti-
of message involvement and positive at-
lize sucli information to help formulate their
Wells, Burnett, and Moriarty (1992, p. 387) titude formation.
attitudes.
h.ive suggested, "an unexpected idea can
be one with a twist, an unexpected associ- Leigh (1991) reviewed the effects of mes- Involvement
ation, or ca tchy phrasing." Surprises or nov- sage congruence among multiple stimulus Consumers often focus their involvement
elty in advertisements are in some way modalities to demonstrate the importance on a primary task (e.g., reading news
unexpected and therefore could enhance of congruence to the subject of compari- articles), thus reducing cognitive resources
consumers' engagement in a way that it sons between TV and radio broadcast me- available to involve in secondary informa-
would not if it were predictable. Advertise- dia. His study was concerned with the tion (e.g., advertisements) surrounding the
ments that elicit positive affect could also manner in which stimulus properties of primary task. Because cognitive resources
draw engagement because liking is be- tasks, the processing strategies adopted by available for attending to secondary infor-
1 ieved to be broadly descriptive of positive an individual, and the processing opera- mation are limited, secondary informa-
/4AD (Seamon, Marsh, and Brody, 1984). Rel- tions performed on each task may influ- tion cannot be explicitly recognized (Leigh,
evance can be defined as the fit between ence overall and individual information 1991; Maclnnis and Jaworski, 1989). In
advertising message and the consumer and processing performance. At the core of this other words, memory traces for this infor-
advertising and the media environment processing is the concept of the schema, be- mation are unlikely to be strong enough
(Ephron, 2005). Overall, relevance speaks ing integrally involved in a perceptual cy- to bp retrievable during search of mem-
to the importance of targeting as a factor cle (Neisser, 1976). A schema is that portion ory, which can cause poor advertising re-
of engagement. Advertisements that fea- of the entire perceptual cycle that is inter- call and brand recognition.
ture utility and relevancy appeals speak to nal to the perceiver, modifiable by experi- Greenwald and Leavitt (1984) used psy-
consumers' interests. This dimension, the ence, and specific to what is perceived. This chological theories of attention and levels
focus of this study, could be highly func- processing is relevant to contextual rele- of processing to establish a framework of
tional in driving engagement among tar- vance because information related to a de- audience involvement in advertising. Tliey
get consumers. veloping or active schema is more likely to outlined four hierarchical levels; preatten-
be interpreted correctly and coherently. Eor tion, focal attention, comprehension, and
The contextual relevance between an ad-
example, highly congruent audio and video elaboration. As a higher level of involve-
vertisement and its surrounding context that
stimuli exhibit comparable positive effects ment Is reached, increased capacity is al-
can initiate engagement Is grounded in the
on processing and memory, compared to located to a message source. Lower levels
theory of information relevancy {Baker and
the cases where the audio and video pro- use relatively little capacity and extract
Lutz, 2000). The information that can help
vide weakly related or even different in- information needed to determine whether
shape a consumer's choice goal tends to be
formation (Leigh, 1992). a higher level will be invoked. The higher
t!ie most relevant information to the con-
sumer (Feldman and Lynch, 1988). Baker Research has preliminarily examined levels require greater involvement and re-
and Lutz (2000) have proposed and tested how contextual relevance performs in mea- sult in increasingly durable cognitive and

December 2 0 0 6 JOUfiOfll OFflDUEHTISIOGHESEflRCH 3 5 7


ADVERTISING ENGAGEMENT

attitudinal effects. These levels then serve personal relevance of the advertising stim- the salient cues (Greenwald and Leavitt,
as the basis for the information consum- ulus itself or the attempt to create a sit- 1984).
ers store in memory (Andrews and Shimp, uation in which consumers are encouraged Engagement initiated by contextual rel-
1990). The final step, elaboration, occurs to process the stimulus personally rele- evance, a salient cue, can cause a shift of
when consumers restate messages and vant to them (Andrews and Durvasula, attention to the source of the message and
summarize their reactions to them. 1991; Laczniak and Muehiing, 1993). increase message involvement, motiva-
Research has suggested that involve- By integrating these two perspectives, tion to process information. When con-
ment is best dealt with when it is concep- message involvement represents an indi- sumers are motivated to process secondary
tualized within a particular domain vidual variable that could indicate the information, secondary information be-
(Andrews and Durvasula, 1991; Andrews, amount of arousal or interest tliat is evoked comes increasingly the focus of consumer
Durvasula, and Ahkter, 1990). One do- by advertising messages (Laczniak, Kempf, attention. Consumers at the focal atten-
main that is highly related to engagement and Muehiing, 1999). Thus, perceived rel- tion stage usfc? "modest capacity to focus
is message involvement. Message involve- evance of the messages, enhanced by on one message source, and to decipher
ment is defined here as "a motivational higher engagement based on contextual the message's sensory content into cat-
construct that influences consumers' mo- relevance, directs consumers' direction of egorical codes" (Greenwald and Leavitt,
tivation to process information at the time attention and the intensity of advertising 1984, p. 584). In other words, consumers
of message exposure" {Baker and Lutz, processing. Then it influences the conse- use perceptual and semantic processing
2000, p. 2). Message involvement con- quences of message effects by the associ- to produce word and object category rep-
cerns the message, not the general prod- ation of levels of involvement with an resentations. Consequently, engagement
uct class that is deemed relevant and is orderly series of attitudinal effects (Green- initiated by contextual relevance may also
elaborated upon (Batra and Ray, 1985). wald and Leavitt, 1984). Consumers who increase message believability (Wang,
Moreover, message involvement exists as are more involved in processing advertis- 2006).
an individual state evoked by a particular ing messages may undergo a semantic Using American Express advertisement
message at a particular point of time (Lacz- analysis in which the memorial represen- (Figure 1) as an example, placing an ad-
niak, Muehling, and Grossbart, 1989). tation of the information is accessed from vertisement that features a tennis star,
Ceisi and Olson (1988) have tested sev- memory, which in turn affects subsequent Andy Roddick, and inviting consumers to
eral hypotheses regarding the effects of attitudes. This seems to suggest the rela- play a tennis game within an online news
message involvement on the amount of tionship between engagement and mes- article about Andy Roddick, creates the
comprehension. Their results have re- sage involvement; engagement can be a engagement effect based on contextual
vealed that higher message involvement driver of message involvement. relevance. Once consumers consider this
increases the amount of the attention and contextual relevance as a salient cue, con-
cognitive effort during comprehension of Relationship between involvement sumers may shift their attention to the
advertisements. and engagement advertisement, which then increases con-
Dominant among the operationaliza- Among the four levels of involvement, sumers' message involvement with the
tions of message involvement are those two stages, the characteristics of preatten- advertisement. In this case, the engage-
that consider message involvement to tion and focal attention, are particularly ment effect becomes the driver of mes-
be primarily an attentional (focal atten- important to demonstrate the relationship sage involvement.
tion, direction/intensity) construct and a between engagement and message involve-
personal/situational (message/personal ment. Consumers in the preattention level RESEARCH HYPOTHESES AND
relevance) construct (Laczniak and Mueh- allocate little capacity to pr(.x:essing incom- QUESTION
ling, 1993). When viewed from an atten- ing messages. In this level of audience Because the engagement effect increases
tional perspective, message involvement involvement, sensory buffering and fea- message involvement, it is likely that the
is often concerned with the focus on a ture analysis are the processing criteria engagement effect will increase the like-
particular aspect of the advertisement (Lacz- that consumers may employ. Salient cues lihood of stronger advertising recall.
niak and Muehiing, 1993). When viewed in the unattended information may be Studies have shown that information rele-
from a relevance perspective, message in- detected and cause a shift of attention to
vancy facilitates naming and categorizing
volvement is often concerned with the the source of the message that contains
semantically related information (Fuentes,

358 OF RDiEflTISinG flESEfll CH December 2006


ADVERTISING ENGAGEMENT

Engagement initiated by contextual reievance, a salient relevance in which a brand's messages


are framed and presented in an advertise-
cue, can cause a shift of attention to the source of the metit related to its surrounding context.
Because engagement is to turn on a pros-
message and increase message Involvement, motivation pect to a brand idea in an advertisement
related to its surrounding context, contex-
tual relevance between a primary task
to process information the engagement effect be-
and an online advertisement was manip-
ulated to form the lower and higher en-
comes the driver of message involvement.
gagement condition (Figure 2).
The participants were involved in a pri-
mary task of playing an online game fea-
tured in a website. Next to the online
Carmona, Agis, and Catena, 1994). The Hlc: Higher engagement initiated by game was an online advertisement. The
heightened attention and possible elabo- contextual relevance will gener- same online game was featured in both
ration due to the engagement effect are ate stronger message believabil- higher and lower engagement conditions.
presumed to create cognitive pathways ity than lower engagement. However, in the lower engagement con-
back to the originating message, which dition, the participants were exposed to
then increases the probability of adver- Hid: Higher engagement initiated by
an online advertisement that did not in-
tising recall. Similarly, positive effects contextual relevance will gener-
vite them to play a game. The advertise-
are expected to foster improvement of ate stronger -4M than lower
ment said "Receive Your Free Sony PS3."
evaluations of the advertisement (Leigh, engagement.
The participants in the higher engage-
1994). Hie: Higher engagement initiated by ment condition were exposed to an online
Research has also showed that the pre- contextual relevance will gener- advertisement that invited them to play a
attentive analysis of information can en- ate stronger A,XD than lower game. The advertisement invited the par-
hance liking for it (Janiszewski, 1988) engagement. ticipants to kick the field goal to receive a
because the subconscious analyses can in- freePS3.
fluence preference for a stimulus (Anand, Based on the literature review, engage-
The argument—playing an online game
Holbrook, and Stephens, 1988). As the ment and message involvement are highly
next to an online advertisement that in-
engagement effect increases message in- interrelated constructs. However, no di-
vited the participants to play a game could
volvement, an individual's liking of the rect conclusion has been drawn regarding
attract their attention—was based on higher
message is enhanced because either sub- their relationship and effects on message
engagement initiated by the contextual rel-
conscious or conscious analyses often believability. AM, and A^D- Thus, this study
evance between the advertisement and the
create a feeling of familiarity that Is inter- asks the following research question.
online game. Once encountered, the game
preted as affect or preference for the featured in the advertisement that was as-
RQ: What are the relations among en-
message (Bonanno and Stillings, 1986; Mac- sociated with the surrounding context (the
gagement, message Involvement,
lnnis, Moorman, and Jaworski, 1991). Thus, online game) could initiate the engage-
message believahility. AM, and
this study tests the following hypotheses. ment effect. However, it is not possible to
evaluate this argument with an experimen-
Hla: Higher engagement initiated by tal design that explicitly directs partici-
METHODOLOGY
contextual relevance wiil gener- pants to process advertisements (McQuarrie
Design
ate stronger advertising recall and Mick, 2003). Instead, this study used
One-way experimental design manipulat-
than lower engagement. an experiment representing a realistic ex-
ing two levels of engagement (lower ver-
posure condition in which the participants
Hlb: Higher engagement initiated by sus higher engagement) was employed.
were told to engage in playing the online
contextual relevance will gener- The design of lower versus higher engage-
game to which processing resources were
ate higher message involve- ment condition was based on the study's
applied. In other words, there was no
ment than lower engagement. definition of engagement, the contextual

December 2006 JDIIHOHL OF HDUERTISinG RESEH 3 5 9


^ . . . _p

ADVERTISING ENGAGEMENT

Lower Engagement Condition first page of the booklet informed the


participants that tlie principal imestiga-
tor is interested in their opinions about an
online game. The second page of the book-
let informed the participants of the web
address that manipulated either lower or
higher engagement condition. In eitlier
Receiva Your FHCE' Sony PS3*1 condition, the participants were given
10 minutes to play the online game. After
10 minutes, the participants were asked
to complete a set of questionnaires.

Participants
The participants (N = 239) used in this
study were undergraduate students at a
northeastern university. The participants
were instructed not to discuss this study
with any other students or anyone else
once they were recruited to participate in
Higher Engagement Condition the study. There were 119 participants in
the lower engagement condition and 120
participants in the higher engagement con-
dition. There were 112 male participants
(46.9 percent) and 127 female participants
(53.1 percent) with an averaged age of 20
years old.
Contrary to popular belief that PC games
are for kids, 90 percent of all purchasers
arc over 18 years old (Games-advertising,
2006). Moreover, women are purchasing
just as much video/PC entertainment soft-
ware as men. Because undergraduate stu-
dents are one of the primary target
audiences, this study's samples repre-
sented the populations this study pur-
ported to represent.

Stimuil
This study designed two websites that
manipulated the two experimental condi-
Figure 2 Experimental Conditions
tions (Figure 2). In the lower engagement
condition, the website featured an online
game and an online advertisement that
coaching or mentioning of looking at the and were assigned randomly to one of initiated lower engagement effect. In the
tested advertisement in either condition. the two experimental conditions. AH par- higher engagement condition, the website
The participants were asked to com- ticipants first received an instmction book- featured the same online game and an
plete this study in a computer lahoratory let including a set of questionnaires. The online advertisement that initiated higher

360 December 2006


ADVERTISING ENGAGEMENT

engagement effect based on contextual An advertising recall was labeled as a not ike it/like it {Hailahan, 1999). Tiie
relevance. correct advertising recall, whereas an in- Cronbach's a value for AM {M = 3.12,
This study used two manipulation correct advertising recall, no answer, do SD - 1.48) was 0.93. /^AD was measured
checks to ensure the validity of the study's not know, or do not remember were not by asking the participants to complete
engagement manipulations. The measure labeled as an advertising recall. the sentence, "1 would describe the ad-
reflecting perceived engagement asked the Message invoh ement was measured by vertisement as . . . ," using a 6-item scale
participants to rate the level of perceived asking the participants "how mucli atten- composed of bad/good, unpleasant/
engagement when exposed to the adver- tion you paid to process the advertise- pleasant, low quality/high quality, do not
tisement, where 1 — "not engaged at all" ment," "how engaging it was for you to like it/like it, not desirable/desirable, and
and 7 - "extremely engaged" (Laczniak, process the advertisement," "what was the unfavorable/favorable (Hailahan, 1999),
Kempf, and Muehling, 1999). The partici- overall attention you had with the adver- The Cronbach's a value for A^) {M -
pants in the lower engagement condition tisement," and "how involving it was for 3.29, SD = 1.57) was 0.97.
{M ^ 2.15, SD = 1.35) perceived lower you to process the advertisement" (Lacz- The results indicated that all measures
engagement than the participants in the niak, Kempf, and Muehling, 1999). These were reliable. Consequently, the values of
higher engagement condition {M = 3.23, four items were measured on a 7-point scale the constructs were computed as the mean
SD = 1.86), f (1,155) = 16.4, p < .001, v^ = where 1 == "not at all" and 7 ^ "extremely." of the ratings of the items associated with
0.1. The participants were also asked how The Cronbach's « value for message in- each construct (Table 1). Tliis study mea-
relevant the context of playing an online volvement [M - 2.75, SD - 1.57) was 0.94. sured and used two covariates for data
game was to the advertisement, where Message beiievability was measured by analysis. Participants' gender was asked
1 = "not at all" and 7 = "a lot." The asking the participants to complete the and coded. Among 157 participants who
participants in the lower engagement con- sentence, "to what extent, do you believe noticed the advertisement, there were 70
dition (M = 2.58, SD - 1.45) perceived the messages in the advertisement are ...," male participants (44.6 percent) and 87
lower contextual relevance between the using a 5-item scale composed of not female participants (55.4 percent), indicat-
online game and the advertisement than informtiti\'e/informative, untrustworthy/ ing a fair distribution of gender used in
the participants in the higher engagement trustworthy, inaccurate/accurate, uncon- the data analysis. Participants' age was
condition (M = 3.13, SD - 1.69), f (1,155) - vincing/convincing, and not believable/ also recorded (M = 19.85, SD = 3.22).
4.61, p < .033, 7)^ - 0.03. Thus, the en- believable (Hailahan, 1999; Wang, 2006).
gagement manipulations were successful. The Cronbach's a value for message be- RESULTS
lievability (M - 3.79, SD = 1.5) was 0.95. Engagement effects
Measures AM was measured by asking the par- A chi-squtire test was performed to exam-
A question asked the participants whether ticipants to complete the sentence, "the ine Whether the participants noticed the
they noticed the advertisement while play- messages in the advertisement are ... ," advertisement in the lower versus higher
ing the online game, where 1 = "yes" and using a 5-item scale composed of boring/ engagement condition (Table 2). A second
0 - "no." This measure ensured that the interesting, not attention-getting/attention- chi-square test was performed to test the
study only analyzed the data collected getting, bad/good, not fun/fun, and do relationship between engagement and
from the participants who noticed the ad-
vertisement. It was unlikely that the par-
ticipants could answer questions regarding
the advertisement when they did not even TABLE 1
notice the advertisement. Descriptive Statistics for Measures
Advertising recall was measured by ask-
Measures M SO Reliability
ing the participants to list the advertise-
ment's contents (Cadoppo and Petty, 1981). Message involvement 2.75 1.57 0.94
In other words, the participants needed to Vlessage believability 3.79 1.5 0.95
remember the advertisement to produce a
Attitude tovi/ard the messages {A^ ) 3.12 1.4S 0.93
correct advertising recall. Participants' an-
swers were recorded by two categories. Attitude toward the advertisement 3.29 1.57 0.97

December 200 6 J[ onoRi OF eoo .flTiSlliu Emm 361


ADVERTISING ENGAGEMENT

TABLE 2 percent) generated higher advertising re- Tlie participants in the higher engage-
call than the participants in the lower ment condition (M - 3.11, SD = 1.69)
Engagement and Advertising
engagement condition (43 percent), x^{l) ^ exhibited higher message involvement tlian
Awareness 4.31, p < .038. Thus, Hypothesis Hla was the participants in the lower engagement
Whether noticing supported, condition (M - 2.26, SD - 1.25), f (1, 153) =
the advertisement The multivariate tests based on the 12.935, p < .000, 7]- < 0.078. The partici-
or not? MANCOVA procedure revealed that there pants in the higher engagement condition
was a main effect for the engagement (M - 4.05, SD = 1.41) exhibited stronger
Engagement Yes No message believability than the partici-
effect, Wilks' A = 0.908, f (4, 150) = 3.817,
Higher 90 (75%) 30 (25%) p < .006, Tj^ = 0.092; the mean vectors pants in the lower engagement condition
were not equal and the set of means be- (M = 3.44, SD = 1.54), F(l, 153) - 6.91,
Lower 67 (56%) 52 (44%)
tween conditions (lower versus higher en- p - .009, 7}^ - 0.043. The participants in
gagement) was different. Gender and age the higher engagement condition (M =
were not found to contribute to the model 3.31, SD = 1.56) exhibited stronger A^
advertising recall among 157 participants significantly as covariates. The tests of than [he participants in the lower engage-
who noticed the advertisement (Table 3), between-participants effects based on the ment condition (M = 2.85, SD - 1.34),
The MANCOVA procedure was used with individual univariate tests and means f (1,153) = 4.073, p < .045, T?^ = 0.026. The
message involvement, message believabil- (standard deviations) of key dependent participants in the higher engagement con-
ity. AM, and A^^ as the dependent vari- variables were reported in Tables 4 and 5, dition (M - 3.56, SD = 1.61) exhibited
ables, engagement as the independent respectively. The dependent variables' stronger A,\o than the participants in the
variable, and age and gender as the mean comparisons between lower and lower engagement condition (M = 2.92,
covarlates. higher engagement were depicted in SD - 1.46), F(l, 153) = 6.857, p = .01, TJ^ =
The results based on the first chi-square Figure 3. 0.043.
test revealed that 157 (66 percent) out of
239 participants noticed the advertise-
ment. The participants in the higher en-
gagement condition (75 percent) were more Message Involvement Message Believability
likely to notice the advertisement than Attitude toward the Message Attitude toward the Advertisement
the participants in the lower engagement
4,5 1
condition (56 percent), ,^-^(1) = 9.27, p <
.002. The results based on the second chi-
4- 4.05
square test revealed that the participants
in tbe higher engagement condition (60
3.5- 3.44

3-
TABLE 3
Engagement and Advertising 2.5-
Recall _^ 2.26
2-
Whether recalling
the advertisement
1,5
or not? Lower Engagement Higher Engagement
Engagement Yes No Engagement

?^J^.9^l Figure 3 Mean Comparisons between Lower and Higher


Lower 29jf3%) 38(57%) Engagement

362 JOURfiei Of flOUEflTiSlOB December 2006


ADVERTISING ENGAGEMENT

Mediation effects
-r . r o j . r»4.-- 4- i-ff *-^ To answer the research question asked.
Tests of Between-Participants Effects '
this study examined the relations (Fig-
Source Dependent Variable df F Significant rj^ ure 4) among the engagement effect, mes-
sage involvement, messaee believabilitv.
Corrected model Message involvement 3 6.616 0.000 0.115 ^ *
- AM, and A^u- A series of regression analy-
.'}('.®ssage believability 3 4.244 0.007 0.077 ^^^ ^ ^ ^ ^^^^ ^^ ^^^^ f^^. rnediating effects
AM 3 2.359 0.074 0.044 among the independent and dependent
"A^, 3 3^238 aO24 0^06 ™'*''='' ""'^'^ '"" ''"'"''• ' " ' ' ' • ^'^^^
sage involvement was regressed on the
]ntercept_ Message involvement 1 37-093 0.000 0.195 engagement effect (^ - -0.267, p < .001),
Message believability 1 39.093 0.000 0.204 F(l, 155) = 11.854, p < .001, R~ - 0.071.
Z 1 31^245 0^000 0^17 The r^ults revealed that iower engage-
ment reduced message involvement. Mes-
.^AD 1 H-.^9.? .9:9.99 9:.}:7.^. sage believability was regressed on the
Age Message involvement 1 5.322 0.022 0.034 engagement effect (^ = -0.203, p < .011),
""''"' F(l, 155) = 6.649, p < .011, R^ - 0.041.
Message believability 1 0.092 0.762 0.001 ^
The results revealed that lower engage-
^M ^ h^J^. 9.-3^.^ 9.-.9.9.® ment reduced message believability. Mes-
4^j^ 1 2.179 0.142 0.014 sage believabiiity was regressed on the
^ „.„ „„ engagement effect {fi - -0,078, p = .283)
Gender Message involvement 1 1.344 0.248 0.009 ^ ^ '^ '^
and message involvement (/3 = 0.467, p <
"y.^ssage believability 1 5.504 0.02 0.035 QQQJ^ p^^^ ^54j = 24.818, ;; < ,000, R~ =
^ X ±A72 0.227 0.01 0.244. The results based on the simulta-
'Zo 1 0^599 o!4^O 0^004 neous regression analysis revealed that
the engagement effect was no longer evi-
^".?,^.^?^^."!^. .'y'.^^sage'nyolvenient 1 12.935 0.000 0.078 dent on message believability and higher
Message believability 1 6.91 0.009 9;9:*.3 message involvement enhanced message
AM 1 4^073 0^045 0^026 belieyabilit>^
•"• A3 Baron and Kenny (1986) have sug-
.•^Aci 1 9;,^^.7. .9:9.^ 9,-,9.1? gested, the results of three regression equa-
Error Message involvement 153 tions provided the tests of the linkages of
the mediatins effect. The results revealed
Message believability 153
that there was a perfect mediation be-
.V.M }:^,r. tween message involvement and the en-
fy^^ 153 gagement effect on message believabiiity
because three conditions of mediation held.
Total Message involvement 157
First, the engagement effect affected nies-
beiievability 157 sage, involvement in the first equation.
157 Second, the engagement effect also af-
fected message believability in the second
157
equation. Finally, message involvement af-
,,'.",™,'.y,^.';",?Pf, ^^^. fected message beUevabinty, and the en-
Message believabiiity 156 gagement effect did not aftt-ct message
believabilitv in the third equation.
AM 156 ^ '
AAD was regressed on message believ-
ability {(i - 0.734, p < .000), f ( l , 155) =

December 2 0 0 6 JOURHftL OF HDUEHTiSIHG HESEfiBCH 3 6 3


ADVERTISING ENGAGEMENT

TABLE 5 DiSCUSSION

Means (Standard Deviations) of Key Dependent Measures The evolving tield of interactive and dig-
ital media such as the internet is present-
Lower Higher ing new opportunities to generate higher
Engagement Engagement engagement with online advertisements.
67 90 While limited studies have empirically
examined the engagement effect, this study
M SD M SD Significant Effects F
not only contributes to the literature by
Message involvement 2.26 1.25 3.11 1.69 Engagement 12.94*** examining the definition of engagement
in an interactive medium but also con-
Message believability 3.44 1.54 4.05 1.41 Engagement 6.91**
tributes to the advertising processing re-
2.85 1.34 3.31 1.56 Engagement 4.07* search by investigating the engagement
2.92 1.46 3.56 1.61 Engagement 6.86** effect on advertising recall, message in-
Noir • /' < .05: " p < .07: * " p < .001.
volvement, message believability, A^, and
AAD- First, the results have suggested that
engagement initiated by contextual rele-
vance can be influential (Har\'ey, 2006a).
180.524, p < .000, R- = 0.538. Tlic results the single regression equation (Baron and
The most important finding in this study,
revealed that message believability posi- Kenny, 1986).
however, is that message involvement me-
tively enhanced AAD- -^M was regressed As depicted in Figure 4, two impor- diates the engagement effect on message
on message believability (^ = 0.73, ;) < tant mediators emerged from the exam- believability mediating the effect of AM
.OOO), f(], 155) - 176,436, p < .000, R^ = ined relations among the engagement on AAD- Even though higher engage-
0.532. The results revealed that message effect, message involvement, message be- ment can enliance advertising recall, mes-
bclievabiiity also positively enhiinced AM. lievability, AM, and AAD- While message sage involvement, message believability.
V4AD was regressed on A^\ (fi ^ 0.747, p < involvement mediated the engagement ef- AM, and AAD. the engagement effect in-
.000) aiid message believability (^ - 0.188, fect on message believability, AM medi- fluences message believability by the me-
/' < .000), F(2, 154) - 306.442, p < .000, ated the effect of message believability diation of message involvement. Then
R~ = 0.799. An acceptable mediation held on AAD- In other words, the engagement message believability influences AAD ^y
because the standardized coefficient of effect was a driver of message involve- the mediation of AM- This result is ex-
message believability on A^D was less in ment on message believability, a driver tremely important because it supports the
the multiple regression equation than in of AM on AAr^. study's argument that the engagement
effect is an influential driver of message
involvement. It also identifies a model
' " • • • " • — ; sequence regarding where and how the
1Message .
relations among the engagement effect,
* invoivement i \ /
p = -0.267/ ' '\ P = 0.73/ ''\p-0.747 message involvement, message believabil-
P = 0.467'\ / ity, AM, and AAD operate. Moreover, it
\ / '-« clarifies the highly correlated relation-
Message
Engagement AAD ship between engagement and message
p--0.203 Beitevabiiity p = 0.734
involvement.

-•• Nlediation Effect


Practlcai impiications
irect Effect
Advertisers should be encouraged by the
lediator
results presented here because the results
confirm the importance of engagement in
Figure 4 The Relations among Message Involvement, many aspects of advertising processing.
Message Believability, A^, and /AAD Tlie immediate implication of this study

3 6 4 JOyROeL OF HDyEftTISIflG December 2006


ADVERTISING ENGAGEMENT

is that using contextual relevance to ini- Advertisers can generate the engagement effect by using
tiate the engagement effect not only can
help advertisers have their online adver- internet technologies to reach consumers with highly
tisements actually noticed by consumers
but also can create higher consumer mes- relevant advertisements based on what they read or do
sage involvement. Advertisers can gener-
ate the engagement effect by using internet
(e.g-r a primary task) while placing advertisements
technologies to reach consumers with
highly relevant advertisements based on
adjacent to editorial contents expected to be visited by
what they read or do (e.g., a primary
task) while placing advertisements adja-
target consumers.
cent to eciitorial contents expected to be
visited by target consumers. Moreover,
the same strategy can be used to readi
consumers based on where they navigate.
ican Express used several TV spots dur- messages. Thus, advertisers should strat-
In the same line of reasoning, manip- ing the U.S. Open broadcast to promote egize to enhance advertising engagement
ulating engagement based on contextual the tennis game, "Roddick vs. Pong," and by integrating various dimensions or driv-
relevance should also be implemented in invite tennis fans to go to the website and ers of engagement. Consistent with past
traditional media. In newspaper or mag- play die game. Once tennis fans are mo- research on message effects, message nov-
azine context, engagement based on con- tivated to go to the site and play the elty, utility, and affect are factors that may
textual relevance can be utilized by game, they are surrounded by adverti.sing enhance advertising engagement in addi-
reaching consumers with highly relevant stimulus or product information about tion to engagement initiated by contex-
advertisements based on what they read American Express. This example demon- tual relevance (Harvey, 2006a; Seamon,
while placing advertisements adjacent to strates the use of behavioral engagement Brody, and Kauff, 1983a, 1983b ). Conse-
editorial contents. In fact, the contextual and advertising engagement based on con- quently, enhancing advertising engage-
relevance between an advertisement and textual relevance to achieve advertising ment by integrating various drivers of
a primary task has been employed in and branding e0ectiveness. eng^ement can create complementary ef-
formulating advertorial contents. It is com- By integrating the study's results and fects of these drivers and achieve the great-
mon to see an advertisement that is placed Han.'ey's (1997) expanded ARF model, est advertising and branding effectiveness.
next to a magazine article and looks like engagement should be directly related to
a magazine article, This approach is used advertising exposure, vehicle exposure, Engagement going forward
as a contextual targeting approach by ad- advertising interaction, advertising com- While this study has mainly examined
vertising practitioners (Harvey, 2006a), For munication, advertising attentiveness, re- contextual relevance as the driver of en-
example, airline in-flight magazines of- tention, and advertising persuasion. gagement, the effects of engagement ini-
ten feature reports about travel destina- Ad\'ertising attentiveness such as mes- tiated by other drivers such as affect have
tions to which the airline flies and place sage involvement mediates the engage- not been fully examined and addressed.
their advertisements next to the reports. ment effect on advertising retention and Future research should examine the effec-
The results have also suggested a solu- persuasion such as message believability. tiveness of this driver of engagement on
tion to advertising clutter. Because of too After higher engagement is generated by message effects. Future research can also
many advertisements for the same prod- advertising exposure, vehicle exposure, examine whether the same relations among
uct category fighting for consumers' at- advertising interaction, and advertising the engagement effect initiated by affect,
tention, consumers may tend to avoid communication, message believability in- message involvement, message believabil-
looking at any of them, Tlius, engage- fluences /4AD by die mediation of AM ity. AM, iind /4AU will materialize.
ment based on behavioral engagement and As Ephron (2006) and Harvey (2006b) Although this study takes an important
engagement in,itiatcd by contextual rele- have suggested, engagement does not just step to examine the effects of engage-
vance may help advertisements escape depend on the media engagement, it de- ment, additional research is needed to
the advertising clutter. For example, Amer- pends even more on the creativity and determine the underlying process and

December 2006 M M ] . OF HDIIERTISIIIG ReS[l1l1CII 3 6 5


ADVERTISING ENGAGEMENT

generalizability of the findings. The par- As interest in engagement continues to grow, new
ticipants might try to stay focused on
playing the online game, resisting the ad- research and in-maricet experience wiii increase the
vertising messages even though the mes-
sages were relevant, Future study should understanding of engagement's role in improving
examine whether consumers may react to
the surprise of unexpected event by look- advertising resuits. Findings in engagement research can
ing at an advertisement when finding the
ad\ ertisement in a completely unrelated be used to change the way that advertisers think about
site. Future research should also examine
the robustness of engagement effects in the relations among consumers, advertising messages,
cross-media contexts. For example, adver-
tisers have used TV spots to invite audi- and advertising environments.
ences to go to their websites for voting or
other activities. The comparisons of cross-
media use in terms of engagement effects
warrant future investigations.
As interest in engagement continues to evolution to a new strategy or measure- ment. As advertisers' knowledge about
grow, new research and in-market experi- ment that reflects the complexity of ttxlay's engaging consumers continues to grow,
ence wilt increase the understanding of media choices and consiuner-empowcred advertisers' ability to reach consumers with
engagement's role in improving advertis- media consumption can be laid. The most targeted, engaging messages will lead to
ing results. Findings in engagement re- obvious factor that underpins this re- better communication, and ultimately, bet-
search can be used to change the way that form is the challenge of advertising ter advertising results.
advertisers think about the relations among clutter. In reality, consumers do not
consumers, advertising messages, and ad- usually look for advertisements to pro- ALEX WANQ (Ph.D., University of Texas at Austin) Is an
vertising environments. Future studies cess. Thus, this study used a realistic on- assistant professor of communication sciences at the
should identify and examine a variety of line environment to test engagement University of Connecticut-Stamford. His research fo-
exposure and relationship factors that af- effects initiated by contextual relevance. cuses on integrated marketing communications, Infor-
fect engagement. In the same tine of rea- The results certainly confirm the new mation processing psychology, and website promotion

soning, future studies should examine a definition of engagement offered by the management. Dr. Wang's industry experience in adver-

variety of factors that affect targeting based ARF and demonstrate the importance of tising, public relations, and electronic publishing took

on contextual relevance by examining the engagement as a driver of message in- place in Taipei, Taiwan and in Boslon. His works have
volvement and a metric for advertising been published in the Journal ofAavertising Research.
fit among consumers, advertising mes-
effectiveness. the Journal ol Public Relations Research, the Journal of
sages, and brands. Drivers of engagement
may vary by demographic, product cat- The results have suggested that engage- Marketing Communications, the intemationai Journa/ of

egory, medium, and genres within media, ment should not be regarded as a direct Strategic Communication, the Journal ot Promotion

which should be studied further in the indicator of advertising results. Rather, Managenient. the Journal of Website Promotion, and

advertising engagement is the goal. This elsewhere.


future. Finally, future studies should ex-
plore how creativity placed in various is to say advertising engagement is ex-
advertising environments differs, perhaps plained by the message and surrounding
resulting in the development of different context. Engagement initiated by contex- REFERENCES
creative messages for each medium. tual relevance works as the door atten-
dants of message involvement. They direct ANAND, P., M. B. HOLBROOK, and D. STEPHENS.
CONCLUSION consumers to see and respond to the ad- 'Hie Formation of Affective Judgments: The
By understanding how engagement initi- vertisement. In this case, the key to adiieve Cognitive-Affective Model Versus the Indepen-
ated by contextual relevance affects mes- advertising engagement is to understand dence Hypothesis." Journal of Consumer Re-
sage effects, the groundwork for the how to materialize the drivers of engage- sairch 15, 4 (1988): 386-91.

366 HDUEHTISIOG RESEBRCH December 2 0 0 6


ADVERTISING ENGAGEMENT

ANDREWS, J. C. "Motivation, Ability, and Op- BoNANNO, C. A., and N. A. STILLINGS. "Prefer- s, L. J., E. CARMONA, 1. R AGIS, and A.
portunity to Process Information; Conceptual ence, Familiarity, and Recognition After Re- . "The Role of the Anterior Attention
and Experimental Manipulation Issues." Ad- peated Exposures to Random Geometric Systein in Semantic Processing of Both Foveal
vances in Consumer Research 15 (1988): 219-25. Shapes." American ]ourm\ of Psychology 99, 4 and Parafoveal Words." Journal of Cognitive
(1986): 403-15. Neuroscience 6, 4 (1994): 17-25.
, and S. DuRVA&ULA. "Suggestions for
Manipulating and Measuring Involvement in BURNKRANT, R. E., and A. G. SAWYER. "Effects GAMES-ADVERTISING. "Who's Playing Games?
Advertising Message Content." AdxHinces in Con- of Involvement and Message Content on Infor- Surprise—It's Not Just Who You Think": jURL:
turner Reu'iuch 18 (1991): 194-201. mation Processing Intensity." In Injornmlion Pro- http: //www.games-advertising.com /assets/ Fast
cessing Research in Advertising, R. ]. Harris, ed. 7o20Pacts%20%20About7..20the"/o20Consumer.
, , and S. H. AIIKTER. "A Frame- Hillsdale, NJ: Lawrence Eribaum, 1983. html, 2006 (accessed June 2006).
work for Conceptualizing and Measuring the
Involvement Construct in Advertising Re- CACIOPPO, J. T., and R. E. PETTY. "Social Psy-
GKEENWALD, A. G., and C. LEAVITT. "Audience
search." lourml af Advertising T*. 4 (1990): 27-40. chological Procedures for Cognitive Response
Involvement in Advertising: Four Levels." jour-
.Assessment: The Thought-Listing Technique."
nal of Consumer Research 11, 2 (1984): 581-92.
, and T. A. SHIMP. "Effects of Involve- In Cognitive Responses in Persuasion, R. E. Petty,
ment Argument Strength, and Source Charac- T. Ostrom, and T. C. Brock, eds. Hillsdale, NJ:
teristics on Central and Peripheral Processing Lawrence Eribaum Associates, 1981. HALEAHAN, K. "Content Class as a Contextual
of Advertising." Psx/cholo^if mul Mnrkctiitg 7, 3 Cue in the Cognitive Processing of Publicity
(1990): 195-214. versus Advertising." Journal of Public Relations
CELS], R. L., and J. C. OLSON. "The Role of
Research 11, 4 (1999): 293-320.
Involvement in Attention and Comprehension
BAKEK, W. E., and R. J. LUT/. "An Empirical Processes." Journal of Consumer Research 15, 3
Test of an Updated Relevance Accessibility (1988): 210-24. HARVEY, B. "ARF Engagement Recipe: Surprise,
Model of Advertising Effectiveness." loiinial af Utililty, and Emotion": (URL: http://www.next
Adirrtisin'i 29, 1 (2000): 1-14. ELLIOTT, S. "New Rules of Engagement." Tlie Neiv centurymedia.com/2006/03/arf-engagement-
York Times, March 21, 2006: [URL: http://www. recipe-surprise-utLlity.htmlJ, March 3, 2006a
BARUCCI, Btm. "Advertising Industry "Turned nytimes.com/2006/03/21 /business/media/ (accessed June 2006).

On' by New Measurement Model." The Ad- 21adco.html?ex=1300597200&en=ec7491e84797


vertising Research Foundation: [LJRL: http:// 252e&ei = 5088&partner=rssnyt&emc=rssJ (ac-
-. "Engagement and the ARF Model": [URL:
vvwvi'.thearf.org/dovvnloads/2006-03"21_PR^ cessed June 2006).
http://wv/w.nextcenturymedia.com/2006/03/
Engagement.pdf#seiirch ='Turned %20on%E2% engagement-and-arf-model.htnill, March 14,
80%99%20bY7,.20New%20Measurement%20 EPHRON, E. "Delivering the Message: How Con- 2006b (accessed June 2006).
Model], 2006. sumer Involvement Flows from Magazine Edit
to Advertising." Mediawcek, June 20, 2005-
. "The Expanded ARF Model: Bridge to
BARON, R. M., and D. A. KRNNY. "The
Accountable Advertising Future." Journal of Ad-
Moderator-Mediator Variable Distinction in So- •. "Want Engagement? Make a More En-
vertising Research 37, 1 (1997): 11-20.
cial Psychtiiogical Research: Conceptual, Stra- gaging Ad": [LJRL: http://ephrononmedia.com/
tegic and Statistical Considerations." }oiirnal of article_archive/articleViewerPubIic-asp?articIe
PersoiiiilitJi imd Social Psychohg}/ 51, 6 (1986): 1D = ]48], January 11, 2006 (accessed June JANISZEWSKI, C. "Preconscious Processing Ef-
1173-82. 2006). fects: The independence of Attitude Formation
and Conscious Thought." journul of Consumer
Research 15, 3 (1988): 199-209.
, R., and M. RAY. "HOW Advertising FhLDMAN, J. M., and J. G. LYNCH. "Self-
Works at Contact." In Psychological Processes generated Validity and Other Effects of Mea-
Olid Advertising Effects, L. F. Alwitt and A. A. surement on Belief, Attitude, Intention, and JuDi>, C. M., and D. A. KENNY. "Process Anal-
Mitchell, eds. Hillsdale, Nj: Lawrence Eribaum Behavior." joiirnal of Applied Psi/chology 73, 3 ysis: Estimating Mediation in Evaluation Re-
Associates, 1985. (1988): 421-35. seandi." Evaluatkvi Research 5. 3 (1981); 602-19.

December 2006 ME^l OF [1ESERRCH 3 6 7


ADVERTISING ENGAGEMENT

LACZNIAK, R. N., D. S. KEMPF, and D. D. tions for Research." Journal of Advertising 20, 2 and Cerebral Laterality." Journal of Expci'imental
MUEHLING. "Advertising Message Involve- (1991): 71-75. Psychology: Learning, Memory, and Cognition 9, 3
ment; Tlie Role of Enduring and Situalional (1983a); 544-55.
Factors." Journal of Current hf^ues and Research MACINNIS, D. J., and B. J. JAWORSKI. "Informa-
in Advertiiiing 21, 1 (1999): 51-61. tion Processing from Advertisements: Toward -, and -. "Affective Discrimina-
an Integrative Framework." Journal of Market- tion of Stimuli That Are Not Recognized: Ef-
, and D. D. MUEHLING. "The Relationship ing 53, 4 (1989): 1-23. fects of Delay between Study and Test." Bulletin
between Experimental Manipulations and Tests of the Psychonomic Society 21, 2 (1983b): 187-89.
of Theory in an Advertising Message Involve- , C. MOORMAN, and B. J. JAWORSKI. "En-
ment Context." lourmt of Advertising 22,3 (1993): hancing and Mcasurijig Consumers' Motiva- , R. L. MARSH, and N. BRODY. "Critical
59-74. tion, Opportunity, and Ability to Process Brand Importance of Exposure Duration for Affective
Information from Ads." Journal of Marketing 55, Discrimination of Stimuli That Are Not Recog-
, , and S. GROSSBART. "Manipulat- 4 (1991): 32-53. nized." Journal of Experimental Psychology: Learn-
ing Message bivolvement in Advertising Re- ing. Memory, and Cognition 10, 3 (1984): 465-69.
search." Journal of Advertising 18, 2 (1989): 28-38. McQuARRiE, E. F., and D. G. MICK. "Visual and
Verbal Rhetorical Eiguros under Directed Pro- WANG, A. "When Synergy in Marketing Com-
, J. H. "The Use of Figures of Speech in cessing versus Incidental Exposure to Adver- munication Online Enhances Audience Re-
Pr-iat Ad Headlines." journal of Advertising 23, 2 tising." Journal of Consumer Research 29,1 (2CK)3): sponse: The Effects of Varying Advertising and
(1994): 17-34. 579-87. Product Publicity Messages." Journal of Adver-
Using Research 46, 2 (2006): 160-70.
• -. "Modality Congruence, Multiple Re- NEISSER, U . Cognition and Reality. Principles and
source Theory and Intermedia Broadcast Com- Implications of Cognitive P9[/cJwhgi/. San Fran- WEBB, P , and M. RAY. "Effects of TV Clutter."
paristms; An Elaboration." loiirnal of Advertising cisco. CA: W. H. Freeman, 1976. Journal of Advertising Research 19, 2 (1979): 7-12,
IX 2 (1992): 55-62.
SEAMON, J. G., N. BRODY, and D. KAUIT. "Af- WELLS, W., J. BuRNcrr, and S. MORIARTY. Ad-
. "Information Processing Difference fective Discrimination of Stimuli That Are Not zvrtising: Principles ami Practice. Englewood Cliffis,
among Broadcast Media: Review and Sugges- Recognized: Effects of Shadowing, Masking, NJ: Prt?ntice-Hall, 1992.

3 6 8 JDURDHL DF HDUEHTISIHG December 2006

You might also like