Content uploaded by Martin Eisend
Author content
All content in this area was uploaded by Martin Eisend on Nov 26, 2018
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ujoa20
Download by: [FU Berlin] Date: 20 October 2015, At: 02:12
Journal of Advertising
ISSN: 0091-3367 (Print) 1557-7805 (Online) Journal homepage: http://www.tandfonline.com/loi/ujoa20
Advertising Repetition: A Meta-Analysis on
Effective Frequency in Advertising
Susanne Schmidt & Martin Eisend
To cite this article: Susanne Schmidt & Martin Eisend (2015) Advertising Repetition: A Meta-
Analysis on Effective Frequency in Advertising, Journal of Advertising, 44:4, 415-428, DOI:
10.1080/00913367.2015.1018460
To link to this article: http://dx.doi.org/10.1080/00913367.2015.1018460
View supplementary material
Published online: 25 Mar 2015.
Submit your article to this journal
Article views: 249
View related articles
View Crossmark data
Literature Review Corner
Advertising Repetition: A Meta-Analysis on Effective
Frequency in Advertising
Susanne Schmidt
Technical University Dortmund, Dortmund, Germany
Martin Eisend
European University Viadrina, Frankfurt, Germany
This study uses meta-analytic techniques to examine the
number of exposures that maximize consumer response to an ad.
The results show that in an experimental setting maximum
attitude is reached at approximately ten exposures, while recall
increases linearly and does not level off before the eighth
exposure. The findings are of interest for two opposing schools of
thought in the advertising literature on effective frequency. They
support the repetitionists’ beliefs over the minimalists’ beliefs on
the number of ad exposures needed for maximum consumer
response. The study further investigates whether the repetition
effects depend on contingent factors. Low involvement and
spaced exposures enhance repetition effects on attitude toward
the brand. Embedded advertising and massed exposures enhance
the repetition effects on recall. Repetition effects decay over time
for both attitude toward the brand and recall. The study provides
important implications for researchers by contributing to the
discussion on effective frequency and providing support for the
repetitionists’ view. This view has implications for practitioners
who try to optimize advertising frequency.
Multiple exposures to an advertisement increase consumer
awareness of the advertising message and facilitate consumer
processing of the included information (Vuokko 1997). In this
way, advertising repetition enhances consumers’ attitude and
recall. The importance of advertising repetition effects is
reflected by the vast amount of research conducted in this area
in recent decades. Many studies have researched the general
effects of advertising repetition on various consumer-related
variables and have examined which factors drive or hamper
advertising repetition effects (Pechmann and Stewart 1988).
The variety of studies resulted in divergent findings on the
effects of advertising repetition, thus leaving two main research
gaps in the understanding of the effects of advertising repetition.
First, while there is consensus on the general course of
advertising repetition effects on attitude and recall (i.e., an
inverted U-shaped curve for the advertising repetition effect
on attitude and a logarithmic course of effect for recall), the
number of exposures that maximizes consumer response is still
subject to continuous scientific debate. This is because the
research findings differ considerably and indicate different
exposure levels at which maximum attitude is reached (e.g.,
Kohli, Harich, and Leuthesser 2005; Nordhielm 2002). Tellis
(1997) calls the search for the optimum number of exposures
the quest for the “holy grail of effective frequency” (p. 75) and
distinguishes the literature on this topic into two schools of
thought: the minimalists, who believe that a few (usually one
to three) exposures achieve the maximum response, and the
repetitionists, who argue that repetition is necessary for opti-
mal consumer response.
Second, while a variety of studies examines factors that
strengthen or weaken advertising repetition effects, they
typically investigate only one or two factors at a time. A com-
prehensive and simultaneous analysis of various and interde-
pendent factors does not exist. To date, there is a qualitative
literature synthesis (Pechmann and Stewart 1988) examining
stimuli, respondent, as well as methodological characteristics
that may moderate the effect of advertising exposure on atti-
tude and recall. Reviews and meta-analyses from outside of
the advertising field provide additional factors that advertising
scholars have not yet considered, such as stimulus novelty
(Harrison 1977; Janiszewski, Noel, and Sawyer 2003). Thus, a
thorough quantitative review that takes into account additional
moderating factors benefits both research and practice, as
it shows whether and how the repetition effects depend on
contingent factors.
Address correspondence to Susanne Schmidt, Technical Univer-
sity Dortmund, Martin-Schmeißer-Weg 12, 44227 Dortmund, Ger-
many. E-mail: susanne2.schmidt@tu-dortmund.de
Susanne Schmidt (PhD, European University Viadrina) is a
research assistant, Technical University Dortmund.
Martin Eisend (PhD, Free University Berlin) is a professor of market-
ing, European University Viadrina.
415
Journal of Advertising, 44(4), 415–428
Copyright Ó2015, American Academy of Advertising
ISSN: 0091-3367 print / 1557-7805 online
DOI: 10.1080/00913367.2015.1018460
Downloaded by [FU Berlin] at 02:12 20 October 2015
We try to close these research gaps by conducting a meta-
analysis with 312 effect sizes of experimental studies. By
doing so, this study represents the first quantitative synthesis
of advertising repetition effects. We provide generalizable
conclusions on the number of advertising exposures that maxi-
mizes consumer response and on factors that moderate the
effects of advertising repetition. By integrating advertising
repetition effects of previous experimental studies, we com-
pare findings from forced exposure settings. These settings
minimize confounding effects of respondents’ inattentiveness
or of random exposure.
We structure the remainder of the study as follows. First,
we provide the theoretical foundation for advertising repetition
effects by discussing the two-factor theory and learning the-
ory. We derive hypotheses for several factors that potentially
moderate advertising repetition effects. Next, we explain the
database development and the steps of our analysis. We pres-
ent the results and discuss them in terms of their implications
for theory and practice.
THEORETICAL BACKGROUND
Pechmann and Stewart (1988) define advertising repetition
effects as the “differential effects of each successive advertis-
ing exposure, i.e., [...] the differential effects of a given expo-
sure within a sequence of exposures” (p. 287). This view
includes wear-in effects (i.e., an advertisement has a signifi-
cant positive effect on consumers at a certain level of expo-
sure) and wear-out effects (i.e., an advertisement has no
significant effect on consumers or may even have a negative
effect at a certain level of exposure). The definition does not
refer to cumulative effects of successive advertising exposure
(carryover effects). In this study, we examine complete and
forced exposures in experimental settings and we focus on the
effects of advertising repetition on two variables: attitude and
recall. These two variables are major outcome variables in
advertising research and are both able to predict behavior as a
result from advertising effects. The majority of studies dealing
with advertising repetition effects have analyzed these varia-
bles (Kirmani 1997), which is an important criterion to obtain
a sufficient number of effect sizes in a meta-analysis.
Repetition Effects on Attitudes
We suggest that advertising repetition has a negative curvi-
linear effect on attitudes. The two-factor theory (Berlyne
1970; Cacioppo and Petty 1979; Stang 1975) provides the the-
oretical basis for research about repetition effects on attitude.
As an extension of the mere-exposure effect (Zajonc 1968),
which argues that mere repetition enhances stimuli evalua-
tions, the theory states that there are two factors—a positive
and a negative factor—that interact with each other and influ-
ence the effect of advertising repetition on attitude. Positive
factors include habituation (Berlyne 1970) and learning (Stang
1975) and result in positive thoughts (Cacioppo and Petty
1979). A negative factor is redundancy or boredom (Berlyne
1970), which result in negative thoughts (Cacioppo and Petty
1979).
Although researchers use different terms for both factors,
all explanations indicate an inverted U-shaped curve as the net
effect of the positive and negative factors of repetition. That
is, the course of the repetition effect is initially positive. Atti-
tude increases with exposure until familiarity and learning is
saturated, which is indicated by the highest level of attitude.
Additional exposures past this point lead to boredom and to
negative thoughts that outweigh positive ones. Attitude
decreases and the course of effect becomes negative.
H1: Advertising repetition has an inverted U-shaped course of
effect on attitude.
Repetition and Recall
We suggest that advertising repetition has a logarithmic
course of effect on recall. The explanation of the advertising
repetition effect on recall is based on learning theory, that is,
on implicit or explicit learning. Stimulus-reaction theories pro-
vide the basis for learning information consciously or subcon-
sciously (Blythe 2013; Pavlov 2003). When processing
information, respondents establish associations by which they
link new information to objects and experiences that are
already stored in memory (Schacter 1996). Recipients recall
information more easily the more often the information is
repeated, because the number of associations with already-
stored information increases (Fuentes et al. 1994). Further-
more, recall for nonadvertised objects is inhibited through rep-
etition, thereby enhancing memory of the advertised object
(Jin, Suh, and Donavan 2008).
In 1885, Ebbinghaus noted that repetition is key for suc-
cessful learning (both subconsciously and consciously) and
thus for recalling information. He established the concept of
the learning curve, explaining how repetition increases recall.
In his experiments, Ebbinghaus determined that learning
develops quickly within the first exposures, leading to a rapid
increase in recall. The learning effect weakens with additional
exposures but still increases until the information is fully
learned and recalled. This implies a logarithmic course of
effect for advertising repetition on recall with the strongest
learning effect within the first exposures.
H2: Advertising repetition has a logarithmic-shaped course of
effect on recall.
The Effective Repetition Level: Minimalists’ Versus
Repetitionists’ View
While researchers agree on the course of advertising repeti-
tion effects on attitude and recall, they still debate effective
416 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
repetition level. Tellis (2004) separates the literature on this
topic into two schools of thought: the minimalists, who believe
that a few (usually one to three) exposures achieve maximum
response, and the repetitionists, who argue that repetition is
necessary for optimal consumer response. A prominent propo-
nent of the minimalist view is Krugman (1972), who argued
that three exposures may be enough, because “fours, fives, etc.
are repeats of the third exposure effect” (p. 13). Tellis (1997)
lists several studies that support the minimalist view. For
instance, McDonald (1971) found that response peaked at two
exposures. Gibson (1996) analyzed experimental data and
found that one exposure sufficed to achieve big changes in atti-
tudes and coupon usage.
The repetitionists argue that several exposures are neces-
sary for optimal consumer response. An early proponent of
this view is Zielske (1959), who found that repetition by as
many as 13 exposures continued to increase recall of the mes-
sage, independently of whether repetitions were massed or
spaced. Several other studies have found support for a repeti-
tionists’ view in both field and laboratory situations. For
instance, studies have shown that attitudes may still increase
upon the fifth exposure (Kohli, Harich, and Leuthesser 2005)
or even the 10th or 25th exposure to the product (Nordhielm
2002).
As a result of this discussion, we distinguish between the
minimalists’ view, which argues for one to three repetitions,
and the repetitionists’ view, which argues for more repetitions
in order to achieve maximum impact on attitudes and/or before
recall levels off. By means of a meta-analysis, we try to find
support for either school of thought in advertising research.
Moderating Effects
The advertising literature and research in neighboring fields
suggest several moderator variables of advertising repetition
effects. These moderators refer to characteristics of the mes-
sage, brand, consumer, communication context, repetition,
and method. Figure 1 provides our research model with the
moderator variables that we investigate in our study. We
derive these variables from the literature review on advertising
repetition by Pechmann and Stewart (1988), the literature
review on stimulus repetition by Harrison (1977), the meta-
analysis on stimulus repetition by Bornstein (1989), and the
meta-analysis on message spacing by Janiszewski, Noel, and
Sawyer (2003). As is common practice in meta-analysis, the
selection of moderator variables is confined to the information
provided in the primary studies that are included in a meta-
analysis. As a result, some variables that are suggested by
Pechmann and Stewart (1988)—for example, variation of
advertisement or advertising content (Chang 2009)—or that
would be intriguing in today’s context—such as the use of
humor in advertising (Gelb and Zinkhan 1985, 1986)—could
not be included due to missing information in the primary
studies. Table 1 provides definitions and operationalization of
the variables in our meta-analysis.
Message spacing refers to the duration between each adver-
tising exposure. The studies included in this meta-analysis use
repeated stimuli advertisements either with or without a break
between exposures. We assume that advertising repetition
effects are generally stronger for spaced advertising repetition
than for messages that are repeated without spacing (i.e.,
massed advertising). We base our hypothesis on the findings
of reviews and meta-analyses outside the advertising literature,
such as Bornstein (1989), who shows in his meta-analysis that
message spacing positively influences attitudes. The two-fac-
tor theory (Berlyne 1970; Cacioppo and Petty 1979; Stang
1975) accounts for the attitude effect. Due to spacing between
exposures, it takes longer for the recipients to get familiar with
the stimulus before the negative effect, boredom, develops
and outweighs the positive familiarity effect. If advertising is
FIG. 1. Conceptual framework.
LITERATURE REVIEW CORNER 417
Downloaded by [FU Berlin] at 02:12 20 October 2015
TABLE 1
Variables Used for the Meta-Analysis
Variable
Related
Hypotheses Definition
Operationalization
(Information Provided in
Primary Studies)
Coding Scheme/Data
Description
Dependent
Attitude toward
the brand
1 Attitude toward the brand
refers to the respondents’
affective reaction toward
the brand in the stimulus
advertisement.
Brand evaluation, attitude,
perceived brand quality,
overall liking.
166 ES and 128 standardized
means in 19 studies
Recall 2 Recall refers to aided and
unaided memory of the
stimulus, i.e., brand name,
advertising content.
Product recall, brand recall,
claim recall, ad feature
recall, specific feature
recall, product categories
recall, message recall, ad
recall.
146 ES and 226 standardized
means in 18 studies
Moderating
Exposure level*1, 2 Exposure level refers to the
number of times the
respondent has been
exposed to the stimulus
advertisement when
measuring attitude and
recall.
Information about the
exposure level provided in
primary studies; exposure
level coded as continuous
variable.
attitude: MD5.30, SD D
3.37; recall: MD4.60, SD
D2.07
Message spacing* 3 Message spacing refers to
whether the exposures of
the stimulus advertisement
were spaced or massed.
Information about spaced or
massed exposures was
provided in the primary
studies. Because only a
small number of studies
reported the length of each
spacing period, we
dummy-coded this variable
as massed versus spaced
exposures.
1Dspaced exposures
(attitude toward the brand:
127 ES; recall: 139 ES)
0Dmassed exposures
(attitude toward the brand:
39 ES; recall: 7 ES)
Advertising length* 4 Advertising length refers to
the duration for which the
respondents were exposed
to the stimulus advertising.
Since only few studies
reported the exact length of
each advertising exposure,
we dummy-coded
advertising length as long
versus short-duration
advertising length.
1Dlong exposure to
advertising (attitude
toward the brand: 99 ES;
recall: 106 ES); 0 Dshort
exposure to advertising
(attitude toward the brand:
67 ES; recall: 40 ES)
Brand novelty* 5 Brand novelty refers to the
level of brand awareness
respondents have when
they are first exposed to the
stimulus advertisement.
Information about whether
the brand was new to the
respondents or not was
provided in the primary
studies. We dummy coded
brand novelty as known
versus unknown brand.
1Dthe test brand is unknown
to the test persons (80 ES
attitude; 90 ES recall); 0 D
the test brand is known to
the test persons (86 ES
attitude, 56 ES recall)
Advertising novelty** 6 Advertising novelty refers to
the level of awareness
respondents have of the
Information about whether
the advertising was new to
the respondents or not was
1Dthe stimulus
advertisement is unknown
to the test persons (attitude
(Continued on next page)
418 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
TABLE 1
Variables Used for the Meta-Analysis (Continued)
Variable
Related
Hypotheses Definition
Operationalization
(Information Provided in
Primary Studies)
Coding Scheme/Data
Description
stimulus advertisement,
when they are first exposed
to it.
provided in the primary
studies. We dummy coded
advertising novelty as
known versus unknown
advertisement.
toward the brand: 15
studies; recall: 14 studies);
0Dthe stimulus
advertisement is known to
the test persons (attitude
toward the brand: 4
studies; recall: 3 studies)
Measurement delay** 7 Measurement delay refers to
the point in time when
attitude and recall are
measured after respondents
have been exposed to the
advertisement. The
measurement occurs either
immediately or with a
delay.
Information about immediate
or delayed measurement of
attitude and recall was
provided in the primary
studies. Only few studies
reported the exact length
for each delay and we
therefore dummy-coded
this variable as immediate
versus delayed
measurement.
1Dmeasurement of
dependent variables at a
later point in time after
exposure (attitude toward
the brand: 9 studies; recall:
9 studies) 0 D
measurement of dependent
variables immediately after
exposure (attitude toward
the brand: 10 studies;
recall: 9 studies)
Involvement** 8 Involvement refers to the
degree of personal
references to a certain
stimulus.
Information about low and
high involvement was
either manipulated in the
primary studies or was
inferred from information
in the method section, i.e.,
the experimental
procedure, the test product,
and the respondent
characteristics. This
variable was thus dummy-
coded as low versus high
involvement.
1Dthe involvement of the
test persons is high
(attitude toward the brand:
6 studies; recall: 4 studies)
0Dthe involvement of the
test persons is low (attitude
toward the brand: 13
studies, recall: 12 studies)
Embedded advertising** 9 Embedded advertising refers
to whether the stimulus
advertisement is embedded
in a television show,
magazine, etc., or whether
respondents are exposed to
the stimulus advertisement
only.
Information about embedded
stimulus advertisement
was provided in the
primary studies. We
dummy-coded the variable
as embedded versus
stimulus advertisement
only.
1Dthe stimulus
advertisement shown to the
test persons was embedded
(attitude toward the brand:
10 studies; recall: 13
studies) 0 Donly the
stimulus advertisement
was shown to the test
persons (attitude toward
the brand: 9 studies; recall:
5 studies)
Year** Control Year refers to the age of a
study and thus controls for
developments over time.
We examined the publication
year of the article.
attitude: MD1990, SD D
10.77; recall: MD1991,
SD D8.31
Note.ESDnumber of effect sizes; MDmean; SD Dstandard deviation. *Effect sizes vary at effect size level; **Effect sizes vary at
study level.
LITERATURE REVIEW CORNER 419
Downloaded by [FU Berlin] at 02:12 20 October 2015
repeated continuously without a break between repetitions,
boredom would develop quickly and outweigh the positive
effect of familiarity much faster. In fact, prior research sug-
gests that highly concentrated and massed repetition schedules
lead to negative response such as irritation and unfavorable
impressions (Heflin and Haygood 1985).
As for the effect on recall, the meta-analysis by Janiszew-
ski, Noel, and Sawyer (2003) shows that message spacing pos-
itively influences recall. Learning theory explains these
findings. Due to spacing between exposures, recipients have
time to process information and form associations with infor-
mation stored in their memory, which, in turn, increases recall.
If advertising repetition is massed, recipients do not have suffi-
cient time to process the information and create links to stored
information, which reduces the repetition effect on recall.
Thus, message spacing increases the size of the positive adver-
tising repetition effect on attitude and recall.
H3: The effect of advertising repetition on (a) attitude and (b)
recall is stronger when message spacing occurs between advertising
exposures.
Advertising length refers to the duration of exposure to a
stimulus advertisement. Long advertising length refers to a
long time of exposure to the stimulus advertisement, whereas
short advertising length means that recipients were exposed to
the advertisement for a short period of time. We propose that
advertising repetition effects are generally stronger for short
compared to long advertising length. We base our assumption
on the findings by Bornstein (1989) who shows that positive
repetition effects on attitude become weaker the longer a
recipient is exposed to a stimulus. Pechmann and Stewart
(1988) support this finding in an advertising context by show-
ing that negative repetition effects on attitude develop more
quickly when being exposed to a long compared to a short
advertisement. The two-factor theory provides an explanation
for this finding. When being exposed to a short-duration adver-
tisement, familiarity prevails and boredom develops later.
Consequently, the effect size of advertising repetition is bigger
for a short exposure than for a long one. If the advertisement
exposure is long, boredom and fatigue develop quickly and
outweigh the positive effect of familiarity much sooner.
As for the effect on recall, learning theory suggests that a
short-duration advertisement can create only few semantic
links and, thus, recipients need a higher number of exposures
to create a semantic network. Accordingly, the effect of adver-
tising repetition becomes stronger for such advertisements.
When being exposed to a long-duration advertisement, recipi-
ents are more likely to process the advertisement’s entire con-
tent and to build strong associations to already-stored
information within only a few exposures. Singh and Cole
(1993) have compared the effectiveness of 15-second and 30-
second commercials and found that longer commercials are
superior in influencing consumers’ learning and attitude. Thus,
short-duration advertisements need more repetitions to be
effective, which increases the repetition effect on attitude and
recall of short-duration advertisement compared to long-dura-
tion advertisement.
H4: The effect of advertising repetition on (a) attitude and (b)
recall is weaker for the exposure to a long-duration advertisement
than for the exposure to a short-duration advertisement.
Brand novelty refers to the level of brand awareness, and
advertising novelty refers to the level of advertisement aware-
ness recipients have when they are exposed to advertisements
for the first time. The brand or advertisement can be either
unknown or known by the recipients. We propose that adver-
tising repetition effects are generally stronger for an unknown
brand and advertising than for known ones. The proposition
related to attitudes follows the two-factor theory. When expos-
ing recipients to an unknown brand or ad, it takes longer for
recipients to get familiar with the brand or ad, because recipi-
ents have no prior information about the ad and brand
(Machleit and Wilson 1988). Hence, it takes longer until bore-
dom develops and outweighs the positive effect. When being
exposed to a known brand or ad, the positive factor of familiar-
ity has already been established. With every additional repeti-
tion the negative factor of boredom increases and the
repetition effect weakens.
As for recall, we base our hypothesis on Stang (1975), who
posits that wear-out occurs as soon as recipients have proc-
essed and learned everything about a stimulus. Thus, consum-
ers process unknown brands or ads until they have learned
everything about them. Because known brands and ads have
been learned before, related knowledge is already stored in the
semantic network. The knowledge knots within this network
trigger continuous processing of the brand and/or ad (Britton
and Tesser 1982), and repetition does not add to this effect.
Thus, novelty of the brand or ad increases the size of the posi-
tive advertising repetition effect on attitude and recall.
H5: The effect of advertising repetition on (a) attitude and (b)
recall is stronger when recipients are exposed to a novel brand ver-
sus a familiar brand.
H6: The effect of advertising repetition on (a) attitude and (b)
recall is stronger when recipients are exposed to a novel advertise-
ment versus a familiar advertisement.
Measurement delay refers to the point in time when atti-
tude and recall are measured after exposure. The measure-
ment occurs either immediately or with a delay. We propose
that advertising repetition effects are generally stronger when
immediately measured instead of when measurement is
delayed. This suggestion opposes the findings in Harrison’s
(1977) literature review and Bornstein’s (1989) meta-analy-
sis, which both show that measurement delay increases
the effect of repetition on attitude. We suggest that attitude
420 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
decays over time, following research by Chattopadhyay and
Nedungadi (1992) and Haugtvedt and colleagues (1994).
When measuring attitude immediately after exposure, the
effect of advertising repetition is high because it captures the
highest degree in attitude change achieved by repeated expo-
sure. However, the effect of advertising repetition decreases
once recipients are not exposed to the advertising. Attitude
effects suffer from decays and are weaker when measured
with a delay.
As for recall, Pechmann and Stewart (1988) suggest a simi-
lar positive effect of immediate measurement of advertising
repetition, because forgetting effects occur when measured
with a delay (Zielske 1959). When measuring recall immedi-
ately after exposure, forgetting effects have not yet occurred
and the recipient is able to access the complete semantic net-
work. When measuring recall with a delay, however, for-
getting effects may have already developed (Ebbinghaus
1885), which reduces the positive effect of advertising repeti-
tion on recall.
H7: The effect of advertising repetition on (a) attitude and (b)
recall is weaker when attitude is measured after a delay than when
it is measured immediately after exposure.
Involvement refers to the degree of personal references to a
certain stimulus. High involvement means that recipients have
a high number of personal references to a stimulus, whereas
low involvement implies that recipients have few or no per-
sonal references to a stimulus. We propose that advertising
repetition effects are weaker for recipients with high involve-
ment compared to low involvement. In terms of attitude, low-
involved recipients need more repetitions to become familiar
with the stimulus and boredom occurs. Thus, the positive repe-
tition effect prevails longer and advertising repetition is more
influential, a finding that aligns with the literature review by
Pechmann and Stewart (1988). However, when being highly
involved, only a small number of exposures is needed to
become fully familiar with the stimulus and boredom occurs.
Thus, the positive effect of advertising repetition is soon out-
weighed by the negative effect (Anand and Sternthal 1990).
As for recall, we base our hypothesis on the elaboration-
likelihood model of persuasion (Petty and Cacioppo 1981),
which shows that the level of involvement determines the
intensity with which the stimulus is processed. When involve-
ment is low, the stimulus is not processed intensively and thus
more exposures are required to establish associations with
already-stored information and to recall the stimulus. High
involvement, however, leads to selective attention to the
respective stimulus and thus increases resistance to forgetting
(Burke and Srull 1988). High involvement leads to intensive
information processing and provokes fast and intensive learn-
ing, leading to faster wear-in and wear-out (Cauberghe and De
Pelsmacker 2010). Thus, repetition under high involvement
leads to small effects on recall.
H8: The effect of advertising repetition on (a) attitude and (b)
recall is weaker when involvement of the recipients is high.
Embedded advertising refers to the context in which the
advertisement is being presented to the recipients. Test adver-
tisements may be embedded in a television show, print maga-
zine, and so on, with other stimuli competing for attention.
Nonembedded advertisements are presented without any com-
peting context. We suggest that advertising repetition effects
become bigger when advertisements are embedded, because
consumers are exposed not only to the stimulus brand but also
to competitive brands in other advertisements and to nonre-
lated context that distracts the consumer from the stimulus
message and can lead to diminished advertising effectiveness
(Kent 1993). Due to this distraction, familiarity develops more
slowly, boredom occurs later, and the positive effect of famil-
iarity outweighs the negative effect of boredom at a higher
number of repetitions. When advertisement is not embedded,
consumers are not distracted and can focus on the test adver-
tisement. Familiarity is reached quickly and the negative fac-
tor develops sooner, outweighing the positive factor.
As for recall, the distraction caused by embedded advertise-
ments leads to more information that has to be remembered
and to be connected to already stored information in order to
recall the stimulus. Because recipients have limited resources
for message retention, repeating advertising clears resources
for processing, but competing ads or cluttered environments
will reduce resource availability (Malaviya, Meyers-Levy, and
Sternthal 1999). Cluttered environments lead to confusion of
claims and messages, interference from other brands and infor-
mation, inattention to the focal ad, and thereby lower recall
(Yaveroglu and Donthu 2008). Thus, more repetition is needed
to build strong associations with the test brand and to enhance
recall. When test advertisements are shown without context, it
is easier and faster to build associations and to recall the
stimulus.
H9: The effect of advertising repetition on (a) attitude and (b)
recall is stronger when the advertisement is embedded.
METHOD
Database Development
We use a meta-analytic data set to assess the advertising
repetition effect on attitude and recall. We choose attitude
toward the brand as the attitude variable for our meta-analysis,
as it relates closely to behavioral outcomes. Similar to other
meta-analyses in the advertising literature (Amos, Holmes,
and Strutton 2008; Brown, Homer, and Inman 1998), we sub-
sume different recall types (e.g., brand or ad recall) under one
category.
1
The meta-analytic data set contains data from
experimental studies only. We focus on experimental studies
to control for an equal number of exposures within treatment
LITERATURE REVIEW CORNER 421
Downloaded by [FU Berlin] at 02:12 20 October 2015
groups and to ensure a high level of attention of respondents at
each exposure. Attention of respondents in experimental
settings varies less than attention of respondents during field
studies. This way, we reduce confounding effects of inatten-
tiveness or random exposure when comparing series of expo-
sures across respondents.
We compile the meta-analytic database using an approach
consistent with recommendations in the literature (e.g., Hunter
and Schmidt 2004) and closely following methods used in cur-
rent meta-analyses in the literature (e.g., Rubera and Kirca 2012;
Zablah et al. 2012). To identify relevant studies, we first perform
a keyword search of electronic databases (e.g., EBSCO Business
Source Premier, ISI Web of Knowledge, PsychInfo, ABI/
INFORM) using, for example, “ad repetition,” “message repeti-
tion,” “ad exposure,” “ad frequency,” and “ad wear-in” and “ad
wear-out” as keywords, followed by an Internet search on Goo-
gle Scholar. We then search journal outlets that turned out to be
major sources for journal articles dealing with advertising repeti-
tion: International Journal of Advertising, Journal of Advertis-
ing, Journal of Advertising Research, Journal of Consumer
Research, Journal of Marketing, Journal of Marketing Research,
Journal of the Academy of Marketing Science, Marketing Sci-
ence,andPsychology and Marketing. Once a study is identified,
references are examined in a search for further studies.
To be included in the meta-analysis, a manuscript has to
provide an experimental study with controlled settings and
forced exposures that measure the effect of advertising repeti-
tion, in other words, multiple exposures of consumers to
advertisements at different points in time, on either attitude
toward the brand or recall. Note that to be included a study
had to provide data for at least two repetition levels (e.g., one
repetition versus three repetitions or five versus ten repeti-
tions). By comparing the outcomes at different levels of repeti-
tion across studies, it is possible to estimate a course of effect
even if the primary study did not intend to find an optimal
level of repetition. This approach has been successfully
applied in other meta-analyses that have estimated the optimal
amount of negative information in two-sided advertising
(Eisend 2006) or the curvilinear relationship between portion
size and consumption (Zlatevska, Dubelaar, and Holden
2014). While the resulting estimate comes with sampling vari-
ation and is not generalizable across any kind of research con-
text, the estimate provides us with sufficient evidence to
decide between the minimalists’ and repetitionists’ view.
The studies have to provide sufficient information to com-
pute effect sizes at different exposure levels. Therefore, we
excluded studies that did not provide sufficient data and stud-
ies for which necessary data could not be retrieved from the
authors or calculated otherwise. The final data set comprises
37 studies: 19 studies provide data on attitude toward the brand
and 18 studies on recall (see Online Appendix). The number of
studies in our meta-analysis equals the average number of
studies (37.7) that has been included in other advertising meta-
analyses (Eisend 2014).
Since we use experimental studies for our analysis, the
effect size metric selected for the meta-analysis is the stan-
dardized mean difference that describes the effect of repetition
on any of the dependent variables. The standardized mean dif-
ference is computed as follows (Hedges 1981): the mean of
group 2 (i.e., dependent variable scores at second exposure) is
subtracted from the mean of group 1 (i.e., dependent variable
scores at first exposure). This result is divided by the pooled
standard deviation.
To display the course of exposure effects on attitudes, we
additionally calculate the standardized mean of the attitude
measure at each exposure level. The standardized mean for
attitude is computed by taking the mean value of attitude at a
particular exposure level divided by the number of scale inter-
vals. As for recall, all values are based on open questions that
are not standardized.
We include the following moderating variables in our meta-
analysis: exposure level, message spacing, advertising length,
brand novelty, advertising novelty, measurement delay,
involvement, embedded advertising, and year of publication.
Table 1 gives an overview of the definition, coding scheme,
and the data description of each variable. Some moderator var-
iables were manipulated and tested in primary studies: brand
novelty is examined by Campbell and Keller (2003); measure-
ment delay is examined by Craig, Sternthal, and Leavitt
(1976), Haugtvedt and colleagues (1994), Singh and col-
leagues (1994), and Singh, Rothschild, and Churchill (1988);
message spacing is examined by Malaviya (2007), Malaviya
and Sternthal (1997), and Singh and colleagues (1994); adver-
tising length is examined by Malaviya and Sternthal (1997),
Singh and Cole (1993), Singh, Rothschild, and Churchill
(1988), Swasy and Rethans (1986), and Rethans, Swasy, and
Marks (1986). As for the remaining studies, we inferred infor-
mation about the moderator variables from the context of the
study (see description in Table 1). By dummy-coding most of
the moderating variables, we could retain almost all studies in
the moderator analysis. Only two studies from our data set
could not be included in the moderator analysis due to missing
data: Craig, Sternthal, and Leavitt (1976); and Singh and col-
leagues (1994).
The moderating variables have a nested structure, that is,
they are measured either at the effect size level or the study
level. We indicate in Table 1 the level at which these variables
are measured and vary. The moderating variables that vary at
the effect size level have various manifestations within one
study. The moderating variables that vary at the study level
have the same manifestation within the study but various man-
ifestations across different studies. For instance, message spac-
ing is measured at the effect size level because spaced versus
massed exposure is manipulated and tested within some stud-
ies. Thus, message spacing varies within studies. Embedded
advertising is measured at the study level because none of the
studies manipulates or tests embedded versus nonembedded
advertising in one study, but the studies use either embedded
422 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
or nonembedded advertising. Therefore, the variable advertis-
ing type varies between (across) studies. Note that the level
varies between brand and advertising novelty. Most of the
studies considered in this meta-analysis test both new brands
and familiar brands in one study, whereas advertising novelty
does not vary within studies. The authors use either new or
familiar advertising in their studies.
Data Analysis
We were able to calculate 128 standardized means and 166
standardized mean differences for advertising exposure effects
on attitude toward the brand.
2
For advertising exposure effects
on recall, the selected manuscripts provided information to
calculate 226 means and 146 standardized mean differences.
Course of effect. We analyze the course of advertising
exposure effects by regressing the number of exposures on the
standardized means of attitude toward the brand and the means
of recall. We apply a variance weighted regression model to
account for the variation in sample size and in the resulting
variance of each effect size. To test for nonlinear effects, we
compare two models. The first model is the baseline model
that includes a linear term for the relationship between expo-
sure level and the dependent variables. In the second model,
we add a nonlinear term, in this case, the quadratic term for
attitude toward the brand and a logarithmic term for the recall.
The change in Fvalue indicates whether the nonlinear term
significantly explains more variance than does the linear term.
Moderator effects. Other than in prior studies, we exam-
ine the moderating effects simultaneously. That is, we apply a
multivariate analysis where the influence of all moderator vari-
ables on the effect sizes are tested jointly and not separately as
has been done in previous research. We model the effect sizes
as a linear function of the moderating variables. Because our
data have a nested structure, we use a variance-known hierar-
chical linear model (HLM) estimation procedure to account
for nested data and within-study error correlation between esti-
mates. This procedure is recommended for meta-analysis with
a nested data structure (Bijmolt and Pieters 2001). As our sam-
ple is rather small (19 studies provide estimates for attitude
and 18 studies provide estimates on recall), we follow the
methodological approach by Krasnikov and Jayachandran
(2008) and specify an intercept-as-outcome model:
1. (1a) Level 1: ES
ij
Db
0j
+b
1j
* exposure level
ij
+b
2j
*
message spacing
ij
+b
3j
* advertising length
ij
+b
4j
* brand
novelty
ij
+e
ij
2. (1b) Level 2: b
0j
Dg
00
+g
01
* advertising novelty
j
+g
02
*
measurement delay
j
+g
03
* involvement
j
+g
04
* embedded
advertising
j
+g
05
* year
j
+u
0j
3. (1c) Mixed model: b
1j
Dg
10
;b
2j
Dg
20
;b
3j
Dg
30
;b
4j
D
g
40
ES
ij
Dg
00
+g
01
* advertising novelty
j
+g
02
* measure-
ment delay
j
+g
03
* involvement
j
+g
04
* embedded adver-
tising
j
+g
05
* year
j
+g
10
* exposure level
ij
+g
20
*
message spacing
ij
+g
30
* advertising length
ij
+g
40
* brand
novelty
ij
+u
0j
+e
ij
,
where ES
ij
represents the effect sizes (i.e., the effect of repeti-
tion on either attitude or recall) given within the sample; b
0j
the intercept at Level 1; and b
1j
to b
4j
the slope estimates for
the variables which vary within studies (exposure level, mes-
sage spacing, advertising length, and brand novelty). On Level
2, g
00
describes the intercept in Level 2; g
01
to g
05
the slope
estimates for variables which vary between studies (advertis-
ing novelty, measurement delay, involvement, embedded
advertising, and year); and u
0j
the unexplained variance
between studies.
To ensure robustness, we examine the influence of the mod-
erating variables incrementally by running four different mod-
els. Model 1 contains the control variable (year), Model 2
contains the control and the effect size level moderators (Level
1), Model 3 contains the control and the study level modera-
tors (Level 2), and Model 4 is the full model that contains all
control and moderating variables.
Collinearity is always an issue in meta-analytic regres-
sion analysis. Because there is no direct diagnostic for
multicollinearity in HLM, we examine the correlations
among the variables and the variance inflation factor (VIF)
of the regression models. The highest correlation in the
data set of .509 and the low VIFs indicate that multicolli-
nearity is not a problem. We report the correlation matrix
separately for study level variables and effect size level
variables in the Online Appendix.
RESULTS
Course of Effect
Table 2 reports the regression results for the course of
advertising exposure effects. The findings show that attitude
toward the brand does not take a linear course with increasing
advertising repetition (DFD1.73, p>.05). Instead, the effect
of advertising repetition on attitude toward the brand can be
modeled by a nonlinear quadratic course of effect (DFD3.12,
p<.05) in shape of an inverted Ucurve. When calculating the
maximum level of attitude by setting the first derivation to
zero, we find an optimum at 10 complete exposures. Thus, our
study shows an inverted U-shaped course of effect as agreed
on in the literature, but with a maximum level of more than
three exposures. Regarding the exposure effects on recall, we
do not find a logarithmic course of effect (DFD.71, p>.05)
as proposed in the literature, but a linear course of effect
(DFD132.28, p<.01).
Moderator Effects
A homogeneity test reveals variability across effect sizes
for both dependent variables (p<.01), supporting the need to
LITERATURE REVIEW CORNER 423
Downloaded by [FU Berlin] at 02:12 20 October 2015
apply moderator variables to explain this variance. Table 3
and Table 4 present the findings of the HLM related to both
dependent variables. While we find some changes of the coef-
ficients at the effect size level across the four models related to
attitude, the major tendency of the significance levels, direc-
tion, and size of effects remains rather stable. We refer to the
findings of the full model (Model 4). The model shows that
five moderating factors influence the effect of advertising rep-
etition on attitude: exposure level, message spacing, measure-
ment delay, involvement, and year. Repetition has a stronger
effect on attitude toward the brand when the exposures to the
test advertising are spaced and a weaker effect when they are
massed (bD.268, SE D.139 p<.05), in line with hypothesis
3a. The effect of advertising repetition is stronger when atti-
tude toward the brand is measured immediately after exposure
rather than with a delay (bD¡.618, SE D.150, p<.01), and
when the participants show low involvement for the advertised
product rather than high involvement (bD¡.497, SE D.151,
p<.01). The findings are in line with hypothesis 7a and
hypothesis 8a. The positive effect of advertising repetition
decreases with increasing exposure level (bD¡.101, SE D
.035, p<.01), which matches the results of the course of
effect analysis.
With respect to recall, we find that four factors influence
the effect size of advertising repetition: exposure level, mes-
sage spacing, measurement delay, and embedded advertising.
We find that the effect of advertising repetition is stronger
when the exposures to the test advertising are massed and not
spaced (bD¡.124, SE D.034, p<.01), which is opposite
of the moderating effect found for attitude toward the brand,
and not in line with hypothesis 3b. Recall seems to need con-
tinuous and massed stimulation to be effective. As soon as
TABLE 2
Course of Advertising Effects: WLS Regression Results
Attitude toward the Brand
(Standardized Mean)
Recall (Mean)
Model 1 Model 2 Model 1 Model 2
Repetition .851 (.646) 3.065 (1.409) .326 (.028) .373 (.063)
Repetition
2
¡.162 (.092)
Ln repetition ¡.308 (.365)
R
2
.014 .038 .384 .386
DR
2
.014 .024 .384 .002
DF1.732 3.115 132.277 .713
Sig. DF.100 .040 .001 .200
Note. Unstandardized coefficients (standard error in parentheses). One-tailed significance testing.
TABLE 3
Moderator Analysis: HLM Results for Attitude Toward Brand
Model 1 Model 2 Model 3 Model 4
Control
Year, g
05
.001 (.010) .005 (.008) ¡.012* (.006) ¡.010 * (.006)
Effect Size Level
Exposure level, g
30
¡.062 (.050) ¡.101** (.035)
Message spacing, g
10
¡.012 (.144) .268 * (.139)
Advertising length, g
20
.074 (.170) ¡.346 (.229)
Brand novelty, g
40
¡.151 (.157) .037 (.132)
Study Level
Advertising novelty, g
01
.224 (.190) .204 (.170)
Measurement delay, g
02
¡.447** (.113) ¡.618** (.150)
Involvement, g
03
¡.311** (.116) ¡.497** (.151)
Embedded advertising, g
04
¡.087 (.208) .116 (.233)
Note. Unstandardized coefficients and robust standard errors (SE) are provided.
**p<.01; *p<.05, one-tailed test.
424 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
exposures are spaced, recall effects of advertising repetition
become weaker. The effect of measurement delay is similar
to the one observed for attitude toward the brand: advertising
repetition has a stronger effect when recall is measured
immediately after exposure rather than with a delay (bD
¡.320, SE D.126, p<.05). The finding is in line with
hypothesis 7b. Finally, advertising repetition enhances recall
more strongly when the test advertising is embedded in a TV
program or magazine than when it is not embedded (bD
.360, SE D.152, p<.05), supporting hypothesis 9b. The
positive repetition effect decreases with increasing exposure
level (bD¡.029, SE D.011, p<.01), which indicates that
the repetition effect levels off. The remaining variables in
our analysis do not influence the effect size of advertising
repetition on recall.
DISCUSSION
This study offers a comprehensive synthesis of advertising
repetition effects. By conducting a meta-analysis, we support
theoretical models that explain the course of advertising expo-
sure effects. The study finds an inverted U-shaped course of
advertising exposure effects on attitude and thus supports the
two-factor theory (Berlyne 1970; Cacioppo and Petty 1979;
Stang 1975). The results only partly support the learning curves
of Ebbinghaus (1885) by showing a linear relationship between
repetition and learning. This effect might be due to the low varia-
tion in repetitions in the studies that provide mean values for
recall: the range is between one and eight exposures, with only
one study reporting exposure levels of 14 and 21. The average
exposure rate is 3.5. Hence, it is not unlikely that the logarithmic
learning curve eventually appears, but it occurs at a high repeti-
tion rate. The exposure effect in HLM for recall indicates that
the repetition effect for recall indeed levels off.
The findings provide new insights on effective frequency
and the number of advertising exposures that leads to optimal
consumer response regarding attitudes and recall. They clearly
support the repetitionists’ view in the literature over the mini-
malists’ view: few exposures are not enough to achieve maxi-
mum response, but repetition is essential for consumer
response.
The repetitionists’ view does not necessarily constitute a
universal generalization, as repetition effects depend on con-
tingent factors. To explain under which conditions advertising
repetition works best, this study further examines moderator
effects. Moderator variables that influence the repetition effect
do not question the nonlinear effects, but show that the maxi-
mum response or the level-off occurs with fewer or more repe-
titions. For instance, the intercept in the HLM model referring
to attitudes is nonsignificant, which is due to the curvilinear
effect (i.e., tradeoff in effects of the positive and the negative
factor that lead to a zero effect). Because the moderator
involvement is significant and has a negative sign in the HLM
model, high involvement needs less repetitions to achieve the
zero effect (i.e., the point where the negative effect outweighs
the positive one) and as a result high involvement needs less
repetitions to achieve maximum impact.
We find that study design and respondent characteristics
increase the repetition effect. Specifically, we find that low
involvement and spaced exposures enhance advertising
exposure effects on attitude toward the brand, whereas
embedded advertising and massed exposures enhance the
advertising exposure effects on recall. We also find that
advertising repetition effects decay over time for attitude
toward the brand and recall, in line with the findings of
Chattopadhyay and Nedungadi (1992) and Haugtvedt and
colleagues (1994), and also in line with forgetting effects
(Zielske 1959).
TABLE 4
Moderator Analysis: HLM Results for Recall
Model 1 Model 2 Model 3 Model 4
Control
Year, g
05
¡.004 (.005) ¡.000 (.004) .003 (.004) .003 (.005)
Effect Size Level
Exposure level, g
30
¡.031* (.018) ¡.029** (.011)
Message spacing, g
10
¡.111** (.041) ¡.124** (.034)
Advertising length, g
20
.021 (.057) .030 (.070)
Brand novelty, g
40
.166* (.090) .123 (.081)
Study Level
Advertising novelty, g
01
.345* (.176) .249 (.182)
Measurement delay, g
02
¡.373* (.158) ¡.320* (.126)
Involvement, g
03
.043 (.061) .032 (.090)
Embedded advertising, g
04
.423* (.162) .360* (.152)
Note. Unstandardized coefficients and robust standard errors (SE) are provided.
**p<.01; *p<.05, one-tailed test.
LITERATURE REVIEW CORNER 425
Downloaded by [FU Berlin] at 02:12 20 October 2015
Theoretical and Practical Implications
These findings contribute to the discussion of two main
schools of thought in advertising repetition by supporting the rep-
etitionists’ view. Our findings on the advertising repetition course
of effect are in line with theories that explain repetition, and they
provide important implications for researchers who design stud-
ies to learn the general effects of repetition. They show that small
number of repetitions will lead to small effects only and that up to
10 exposure levels should be chosen to maximize effects on atti-
tudes; for effects on recall, even more exposure levels are needed.
The results show that advertising repetition is a valuable
tool for reaching the consumer and effectively conveying an
advertising message. Our meta-analysis is based on experi-
mental studies that ensure comparability of findings across
treatments and respondents. Therefore, the generalizability of
these findings to real-world settings is somewhat limited,
because experimental studies allow for better control of con-
founding effects and tend to lead to stronger effect sizes (Field
and Hole (2002). As a result, wear-in should occur faster and
the optimum level should be reached with fewer repetitions in
an experimental setting compared to a field setting. More real-
istic advertising stimuli in field studies show creative varia-
tions of styles and plots (Chang 2009). Variation strategies
keep attention and interest high and might therefore even delay
the optimum frequency effect. Furthermore, many exposures
in real-world settings are not completed, and higher exposure
rates are necessary to reach optimum response. Hence, the fig-
ures we found in our meta-analytic data set might underesti-
mate the optimum exposure level needed in a real-world
setting. This is why we conclude that the repetitionists’ view
is a meaningful approach for practitioners, too.
The findings of the moderator analysis provide further
information for practitioners as they show that repetition levels
need to be adapted to specific situations. For instance, the
moderating factors indicate that more advertising repetition is
needed in low-involvement situations. That is, fast-moving
consumer goods benefit from heavy advertising penetration.
Limitations and Avenues for Future Research
As described, a major limitation refers to the generalizabil-
ity of our findings, in particular, in terms of external validity.
While we cannot directly transfer the figures of optimum expo-
sure level to a real-world setting, we conclude that these fig-
ures might be even higher in these environments. Although
our findings suggest following the repetitionists’ view, further
research is needed to establish exact figures on an optimum
repetition level for practitioners.
Further limitations are common for the use and application
of meta-analytic data and techniques. First, meta-analysis
works with aggregated and generalized findings. Thus, our
results provide a broad guideline and a better understanding
of advertising effectiveness but do not give detailed advice
for any situation marketers and researchers may encounter.
Second, our findings are limited to the information provided in
primary studies. We were able to examine only the variables
for which information was provided. Other advertising varia-
bles, such as advertising content elements (e.g., humor that
has been shown to lead to varying repetition effects; Gelb and
Zinkhan 1985, 1986), could not be examined. Furthermore,
some variables had to be dummy-coded due to missing infor-
mation in the primary studies, resulting in cutoffs and possible
loss of significant effects. Other variables analyzed in the pri-
mary studies, such as age or gender of respondents, could not
be included in our moderator analysis due to low data varia-
tion. That is, most of the experiments were based on student
samples with an equal proportion of male and female respond-
ents. Moreover, we were not able to examine advertising expo-
sure effects on variables such as choice sets or reaction time
due to only a small number of primary studies that examined
these dependent variables. Building on these limitations,
future studies may examine the effect of advertising repetition
on direct behavioral outcomes and may examine advertising
content factors as moderators for this relationship.
Conclusion
In summary, we meta-analyze prior studies on advertising
repetition effects and confirm a non-linear course of advertis-
ing exposure effects. Our findings on the most effective expo-
sure level for maximum attitude and for the strongest increase
in recall support the repetitionists’ view over the minimalists’
view. We find that maximum attitude is reached after 10 expo-
sures, and the increase in recall does not level off at lower
exposure rates either. Our results indicate that moderating
effects found in prior literature reviews and meta-analyses out-
side the advertising field cannot be readily transferred to an
advertising context and that particularly study characteristics,
such as message spacing and measurement delay, influence
the advertising repetition effect.
NOTES
1. We performed several tests of equivalence of effect sizes between
recall types (e.g., we compared effect sizes related to brand recall
with effect sizes related to ad recall). The effect sizes did not differ
across recall types, except for attribute recall, which we excluded
from this category.
2. Manuscripts that did not report scale intervals could not be consid-
ered for the calculation of standardized means.
SUPPLEMENTAL DATA
Supplemental data for this article can be accessed at www.
tandfonline.com/ujoa.
REFERENCES
Amos, Clinton, Gary Holmes, and David Strutton (2008), “Exploring the
Relationship Between Celebrity Endorser Effects and Advertising
Effectiveness,” International Journal of Advertising, 27 (2), 209–34.
426 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
Anand, Punam, and Brian Sternthal (1990), “Ease of Message Processing as a
Moderator of Repetition Effects in Advertising,” Journal of Marketing
Research, 27 (3), 345–53.
Berlyne, D. E. (1970), “Novelty, Complexity, and Hedonic Value,” Perception
and Psychophysics, 8 (5), 279–86.
Bijmolt, Tammo H. A., and Rik G. M. Pieters (2001), “Meta-Analysis in Mar-
keting When Studies Contain Multiple Measurements,” Marketing Letters,
12 (2), 157–69.
Blythe, J. (2013), Consumer Behaviour, 3rd ed., Thousand Oaks, CA: Sage.
Bornstein, Robert F. (1989), “Exposure and Affect: Overview and Meta-Anal-
ysis of Research, 1968–1987,” Psychological Bulletin, 106 (2), 265–89.
Britton, Bruce K., and Abraham Tesser (1982), “Effects of Prior Knowledge
on Use of Cognitive Capacity in Three Complex Cognitive Tasks,” Journal
of Verbal Learning and Verbal Behavior, 21 (4), 421–36.
Brown, Steven P., Pamela M. Homer, and J. Jeffrey Inman (1998), “A Meta-
Analysis of Relationships Between Ad-Evoked Feelings and Advertising
Responses,” Journal of Marketing Research, 35 (1), 114–26.
Burke, Raymond R., and Thomas K. Srull (1988), “Competitive Interference
and Consumer Memory for Advertising,” Journal of Consumer Research,
15 (June), 55–68.
Cacioppo, John T., and Richard E. Petty (1979), “Effects of Message Repeti-
tion and Position on Cognitive Responses, Recall, and Persuasion,” Jour-
nal of Personality and Social Psychology, 37 (1), 97–109.
Campbell, Margaret C., and Kevin Lane Keller (2003), “Brand Familiarity and
Advertising Repetition Effects,” Journal of Consumer Research, 30 (2),
292–304.
Cauberghe, Verolien, and Patrick De Pelsmacker (2010), “Advergames. The
Impact of Brand Prominence and Game Repetition on Brand Responses,”
Journal of Advertising, 39 (1), 5–18.
Chang, Chingching (2009), “Repetition Variation Strategies for Narrative
Advertising,” Journal of Advertising, 38 (3), 51–65.
Chattopadhyay, Amitava, and Prakash Nedungadi (1992), “Does Attitude
Toward the Ad Endure? The Moderating Effects of Attention and Delay,”
Journal of Consumer Research, 19 (1), 26–33.
Craig, C. Samuel, Brian Sternthal, and Clark Leavitt (1976), “Advertising
Wearout: An Experimental Analysis,” Journal of Marketing Research,13
(4), 365–72.
Ebbinghaus, H. (1885), €
Uber Das Ged €
achtnis: Untersuchungen Zur Experi-
mentellen Psychologie [Memory: A contribution to experimental psychol-
ogy], Leipzig: Duncker & Humber.
Eisend, Martin (2006), “Two-Sided Advertising: A Meta-Analysis,” Interna-
tional Journal of Research in Marketing, 23 (2), 187–98.
Eisend, Martin (2014), “How Effective Is Advertising? A Meta-Meta-Analysis
of Effect Sizes in Advertising Research,” Proceedings of the 13th Interna-
tional Conference on Research in Advertising, Amsterdam, the
Netherlands.
Field, Andy, and Graham J. Hole (2002), How to Design and Report Experi-
ments, Los Angeles, CA: Sage.
Fuentes, Luis J., Encarna Carmona, Inmaculada F. Agis, and Andr
es Catena
(1994), “The Role of the Anterior Attention System in Semantic Process-
ing of Both Foveal and Parafoveal Words,” Journal of Cognitive Neurosci-
ence, 6 (1), 17–25.
Gelb, Betsy D., and George M. Zinkhan (1985), “The Effect of Repetition on Humor
in a Radio Advertising Study,” Journal of Advertising, 14 (4), 13–20,68.
Gelb, Betsy D., and George M. Zinkhan (1986), “Humor and Advertising
Effectiveness after Repeated Exposures to a Radio Commercial,” Journal
of Advertising, 15 (2), 15–20, 34.
Gibson, Lawrence D. (1996), “What Can One TV Exposure Do?,” Journal of
Advertising Research,” Journal of Advertising Research, 36 (2), 9–18.
Harrison, Albert A. (1977), “Mere Exposure,” in Advances in Experimental
Social Psychology, Berkowitz Leonard, ed., San Diego, CA: Academic
Press, 39–83.
Haugtvedt, Curtis P., David W. Schumann, Wendy L. Schneier, and Wendy L.
Warren (1994), “Advertising Repetition and Variation Strategies:
Implications for Understanding Attitude Strength,” Journal of Consumer
Research, 21 (1), 176–89.
Hedges, Larry V. (1981), “Distribution Theory for Glass’s Estimator of Effect
Size and Related Estimators,” Journal of Educational and Behavioral Sta-
tistics, 6 (2), 107–28.
Heflin, Debbora T.A., and Robert C. Haygood (1985), “Effects of Scheduling
on Retention of Advertising Messages,” Journal of Advertising, 14 (2),
41–47, 64.
Hunter, J.E., and F.L. Schmidt (2004), Methods of Meta-Analysis: Cor-
recting Error and Bias in Research Findings, Thousands Oaks, CA:
Sage.
Janiszewski, Chris, Hayden Noel, and Alan G. Sawyer (2003), “A MetaAnal-
ysis of the Spacing Effect in Verbal Learning: Implications for Research on
Advertising Repetition and Consumer Memory,” Journal of Consumer
Research, 30 (1), 138–49.
Jin, Hyun Seung, Jaebeom Suh, and D. Todd Donavan (2008), “Salient Effects
of Publicity in Advertised Brand Recall and Recognition: The List-
Strength Paradigm,” Journal of Advertising, 37 (1), 45–57.
Kent, Robert J. (1993), “Competitive Versus Non-Competitive Clutter in
Television Advertising,” Journal of Advertising Research,33(2),40–
46.
Kirmani, Amna (1997), “Advertising Repetition as a Signal of Quality: If It’s
Advertised So Much, Something Must Be Wrong,” Journal of Advertising,
26 (3), 77–86.
Kohli, Chiranjeev S., Katrin R. Harich, and Lance Leuthesser (2005),
“Creating Brand Identity: A Study of Evaluation of New Brand Names,”
Journal of Business Research, 58 (11), 1506–15.
Krasnikov, Alexander, and Satish Jayachandran (2008), “The Relative Impact
of Marketing, Research-and-Development, and Operations Capabilities on
Firm Performance,” Journal of Marketing, 72 (4), 1–11.
Krugman, Herbert E. (1972), “Why Three Exposures May Be Enough,” Jour-
nal of Advertising Research, 12 (6), 11–14.
Machleit, Karen A., and R. Dale Wilson (1988), “Emotional Feelings and Atti-
tude Toward the Advertisement: The Roles of Brand Familarity and Repe-
tition,” Journal of Advertising, 17 (3), 27–35.
Malaviya, Prashant (2007), “The Moderating Influence of Advertising Context
on Ad Repetition Effects: The Role of Amount and Type of Elaboration,”
Journal of Consumer Research, 34 (1), 32–40.
Malaviya, Prashant, Joan Meyers-Levy, and Brian Sternthal (1999), “Ad Repe-
tition in a Cluttered Environment: The Influence of Type of Processing,”
Psychology and Marketing, 16 (2), 99–119.
Malaviya, Prashant, and Brian Sternthal (1997), “The Persuasive Impact of
Message Spacing,” Journal of Consumer Psychology, 6 (3), 233–55.
McDonald, Colin (1971), “What Is the Short-Term Effect of Advertising?,”
Working Paper 71–142, Marketing Science Institute.
Nordhielm, Christie L. (2002), “The Influence of Level of Processing on
Advertising Repetition Effects,” Journal of Consumer Research, 29 (3),
371–82.
Pavlov, I.P. (2003), Conditioned Reflexes, Mineola, NY: Dover.
Pechmann, Cornelia, and David W. Stewart (1988), “Advertising Repetition:
A Critical Review of Wearin and Wearout,” Current Issues and Research
in Advertising, 11 (1–2), 285–329.
Petty, R.E., and J.T. Cacioppo (1981), Attitudes and Persuasion: Classic and
Contemporary Approaches, Dubuque, IA: W.C. Brown.
Rethans, Arno J., John L. Swasy, and Lawrence J. Marks (1986), “Effects of
Television Commercial Repetition, Receiver Knowledge, and Commercial
Length: A Test of the Two-Factor Model,” Journal of Marketing Research,
23 (1), 50–61.
Rubera, Gaia, and Ahmet H. Kirca (2012), “Firm Innovativeness and Its Per-
formance Outcomes: A Meta-Analytic Review and Theoretical Integra-
tion,” Journal of Marketing, 76 (3), 130–47.
Schacter, Daniel L. (1996), “Illusory Memories: A Cognitive Neuroscience
Analysis,” Proceedings of the National Academy of Sciences, 93 (24),
13527–33.
LITERATURE REVIEW CORNER 427
Downloaded by [FU Berlin] at 02:12 20 October 2015
Singh, Surendra N., and Catherine A. Cole (1993), “The Effects of Length,
Content, and Repetition on Television Commercial Effectiveness,” Journal
of Marketing Research, 30 (1), 91–104.
Singh, Surendra N., Sanjay Mishra, Neeli Bendapudi, and Denise Linville
(1994), “Enhancing Memory of Television Commercials through Message
Spacing,” Journal of Marketing Research, 31 (3), 384–92.
Singh, Surendra N., Michael L. Rothschild, and Gilbert A. Churchill, Jr.
(1988), “Recognition versus Recall as Measures of Television Commercial
Forgetting,” Journal of Marketing Research, 25 (1), 72–80.
Stang, David J. (1975), “Effects of ’Mere Exposure’ on Learning and Affect,”
Journal of personality and Social Psychology, 31 (1), 7–12.
Swasy, John L., and Arno J. Rethans (1986), “Knowledge Effects on Curiosity
and New Product Advertising,” Journal of Advertising, 15 (4), 28–34.
Tellis, Gerard (1997), “Effective Frequency: One Exposure or Three
Factors?,” Journal of Advertising Research, 37 (4), 75–80.
Tellis, Gerard (2004), Effective Advertising: Understanding When, How, and
Why Advertising Works, Thousand Oaks, CA: Sage.
Vuokko, Pirjo (1997), “The Determinants of Advertising Repetition Effects,”
in Measuring Advertising Effectiveness, William D. Wells, ed., Mahwah,
NJ: Erlbaum, 239–60.
Yaveroglu, Idil, and Naveen Donthu (2008), “Advertising Repetition and
Placement Issues in On-Line Environments,” Journal of Advertising,37
(2), 31–43.
Zablah, Alex R., George R. Franke, Tom J. Brown, and Darrell E. Bartholo-
mew (2012), “How and When Does Customer Orientation Influence Front-
line Employee Job Outcomes? A Meta-Analytic Evaluation,” Journal of
Marketing, 76 (3), 21–40.
Zajonc, Robert B. (1968), “Attitudinal Effects of Mere Exposure,” Journal of
Personality and Social Psychology, 9 (2, pt. 2), 1–27.
Zielske, Hubert A. (1959), “The Remembering and Forgetting of Advertising,”
Journal of Marketing, 23 (3), 239–43.
Zlatevska, Natalina, Chris Dubelaar, and Stephen S. Holden (2014), “Sizing up
the Effect of Portion Size on Consumption: A Meta-Analytic Review,”
Journal of Marketing, 78 (3), 140–54.
428 LITERATURE REVIEW CORNER
Downloaded by [FU Berlin] at 02:12 20 October 2015
