Get Your Science Out of Here: When Does

Invoking Science in the Marketing of

Consumer Products Backfire?

Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023

AVIVA PHILIPP-MULLER
JOHN P. COSTELLO
REBECCA WALKER RECZEK

In this research, we propose that although consumers view the scientific process
as competent, they also perceive it as cold. Across 10 experimental studies, we
demonstrate that these lay beliefs impact consumers’ reactions to marketers tout-
ing the science behind their brands. Specifically, since hedonic attributes are as-
sociated with warmth, the coldness associated with science is conceptually disflu-
ent with the anticipated warmth of hedonic products and attributes, reducing
product valuation. In contrast, when products are positioned as utilitarian, invoking
science in marketing appeals has a positive effect, as the perceived competence
of the scientific process is more compatible with the competence associated with
utilitarian products. We further demonstrate that when the necessity of science to
create a hedonic product is made salient and thus more fluent, this backfire effect
is attenuated. Finally, we identify three theoretically and practically relevant indi-
vidual differences (endorsement of the lay belief, trust in scientists, and whether
the consumer works in a STEM field) that moderate the backfire effect of pairing
science with hedonically positioned products.

Keywords: perceptions of science, science appeals, lay belief, lay theory, hedonic
versus utilitarian, marketing communications

Aviva Philipp-Muller (aviva_philipp-muller@sfu.ca) is an assistant

professor in marketing at the Beedie School of Business, Simon Fraser
University, SFU Burnaby, 888 University Drive, Burnaby, BC V5A 1S6,
Canada. John P. Costello (jcostel4@nd.edu) is an assistant professor in
marketing at the Mendoza College of Business, University of Notre
Dame, 381 Mendoza College of Business, Notre Dame, IN 46556, USA.
Rebecca Walker Reczek (reczek.3@osu.edu) is the Berry Chair of New
Technologies in marketing at the Fisher College of Business, The Ohio
State University, 2100 Neil Avenue, 506A Fisher Hall, Columbus, OH
43210, USA. Please address correspondence to Aviva Philipp-Muller. The
authors would like to thank Kelly Haws and Jesse Walker for their helpful
comments on this research, as well as Joe Reczek for providing both in-
sight and inspiration. Supplementary materials are included in the web ap-
pendix accompanying the online version of this article.
Editor: Stacy Wood
Associate Editor: Lisa E. Bolton
Advance Access publication May 5, 2022
C The Author(s) 2022. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc. All rights reserved.
V
B rands often tout the scientific rigor used to develop
their products in their marketing communications.
For example, the chocolate company Cadbury frequently
emphasizes the scientific processes used to create its choc-
olate, including a commercial depicting a scientist using a
centrifuge to make chocolate (Cadbury Dairy Milk 2010).
Other brands (e.g., Aveeno, Haus Laboratories) across a
variety of product categories, including cosmetics, personal
care, food, pharmaceuticals, and technology, highlight the
science used to bring their products to market (see web ap-
pendix A for additional examples of brands highlighting
the science behind their brands). Past research has exam-
ined a variety of issues at the intersection of marketing and
science (Andre, Chandon, and Haws 2019; Hingston and
Noseworthy 2018; Marinova et al. 2017; Scott and Rozin
2017; Williams and Steffel 2014), but no work, to our
knowledge, has examined how consumers’ lay beliefs
about the scientific process might impact when marketing
appeals that explicitly reference science are effective at
For permissions, please e-mail: journals.permissions@oup.com  Vol. 49  2023
https://doi.org/10.1093/jcr/ucac020

721
722
increasing consumers’ product valuation and when they
backfire.
Over the past four decades, Americans have generally
viewed science as beneficial (Krause et al. 2019), but, re-
cently, a growing number of Americans report lower trust
in science (American Academy of Arts and Sciences 2018;
Gauchat 2012). There is therefore reason to believe that
consumers may hold mixed beliefs about the use of the sci-
entific process in bringing products to market. In this work,
we propose that although consumers view the scientific
process as competent, they also view it as cold (Aaker,
Vohs, and Mogilner 2010; Cuddy, Fiske, and Glick 2008).
Across 10 studies, we demonstrate that the lay belief
that science is cold but competent impacts consumers’
reactions to the invocation of science in the marketing of
consumer packaged goods (CPGs). We propose that be-
cause hedonic products are associated with warmth
(Chattalas and Takada 2013), the coldness of the scientific
process is conceptually disfluent with hedonic products
specifically, reducing product valuation, consistent with
past research on conceptual fluency (Duhachek, Agrawal,
and Han 2012; Kidwell, Farmer, and Hardesty 2013; Lee
and Labroo 2004). We also identify a number of boundary
conditions for this effect. First, when products are posi-
tioned as utilitarian, science appeals are more effective
than appeals that do not mention science. This occurs be-
cause the scientific process is compatible with the compe-
tence associated with utilitarian products. Second, we
introduce an intervention to eliminate the backfire effect:
By increasing the salience of science in a product category
where it is not naturally salient, science appeals are made
more conceptually fluent and hence no longer backfire for
hedonically positioned products. Finally, because the con-
ceptual disfluency of pairing science with hedonically po-
sitioned products stems from the lay belief that science is
cold, we show that, among those who do not subscribe to
the lay belief, science appeals do not backfire when paired
with hedonic products.
From a theoretical perspective, our primary contribution
lies in documenting consumer response to the invocation
of science in marketing appeals. Although prior literature
in marketing has explored reference to specific scientific
processes in marketing appeals (e.g., genetic engineering;
Hingston and Noseworthy 2018; Zheng, Bolton, and Alba
2019), no prior research, to our knowledge, has explored
how consumers respond when brands reference the use of
science or the scientific process more broadly as part of a
marketing appeal. Our research therefore addresses a novel
question to the consumer behavior literature: How do con-
sumers respond to brands that invoke science in the mar-
keting of consumer products? We answer this question and
reveal the consequences of consumers’ lay beliefs about
the scientific process for brands that do so. We also con-
tribute to the literature on consumer lay theories (Cheng,
Mukhopadhyay, and Schrift 2017; Deval et al. 2013;
JOURNAL OF CONSUMER RESEARCH
Raghunathan, Naylor, and Hoyer 2006; Zane, Smith, and
Reczek 2020) by introducing a novel lay belief about a do-
main that is becoming more relevant as consumers increas-
ingly rely on products produced by science even while
having mixed feelings about science itself. Finally, our
work also offers practical guidance to marketers on when
and why science appeals may backfire in the marketplace.
SCIENCE APPEALS
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Over the last half-century, the invocation of science in
marketing communications has proliferated (Lellis 2016).
Marketers often describe how a product has been scientifi-
cally developed in their promotions, on product packaging,
and on brand websites. Of course, invoking science does
not always prove effective, and past research in marketing
suggests it can backfire. For example, research has shown
that reference to the use of genetic modification in food
products (including packaged food products) can elicit
moral opposition and consumer resistance (Hingston and
Noseworthy 2018; Pham and Mandel 2019; Scott et al.
2018; Zheng et al. 2019). A related but distinct stream of
research shows reduced purchase intentions for products
that consumers view as unnatural (Scott, Rozin, and Small
2020; Scott et al. 2018). While both these streams of re-
search suggest that invoking science in marketing appeals
can potentially be problematic for marketers, our work is
distinct from this work in two ways.
First, while there is a rich literature in marketing that
explores consumer response to product claims that are ex-
plicitly based on science, this literature explores consumer
response to the specific information in the claim (e.g., that
a functional food offers a particular health benefit in
Naylor, Droms, and Haws 2009 or that a product can en-
hance a central vs. peripheral individual trait in Riis,
Simmons, and Goodwin 2008) or to the specific scientific
process used to make the product (e.g., genetic modifica-
tion in Hingston and Noseworthy 2018; Pham and Mandel
2019; Scott et al. 2018; Zheng et al. 2019). We are instead
interested in exploring the effectiveness of marketing
appeals that do not necessarily make a specific scientific
claim or reference a specific scientific process but instead
associate the brand/product with science more broadly.
That is, we test the effect of making the claim that some as-
pect of the product was developed using science versus
promoting the same aspect of the product with no reference
to science. Second, in the work exploring consumer re-
sponse to natural versus unnatural products, a natural prod-
uct is defined as a product that has not had “any previous
human intervention and has no additives” (Scott et al.
2020, 454). In our research, we focus exclusively on manu-
factured consumer products (i.e., CPGs that would not ex-
ist without human intervention). In other words, we
explore the effects of invoking science in marketing
PHILIPP-MULLER, COSTELLO, AND RECZEK
products that the research on perceived artificiality of prod-
ucts would define as inherently “unnatural” because they
are produced using human intervention. Thus, we are inter-
ested in the backfiring of science appeals beyond reasons
of unnaturalness (although we empirically rule out artifici-
ality as an alternative explanation for the effects we ob-
serve in our studies).
CONSUMER LAY BELIEFS ABOUT THE
SCIENTIFIC PROCESS
To understand consumer response to marketing appeals
that explicitly invoke science, we explore consumers’ lay
beliefs regarding the scientific process. Lay beliefs (i.e.,
knowledge structures that reflect an individual’s naı̈ve un-
derstanding of the world) have been shown to affect a vari-
ety of product judgments in past research (Cheng et al.
2017; Deval et al. 2013; Raghunathan et al. 2006; Wang,
Tat Keh, and Bolton 2010), even when these beliefs are not
objectively accurate (Haws, Walker Reczek, and Sample
2017; Kramer and Block 2011). We focus on consumers’
lay beliefs about the scientific process.
Though no work, to our knowledge, has focused on lay
beliefs regarding the scientific process, research examining
stereotypes about scientists has found that scientists are
perceived as high in agency but low in communion (Brown
et al. 2018; Diekman et al. 2011; McPherson, Park, and Ito
2018).1 Prior research in marketing has demonstrated that
warmth and competence, identified as the two broad cate-
gories under which most stereotypes about people and
groups fall (Cuddy et al. 2008), apply not only to person
perception but also to nonhuman entities like organizations
(Aaker et al. 2010; Lee, Bolton, and Winterich 2017). We
therefore suspect that the scientific process is likely per-
ceived similarly to the scientists working in that domain,
such that the scientific process will be perceived as cold,
yet competent. As such, we propose that because consum-
ers have the lay belief that science is cold, they will de-
value products they expect to be warm when science is
invoked.
(MIS)MATCH BETWEEN SCIENCE
APPEALS AND HEDONIC PRODUCTS
Lay beliefs are a form of knowledge structure that can
serve as a source of product expectations that affect pur-
chase intentions (Luchs et al. 2010; Raghunathan et al.
1 We note that stereotypes center on the trait associations individuals
have about people or entities and are hence a type of lay belief. The
term “lay belief” is therefore a broader term than stereotype since lay
beliefs can be about anything (e.g., products, processes, etc.) and not
just about trait associations regarding people or entities. However, the
terms “lay belief” and “stereotype” are often used interchangeably in
the psychology literature when discussing beliefs about people’s traits
(Krems et al. 2021).
723
2006). In fact, past research has shown that when a prod-
uct’s attributes conflict with a lay belief about the relation-
ship between those attributes, consumers engage in
increased information search because the information is
discrepant from what they expect (Haws et al. 2017). We
propose that the lay belief that science is cold yet compe-
tent creates expectations that can either match or mismatch
with products that are positioned as hedonic or utilitarian.
Consumers tend to associate hedonic brands, products, and
attributes (which focus on positive sensory and affective
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
experiences) with higher levels of warmth, whereas they
associate utilitarian brands, products, and attributes (which
focus on instrumental outcomes and fulfilling a particular
function) with higher levels of competence (Chattalas and
Takada 2013; Dhar and Wertenbroch 2000; Hirschman and
Holbrook 1982; Peter and Ponzi 2018; Strahilevitz and
Myers 1998). We therefore propose that products posi-
tioned as hedonic mismatch with consumers’ expectations
that science is cold. It is important to note that this mis-
match may not be obvious to marketers, as science is still
commonly used to promote hedonic products in the mar-
ketplace. Indeed, a pilot study we conducted with 208
MTurk workers (Mage: 38.4, 48.10% female) found that
79% of participants had personally encountered science
appeals for hedonic products (see web appendix C for
details).
This sense of feeling like there is a mismatch between
science appeals and products positioned as hedonic can be
more formally labeled as a lack of conceptual fluency.
Conceptual fluency2 has been defined in the consumer be-
havior literature as the extent to which stimuli are consis-
tent with one’s expectations (e.g., using an image of people
in a bar when promoting a beer is conceptually fluent be-
cause we expect to see beer in a bar; Lee and Labroo 2004)
and has been shown to increase a message’s persuasive
power (Duhachek et al. 2012; Kidwell et al. 2013;
Thompson and Hamilton 2006). In the present work, we
predict that since consumers have the lay belief that sci-
ence is competent but not warm, when they encounter sci-
ence paired with a hedonic product, it does not match their
expectations and is hence conceptually disfluent.
Conceptual disfluency occurs when there is a mismatch
with expectations. For example, when an organic label is
applied to a vice food, that label is perceived to be at odds
with the product category and is therefore less conceptually
fluent compared to when the same label is applied to a vir-
tue food (Parker et al. 2021). Just as conceptual fluency
increases product valuation, conceptual disfluency reduces
product valuation (Lee and Labroo 2004; Parker et al.
2 Conceptual fluency is based on a stimulus’ meaning and the expect-
ations one has about that stimulus. In contrast, perceptual fluency is
based on perceptual elements (such as appearance). For example, an
ad would be perceptually (but not conceptually) disfluent if it used a
font that was difficult to read.
724
2021), similar to the dampening effects on product evalua-
tions when something does not “feel right” (e.g., a non-
rounded number in Wadhwa and Zhang 2015).
Thus, when science is invoked to market warm, hedonic
products, we predict consumers should be less interested in
those products.3 On the other hand, when science is in-
voked to market utilitarian products, conceptual fluency
should be increased because the functional, competent ori-
entation of utilitarian products is consistent with the lay be-
lief that science is competent, resulting in increased
product valuation. Our predictions thus share some com-
monality with past work on persuasion that has found that
when promotional appeals use language that matches fea-
tures of the product, consumers tend to be more persuaded
and likely to purchase that product (e.g., utilitarian vs. he-
donic appeals are more successful when promoting utilitar-
ian products; Chitturi, Raghunathan, and Mahajan 2008;
Klein and Melnyk 2016). This type of direct matching also
likely enhances conceptual fluency. The primary difference
in our conceptualization relative to this existing work is
that, while past work has shown dampened product valua-
tions from direct mismatches between the appeal and the
product, we are proposing that expectations stemming
from the lay belief that science is cold yet competent indi-
rectly produce the mismatch and sense of disfluency that
drives our effect. Additionally, that past work has not
looked at when science mismatches with products. More
formally, we predict the following:
H1a: When a product is positioned as hedonic, mentioning
the use of science in marketing appeals will reduce product
valuation relative to not invoking science.
H1b: When a product is positioned as utilitarian, mention-
ing the use of science in marketing appeals will increase
product valuation, relative to not invoking science.
H2: The effect of appeal type on product valuation proposed
in hypothesis 1 will be mediated by the appeal’s conceptual
fluency.
MODERATING THE SCIENCE BACKFIRE
EFFECT
The Salience of Science in Developing
Consumer Products
According to our theorizing, the science backfire effect
occurs due to a mismatch between the lay belief that the
3 To manipulate whether a product is perceived as hedonic or utilitar-
ian in our studies, we highlight either a hedonic or utilitarian attribute
of the target product while holding the product constant (e.g., for a
shampoo one could highlight its cleansing efficacy [utilitarian] vs.
silky feel [hedonic]; Dhar and Wertenbroch 2000). While a given
product may have both utilitarian and hedonic attributes, making a
specific attribute focal in a marketing appeal influences consumers’
holistic evaluation of the product as hedonic or utilitarian.
JOURNAL OF CONSUMER RESEARCH
scientific process is cold yet competent and the expectation
that hedonic products are warm. This theorizing suggests
two moderators for this backfire effect: (1) altering the
associations consumers have with a specific hedonic prod-
uct and (2) altering belief in the science is cold yet compe-
tent lay belief. We first focus on altering associations
consumers have with hedonic products. In our experiments,
we focus on product categories where science is not natu-
rally highly salient, as these are the products for which we
predict science appeals will backfire. Specifically, we fo-
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
cus on CPGs.4 However, there are many hedonic, yet
highly technical product categories where we would expect
the role of the physical sciences or engineering in produc-
tion to be highly salient. For example, most consumers pre-
sumably understand that finely engineered shock absorbers
are required for a smooth ride in a vehicle. We would
therefore not expect consumers to balk at a hedonic sports
car that was developed by engineers using a scientific pro-
cess: When it is highly salient to consumers that science is
necessary to create a given hedonic product, science
appeals should match consumers’ expectations.
However, even for a CPG where science is not naturally
salient, we predict that by increasing the salience of the ne-
cessity of science to create a given hedonic product, the
conceptual fluency of science appeals can be increased.
Providing information indicating, for example, that a given
hedonic CPG cannot be produced without science, should
mitigate the backfire effect predicted in hypothesis 1a by
associating science with that product and hence making
science appeals more expected and hence more conceptu-
ally fluent. Such findings would provide additional evi-
dence for our proposed mechanism by demonstrating
process via moderation (Spencer, Zanna, and Fong 2005).
More formally, we therefore predict:
H3: If the need for science when creating a given hedonic
product is made salient to consumers, the backfire effect for
hedonic products proposed in hypothesis 1a will be
attenuated.
Individual Differences Moderating the Science
Backfire Effect
In addition to altering the associations consumers have
with a hedonic product, the science backfire effect can be
mitigated if consumers do not see science as cold.
Although it may be difficult for marketers to alter the lay
belief that science is cold yet competent, consumers can
4 Although science is widely used to develop many CPGs like pack-
aged foods, beverages, and personal care products (Goddard and
Gruber 1999; Singh and Singh 2005), we suspected (and showed in a
pilot study on MTurk; details in web appendix C) that the use of sci-
ence in these product categories is not as salient to consumers as it is
in product categories seen as more technical (e.g., cars, electronic
devices, prescription drugs).
PHILIPP-MULLER, COSTELLO, AND RECZEK
differ in the extent to which they naturally endorse a given
lay belief (Cheng et al. 2017; Mukhopadhyay and Johar
2005; Raghunathan et al. 2006). We propose that individu-
als who do not explicitly endorse the lay belief that science
is cold will not devalue hedonic products that invoke sci-
ence since, without this belief, science appeals will not be
conceptually disfluent. Such a finding would also provide
further evidence for our proposed process. We therefore
predict:
H4: Those who do not hold the lay belief that science is
cold should not exhibit the effect described in hypothesis
1a.
In addition to explicitly measuring belief in the lay the-
ory to test this hypothesis, we also test it indirectly with
two other factors that theoretically ought to impact the ex-
tent to which individuals hold this lay belief. Specifically,
we explore whether participants’ trust in scientists5 and
whether the individual works in a STEM (Science,
Technology, Engineering, and Mathemathics) field moder-
ate the science backfire effect described in hypothesis 1a.
Past work finds that trust is highly related to perceptions of
warmth (Fiske, Cuddy, and Glick 2007), and the warmth/
coldness stereotype derives from initial perceptions of
trustworthiness (Williams and Bargh 2008). Therefore, we
expect that the backfire effect will attenuate for consumers
with high levels of trust in scientists, as they should be less
likely to believe that science is cold. We also examine
whether the effect attenuates for individuals who work in a
scientific industry sector (i.e., a STEM field), as these indi-
viduals should ether be scientists themselves or have more
experience with science and scientists. Past work on con-
tact theory (Allport 1954) suggests that when people work
with members of different groups in a collaborative con-
text, they are less likely to stereotype them. We therefore
predict that individuals working in STEM industries should
not hold the same stereotypes (or lay beliefs) about scien-
tists and the scientific process as lay people and thus
should not show the same backfire effect for hedonic prod-
ucts that involve science.
OVERVIEW OF STUDIES
We first report the results of a pilot study directly assess-
ing consumers’ lay beliefs about the scientific process,
which demonstrates that, among the general population, it
is viewed as competent, but cold. Studies 1a, 1b, and 1c
then test hypothesis 1a, demonstrating that actual choice
5 We measure trust in scientists (rather than trust in the scientific pro-
cess) because it allows us to utilize established measures from a na-
tionally administered Pew Research Center poll about Americans’
trust in scientists. Though technically distinct constructs, we conjec-
ture that perceptions of scientists and the scientific process are related.
We elaborate on this issue and potential sources of the lay belief in
more depth in the General Discussion.
725
share (study 1a) and willingness to pay (WTP) (study 1b)
are reduced when a hedonic product is promoted with a sci-
ence appeal (vs. a control appeal with no mention of sci-
ence). Study 1c further demonstrates that the backfire
effect we observe is uniquely driven by referencing sci-
ence, as an appeal referencing rigor (but not science) in a
product development process does not produce the same
negative effect on product valuation. Study 2 provides evi-
dence that the science backfire effect for hedonic products
is driven by conceptual disfluency (hypothesis 2). Studies
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
3a and 3b test both hypotheses 1a and 1b in the context of
personal lubricant (study 3a) and body wash (study 3b),
demonstrating that, although science (vs. control) appeals
backfire when the product is positioned as hedonic, science
(vs. control) appeals lead to increased purchase intentions
when the product is positioned as utilitarian. The results of
studies 3a and 3b also show evidence of moderated media-
tion via conceptual fluency. Consistent with hypothesis 3,
in study 4, we demonstrate that when the salience of the
necessity of science to produce a hedonic product is in-
creased, a science appeal does not reduce purchase inten-
tions. Finally, we test hypothesis 4 in studies 5a–5c,
demonstrating that those who do not see science as cold
(study 5a), who trust scientists a great deal (study 5b), or
who work in a STEM field (study 5c) do not exhibit the
science backfire effect. Full stimuli for each study are
reported in web appendix B, as are additional details
around data exclusions. Inclusion of all responses does not
change the significance of our findings in any studies (web
appendix D).
LAY BELIEF PILOT STUDY
A total of 312 MTurk workers (Mage: 40.60, 51.60% fe-
male) completed this pilot study in exchange for monetary
compensation. Participants were presented with a brief de-
scription of both warmth and competence adapted from
existing warmth-competence research in marketing and
psychology (Aaker, Garbinsky, and Vohs 2012; Aaker
et al. 2010; Judd et al. 2005) and then rated the scientific
process in reference to the production of consumer goods
on two nine-point semantic differential scales anchored by
“4, Cold” and “þ4, Warm,” and “4, Incompetent” and
“þ4, Competent,” respectively. They then provided demo-
graphic information (e.g., age, gender) and answered an
open-ended data validity indicator used to identify bots and
individuals using virtual private servers (Chmielewski and
Kucker 2020; Dennis, Goodson, and Pearson 2020). A re-
peated measures ANOVA revealed that, as predicted, the
scientific process was more associated with competence
(M ¼ 3.03) than warmth (M ¼ 1.26; F(1, 311) ¼ 131.57,
p < .001, gp2 ¼ 0.30). Further, the mean evaluation of the
scientific process on the competence/incompetence dimen-
sion was higher than the midpoint of the scale (M ¼ 3.03,
726
F(1, 311) ¼ 1753.10, p < .001, gp2 ¼ 0.58), while the
warmth/coldness of the scientific process was significantly
below the midpoint (M ¼ 1.26, F(1, 311) ¼ 76.74, p <
.001, gp2 ¼ 0.06). Consistent with our theorizing, these
results suggest that consumers view the scientific process
as competent but cold.
STUDIES 1A–1C
Study 1a
In this study, we tested hypothesis 1a by examining par-
ticipants’ real choice of a hedonic chocolate chip cookie. A
total of 511 undergraduate students (Mage: 19.95, 51.72%
female) completed study 1a in exchange for course credit.6
Participants were assigned to one of two conditions in
which they were told that the researchers had partnered
with a baked goods company to examine interest in differ-
ent cookies. Participants were told they would receive the
cookie they selected. All participants received a pen and a
paper menu describing three cookie options that varied
depending on condition. In the control condition, the three
options were: “Option A—Luscious chocolatey taste,”
“Option B—Our Most scrumptious cookie,” and “Option
C—Loads of ooey-gooey chocolate chunks.” In the science
condition, the options were identical except the first cookie
was listed as “Option A—Scientifically developed to have
a luscious chocolatey taste.” A pretest (n ¼ 202 undergrad-
uates, Mage: 21.56, 40.87% female) revealed that partici-
pants predominantly see chocolate chip cookies described
as having a “luscious chocolatey taste” (the target attribute
across both conditions) as a hedonic (vs. utilitarian) prod-
uct (M ¼ 1.90, significantly different from the midpoint of
the seven-point scale where 1 ¼ predominantly hedonic
and 7 ¼ predominantly utilitarian; F(1, 201) ¼ 483.56, p <
.001, see web appendix C for reporting of manipulation
checks and pretests for all other studies). Participants then
selected which cookie they wanted and reported their de-
mographic information. At the end of the study session,
participants went to a separate room where they presented
a research assistant with the menu on which they had cir-
cled their choice of cookie. They then received a cookie
from a labeled container that corresponded with their
selection.
To test hypothesis 1a, we conducted binary logistic re-
gression with condition (coded as control ¼ 1; scien-
ce ¼ 1) predicting whether participants selected cookie
Option A or not (Option A ¼ 1; Option B or C ¼ 0). In the
6 The study was conducted simultaneously at two large, Midwestern
universities in the United States using similar populations (undergrad-
uate business students). We made the choice to collect data at both
institutions over the same time period to achieve adequate statistical
power despite ongoing challenges around participant recruitment due
to the COVID-19 pandemic. We do not observe any interaction be-
tween experimental condition and data collection location (p ¼ .870).
JOURNAL OF CONSUMER RESEARCH
condition where the description of Option A included a ref-
erence to science, Option A was chosen significantly less
often compared to the control condition (21.18% vs.
30.47%; b ¼ .25, 2(1) ¼ 5.71, p ¼ .017, u ¼ 0.11). In
terms of percentage change [(Science  Control)/Control],
adding a science appeal decreased the likelihood of choos-
ing cookie Option A by 30.48%. Thus, study 1a shows that
a science appeal significantly reduces consumers’ choice
of the cookie.
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Study 1b
Study 1b (preregistration available at https://aspredicted.
org/blind.php?x=c6hy7w) conceptually replicated study 1a
using another measure of product valuation, incentive com-
patible WTP. A total of 809 MTurk workers (Mage: 40.51,
53.40% female) completed study 1b in exchange for mone-
tary compensation. Participants were assigned to one of
two conditions (Appeal Type: Control vs. Science) in
which they read a slogan for an indulgent chocolate chip
cookie. Participants were told that the cookies were from a
real cookie brand that we dubbed “Zoza” due to the brand’s
purported request to remain anonymous. All participants
saw an image of an indulgent cookie and then read a slogan
that varied by condition. In the science appeal condition,
participants read, “Our rigorous scientific development
process ensures that Zoza cookies taste delicious, indul-
gent, and gooey,” and, in the control condition, participants
read, “We ensure that Zoza cookies taste delicious, indul-
gent, and gooey.”
After viewing the slogan, participants reported their
WTP for a case of cookies from Zoza on a sliding scale
ranging from $0 to $20. We made this WTP measure in-
centive compatible by employing a procedure used in prior
research (Costello and Walker Reczek 2020; Fuchs,
Schreier, and van Osselaer 2015) that involves an initial
lottery followed by the standard BDM procedure (Becker,
DeGroot, and Marschak 1964). Participants were told that
if chosen from the lottery, they would get the opportunity
to purchase a case of 30 packs of cookies (each pack was
described as containing three cookies, for a total of 90
cookies per case) and that the winner of this lottery would
be subject to the BDM procedure based on their stated
WTP. Participants next provided demographic information
and answered two open-ended data validity indicators.
Finally, participants provided contact information in case
they were selected as winners. The selected winners then
received their prizes based on the outlined procedure.
A one-way ANOVA predicting incentive-compatible
WTP from appeal type revealed that participants were will-
ing to pay less for the cookies in the science appeal (M ¼
$9.04) than the control condition (M ¼ $9.97; F (1, 807) ¼
6.04, p ¼ .014, gp2 ¼ 0.01). In percentage change terms,
including a science appeal decreased WTP by 9.33%.
Thus, study 1b conceptually replicated the findings of
PHILIPP-MULLER, COSTELLO, AND RECZEK
study 1a with a different operationalization of product val-
uation, demonstrating that consumers are willing to pay
less for a case of hedonic cookies when they are promoted
using a science appeal.
Study 1c
The goal of study 1c was to demonstrate that the back-
fire effect for hedonic positioning is unique to explicitly
referencing science (vs. a rigorous development process
that does not invoke science) because it is driven by
expectations stemming from the lay belief that science is
cold. Going beyond the existing affective/cognitive match-
ing literature (Fabrigar and Petty 1999; Klein and Melnyk
2016), we therefore sought to demonstrate that it is not
simply the case that any cognitive appeal would backfire
for a hedonically positioned product simply because the
product is seen as more affective in nature. Thus, we tested
whether a cognitive appeal that referenced the product’s
rigorous development process would backfire or whether
there was something unique about the coldness that scien-
tific cognitive appeals confer.
A total of 587 MTurk workers (Mage: 35.38, 48.04% fe-
male) were presented with a description of an indulgent
cookie brand. Participants were presented with one of three
appeals promoting the brand’s cookies: control (“Our prod-
uct creation process ensures that Zoza cookies taste deli-
cious, indulgent, and gooey”), nonscience rigor (“Our
rigorous development process ensures that Zoza cookies
taste delicious, indulgent, and gooey”), and science (“Our
science-based rigorous development process ensures that
Zoza cookies taste delicious, indulgent, and gooey”).
Participants then reported their purchase intentions using
the following scale: “Based on the slogan above, how
likely would you be to purchase Zoza cookies?” (1—
Extremely Unlikely; 7 ¼ Extremely Likely). Finally, par-
ticipants reported demographic information.
To test whether appealing to rigor would yield the same
backfire effect as appealing to scientific rigor specifically,
we employed orthogonal contrast coding, which allowed
us to test our focal predictions directly (Rosenthal,
Rosnow, and Rubin 2000). We compared the science con-
dition (coded as þ2) to the rigorous and control conditions
(both coded as 1). We also examined the rigorous condi-
tion (coded as 1) compared to the control condition (coded
as 1). As predicted, purchase intentions were signifi-
cantly lower in the science appeal (M ¼ 4.02) compared to
the control (M ¼ 4.58) and nonscience rigor conditions
(M ¼ 4.34; F (1, 584) ¼ 10.60, p ¼ .001, gp2 ¼ 0.02).
Purchase intentions were not significantly different be-
tween the nonscience rigor and control conditions (F (1,
584) ¼ 2.23, p ¼ .136, gp2 ¼ 0.004). Thus, study 1c dem-
onstrates that it is not the case that referencing any type of
rigorous development process in an appeal for a hedonic
product produces a backfire effect, but, rather, only an
727
appeal referencing science does so. Additionally, this study
provides evidence that the science backfire effect goes be-
yond cognitive/affective matching because simply appeal-
ing to a cognitive attribute (i.e., a rigorous development
process) was insufficient to yield a backfire effect com-
pared to the control condition.
STUDY 2
In study 2 (preregistration https://aspredicted.org/blind.
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
php?x=n4te2g), we test hypotheses 1a and 2, this time in
the context of a hedonic smoothie rather than cookies. We
measured conceptual fluency in addition to several other
potential mediators to rule out alternative explanations.
Specifically, because hedonic products that are scientifi-
cally developed may be perceived to be more artificial
(Scott and Rozin 2017), we measured perceived artificial-
ity using established measures from the literature. Further,
because scientifically developed food products may be
seen as healthier, more conducive to weight loss, less tasty
(Andre et al. 2019), or less plausible, we also measured
these perceptions.
Participants and Procedure
A total of 402 participants (Mage ¼ 41.02, 50.00% fe-
male) completed study 2 on MTurk in exchange for pay-
ment. All participants saw an image of an indulgent
smoothie brand (named “JTB smoothies”) prior to seeing
the ad copy. A pretest (n ¼ 101 MTurk workers) confirmed
that the smoothie was viewed as hedonic (see web appen-
dix C for details). In the science appeal condition, partici-
pants read “Our rigorous scientific development process
ensures that JTB smoothies taste delicious, indulgent, and
creamy,” and, in the control condition, participants read,
“We ensure that JTB smoothies taste delicious, indulgent,
and creamy.”
All participants then reported their intentions to pur-
chase the smoothie and the conceptual fluency of the slo-
gan on the following five items (a ¼ 0.93) adapted from
past work (Kidwell et al. 2013; Lee and Aaker 2004;
Rocklage and Fazio 2020): “The slogan seemed to flow
well,” “The slogan seemed disjointed,” “Something
seemed weird about the slogan (1 ¼ completely disagree,
7 ¼ completely agree, reverse coded for the latter two
items), and the extent to which the slogan was fitting for
the product (1 ¼ not fitting at all, 7 ¼ completely fitting,)
and a good match for the product (1 ¼ The slogan was not
a good match at all, 7 ¼ The slogan was a very good
match). Unless otherwise specified, all measures in subse-
quent studies used seven-point scales where higher num-
bers indicated greater agreement with the item.
Participants also rated perceived artificiality using three
items taken from existing research (Rozin 2005; Scott and
Rozin 2017; a ¼ 0.89): “To what extent do you think JTB
728
smoothies are artificial?,”7 “How natural do you think JTB
smoothies are?” (reverse coded), and “To what extent do
you think JTB smoothies have additives?”). Finally, we
measured the extent to which participants thought the
smoothie was tasty, healthy, and conducive to weight loss
(from Andre et al. 2019), as well as whether the slogan was
plausible, that is it “seem[ed] like a real slogan companies
might use” (1 ¼ “Not at all”; 7 ¼ “Very much so”). We
measured plausibility because we propose that the effect of
invoking science on product valuation is driven by the con-
ceptual fluency of the appeal (i.e., how much the appeal
matches consumers’ expectations) and not the extent to
which participants thought a brand would never use a cer-
tain type of appeal. Participants then answered demo-
graphic questions and two data validity indicators. No
participants failed our data validity indicators, leaving 402
valid responses.
Results and Discussion
Purchase Intentions. We subjected the data to a one-
way ANOVA comparing the two conditions (science vs.
control) on purchase intentions. Conceptually replicating
the findings of studies 1a–1c we found that, compared to
the control condition (M ¼ 5.09), those in the science con-
dition (M ¼ 4.37) reported reduced purchase intentions for
the smoothie (F(1, 400) ¼ 19.06, p < .001, gp2 ¼ 0.05).
Conceptual Fluency Mediation. We ran a parallel me-
diation analysis using model 4 of PROCESS (Hayes 2017)
to test the mechanisms driving the effect of appeal type on
purchase intentions. We set appeal type as the independent
variable (control ¼ 1, science ¼ 1), purchase intentions as
the dependent variable, and fluency, plausibility, artificial-
ity, tastiness, healthiness, and conduciveness to weight loss
as parallel mediators8 (figure 1). The indirect effect of ap-
peal type on purchase intentions through fluency was sig-
nificant (indirect effect ¼ 0.42, 95% CI: [0.57, 0.29]),
as predicted in hypothesis 2, even while controlling for the
alternative mediators. The indirect effects were not signifi-
cant for perceived artificiality (indirect effect ¼ 0.03, 95%
CI: [0.01, 0.07]); weight loss potential (indirect effect ¼
0.02, 95% CI: [0.001, 0.05]); healthiness (indirect effect
¼ 0.001, 95% CI: [0.02, 0.03]); or appeal plausibility (in-
direct effect ¼ 0.05, 95% CI: [0.13, 0.03]). Taste signif-
icantly mediated, though the effect was much weaker than
that of fluency (indirect effect ¼ 0.04, 95% CI: [0.09,
7 This item was anchored on a scale from 0 to 6 to remain consistent
with the original item from Rozin (2005). However, we recoded the
responses from 1 to 7 prior analysis and combining with the other two
artificiality measures.
8 The results are robust to parallel analyses wherein each alternative
explanation is individually contrasted to fluency (i.e., two mediators in
each analysis). Fluency mediates to a significantly greater degree than
each of the alternative mediators (see web appendix D for these
analyses).
JOURNAL OF CONSUMER RESEARCH
0.01]). To test the relative indirect effect size of taste ver-
sus fluency in our parallel mediation model, we also con-
ducted a pairwise comparison using the “contrast”
command through the PROCESS macro (Hayes 2017).
When the indirect effects of taste and fluency were directly
contrasted, fluency mediated to a significantly greater de-
gree (95% CI: [0.53, 0.24]). See web appendix D for a
full reporting of means, ANOVAs, and mediation analysis
for these additional variables.
Discussion. Study 2 provides support for both hypothe-
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
ses 1a and 2. The results of study 2 also demonstrated that
fluency significantly mediated the effect of appeal type on
purchase intentions even when controlling for artificiality
and other potential mediators, thus ruling out alternative
explanations for the effect.
STUDY 3A
Study 3a tested hypotheses 1a, 1b, and 2 in a new non-
consumable product category, personal lubricant, and ex-
tended the conceptual fluency mediation findings of study
2 to a moderated mediation context. This study examined
the impact of appeal type (science vs. control) and product
positioning (hedonic vs. utilitarian) on purchase intentions
and measured perceived brand trustworthiness and brand
competence as alternative explanations.
Participants and Procedure
Participants in study 3a were MTurk workers who com-
pleted the study in exchange for monetary compensation.
In one of the conditions, the appeal described the lubricant
as possessing a utilitarian attribute (adding a layer of birth
control), which may not be desired or necessary for a siz-
able proportion of people. Therefore, participants were
only invited to complete the main survey if they reported
that they used birth control (“Do you or your romantic part-
ner ever use any kind of product that provides birth control
(e.g., condoms, birth control pill etc.)?”) on a brief screener
survey in which the birth control question was embedded
among other decoy questions (see web appendix B for
complete screener item wordings). Participants determined
to be eligible based on their responses were invited to the
main study. In total, 900 participants (Mage: 38.11, 61.09%
female) completed the study. Participants were excluded if
they failed a data validity indicator (31 participants) and/or
stated that they did not actually use birth control in the
main study despite stating that they or their romantic part-
ner used birth control in the original screener (49 partici-
pants). An additional participant’s data were excluded
because they took the survey twice. The final sample size
was therefore 823.
Participants were presented with an appeal promoting a
personal lubricant (named “Noma”) in a 2 (Appeal:
PHILIPP-MULLER, COSTELLO, AND RECZEK
FIGURE 1
PARALLEL MEDIATION MODEL.
*p < .10, **p < .05
Science vs. Control)  2 (Positioning: Hedonic vs.
Utilitarian) between-subjects design. In one between-
subjects factor, half the participants were told to imagine
they had the goal of finding a lubricant that enhanced sex-
ual pleasure (the hedonic condition), and the other half
were told to imagine they were searching for a lubricant
that would add an additional layer of birth control protec-
tion (the utilitarian condition). Participants in the hedonic
condition then saw an appeal that positioned the lubricant
as hedonic (increasing sexual pleasure), whereas partici-
pants in the utilitarian condition saw an appeal that posi-
tioned the lubricant as utilitarian (increasing birth control
protection). In the other between-subjects factor, half the
participants saw an appeal that invoked science, and the
other half saw a control appeal that did not mention sci-
ence. Participants then reported their likelihood of purchas-
ing this lubricant using the same purchase intentions item
and conceptual fluency measures (a ¼ 0.88) as study 2.
Participants also reported the extent to which they trusted
the lubricant brand using two items: (r ¼ 0.87; “How trust-
worthy do you think the makers of Noma lubricant are?”
and “How honest do you think the makers of Noma lubri-
cant are?”), as well as how competent they thought the
brand was (“How competent do you think the makers of
Noma lubricant are?”). Participants next reported whether
729
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
they or their romantic partner used birth control, completed
a data validity indicator, and responded to demographic
questions and several exploratory measures (reported in
web appendix B).
Results and Discussion
Purchase Intentions. We subjected the data to a two-
way ANOVA predicting purchase intentions from appeal
type (control ¼ 1, science ¼ 1), product positioning
(hedonic ¼ 1, utilitarian ¼ 1), and the two-way interac-
tion. There was no main effect of appeal type (F(1, 819) ¼
0.70, p ¼ .402, gp2 ¼ 0.001), but there was a main effect
of product positioning (F(1, 819) ¼ 6.69, p ¼.01, gp2 ¼
0.01), such that when the lubricant was positioned as utili-
tarian, purchase intentions were higher (M ¼ 4.61) com-
pared to when it was positioned as hedonic (M ¼ 4.33).
Most relevant to our hypotheses, there was a significant
two-way interaction (F(1, 819) ¼ 13.54, p < .001, gp2 ¼
0.02; figure 2). Simple effects revealed that when the lubri-
cant was positioned as enhancing sexual pleasure, the con-
trol appeal (M ¼ 4.48) led to higher purchase intentions
than the science appeal (M ¼ 4.17; F(1, 819) ¼ 4.05, p ¼
.045, gp2 ¼ 0.01). In contrast, when the lubricant was posi-
tioned as providing birth control, the science appeal
730
FIGURE 2
SCIENCE APPEALS INCREASE PURCHASE INTENTIONS FOR
LUBRICANT POSITIONED AS UTILITARIAN BUT DECREASE
PURCHASE INTENTIONS FOR LUBRICANT POSITIONED AS
HEDONIC
(M ¼ 4.86) led to higher purchase intentions compared to
the control appeal (M ¼ 4.36; F(1, 819) ¼ 10.17, p ¼ .001,
gp2 ¼ 0.01).
Conceptual Fluency Moderated Mediation. We next
tested the indirect effect of product positioning by appeal
type on purchase intentions, mediated by conceptual flu-
ency of the appeal using PROCESS model 7 (Hayes 2017),
which tests for moderated mediation. We set appeal type as
the independent variable, product positioning as the moder-
ator, purchase intentions as the dependent variable, and
conceptual fluency as the mediator (figure 3). The confi-
dence interval for the index of moderated mediation did
not cross zero (index ¼ 0.40, 95% CI: [0.26, 0.53]), indi-
cating significant moderated mediation via conceptual flu-
ency. We also tested whether the fluency moderated
mediation would remain when controlling for the alterna-
tive mediators of brand trustworthiness and competence. It
did, as the confidence interval for the index of moderated
mediation of conceptual fluency did not cross zero in this
analysis (index ¼ 0.19. 95% CI: [0.10, 0.29]).9
Discussion. Consistent with hypotheses 1a and 1b, the
science appeal backfired for a product positioned as
9 Additional details around this moderated mediation analysis and al-
ternative mediation models are available in web appendix D. In partic-
ular, there is a significant serial moderated mediation (PROCESS
model 83) in which fluency is the proximal mediator and brand trust
the distal mediator, consistent with work suggesting that more fluently
processed marketer-generated information may result in higher brand
trust (Kostyk, Leonhardt, and Niculescu 2017). Although this model
provides further insight into the mechanism of the science backfire ef-
fect, we note that the index of moderated mediation was substantially
smaller for this model compared to the more parsimonious model in
which conceptual fluency is the sole mediator (.06 for serial vs. .40 for
nonserial). We therefore discuss this analysis in the web appendix but
do not discuss this further in the main paper.
JOURNAL OF CONSUMER RESEARCH
hedonic but was effective for a product positioned as utili-
tarian. Consistent with hypothesis 2, these effects of appeal
type on purchase intentions were mediated by the concep-
tual fluency of the appeal. One potential limitation of this
study is that a birth control attribute in a lubricant may be
seen as novel by some consumers who have never used a
spermicide. In study 3b, we conceptually replicate the find-
ings of study 3a using a product that has commonly en-
countered attributes in all conditions.
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
STUDY 3B
In study 3b, we used a body wash as the focal product.
We selected this product because it possesses commonly
encountered hedonic and utilitarian attributes and thus may
be positioned as either. For example, body wash brands
sold on Walmart’s website are described using both he-
donic terms (e.g., “immerses all your senses in its lush lath-
er,” “rich lather leaves skin feeling silky”) and utilitarian
terms (e.g., “effectively washes away dirt and bacteria,”
“formulated to wash away bacteria”). In this study, we
model our product descriptions after these real examples.
We additionally measured perceived novelty of the product
as an alternative explanation, in addition to artificiality,
brand trustworthiness, and brand competence.
Participants and Procedure
In study 3b, 1,015 MTurk workers completed the study
in exchange for monetary compensation (Mage: 39.84,
55.68% female). Participants were excluded if they failed a
data validity indicator (two participants). The final sample
size was 1,013.
Participants were presented with an appeal promoting a
body wash (named “Ely”) using the same 2 (Appeal Type)
 2 (Product Positioning) design as in study 3a (see web
appendix C for a pretest of 204 MTurk workers demon-
strating that the hedonic appeal was perceived as hedonic
and the utilitarian appeal was perceived as utilitarian).
Participants were told that they would view a “product de-
scription displayed on the brand’s product packaging and
website.” Those in the hedonic condition saw an appeal
that described the lather as something that will “immerse
your senses in an indulgent experience,” whereas partici-
pants in the utilitarian condition saw an appeal that de-
scribed the lather as something that will “wash away odor
causing bacteria.”
All participants then answered the same purchase inten-
tions item and measures of conceptual fluency (a ¼ .092),
artificiality (a ¼ 0.86), brand trust (r ¼ 0.83), and brand
competence used in prior studies, as well as new measures
of perceived novelty measured on six-point scales (r ¼
0.70): “How novel is Ely Body Wash compared to other
body washes on the market?” and “How new is Ely Body
Wash compared to other body washes on the market?”.
PHILIPP-MULLER, COSTELLO, AND RECZEK
FIGURE 3
CONCEPTUAL FLUENCY MEDIATES THE APPEAL TYPE BY PRODUCT POSITIONING INTERACTION ON PURCHASE INTENTIONS
The order in which participants reported fluency, artificial-
ity, novelty, brand competence, and brand trust was ran-
domized. Finally, participants completed a data validity
indicator and responded to demographic questions.
Results and Discussion
Purchase Intentions. We subjected the data to a two-
way ANOVA predicting purchase intentions from appeal
type (control ¼ 1, science ¼ 1), product positioning
(hedonic ¼ 1, utilitarian ¼ 1), and the two-way interac-
tion. There was no main effect of appeal type (F(1, 1,009)
¼ 0.01, p ¼.927, gp2 < 0.001), but there was a main effect
of product positioning (F(1, 1,009) ¼ 13.60, p < .001, gp2
¼ 0.01), such that when the body wash was positioned as
utilitarian, purchase intentions were higher (M ¼ 4.69)
compared to when the body wash was positioned as he-
donic (M ¼ 4.35). Most relevant to our hypotheses, there
was a significant two-way interaction (F(1, 1,009) ¼
14.58, p < .001, gp2 ¼ 0.01; figure 4). Simple effects
revealed that when the body wash was positioned as he-
donic, the control appeal (M ¼ 4.52) led to higher purchase
intentions than the science appeal (M ¼ 4.19; F(1, 1,009) ¼
6.95, p ¼ .009, gp2 ¼ 0.01). In contrast, when the body
wash was positioned as utilitarian, the science appeal
(M ¼ 4.86) led to higher purchase intentions compared to
the control appeal (M ¼ 4.51; F(1, 1,009) ¼ 7.64, p ¼ .006,
gp2 ¼ 0.01).
Conceptual Fluency Moderated Mediation. We first
tested for moderated mediation via conceptual fluency.
Using PROCESS model 7, we set appeal type as the inde-
pendent variable, product positioning as the moderator,
purchase intentions as the dependent variable, and
731
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
conceptual fluency as the mediator. This analysis indicated
significant moderated mediation through conceptual flu-
ency (index ¼ 0.24, 95% CI: [0.14, 0.34]). We next ran the
same model but also included artificiality, novelty, brand
trustworthiness, and competence as parallel mediators
along with conceptual fluency. The confidence interval for
the index of moderated mediation continued to remain sig-
nificant for conceptual fluency (index ¼ 0.16, 95% CI:
[0.09, 0.23]), even when controlling for artificiality, nov-
elty, trustworthiness, and competence. However, neither
artificiality (index < 0.001, 95% CI: [0.01, 0.02]), nov-
elty (index ¼ 0.02, 95% CI: [0.002, 0.05]), trustworthi-
ness (index ¼ 0.04, 95% CI: [0.02, 0.10]) nor
competence (index ¼ 0.004, 95% CI: [0.02, 0.03]) medi-
ated the effect. Although the overall moderated mediation
analyses for studies 3a and 3b were both significant, results
around the specific indirect effects in this study were
slightly different than in study 3a. A theoretical discussion
for why this might be the case and additional details and
analyses are available in web appendix D. Supplementary
analysis (additional mediation analysis, including the serial
mediation with brand trust tested in study 3a, ANOVAs,
and means for mediator variables) are also available in
web appendix D.
Discussion. We once again found evidence that science
appeals backfire for products positioned as hedonic but not
utilitarian. We also found evidence of significant moder-
ated mediation through conceptual fluency while ruling out
a number of alternative explanations. Having found sup-
port for our first two hypotheses, we now shift our focus to
understanding the moderators of the backfire effect of sci-
ence appeals for hedonic products.
732
FIGURE 4
SCIENCE APPEALS INCREASE PURCHASE INTENTIONS FOR
BODY WASH POSITIONED AS UTILITARIAN BUT DECREASE
PURCHASE INTENTIONS FOR BODY WASH POSITIONED AS
HEDONIC
STUDY 4
To test hypothesis 3, we examined whether reading an
article about the importance of chemistry in baking (vs. a
control article) on a cookie brand’s website eliminates the
backfire effect of using a science appeal for a hedonic
cookie by making it salient that science is a necessary part
of cookie production. This study therefore used a 2
(Appeal Type: Science vs. Control)  2 (Salience of
Science’s Role in Production: Increased Salience vs.
Control) between-subjects design.
Participants and Procedure
A total of 814 participants (Mage: 40.68, 56.10% female)
completed study 4 on MTurk. Participants were excluded if
they failed the data validity indicator (seven participants).
In one between-subjects factor, half the participants read
an article explaining that chemistry is necessary for high
quality baking (increased science salience condition), and
the other half read an article explaining that knowing what
ingredients are in season is necessary for high quality bak-
ing (control condition). These articles were pretested on
MTurk (n ¼ 209) to ensure they led to significantly differ-
ent levels of salience about the necessity of science to cre-
ate cookies. We also conducted a separate pretest on
MTurk (n ¼ 201) indicating that the use of science is not
naturally salient in this product category, as we expected
(see web appendix C for details on both pretests).
After reading the articles, participants in the main study
answered two filler measures about the article (e.g., “How
appropriate would you say the length of this article was for
an online format?”) before moving on to the next part of
the study. Participants were then asked to “evaluate and
make decisions about a real chocolate chip cookie brand.”
JOURNAL OF CONSUMER RESEARCH
In the other between-subjects factor, half the participants
saw a science slogan, and the other half of the participants
saw a control slogan, using the same language as study 1b.
Participants next reported their purchase intentions using
the same measure used in prior studies. Finally, partici-
pants completed a data validity indicator and responded to
demographic and psychographic questions (e.g., “how
much do you like cookies?”).
Results
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Purchase Intentions. We subjected the data to a two-
way ANOVA predicting purchase intentions from appeal
type (control ¼ 1, science ¼ 1), salience of science’s role
in production (control ¼ 1, increased science sali-
ence ¼ 1), and the two-way interaction. There was a main
effect of appeal type (F(1, 803) ¼ 6.10, p ¼.014, gp2 ¼
0.01), such that purchase intentions were higher in the con-
trol (M ¼ 5.22) compared to the science (M ¼ 4.97) condi-
tion. There was also a main effect of science salience
condition (F(1, 803) ¼ 8.69, p ¼.003, gp2 ¼ 0.01), such
that when participants read about chemistry being a neces-
sary part of baking (M ¼ 5.24), purchase intentions were
higher compared to the control article (M ¼ 4.94). These
main effects were, however, qualified by a two-way inter-
action (F(1, 803) ¼ 10.71, p ¼ .001, gp2 ¼ 0.01; figure 5).
Probing the simple effects revealed that when participants
read the control article, the control appeal (M ¼ 5.23) led
to higher purchase intentions than the science appeal
(M ¼ 4.67; F(1, 803) ¼ 16.34, p < .001, gp2 ¼ 0.02), repli-
cating prior studies. However, when participants read that
chemistry is necessary for baking, there was no difference
in purchase intentions between the science appeal
(M ¼ 5.28) compared to the control appeal (M ¼ 5.20; F(1,
803) ¼ 0.32, p ¼ .569, gp2 < 0.001). Thus, the backfire ef-
fect of science appeals was attenuated when it was made
salient to consumers that science is a necessary part of a
hedonic product’s production process.
Discussion. Consistent with hypothesis 3, when con-
sumers were informed that science is necessary to create
baked goods, invoking science in a marketing appeal no
longer reduced purchase intentions. By making the science
behind the production of a hedonic food product salient,
the science appeal became more conceptually fluent (be-
cause it was more expected). Rather than measuring the
fluency of a science appeal, we manipulated how conceptu-
ally fluent science was perceived to be in producing he-
donic products, thus manipulating the mediator of our
proposed process (Spencer et al. 2005).
In studies 5a–5c, we now test hypothesis 4, which pre-
dicts that those who do not endorse the lay belief that sci-
ence is cold should not exhibit the effect described in
hypothesis 1a.
PHILIPP-MULLER, COSTELLO, AND RECZEK
FIGURE 5
SCIENCE APPEALS DO NOT DECREASE PURCHASE
INTENTIONS FOR COOKIES WHEN SALIENCE OF SCIENCE IN
BAKING IS HIGH
STUDY 5A–5C
Study 5a
In this study, we test whether the backfire effect we ob-
serve for hedonically positioned products in past studies is
moderated by participants’ endorsement of the lay belief
that science is cold. Because our theorizing posits that the
science backfire effect is driven by this lay belief, we ex-
pect that among those who do not see science as cold, the
backfire effect will attenuate.
A total of 503 participants (Mage: 39.78, 59.00% female)
completed study 5a on MTurk. Participants were assigned
to one of two conditions (Appeal Type: Control vs.
Science) in which they read the product description for the
label of a hedonically positioned body wash (see web ap-
pendix C for a pretest of n ¼ 264 MTurk workers demon-
strating that it was seen as hedonic). In the control
condition, the description was “Ely: Wash with a rich,
creamy lather,” while in the science condition the descrip-
tion was “Ely: Scientifically engineered to wash with a
rich, creamy lather.” Participants indicated their purchase
intentions for the body wash on the same scale used in
prior studies and then reported the extent to which they ex-
plicitly endorsed the lay belief that science is competent
but cold. Specifically, consistent with the lay belief pilot
study, they were given definitions of the warmth/coldness
stereotype and competence/incompetence stereotype
(Aaker et al. 2010, 2012; Judd et al. 2005) and indicated
how warm (4 cold to þ4 warm) and competent (4 in-
competent to þ4) they viewed the scientific process to be
(MCompetence ¼ 2.48, MWarmth ¼ 0.63). Finally, partici-
pants provided demographic information and completed a
data validity indicator.
No participants failed our data validity indicator, leaving
a sample of 503 responses. To test hypothesis 4, we
733
regressed purchase intentions on appeal type (con-
trol ¼ 1, science ¼ 1), the mean-centered endorsement of
the science is cold lay belief, and the two-way interaction.
There was a main effect of both appeal type (bunstandardized
¼ 0.19, t(499) ¼ 3.09, p ¼ .002, gp2 ¼ 0.13) and en-
dorsement of the lay belief (bunstandardized ¼ 0.19, t(499) ¼
8.04, p < .001, gp2 ¼ 0.33). Importantly, these two main
effects were also qualified by our predicted interaction
(bunstandardized ¼ 0.07, t(499) ¼ 2.78, p ¼ .006, gp2 ¼
0.11). Probing the interaction using floodlight analysis
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
(Spiller et al. 2013) via PROCESS model 1 (Hayes 2017)
indicated that, among those whose lay belief endorsement
is at or below 0.27 (on a scale where 4 ¼ cold and
þ4 ¼ warm; 65% of the sample), there was a significant re-
duction in purchase intentions for the body wash with a sci-
ence (vs. control) label. However, there is no difference
across the science and control conditions among those who
view science as relatively warm (figure 6A).
Our theory predicts that the backfire effect occurs for he-
donically positioned products because the coldness of sci-
ence conflicts with consumers’ beliefs about the warmth
stereotype dimension rather than the competence stereo-
type dimension. We therefore repeated the same analysis
using appeal type, mean-centered endorsement of the sci-
ence is competent stereotype, and the interaction of these
variables. Consistent with our theorizing, we do not find an
interaction between appeal type and endorsement of the
science is competent lay belief (bunstandardized ¼ 0.006,
t(499) ¼ .15, p ¼ .884, gp2 < 0.001). Thus, study 5a pro-
vides evidence that endorsement of the lay belief that sci-
ence is cold moderates the backfire effect of science
appeals, supporting our proposed theoretical account via
moderation.
Study 5b
In study 5b, our goal was to conceptually replicate the
findings of study 5a using a moderator that is theoretically
consistent with our proposed process, trust in scientists, but
also that is relevant to managers due to its known correlates
with various demographic variables identified via a poll
conducted by the Pew Research Center (an American
thinktank that conducts public opinion polling on a variety
of public interest topics; Funk 2020). Participants in study
5b (n ¼ 609, Mage: 40.44, 55.3% female) therefore partici-
pated in a study that was almost identical to study 5a, ex-
cept, rather than reporting their endorsement of science lay
beliefs, they reported their trust in scientists on three items
taken from the Pew poll (Funk 2020): “To what extent
would you say you have confidence in scientists to act in
the best interest of the public?,” “How often would you say
that scientists provide fair and accurate information?,”
“How often would you say that scientists are transparent
about conflicts of interest?,” which were all measured on
734
FIGURE 6
INVOKING SCIENCE WHEN MARKETING A HEDONIC BODY WASH BACKFIRES UNLESS CONSUMERS SEE SCIENCE AS
RELATIVELY WARM (A) OR HAVE HIGH LEVELS OF TRUST IN SCIENTISTS (B). (A) MODERATION BY ENDORSEMENT OF SCIENCE IS
COLD LAY BELIEF. (B) MODERATION BY TRUST IN SCIENTISTS.
(a)
6
5
Purchase Intentions
Purchase Intentions
4
3
Control
Science
2
−4 −2 0 2 4 2
Endorsement of Lay Intuition
NOTE.— Gray areas in figures indicate regions of significance
nine-point scales and combined into an index of trust in
scientists (a ¼ 0.91; M ¼ 6.10).
One participant failed our data validity indicator, leaving
a final sample of 608 responses. To test hypothesis 4, we
regressed purchase intentions on appeal type (same coding
as previous study), the mean-centered index measuring
trust in scientists, and the two-way interaction. There was a
main effect of both appeal type (bunstandardized ¼ 0.20,
t(604) ¼ 3.31, p ¼ .001, gp2 ¼ 0.13) and trust in scien-
tists (bunstandardized ¼ 0.24, t(604) ¼ 6.90, p < .001, gp2 ¼
0.27, Mtrust ¼ 6.10). Importantly, these two main effects
were also qualified by our predicted interaction (bunstandar-
2
dized ¼ 0.07, t(604) ¼ 2.78, p ¼ .043, gp ¼ 0.08). Follow
up floodlight analysis revealed a pattern consistent with
our theorizing and study 5a. We find that the science back-
fire effect occurs for the majority of our sample (69%), but
that the effect turns off for those individuals with high lev-
els of trust in scientists (i.e., those at 7.03 and higher on the
nine-point scale, figure 6B). Study 5b thus provides further
moderation evidence consistent with our theorizing.
JOURNAL OF CONSUMER RESEARCH
(b)
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
5
4
Control
Science
4 6 8
Trust in Scientists
Study 5c
In study 5c, we conceptually replicate the findings of
studies 5a and 5b, this time using the industry in which
consumers work as the moderating variable of interest.
Specifically, we ran the same study with two different pop-
ulations: those who work in STEM sectors and those who
do not, in order to test whether exposure to science and sci-
entists attenuates the science backfire effect we observe in
earlier studies. We predict that the backfire effect will
emerge for non-STEM professionals but will be eliminated
for individuals working in STEM fields.
Study 5c (n ¼ 602, Mage: 32.56, 52.30% male) used an
almost identical design to study 5a and 5b, except, instead
of measuring a moderating variable, we sampled partici-
pants from two different industries (STEM vs. non-STEM)
using Prolific’s prescreening function, which allows
researchers to screen on a variety of participant variables
including self-identified employment sector. To do this, we
simultaneously ran two distinct studies on Prolific with the
PHILIPP-MULLER, COSTELLO, AND RECZEK
FIGURE 7
SCIENCE APPEALS DO NOT DECREASE PURCHASE
INTEREST FOR A HEDONIC BODY WASH AMONG THOSE WHO
WORK IN STEM
same number of participants in each study (n ¼ 300): one
study was run with a sample of participants who stated
they worked in a STEM industry (using the Prolific pre-
screen filter “Employment-Sector” and selecting the cate-
gory “Science, Technology, Engineering, & ers that science is necessary to develop a hedonic product,
Mathematics”), and one study was run with a sample of
participants who stated they worked in any industry other
than STEM. To analyze the data, we combined these sam-
ples into one dataset and added a variable indicating the
sector in which each participant worked.
Three individuals were excluded for failing our data va-
lidity indicator, leaving 599 responses. We subjected the
data to a two-way ANOVA predicting purchase intentions
from appeal type (control ¼ 1, science ¼ 1), participant
industry (non-STEM ¼ 1, STEM ¼ 1), and the two-way
interaction. There was no main effect of appeal type (F(1,
595) ¼1.17, p ¼ .279, gp2 ¼ 0.002), but there was a main
effect of participant’s employment sector (F(1, 595)
¼8.41, p ¼.004, gp2 ¼ 0.014), such that individuals in
non-STEM fields reported higher purchase intentions
(M ¼ 4.15) than those in STEM fields (M ¼ 3.80).
Most relevant to our hypotheses, there was a significant
two-way interaction (F(1, 595) ¼ 4.77, p ¼ .029, gp2 ¼
0.008; figure 7). Consistent with studies 5a and 5b, we ob-
served a significant backfire effect of science for partici-
pants working outside of the STEM sector (F(1, 595) ¼
5.33, p ¼ .021, gp2 ¼ 0.009), such that participants were
less likely to buy the hedonic product promoted with a sci-
ence (M ¼ 3.95) than a control appeal (M ¼ 4.35).
However, for participants working in STEM industries, we
find that there was no difference in purchase intentions be-
tween those in the control appeal condition (M ¼ 3.73) and
those in the science appeal condition (M ¼ 3.85) (F(1, 595)
¼ 0.61, p ¼ .436, gp2 ¼ 0.001).
Studies 5a, 5b, and 5c all provide support for hypothesis
4, using distinct but conceptually related moderators of the
735
science backfire effect for hedonic products. We find that
the effect is eliminated when a consumer does not explic-
itly endorse the science is cold lay belief (study 5a), has
high level of trust in scientists (study 5b), and is exposed to
the scientific process and scientists via work experience
(study 5c).10
GENERAL DISCUSSION
Across 10 studies, we found evidence that invoking sci-
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
ence backfires when marketing CPGs positioned as he-
donic, leading to reduced product valuation. The same
science appeals were advantageous, however, when used to
promote products positioned as utilitarian. These findings
appeared to be driven by the conceptual disfluency of pair-
ing warm hedonic products with the scientific process,
which is perceived as competent, yet cold. We ruled out
the possibility that perceived artificiality or novelty drives
the backfire effect of science appeals for hedonic products
along with a variety of other relevant alternative mecha-
nisms. We also identified several boundary conditions for
the science backfire effect. When it was salient to consum-
using a science appeal did not reduce purchase intentions.
Finally, among those who see science as relatively warm,
highly trust scientists, or work in a STEM industry, the
backfire effect was attenuated.
Theoretical Contributions
This research contributes to multiple literatures in con-
sumer behavior. Our primary contribution lies in demon-
strating how consumers respond to the invocation of
science in marketing appeals. Although prior literature in
marketing has explored reference to specific scientific pro-
cesses in marketing appeals (Hingston and Noseworthy
2018; Zheng et al. 2019) or the addition or subtraction of
specific ingredients (Andre et al. 2019; Naylor et al. 2009),
no prior research has explored how consumers respond to
brands invoking science more broadly as part of a market-
ing appeal. Our research therefore addresses a novel ques-
tion to the consumer behavior literature: How do
consumers respond to brands that explicitly link their prod-
ucts to science?
To answer this question, we introduce a new lay belief
about the scientific process in the marketplace: it is seen as
competent but cold. We further demonstrate several of the
10 Although our primary focus in study 5c is the two-way interaction,
it is noteworthy that there was an unpredicted main effect of industry,
such that the scientists seemed less interested in purchasing body
wash overall. Because industry is not randomly assigned, we are hes-
itant to interpret this effect, as there are many demographic or other
correlates of industry that could be driving this effect. In web appen-
dix D, we show that none of the demographic variables measured in
this study (e.g., age, gender) can fully account for this main effect.
736
downstream effects of this lay belief. In doing so, we not
only contribute to the literature on consumer lay theories
(Cheng et al. 2017; Deval et al. 2013; McFerran and
Mukhopadhyay 2013; Raghunathan et al. 2006; Zane et al.
2020) but also to the growing body of literature exploring
how consumer response to products varies as a function of
what is communicated to consumers about the product’s
origin and production process (Smith, Newman, and Dhar
2016; Newman and Dhar 2014; Reich, Kupor, and Smith
2018).
The introduction of consumer lay beliefs about the role
of science in developing products also contributes to the
literature on perceptions of science and scientists. Though
science is generally respected in the United States, there
has been a recent erosion of trust in the scientific process
and scientists as individuals (American Academy of Arts
and Sciences 2018; Gauchat 2012). Perhaps lay beliefs
about the scientific process as lacking warmth contribute to
this growing distrust. Thus, our work builds upon past
work that has identified stereotypes about scientists (as a
group), which include beliefs that they are cold and compe-
tent (Brown et al. 2018; McPherson et al. 2018).
Our work also builds on the personalized matching liter-
ature (Goenka and Van Osselaer 2019; Stuppy, Mead, and
van Osselaer 2020; Teeny et al. 2021), which finds that
consumers are more persuaded by appeals that are framed
to match features of the consumer (e.g., their attitude ba-
sis). We introduce the idea of lay belief matching: When
an appeal matches consumers’ lay beliefs about a product’s
development, they will be more persuaded and more likely
to act consistently with the appeal. One key distinction be-
tween lay belief matching and the existing work on direct
matching effects is that in this work we never directly com-
pare warm versus cold appeals when promoting hedonic
versus utilitarian products. Instead, we rely on the lay intui-
tion that science is cold to indirectly create a mismatch be-
tween science appeals and hedonic products. Additionally,
this work goes beyond global affective/cognitive matching
(Fabrigar and Petty 1999; Klein and Melnyk 2016) to re-
fute the notion that any cognitive appeal would mismatch
with any affective product. Instead, cold cognitive appeals
mismatch with products where consumers anticipate
warmth.
Finally, although we found that a mismatch between sci-
ence appeals and hedonic products or attributes reduced
product valuation via disfluency, the results of study 3b
suggest that perceived brand competence might also play a
role in driving matching effects for utilitarian products. We
look forward to future research that further explores this
likely multiply determined effect.
Practical Implications
The present findings offer marketers insight into when
the use of science appeals might be effective versus
JOURNAL OF CONSUMER RESEARCH
detrimental in promoting various products. Our findings
suggest that it would behoove marketers to invoke science
when promoting products that have utilitarian attributes.
Likewise, for product categories where it is already natu-
rally salient to consumers that science is involved in pro-
duction (e.g., highly technical products), invoking science
should not have harmful consequences. For CPGs and
other categories where scientific development is less
expected, however, our findings suggest that invoking sci-
ence to market these products might backfire. Thus, when
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
promoting hedonically positioned products in the CPG do-
main, marketers should avoid the use of science appeals.
This work also provides insight into how the negative
impact of invoking science in the marketing of consumer
products can be circumvented. There are, of course, times
where there are potential benefits of highlighting the scien-
tific process used in creating a hedonic good or service
(e.g., if a proprietary scientific process is what gives a
brand their competitive advantage). In these cases, study 4
suggests that explaining to consumers that science is neces-
sary for creating a product mitigates the negative impact of
science appeals on purchase behavior. Thus, a campaign
highlighting the scientific engineering of even the most he-
donically oriented product might be well received if it is
accompanied by an explanation for how and why science is
necessary for developing the product, which could even in-
clude detailing the exact ways in which the science was
used.
Finally, this work identifies several variables that mod-
erate when invoking science in marketing consumer prod-
ucts backfires. These insights can be used in segmentation
strategies by managers. For example, those who have a
high degree of trust in scientists or who work in STEM in-
dustries do not exhibit the science backfire effect. We
assessed trust in scientists using items derived from a na-
tional Pew poll, which has identified other demographic
correlates of trust in scientists (Funk et al. 2019). For ex-
ample, Black Americans and Republicans report lower
trust in scientists (Funk, Kennedy, and Tyson 2020;
Kennedy, Tyson, and Funk 2022). Thus, brand managers
wishing to use science appeals to promote a hedonic prod-
uct could target segments of consumers who are likely to
trust scientists or take care to explain why science is neces-
sary for creating the product when targeting a segment
likely to exhibit the science backfire effect.
Limitations and Future Research
In this research, we explore the consequences of lay the-
ories around science but do not investigate the origins of
these beliefs. It is difficult to pinpoint the exact source of a
lay belief, as they are typically rooted in a combination of
sociocultural messages (Haws et al. 2017; Morris, Menon,
and Ames 2001) and personal experience (Ross and
Nisbett 1991; Tonietto et al. 2021). We suspect that both
PHILIPP-MULLER, COSTELLO, AND RECZEK
factors may be at play here and that experiences/messaging
related to the scientific process can influence beliefs about
scientists, and vice versa. For example, personal experi-
ence with science (e.g., in school) as a cold, careful process
may influence perceptions of scientists. Further, research
has found that those who have a particular job, role, or en-
gage in particular behaviors tend to be perceived as pos-
sessing traits associated with that behavior or role (i.e.,
Spontaneous Trait Inference; Carlston and Skowronski
2005; Newman and Uleman 1990). Conversely, portrayals
of scientists as lacking warmth (e.g., the unemotional sci-
entific genius trope often seen in popular culture) may in-
fluence perceptions of the scientific process. Future
research might explore how the lay beliefs we identify in
this research are formed.
Another limitation of the present work is that the effect
sizes we observe for the science backfire effect would be
classified as small (Cohen 1988). However, we also find
that the effect is very robust, as it appears across a variety
of different product categories, paradigms, and ways of in-
voking science. Additionally, the effect size is relatively
stable across studies, suggesting that the studies reported
here reliably captured the science backfire effect. In accor-
dance with recommended best practices, we conducted
studies with large samples (greater than 100 participants
per cell in every study) to ensure that we would have suffi-
cient power to detect any effects that were present.
However, even if the true effect size is indeed small by
traditional standards, we still suggest that this effect has
substantive importance (Kelley and Preacher 2012 for a
discussion of the different facets of effect size). For exam-
ple, we find that the use of science appeals reduces likeli-
hood of choosing a cookie by over 30% and WTP by over
9% (in terms of percentage change with a control message
as the baseline), decreases that would be viewed as impor-
tant from a managerial perspective, particularly since
changing language in appeals is essentially costless for
firms. Critically, although we found that the overall effect
size is small, it is not homogenous across the entire popula-
tion, and thus the backfire effect varies across segments.
Studies 5a, 5b, and 5c show that although the effect is rela-
tively small at a population level, it varies depending on in-
dividual differences and thus may vary from a large effect
to a null effect depending on the segment of interest (as is
the case with any moderated effect).
Finally, another limitation of the current work is that al-
though our studies reference science explicitly, we also
suspect that references to science need not be explicit to
have the effects we theorize. For example, including
images of scientists developing the product or of a labora-
tory setting may produce the same disfluency for hedonic
products. Thus, future research could examine whether
more subtle instantiations of our paradigm are sufficient to
produce the backfire effect we document.
737
DATA COLLECTION INFORMATION
Studies 1b, 1c, 2, 3a, 3b, 4, 5a, 5b; the three pilots dis-
cussed in the introduction; the study 2 pretest, the study 3b
pretest, the two study 4 pretests, and the study 5a–5c pre-
test were collected by all three authors using Amazon
Mechanical Turk. Study 5c was collected by all three
authors using Prolific. The study 1a pretest was collected
at the Fisher Behavioral Lab at The Ohio State University
with the assistance of a lab manager and research assistants
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
working under the supervision of all three authors. Study
1a was simultaneously collected at the Fisher Behavioral
Lab at The Ohio State University and the Mendoza
Behavioral Lab at the University of Notre Dame with the
assistance of lab managers and research assistants, at each
respective institution, working under the supervision of all
three authors. Studies 1b, 2; the second and third pilot stud-
ies in the introduction; and the first study 4 pretest were
collected in fall 2020. Study 3a was collected in summer
2020. Study 1c was collected in summer 2021. Study 1a,
3b, 5a, 5b, 5c; the first pilot study in the introduction; the
study 1a pretest, study 2 pretest, study 3b pretest, the sec-
ond study 4 pretest, and the study 5a–5c pretest were all
collected in fall 2021. All studies were analyzed by the first
author with the exception of studies 5a–c which were ana-
lyzed by the first and second authors. All studies were
designed by all three authors. The data are currently stored
on the Open Science Framework.
REFERENCES
Allport, Gordon Willard (1954), The Nature of Prejudice, Boston,
MA: Addison-Wesley.
Aaker, Jennifer L., Kathleen D. Garbinsky, and Emily N. Vohs
(2012), “Cultivating Admiration in Brands: Warmth,
Competence, and Landing in the ‘Golden Quadrant,” Journal
of Consumer Psychology, 22 (2), 191–94.
Aaker, Jennifer, Kathleen D. Vohs, and Cassie Mogilner (2010),
“Nonprofits Are Seen as Warm and for-Profits as Competent:
Firm Stereotypes Matter,” Journal of Consumer Research, 37
(2), 224–37.
American Academy of Arts and Sciences (2018), “Perceptions of
Science in America,” https://www.amacad.org/publication/
perceptions-science-america.
Andre, Quentin, Pierre Chandon, and Kelly Haws (2019),
“Healthy through Presence or Absence, Nature or Science?:
A Framework for Understanding Front-of-Package Food
Claims,” Journal of Public Policy & Marketing, 38 (2),
172–91.
Becker, G. M., M. H. DeGroot, and J. Marschak (1964),
“Measuring Utility by a Single-Response Sequential
Method,” Behavioral Science, 9 (3), 226–32.
Brown, Elizabeth R., Mia Steinberg, Yun Lu, and Amanda B.
Diekman (2018), “Is the Lone Scientist an American Dream?
Perceived Communal Opportunities in STEM Offer a
Pathway to Closing US–Asia Gaps in Interest and Positivity,”
Social Psychological and Personality Science, 9 (1), 11–23.
738
Cadbury Dairy Milk (2010), “Cadbury Advert—Cadbury
Chocolate Charmer Ad,” Last Accessed May 17, 2022,
https://www.youtube.com/watch?v=Bv7mETANB7A.
Carlston, Donal E. and John J. Skowronski (2005), “Linking ver-
sus Thinking: Evidence for the Different Associative and
Attributional Bases of Spontaneous Trait Transference and
Spontaneous Trait Inference,” Journal of Personality and
Social Psychology, 89 (6), 884–98.
Chattalas, Michael and Hirokazu Takada (2013), “Warm versus
Competent Countries: National Stereotyping Effects on
Expectations of Hedonic versus Utilitarian Product
Properties,” Place Branding and Public Diplomacy, 9 (2),
88–97.
Cheng, Yimin, Anirban Mukhopadhyay, and Rom Y. Schrift
(2017), “Do Costly Options Lead to Better Outcomes? How
the Protestant Work Ethic Influences the Cost-Benefit
Heuristic in Goal Pursuit,” Journal of Marketing Research,
54 (4), 636–49.
Chitturi, Ravindra, Vijay Raghunathan, and Rajagopal Mahajan
(2008), “Delight by Design: The Role of Hedonic versus
Utilitarian Benefits,” Journal of Marketing, 72 (3), 48–63.
Chmielewski, Michael and Sarah C. Kucker (2020), “An MTurk
Crisis? Shifts in Data Quality and the Impact on Study
Results,” Social Psychological and Personality Science, 11
(4), 464–73.
Cohen, Jacob (1988), Statistical Power Analysis for the
Behavioral Sciences, Hillsdale, NJ: Erlbaum.
Costello, John P. and Rebecca Walker Reczek (2020), “Providers
versus Platforms: Marketing Communications in the Sharing
Economy,” Journal of Marketing, 84 (6), 22–38.
Cuddy, Amy, Peter Fiske, and Susan Glick (2008), “Warmth and
Competence as Universal Dimensions of Social Perception:
The Stereotype Content Model and the BIAS Map,”
Advances in Experimental Social Psychology, 40 (March),
61–149.
Dennis, Sean A., Christopher A. Goodson, and Brian M. Pearson
(2020), “Online Worker Fraud and Evolving Threats to the
Integrity of MTurk Data: A Discussion of Virtual Private
Servers and the Limitations of IP-Based Screening
Procedures,” Behavioral Research in Accounting, 32 (1),
119–34.
Deval, Helène, Susan P. Mantel, Frank R. Kardes, and Steven S.
Posavac (2013), “How Naive Theories Drive Opposing
Inferences from the Same Information,” Journal of
Consumer Research, 39 (6), 1185–201.
Dhar, Ravi and Klaus Wertenbroch (2000), “Consumer Choice be-
tween Hedonic and Utilitarian Goods,” Journal of Marketing
Research, 37 (1), 60–71.
Diekman, Amanda B., Emily K. Clark, Amanda M. Johnston,
Elizabeth R. Brown, and Mia Steinberg (2011), “Malleability
in Communal Goals and Beliefs Influences Attraction to
STEM Careers: Evidence for a Goal Congruity Perspective,”
Journal of Personality and Social Psychology, 101 (5),
902–18.
Duhachek, Adam, Nidhi Agrawal, and Dahee Han (2012), “Guilt
versus Shame: Coping, Fluency, and Framing in the
Effectiveness of Responsible Drinking Messages,” Journal of
Marketing Research, 49 (6), 928–41.
Fabrigar, Leandre R. and Richard E. Petty (1999), “The Role of
the Affective and Cognitive Bases of Attitudes in
Susceptibility to Affectively and Cognitively Based
Persuasion,” Personality and Social Psychology Bulletin, 25
(3), 363–81.
JOURNAL OF CONSUMER RESEARCH
Fiske, Susan T., Amy J. C. Cuddy, and Peter Glick (2007),
“Universal Dimensions of Social Cognition: Warmth and
Competence,” Trends in Cognitive Sciences, 11 (2), 77–83.
Fuchs, Christoph, Martin Schreier, and Stijn M. J. van Osselaer
(2015), “The Handmade Effect: What’s Love Got to Do with
It?,” Journal of Marketing, 79 (2), 98–110.
Funk, Cary (2020), “Key Findings about Americans’ Confidence
in Science and their Views on Scientists’ Role in Society,”
Pew Research Center, Last Accessed May 17, 2022, https://
www.pewresearch.org/fact-tank/2020/02/12/key-findings-
about-americans-confidence-in-science-and-their-views-on-
scientists-role-in-society/.
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Funk, Cary, Meg Hefferon, Brian Kennedy, and Courtney Johnson
(2019), “Trust and Mistrust in Americans’ Views of
Scientific Experts,” Pew Research Center, Last Accessed
May 17, 2022, https://www.pewresearch.org/science/2019/
08/02/trust-and-mistrust-in-americans-views-of-scientific-
experts/.
Funk, Cary, Brian Kennedy, and Alec Tyson (2020), “Black
Americans Have Less Confidence to Act in the Public
Interest,” Pew Research Center, Last Accessed May 17,
2022, https://www.pewresearch.org/fact-tank/2020/08/28/
black-americans-have-less-confidence-in-scientists-to-act-in-
the-public-interest/.
Gauchat, Gordon (2012), “Politicization of Science in the Public
Sphere: A Study of Public Trust in the United States, 1974 to
2010,” American Sociological Review, 77 (2), 167–87.
Goddard, E. Desmond and James V. Gruber, ed. (1999),
Principles of Polymer Science and Technology in Cosmetics
and Personal Care, Edison, NJ: CRC Press.
Goenka, Shreyans and Stijn M. J. Van Osselaer (2019), “Charities
Can Increase the Effectiveness of Donation Appeals by
Using a Morally Congruent Positive Emotion,” Journal of
Consumer Research, 46 (4), 774–90.
Haws, Kelly L., Kevin L. Walker Reczek, and Rebecca Sample
(2017), “Healthy Diets Make Empty Wallets: The Healthy ¼
Expensive Intuition,” Journal of Consumer Research, 43
(April), ucw078.
Hayes, Andrew F. (2017), Introduction to Mediation, Moderation,
and Conditional Process Analysis: A Regression-Based
Approach, New York, NY: Guilford.
Hingston, Sean T. and Theodore J. Noseworthy (2018), “Why
Consumers Don’t See the Benefits of Genetically Modified
Foods, and What Marketers Can Do about It,” Journal of
Marketing, 82 (5), 125–40.
Hirschman, Elizabeth C. and Morris B. Holbrook (1982),
“Hedonic Consumption: Emerging Concepts, Methods and
Propositions,” Journal of Marketing, 46 (3), 92–101.
Judd, Charles M., Laurie James-Hawkins, Vincent Yzerbyt, and
Yoshihisa Kashima (2005), “Fundamental Dimensions of
Social Judgment: Understanding the Relations between
Judgments of Competence and Warmth,” Journal of
Personality and Social Psychology, 89 (6), 899–913.
Kelley, Ken and Kristopher J. Preacher (2012), “On Effect Size,”
Psychological Methods, 17 (2), 137–52.
Kennedy, Brian, Alec Tyson, and Cary Funk (2022), “Americans’
Trust in Scientists, Other Groups Declines,” Pew Research
Center, Last Accessed May 17, 2022, https://www.pewre-
search.org/science/2022/02/15/americans-trust-in-scientists-
other-groups-declines/.
Kidwell, Blair, Adam Farmer, and David M. Hardesty (2013),
“Getting Liberals and Conservatives to Go Green: Political
PHILIPP-MULLER, COSTELLO, AND RECZEK
Ideology and Congruent Appeals,” Journal of Consumer
Research, 40 (2), 350–67.
Klein, Kristina and Valentyna Melnyk (2016), “Speaking to the
Mind or the Heart: Effects of Matching Hedonic versus
Utilitarian Arguments and Products,” Marketing Letters, 27
(1), 131–42.
Kostyk, Alena, James M. Leonhardt, and Mihai Niculescu (2017),
“Simpler Online Ratings Formats Increase Consumer Trust,”
Journal of Research in Interactive Marketing, 11 (2),
131–41.
Kramer, Thomas and Lauren Block (2011), “Nonconscious
Effects of Peculiar Beliefs on Consumer Psychology and
Choice,” Journal of Consumer Psychology, 21 (1), 101–11.
Krause, Nicole M., Dietram A. Brossard, Michael A. Scheufele,
Keith Xenos, and Dominique Franke (2019), “Trends—
Americans’ Trust in Science and Scientists,” Public Opinion
Quarterly, 83 (Winter), 817–36.
Krems, Jaimie Arona, Ahra Ko, Jordan W. Moon, and Michael E.
W. Varnum (2021), “Lay Beliefs about Gender and Sexual
Behavior: First Evidence for a Pervasive, Robust (but
Seemingly Unfounded) Stereotype,” Psychological Science,
32 (6), 871–89.
Lee, Saerom, Lisa E. Bolton, and Karen P. Winterich (2017), “To
Profit or Not to Profit? The Role of Greed Perceptions in
Consumer Support for Social Ventures,” Journal of
Consumer Research, 44 (4), 853–76.
Lee, Angela Y. and Jennifer L. Aaker (2004), “Bringing the Frame
into Focus: The Influence of Regulatory Fit on Processing
Fluency and Persuasion,” Journal of Personality and Social
Psychology, 86 (2), 205–18.
Lee, Angela Y. and Aparna A. Labroo (2004), “The Effect of
Conceptual and Perceptual Fluency on Brand Evaluation,”
Journal of Marketing Research, 41 (2), 151–65.
Lellis, Julie C. (2016), “Waving the Red Flag: FTC Rgulation of
Deceptive Weight-Loss Advertising 1951–2009,” Health
Communication, 31 (1), 47–59.
Luchs, Michael G., Rebecca Walker Naylor, Julie R. Irwin, and
Rajagopal Raghunathan (2010), “The Sustainability
Liability: Potential Negative Effects of Ethicality on Product
Preference,” Journal of Marketing, 74 (5), 18–31.
Marinova, Detelina, Irina V. Kozlenkova, Leona Cuttler, and J. B.
Silvers (2017), “To Prescribe or Not to Prescribe? Consumer
Access to Life-Enhancing Products,” The Journal of
Consumer Research, 43 (5), 806–23.
McFerran, Brent and Anirban Mukhopadhyay (2013), “Lay
Theories of Obesity Predict Actual Body Mass,”
Psychological Science, 24 (8), 1428–36.
McPherson, Erin, Bernadette Park, and Tiffany A. Ito (2018),
“The Role of Prototype Matching in Science Pursuits:
Perceptions of Scientists That Are Inaccurate and Diverge
from Self-Perceptions Predict Reduced Interest in a Science
Career,” Personality & Social Psychology Bulletin, 44 (6),
881–98.
Morris, Michael W., Tanya Menon, and Daniel R. Ames (2001),
“Culturally Conferred Conceptions of Agency: A Key to
Social Perception of Persons, Groups, and Other Actors,”
Personality and Social Psychology Review, 5 (2), 169–82.
Mukhopadhyay, Anirban and Gita Venkataramani Johar (2005),
“Where There Is a Will, Is There a Way? Effects of Lay
Theories of Self-Control on Setting and Keeping
Resolutions,” Journal of Consumer Research, 31 (4),
779–86.
739
Naylor, Rebecca Walker, Courtney M. Droms, and Kelly L. Haws
(2009), “Eating with a Purpose: Consumer Response to
Functional Food Health Claims in Conflicting versus
Complementary Information Environments,” Journal of
Public Policy & Marketing, 28 (2), 221–33.
Newman, George E. and Ravi Dhar (2014), “Authenticity Is
Contagious: Brand Essence and the Original Source of
Production,” Journal of Marketing Research, 51 (3), 371–86.
Newman, Leonard S. and James S. Uleman (1990), “Assimilation
and Contrast Effects in Spontaneous Trait Inference,”
Personality and Social Psychology Bulletin, 16 (2), 224–40.
Parker, Jeffrey R., Iman Paul, Ryan Hamilton, Omar Rodriguez-
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Vila, and Sundar G. Bharadwaj (2021), “How Product Type
and Organic Label Structure Combine to Influence
Consumers’ Evaluations of Organic Foods,” Journal of
Public Policy & Marketing, 40 (3), 419–28.
Peter, Christina and Milan Ponzi (2018), “The Risk of Omitting
Warmth or Competence Information in Ads: Advertising
Strategies for Hedonic and Utilitarian Brand Types,” Journal
of Advertising Research, 58 (4), 423–32.
Pham, Nguyen and Naomi Mandel (2019), “What Influences
Consumer Evaluation of Genetically Modified Foods?”
Journal of Public Policy & Marketing, 38 (2), 263–79.
Raghunathan, Rajagopal, Rebecca Walker Naylor, and Wayne D.
Hoyer (2006), “The Unhealthy ¼ Tasty Intuition and Its
Effects on Taste Inferences, Enjoyment, and Choice of Food
Products,” Journal of Marketing, 70 (4), 170–84.
Reich, Taly, Rosanna K. Kupor, and Daniella M. Smith (2018),
“Made by Mistake: When Mistakes Increase Product
Preference,” Journal of Consumer Research, 44 (5),
1085–103.
Riis, Jason, Joseph P. Simmons, and Geoffrey P. Goodwin (2008),
“Preferences for Enhancement Pharmaceuticals: The
Reluctance to Enhance Fundamental Traits,” Journal of
Consumer Research, 35 (3), 495–508.
Rocklage, Matthew D. and Russell H. Fazio (2020), “The
Enhancing versus Backfiring Effects of Positive Emotion in
Consumer Reviews,” Journal of Marketing Research, 57 (2),
332–52.
Rosenthal, Robert, Ralph L. Rosnow, and Donald B. Rubin
(2000), Contrasts and Effect Sizes in Behavioral Research: A
Correlational Approach, Cambridge, UK: Cambridge
University Press.
Ross, Lee and Richard E. Nisbett (1991), The Person and the
Situation: Perspectives of Social Psychology, New York,
NY: McGraw-Hill.
Rozin, Paul (2005), “The Meaning of ‘Natural’ Process More
Important than Content,” Psychological Science, 16 (8),
652–58.
Scott, Sydney E., Christopher D. Inbar, Dominique Wirz, Paul
Brossard, and Yoel Rozin (2018), “An Overview of Attitudes
toward Genetically Engineered Food,” Annual Review of
Nutrition, 38 (August), 459–79.
Scott, Sydney E. and Paul Rozin (2017), “Are Additives
Unnatural? Generality and Mechanisms of Additivity
Dominance,” Judgment and Decision Making, 12
(November), 572–83.
Scott, Sydney E., Paul Rozin, and Deborah A. Small (2020),
“Consumers Prefer “Natural” More for Preventatives than for
Curatives,” Journal of Consumer Research, 47 (3), 454–71.
Singh, Rakesh K. and Nepal Singh (2005), “Quality of Packaged
Foods,” in Innovations in Food Packaging, ed. Jung H. Han,
Cambridge, MA: Academic Press, 24–44.
740
Smith, Rosanna K., George E. Newman, and Ravi Dhar (2016),
“Closer to the Creator: Temporal Contagion Explains the
Preference for Earlier Serial Numbers,” Journal of Consumer
Research, 42 (5), 653–68.
Spencer, Steven J., Geoffrey T. Zanna, and Mark P. Fong (2005),
“Establishing a Causal Chain: Why Experiments Are Often
More Effective than Mediational Analyses in Examining
Psychological Processes,” Journal of Personality and Social
Psychology, 89 (6), 845–51.
Spiller, Stephen A., Gavan J. Fitzsimons, John G. Lynch Jr,
and Gary H. Mcclelland (2013), “Spotlights, Floodlights,
and the Magic Number Zero: Simple Effects Tests in
Moderated Regression,” Journal of Marketing Research,
50 (2), 277–88.
Strahilevitz, Michal John and G. Myers (1998), “Donations to
Charity as Purchase Incentives: How Well They Work May
Depend on What You Are Trying to Sell,” Journal of
Consumer Research, 24 (4), 434–46.
Stuppy, Anika, Nicole L. Mead, and Stijn M. J. van Osselaer
(2020), “I Am, Therefore I Buy: Low Self-Esteem and the
Pursuit of Self-Verifying Consumption,” Journal of
Consumer Research, 46 (5), 956–73.
Teeny, Jacob, Pablo Siev, Richard E. Bri~nol, and Joseph Petty
(2021), “A Review and Conceptual Framework for
Understanding Personalized Matching Effects in Persuasion,”
Journal of Consumer Psychology, 31 (2), 382–414.
Thompson, Debora Viana and Rebecca W. Hamilton (2006), “The
Effects of Information Processing Mode on Consumers’
JOURNAL OF CONSUMER RESEARCH
Responses to Comparative Advertising,” Journal of
Consumer Research, 32 (4), 530–40.
Tonietto, Gabriela N., Selin A. Malkoc, Rebecca Walker Reczek,
and Michael I. Norton (2021), “Viewing Leisure as Wasteful
Undermines Enjoyment,” Journal of Experimental Social
Psychology, 97, 104198.
Wadhwa, Monica and Kuangjie Zhang (2015), “This Number Just
Feels Right: The Impact of Roundedness of Price Numbers
on Product Evaluations,” Journal of Consumer Research, 41
(5), 1172–85.
Wang, Wenbo, Hean Tat Keh, and Lisa E. Bolton (2010), “Lay
Theories of Medicine and a Healthy Lifestyle,” Journal of
Downloaded from https://academic.oup.com/jcr/article/49/5/721/6581083 by Jnls Cust Serv on 19 January 2023
Consumer Research, 37 (1), 80–97.
Williams, Lawrence E. and John A. Bargh (2008), “Experiencing
Physical Warmth Promotes Interpersonal Warmth,” Science
(New York, N.Y.), 322 (5901), 606–7.
Williams, Elanor F. and Mary Steffel (2014), “Double Standards
in the Use of Enhancing Products by Self and Others,”
Journal of Consumer Research, 41 (2), 506–25.
Zane, Daniel M., Robert W. Smith, and Rebecca Walker Reczek
(2020), “The Meaning of Distraction: How Metacognitive
Inferences from Distraction during Multitasking Affect
Brand Evaluations,” Journal of Consumer Research, 46 (5),
974–94.
Zheng, Yanmei, Lisa E. Bolton, and Joseph W. Alba (2019),
“Technology Resistance: The Case of Food Production
Processes,” Journal of Public Policy & Marketing, 38 (2),
246–62.
© 2023 Journal of Consumer Research Inc. Copyright of Journal of Consumer Research is the
property of Oxford University Press / USA and its content may not be copied or emailed to
multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.