ArticlePDF Available

Romance, Risk, and Replication: Can Consumer Choices and Risk-Taking Be Primed by Mating Motives?

Authors:

Abstract

Interventions aimed at influencing spending behavior and risk-taking have considerable practical importance. A number of studies motivated by the costly signaling theory within evolutionary psychology have reported that priming inductions (such as looking at pictures of attractive opposite-sex members) designed to trigger mating motives increase males’ stated willingness to purchase conspicuous consumption items and to engage in risk-taking behaviors, and reduce loss aversion. However a meta-analysis of this literature reveals strong evidence of either publication bias or p-hacking (or both). We then report 8 studies with a total sample of over 1,700 participants which sought to reproduce these effects. None of the studies, including one which was fully preregistered, was successful. The results question the claim that romantic primes can influence risk-taking and other potentially harmful behaviors.
Romance, Risk, and Replication: Can Consumer Choices and Risk-Taking
Be Primed by Mating Motives?
David R. Shanks
University College London Miguel A. Vadillo
King’s College London
Benjamin Riedel,Ashley Clymo,Sinita Govind,Nisha Hickin,Amanda J. F. Tamman,
and Lara M. C. Puhlmann
University College London
Interventions aimed at influencing spending behavior and risk-taking have considerable practical impor-
tance. A number of studies motivated by the costly signaling theory within evolutionary psychology have
reported that priming inductions (such as looking at pictures of attractive opposite sex members)
designed to trigger mating motives increase males’ stated willingness to purchase conspicuous consump-
tion items and to engage in risk-taking behaviors, and reduce loss aversion. However, a meta-analysis of
this literature reveals strong evidence of either publication bias or p-hacking (or both). We then report
8 studies with a total sample of over 1,600 participants which sought to reproduce these effects. None of
the studies, including one that was fully preregistered, was successful. The results question the claim that
romantic primes can influence risk-taking and other potentially harmful behaviors.
Keywords: risk, consumer behavior, decision making, priming, meta-analysis
Supplemental materials: http://dx.doi.org/10.1037/xge0000116.supp
Extensive efforts have been made in several areas of psychology
to develop interventions for influencing people’s willingness to
engage in potentially harmful behaviors such as gambling, addic-
tions, other forms of risk-taking, and excessive spending. Although
there have been some notable successes, such as the development
of a range of techniques within the cognitive “debiasing” field
(e.g., Gigerenzer, 1991;Larrick, 2004), new interventions would
have both practical and theoretical utility. Recently, it has been
claimed that risk-taking and spending behavior may be triggered in
part by evolutionarily driven motives (Kenrick & Griskevicius,
2013). Studies designed to support this hypothesis have suggested
that the subtle priming of mating motives can affect these behav-
iors. Certainly, the ubiquitous employment of attractive models
and sexual cues in advertising product categories such as casinos,
fashion, jewelry, cosmetic surgery, cars, cigarettes, and alcohol,
and the evidence for their effectiveness (King, McClelland, &
Furnham, 2015;Reichert, 2002), suggests that controlling such
primes in real-world settings might constitute a valuable interven-
tion.
Evolutionary psychologists have argued that male risk-taking
and conspicuous consumption are costly sexual signals intended to
attract potential mates (Miller, 2000). In brief, a man’s readiness to
tolerate risks and to bear a high cost for certain purchases is a
reliable indicator of his wealth and status. These behaviors are
costly in terms of economic resources and hence exclusive, are
easily perceived by others, and because females are assumed to
place high importance on affluence and prestige in their potential
mates, should increase the prospect of attracting a female mate.
Researchers (see Kenrick & Griskevicius, 2013) have suggested
that men’s willingness to pay elevated prices for particular pur-
chases and to engage in risky behaviors (“young male syndrome”)
is the result of a psychological mechanism designed by sexual
selection as an adaptation to women’s evolved preference for
prosperous and high-status mates.
Several laboratory experiments have been conducted to test this
viewpoint. More specifically, they examined whether males’ risk-
taking and expenditure on publicly consumed goods and services
could be increased by the subtle activation of mating motives.
The results provide what appears to be compelling support for
the evolutionary psychology hypothesis: For instance, priming
of mating motives significantly increased male but not female
participants’ stated willingness to engage in risky behaviors
David R. Shanks, Division of Psychology and Language Sciences,
University College London; Miguel A. Vadillo, Department of Primary
Care and Public Health Sciences, King’s College London; Benjamin Rie-
del, Ashley Clymo, Sinita Govind, Nisha Hickin, Amanda J. F. Tamman,
and Lara M. C. Puhlmann, Division of Psychology and Language Sciences,
University College London.
We thank Daniel Beal, Eugene Chan, Martin Daly, Siegfried Dewitte,
Anouk Festjens, Tobias Greitemeyer, Vladas Griskevicius, Kyu Kim,
Jessica Li, Patrick McAlvanah, Jill Sundie, and Bram Van den Bergh who
kindly provided assistance, including additional procedural or statistical
details or materials for use in the studies reported here. We also thank Ben
Newell for helpful discussions concerning this work.
Correspondence concerning this article should be addressed to David R.
Shanks, Division of Psychology and Language Sciences, University Col-
lege London, 26 Bedford Way, London WC1H 0AP. E-mail: d.shanks@
ucl.ac.uk
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Experimental Psychology: General © 2015 American Psychological Association
2015, Vol. 144, No. 6, 000 0096-3445/15/$12.00 http://dx.doi.org/10.1037/xge0000116
1
AQ: au
AQ: 1
AQ: 2
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
(Greitemeyer, Kastenmüller, & Fischer, 2013) and to pay for
conspicuous but not inconspicuous goods and services
(Griskevicius et al., 2007;Sundie et al., 2011). Consistent with
the general theoretical framework within which they are inter-
preted (Kenrick & Griskevicius, 2013), these effects are not
always restricted to males. Festjens, Bruyneel, and Dewitte
(2014, Study 3) obtained a priming effect in both males and
females on willingness to pay for “reward” items such as
chocolates and wine, and Hill and Durante (2011) found an
effect of mating primes on females’ willingness to take health-
related risks in the service of increasing their attractiveness,
specifically their desire to get a tan or take dangerous dieting
pills.
Since the seminal and highly cited (300 Google Scholar
citations) research of Wilson and Daly (2004), it has been reported
that several decision-making behaviors can be influenced by mat-
ing primes, using a variety of different priming manipulations (the
specific methods of many of these studies will be described in
more detail below). Table 1 summarizes the large space of such
demonstrations. In addition to studies on risk-taking and conspic-
uous consumption, Table 1 lists studies that have explored tem-
poral discounting, loss aversion, and cooperation in the ultimatum
game, all capturing important aspects of decision making. For
example, Wilson and Daly (2004) and subsequent studies found
that male participants discounted monetary rewards more steeply
after viewing pictures of attractive females, and Van den Bergh,
Dewitte, and Warlop (2008) obtained the same result when male
participants physically examined bras compared with t-shirts.
Festjens et al. (2014) obtained a similar effect in both temporal
discounting and loss aversion but with female participants who
physically examined a pair of boxer shorts compared with a t-shirt.
Van den Bergh and Dewitte (2006) observed that male participants
accepted less fair offers in the ultimatum game after viewing
pictures of attractive females.
It is important to emphasize that this space is itself just a small
part of an even larger space in which (a) mating primes have been
reported to influence a range of other behaviors such as creativity,
aggression, the likelihood of noticing conspicuous consumption
products, and the stated importance of wealth (Janssens et al.,
2011;Kenrick & Griskevicius, 2013;Roney, 2003); and (b) the
behaviors listed in the Table have been associated with other types
of evolutionary prime. For instance, Li, Kenrick, Griskevicius, and
Neuberg (2012) found that loss aversion can be affected by “self-
protection” primes. Other research not included here has studied
the effects of the presence of opposite sex individuals on risk-
taking. As an illustration, Ronay and von Hippel (2010) found that
male skateboarders took greater risks in the presence of a female
than of a male observer. Although priming may contribute to such
effects, other processes such as attention oriented to the observer
are likely to play a role.
Here we evaluate whether decision making behaviors can be
influenced by mating primes. In addition to the growing recogni-
tion that psychology must generally devote more efforts toward
replication (e.g., Asendorpf et al., 2013;Simons, 2014), these
particular studies merit attention for several further and more
specific reasons. First, they are motivated by and conceptualized
Table 1
Studies of the Influence of Mating Primes on Various Aspects of Decision Making
Decision-making domain
Prime method
Opposite-sex pictures Romantic text Other formats
Conspicuous consumption Griskevicius et al. (2007, Study 1)
Sundie et al. (2011, Study 1)
Studies 4 and 5
Griskevicius et al. (2007, Studies 2
and 3)
Sundie et al. (2011, Studies 2 and 3)
Studies 1–3
Chan (2015)
Festjens et al. (2014, Study 3)
Benevolence Griskevicius et al. (2007, Study 1) Griskevicius et al. (2007, Studies
2–4)
Study 3
Gambling Baker and Maner (2008)
Greitemeyer et al. (2013, Experiment 2)
Li (2012)
McAlvanah (2009)
Studies 6 and 7b
Social risk-taking Studies 5 and 6
Sexual/health risk-taking Greitemeyer et al. (2013, Experiment 1) Hill and Durante (2011, Study 2)
Hill and Durante (2011, Study 1)
Studies 6 and 7a
Driving risk-taking Greitemeyer et al. (2013, Experiment 3)
Study 5
Substance risk-taking Study 6
Physical risk-taking Baker and Maner (2009)
Loss aversion Li et al. (2012, Studies 1–3) Festjens et al. (2014, Study 2)
Study 8
Temporal discounting Kim and Zauberman (2013) Festjens et al. (2014, Study 1)
Van den Bergh et al. (2008, Study 1A) Van den Bergh et al. (2008, Study 1B)
Wilson and Daly (2004)
Cooperation (ultimatum
game)
Van den Bergh and Dewitte (2006)
Note. Bold indicates studies reported in this article.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2SHANKS ET AL.
T1
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
within a theoretical framework whose plausibility has been ques-
tioned (Newell & Shanks, 2014a,2014b). In this framework,
primes are assumed to have broad and long-term effects, influenc-
ing a wide range of possible downstream behaviors, and these
influences are largely automatic. They assume, for instance, that a
priming induction designed to trigger mating motives, such as
looking at pictures of attractive opposite sex members, can in-
crease an individual’s stated willingness to engage in risky driving
behaviors (Greitemeyer et al., 2013). Transfer and generalization
from the prime induction to the measured behavior would have to
be very broad for such an influence to occur, and a considerable
amount of research suggests that such broad transfer is very much
the exception rather than the rule (Newell & Shanks, 2014a). They
also assume that primes can influence behavior unconsciously, and
again the evidence for such influences is debatable at best (Newell
& Shanks, 2014a,2014b).
Second, these priming effects might be mediated by processes
that are rather different from—and less theoretically novel than—
those envisaged by their proponents. Specifically, the design of
many behavior priming studies leaves open the possibility that
compliant participants are able to infer, and hence behave in
accordance with, the experimenter’s hypothesis (Durgin et al.,
2009;O. Klein et al., 2012;Orne, 1962;Rosenthal & Rubin,
1978). It does not seem far-fetched to imagine that a male partic-
ipant first asked to look at pictures of attractive opposite sex
members, and then to state his willingness to engage in risk-taking
behaviors, might intuit and conform to the experimenter’s hypoth-
esis. Indeed it is even possible, particularly in laboratory experi-
ments, that the experimenter might subtly and unintentionally
convey this expectancy (O. Klein et al., 2012).
Finally, many previous claims of subtle behavior priming ef-
fects, some quite similar to those described above, have failed to
withstand close scrutiny and been identified as possible false
positives (e.g., R. A. Klein et al., 2014;Pashler, Coburn, & Harris,
2012;Shanks et al., 2013). After completing some of the initial
studies reported below, which were designed primarily to explore
the generality and boundary conditions on these priming effects,
we commenced a meta-analysis of the extant studies to try to
understand the discrepant findings (our later experiments, partic-
ularly Studies 3, 6, 7, and 8, are more exact replications). This
meta-analysis speaks to the issue of whether some published
romantic priming effects might be false positives.
Meta-Analysis of Romantic Priming Effects on
Decision Making
The 15 studies included (those listed in Table 1) are all the
reports, whether published or unpublished (e.g., dissertations),
identified by searching ProQuest, PsycINFO, and Web of Science
(search terminated in April 2015) using combinations of the terms
“priming,” “mating/romantic motives/goals,” and “sex(ual) cue,”
which examination revealed used a mating prime-induction
method and a decision-making dependent variable. The reference
lists of those articles were examined and descendancy searches
were conducted via Google Scholar to examine all articles subse-
quently citing them. We also contacted the authors of all identified
reports requesting information about any additional unpublished
data they possessed or knew about. For each publication we
extracted the test statistic, means, SDs, and sample size for the key
experimental prediction, and calculated effect sizes from these. In
some cases the authors kindly provided additional details. The
complete set of data is available at https://osf.io/zubfs/.
In total these articles reported 43 independent effects, all but one
of which were originally described as statistically significant. The
results are depicted in the forest plot in Figure 1 and the black
circles in the funnel plot in Figure 2.
1
In the latter, effect size is
plotted against the SE of the effect size. Studies with larger
samples have lower SEs. The meta-analysis fails to find evidence
of significant heterogeneity among the effects, Q(42) 53.7, p
.11, I
2
19.6%. Although it would be wrong to regard the
different dependent measures as psychologically interchangeable,
they appear to be statistically similar in terms of the effect size
estimates they yield. Given the absence of substantial between-
study heterogeneity, we have not attempted to identify moderator
variables or effect modifiers.
The meta-analysis, using a random-effects model, yields an
effect size of d0.57, 95% confidence interval (CI) [0.49, 0.65],
a medium-sized effect, marked by the rightmost vertical line in
Figure 2. The triangle around this line (the “funnel”) marks the
region where the individual studies are expected to be distributed.
Near to the top of the figure, studies should yield effect size
estimates close to the mean effect size. These large studies are
unlikely to yield estimates that diverge far from the mean effect,
because their sampling error will be low, in the same way that
estimates of the proportion of heads based on 1,000 coin-tosses
will be close to 50%. Toward the bottom of the figure, studies are
expected to be more widely distributed on either side of the mean
effect size, because their sampling error will be high, in the same
way that estimates of the proportion of heads based on four
coin-tosses may be a long way above or below 50%. More impor-
tant, the scatter of points should be symmetrical: studies should be
as likely to deviate below as above the mean effect size.
This is clearly not what Figure 2 reveals. Studies with lower
error do not converge on the mean effect size: they yield estimates
that are consistently lower than the mean, while studies with more
error yield estimates that are consistently greater than the mean (in
other words, the datapoints are not symmetrically distributed).
This asymmetry (captured by the red regression line) is statistically
significant by the Egger test (Egger, Smith, Schneider, & Minder,
1997), t(41) 6.24, p.0001. Moreover, there is a striking lack
of studies falling in the part of the funnel that is shaded gray. This
is the region in which nonsignificant (p.05) results fall (the
darker portion of this area depicts the region of marginal signifi-
cance, 0.10 p.05).
If this pattern is not consistent with how a funnel plot is
expected to appear, then what does it mean? Two obvious possi-
1
Despite only 1 of the 43 contrasts being originally reported as nonsig-
nificant (Griskevicius et al., 2007, Study 1, men), in Figure 1 there are 10
whose lower 95% confidence interval includes zero (and these fall inside
the shaded dark gray area in Figure 2). The reason for this discrepancy is
that for some studies the effect sizes could be computed in different ways
(e.g., from within- vs. between-subjects contrasts; from test statistics vs.
the authors’ own reported effect size estimates). Our aim was to maximize
consistency across the meta-analysis but this sometimes yielded a result
that was nonsignificant even though an alternative method did yield a
significant result. This discrepancy has no bearing on the meta-analysis,
however, which is based on the effect size estimates themselves and not on
their statistical significance.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
MATING MOTIVES AND DECISION MAKING
F1
F2, Fn1
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
bilities (albeit not the only ones: see Sterne et al., 2011, for
discussion of the range of possible causes of funnel plot asymme-
try) are that the published studies are affected by publication bias
(selection) or p-hacking (inflation). On this interpretation, it is
either the case that studies falling in the shaded area to the left of
the vertical line have been conducted but—because they yielded
nonsignificant results—not published, or studies have been artifi-
cially shifted in the funnel plot as a result of p-hacking (or some
combination of both of these). Examples of p-hacking include
continuing to test additional participants until a significant result is
Figure 1. Forest plot from the meta-analysis showing the effect size (ES, Cohen’s d) and 95% confidence
interval (CI) of each of 43 independent studies (from the reports listed in Table 1) and the meta-analytic effect
size from a random-effects analysis. For each study the marker size is proportional to the sample size.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4SHANKS ET AL.
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
achieved (Yu, Sprenger, Thomas, & Dougherty, 2014), or remov-
ing outliers post hoc to attain significance, or collecting several
dependent measures and only reporting the one(s) that yield sig-
nificant results (John, Loewenstein, & Prelec, 2012;Simmons,
Nelson, & Simonsohn, 2011). Whatever the cause (publication
bias or p-hacking), the conclusion is the same: despite 42/43
statistically significant results, the published literature cannot be
taken as providing a sound basis for estimating the true size of the
effect of romantic priming on decision making. The published
studies are consistent with a true effect much smaller than 0.5.
Indeed, extrapolating from the datapoints to a study with zero
standard error (Stanley & Doucouliagos, 2014), the effect size
estimate is close to zero.
Another way of appreciating the extreme bias evident among
these studies is to note that for every additional 30 participants
added per group (prime/control), the effect size is expected to
decrease by approximately 0.1 effect size (Cohen’s d) units. Thus,
while a study with 30 participants per group is expected to yield d
0.7, the same study with 180 per group yields d0.2. (This
estimate comes not from the funnel plot in Figure 2 but rather from
a simple linear regression of effect size onto study sample sizes.)
It is important to stress that we are not concluding from this
analysis that the true effect size is small (or zero). Nor are we
suggesting that selection or inflation are any more prevalent in this
literature than anywhere else in behavioral research. Indeed, many
other examples of publication bias/p-hacking in psychology have
been documented recently (e.g., Bakker, van Dijk, & Wicherts,
2012;Carter & McCullough, 2014;Flore & Wicherts, 2015),
including in a meta-analysis of priming studies in another domain
(religious priming; Shariff, Willard, Andersen, & Norenzayan,
2015). Ferguson and Brannick (2012) estimated that among meta-
analyses that tested for funnel plot asymmetry, it was present in
around 2040%. Rather, our more modest conclusion is that any
firm inference about the true effect size based on the published
research is unjustified. We simply do not know what studies may
have been conducted but not published, nor do we know for certain
that p-hacking has taken place. What we do know is that the funnel
plot is irregular and that therefore the published research does not
validly support any conclusions about the true effect of romantic
priming on decision making.
Overview of Present Studies
The eight studies reported here used combinations of two prime
induction methods and nine dependent measures. The former in-
volved either a text-based or a pictorial procedure for inducing a
mating motive. In the text-based version, participants read about a
romantic episode or about a neutral event. In the pictorial version,
they viewed attractive opposite sex members, or neutral people/
scenes. Some studies (Festjens et al., 2014;Van den Bergh et al.,
2008) have used other methods which are not explored here
(though they are noted in Table 1 for completeness): in these
studies, participants physically examined an item of mating-related
clothing (e.g., a bikini or a pair of boxer shorts).
The dependent measures, based on previous studies in this field,
involved stated willingness either to pay for various conspicuous
consumption goods (Studies 1–5), to engage in risky behaviors in
the domains of sexual or social behavior, substance abuse, driving,
or gambling (Studies 5–7), or to pay for gains or to avoid losses
(Study 8). In Studies 3 and 5 additional dependent measures
(benevolence and nonconspicuous consumption, respectively)
were included to match the procedures of previous studies as
closely as possible. Although all the dependent measures were
based on self-report, such measures appear to be a reliable and
valid method to assess many risk-taking behaviors (see Brener,
Billy, & Grady, 2003).
While some of the studies reported here combined prime induc-
tions and dependent measures in combinations different from those
used in the published experiments on which they are modeled,
others were close replications using identical combinations. For
reference, Table 2 provides details of major differences between
the studies reported in this article and the original studies on which
they are based. As described below, past research has found both
prime induction methods to be effective and has used them inter-
changeably.
Study 1
The first experiment used a text prime induction method, as used
by Griskevicius et al. (2007, Studies 2–4) and Sundie et al. (2011,
Study 2). Spending patterns were measured using a method similar
to that used in Griskevicius et al.’s (2007) Study 1. The key
hypothesis is that priming male participants with mating motives
will lead them (but not female participants) to increase their
willingness to pay for publicly visible goods and services.
Method
Participants. In all experiments reported here we aimed to
collect approximately 40 completed surveys per group for each
gender, slightly above the median sample size (34) in previous
studies. Although our primary approach to sample size was a
Bayesian one (see below), we note that this sample size is adequate
Figure 2. Funnel plot from the meta-analysis. Black circles represent the
effect sizes (Cohen’s d) of each of the 43 independent studies from the reports
listed in Table 1 plotted against the inverse of that study’s SE. The rightmost
vertical line is the effect size estimate from a random-effects model of
these 43 studies, and the red line is the regression line from the Egger test.
Open triangles denote the effect sizes and SEs of the new studies reported
in the present article. The shaded gray area depicts the region in which p
.05 for individual studies. The darker portion of this area depicts the region
of marginal significance, 0.10 p.05. See the online article for the
color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
MATING MOTIVES AND DECISION MAKING
OC
NO
LL
IO
NR
E
T2
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
to detect an effect of size d0.57 (the meta-analytic effect size)
with power (1 - )0.80.
One hundred and seventy-three participants, recruited through
the psychology participant panel at University College London
(UCL), completed the study. Participants filled out a self-
administered online survey for the chance to win one of four £20
online retail vouchers. Sixteen of the submitted surveys were
excluded from subsequent analysis either because they took an
inappropriate amount of time (less than 5 min or more than 1 hr)
to complete the survey or because of unreflective responding such
as giving the same response to all willingness-to-pay (WTP)
questions (note that the statistical inferences below are not altered
by including these datasets). This resulted in a final sample of 157
participants, of whom 83 were female and 74 were male. The mean
age of these participants was 27.8 years (SD 8.8). The survey
was administered online using Qualtrics survey software (www
.qualtrics.com).
Procedure. The study used a 2 (Gender) 2 (Priming Con-
dition: Control vs. Experimental) between-subjects factorial de-
sign. Participants were randomly assigned to one of the two
priming conditions. Seventy-nine participants were assigned to
each condition, with the gender split reported in Table 3.
Participants were first presented with introductory information
about the aim of the study and gave their informed consent. The
introductory information stated that the survey had been designed
to “investigate consumer preferences and memory recall.” Partic-
ipants were told that they were being asked to read a short text as
part of a memory recollection exercise. The consumer preference
questions were placed between the text and the memory questions
so as to appear as a distractor allowing for memory decay.
Priming. The priming materials were identical to those used
by Griskevicius et al. (2007, Studies 2–4). Participants in the
experimental condition were requested to read a romantic scenario
involving meeting a highly desirable person of the opposite sex
during a holiday on a tropical island. Participants imagined spend-
ing a romantic afternoon as well as evening with this person and
being strongly motivated to romantically pursue this relationship.
The text ended with a kiss with the person and a profound feeling
of excitement as to what the rest of the night may bring. Partici-
pants in the control condition read an emotion-laden scenario
about getting ready to go to a concert with a same-sex friend.
Participants imagined not being able to find the tickets for the
show and frantically searching for them everywhere. The text
finished with the friend showing up, tickets in hand, and an elated
mood in anticipation of the entertaining evening ahead. The sup-
plemental materials report a manipulation check confirming that
the prime text significantly activates mating intentions.
Willingness to pay. Participants were presented with a ran-
domized list of 10 items for each of which they had to state the
maximum amount of money they would be willing to pay. The
items on the list consisted of publicly visible goods and services
purchased by women as well as men: a new watch, a dinner with
friends at a restaurant (per person), a new mobile phone (not part
of a contract), a short vacation abroad (transportation and accom-
modation, per person), a new pair of shoes, a new pair of jeans, a
new pair of sunglasses, a new fragrance, a new laptop, and a new
portable media (e.g., MP3) player.
The first four of these items corresponded to goods and services
utilized in Study 1 of Griskevicius et al. (2007). The fifth item (a
new car) from that study was not included in the present research
as it was presumed that most of the survey respondents would be
in their twenties and not able to afford a new automobile. Partic-
ipants provided numerical responses for each item. Average
United Kingdom market prices for the goods and services were
listed in brackets next to each. These averages were calculated on
the basis of the frequency distribution of prices on online retail
sites (www.amazon.co.uk). Participants were given the option to
indicate that they would not purchase a specific item, in which case
Table 2
Major Differences (Features Present in the Replication Compared With the Replicated Experiment) Between the Studies Reported in
this Article and the Original Studies on Which They Are Modeled
Study Replicated experiment(s) Major differences
1 Prime induction: Griskevicius et al. (2007), Studies 2–4 Online sample
Consumption measure: Griskevicius et al. (2007), Study 1 Consumption measure (monetary amount)
2 Prime induction: Griskevicius et al. (2007), Studies 2–4 Online sample
Consumption measure: Griskevicius et al. (2007), Study 1
3Griskevicius et al. (2007) Study 2 Online sample
4 Prime induction: Griskevicius et al. (2007) Study 1 Online sample
Consumption measure: Griskevicius et al. (2007) Studies 2–4
5Griskevicius et al. (2007) Study 1, male participants Online sample
Greitemeyer et al. (2013) Experiments 3 and 4, male participants Control stimuli (people)
Risk attitudes and consumption behavior measured within-subjects
6Greitemeyer et al. (2013) Experiments 1 and 2 Online sample
Risk attitudes measured within-subjects
7a Greitemeyer et al. (2013) Experiment 1, male participants
7b Greitemeyer et al. (2013) Experiment 2, male participants Online sample
8Li et al. (2012), Study 1 Continuous willingness-to-pay scale
Loss aversion (gambling) test added
Gambling question added
Self-protection prime condition added
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6SHANKS ET AL.
T3
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
they were instructed to enter a value of zero (this was rarely
chosen).
After stating their willingness to pay for the target items, par-
ticipants completed questions assessing their purchase motivation
(stating to what extent their decisions to purchase goods and
services depended generally on factors such as price, quality,
opportunity to display wealth, etc.), and their memory for the
priming stories. For the sake of brevity analyses of these responses
are not reported. As described in the supplemental materials,
participants in this and all subsequent experiments also gave
Table 3
Sample Size (N) and Descriptive Statistics for the Experimental (Prime) and Control Groups in Each Study, and Bayes Factors
NPrime M(SD)NControl M(SD) Difference [CI] Bayes factor BF
01
Study 1
Conspicuous consumption
Male 40 10.69 (8.83) 34 9.61 (5.21) 1.08 [2.36, 4.52] 4.70
Female 39 10.18 (4.20) 44 9.89 (4.60) .29 [1.64, 2.22] 5.71
Study 2
Conspicuous consumption (10 items)
Male 27 4.67 (1.37) 27 4.97 (1.91) .30 [1.20, .61] 4.05
Female 32 4.57 (1.74) 39 4.18 (1.60) .38 [.41, 1.17] 3.60
Conspicuous consumption (four items)
Male 27 5.25 (1.81) 27 5.56 (1.92) .31 [1.32, .71] 4.17
Female 32 4.96 (1.89) 39 4.67 (1.93) .29 [.62, 1.20] 4.56
Study 3
Conspicuous consumption
Male 32 5.49 (1.39) 32 5.29 (1.10) .21 [.42, .83] 4.34
Female 40 4.68 (1.27) 47 4.74 (1.08) .06 [.56, .44] 5.91
Inconspicuous consumption
Male 32 4.74 (1.56) 32 4.72 (1.20) .03 [.67, .72] 5.28
Female 40 4.61 (.93) 47 4.65 (.98) .04 [.45, .37] 5.99
Conspicuous benevolence
Male 32 5.46 (1.59) 32 5.16 (1.56) .30 [.49, 1.09] 4.07
Female 40 6.14 (1.39) 47 5.87 (1.33) .26 [.32, .85] 4.17
Inconspicuous benevolence
Male 32 5.00 (1.32) 32 5.12 (1.45) .12 [.81, .57] 5.02
Female 40 5.28 (1.06) 47 4.99 (1.25) .29 [.21, .79] 3.23
Study 4
Conspicuous consumption
Male 33 4.65 (1.40) 35 4.56 (1.19) .09 [.53, .72] 5.22
Female 34 4.58 (1.13) 38 4.82 (1.10) .23 [.76, .29] 3.88
Study 5
Risk-taking (driving)
Male 39 34.19 (15.81) 40 37.19 (15.71) 3.00 [10.06, 4.06] 4.19
Risk-taking (social)
Male 39 45.49 (12.71) 40 43.50 (17.00) 1.98 [4.75, 8.72] 4.97
Conspicuous consumption
Male 39 4.56 (2.02) 40 4.51 (1.63) .06 [.76, .88] 5.77
Inconspicuous consumption
Male 39 4.29 (1.24) 40 4.27 (1.12) .02 [.51, .55] 5.81
Study 6
Risk-taking (sexual)
Male 25 2.73 (1.51) 31 3.25 (1.62) .53 [1.37, .32] 2.48
Female 19 2.01 (1.71) 22 2.19 (1.54) .18 [1.21, .85] 4.10
Risk-taking (gambling)
Male 25 1.16 (1.11) 31 1.45 (1.46) .29 [1.0, .42] 3.65
Female 19 1.26 (1.33) 22 1.50 (1.34) .24 [1.08, .61] 3.77
Risk-taking (substance)
Male 25 3.12 (2.25) 31 3.06 (2.36) .06 [1.19, 1.31] 4.94
Female 19 2.72 (2.08) 22 2.85 (2.01) .13 [1.42, 1.17] 4.26
Study 7a
Risk-taking (sexual)
Male 61 3.91 (1.76) 59 3.89 (1.86) .02 [.63, .68] 7.06
Study 7b
Risk-taking (gambling)
Male 53 .76 (1.11) 53 .83 (1.09) .08 [.50, .35] 6.30
Study 8
Loss aversion
Male 109 3.92 (85.4) 109 4.19 (80.4) .27 [21.9, 22.4] 9.41
Female 109 16.83 (75.9) 108 16.21 (79.4) .61 [21.4, 20.2] 9.38
Note.CI95% confidence interval.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
MATING MOTIVES AND DECISION MAKING
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
demographic information and reported their awareness of the pur-
pose of the experiment via a series of funnel debriefing questions.
Results
All statistical analyses reported in this article were computed in
JASP (Love et al., 2015). The complete set of data for this and all
subsequent studies is available at https://osf.io/ytvj7/.
Willingness to pay. The willingness to pay amounts were
aggregated for each participant by summing across the 10 items.
The sums were calculated on the basis of scaled values calcu-
lated as the ratio between the values entered and the associated
anchor value.
The mean of participants’ scaled values was used to measure
their willingness to pay for conspicuous goods and services. De-
scriptive statistics are presented in Table 3. A 2 (Condition: Prime
vs. Control) 2 (Gender) between-subjects analysis of variance
(ANOVA) revealed no significant interaction of gender and prim-
ing condition on the amount of money participants indicated they
would spend, F(1, 153) 0.17, p.68, p
2.001. In addition,
neither the main effect of prime condition, F(1, 153) 0.51, p
.48, p
2.003, nor of gender, F(1, 154) 0.01, p.91, p
2.00,
was significant. To examine the specific experimental hypotheses,
two planned contrasts were computed. Priming, critically, did not
have a significant effect on male participants’ willingness to pay,
t(72) 0.63, one-tailed p.27 (effect sizes for this and all
subsequent critical contrasts are reported in Figure 3). Female
participants in the experimental condition also did not pay more
than those in the control condition, t(81) 0.30, one-tailed p
.38. Table 3 reports the confidence intervals on the differences.
Bayes factor analysis. Bayes factors were calculated for the
simple contrast analysis on male and female participants’ willing-
ness to pay (see Table 3). Bayes factors (BF
01
) represent the
probability of the data given the null hypothesis versus the prob-
ability of the data given the experimental hypothesis. Put more
simply, the Bayes factor provides an indication of whether, given
the data, the null hypothesis or the experimental hypothesis is more
likely. The method used to quantify the Bayes factor and the
relevant probabilities was the one developed by Rouder, Speck-
man, Sun, Morey, and Iverson (2009) (using the Cauchy distribu-
tion, default scale ron effect size 1.0).
A common yardstick is to interpret BF
01
Bayes factors between
1 and 3 as “barely worth a mention,” ones between 3 and 10 as
providing substantial support for the null hypothesis, and ones
greater than 10 as providing strong support (Jeffreys, 1961;Wet-
zels et al., 2011). The Bayes factors for both female and, more
importantly, male participants in Study 1 were both greater than 4
(all reported Bayes factors are two-sided). These findings suggest
that the data are at least four times as likely under the null
hypothesis compared to the experimental hypothesis. In light of the
data obtained, the hypothesis that priming condition had an effect
on male participants’ willingness to pay is much less likely than
the hypothesis that priming did not have an effect.
Study 2
In Study 1 participants’ willingness to pay was measured by
open-ended, numerical responses for each item, with average
United Kingdom market prices provided for each. This is slightly
Figure 3. Forest plot for the main experimental tests in Studies 1–8, showing the effect size (ES, Cohen’s d)
and 95% confidence interval (CI) of each as well as the meta-analytic effect size from a random-effects analysis.
Except where otherwise noted, the tests are all for male participants. For each study the marker size is
proportional to the sample size.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8SHANKS ET AL.
F3
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
different from any of the methods used by Griskevicius and
colleagues. In Study 2 the method for measuring willingness to pay
was identical to that of Griskevicius et al.’s (2007, Study 1)
original experiment.
Method
Participants. One hundred and twenty-eight UCL partici-
pants completed the survey for a chance to win one of four £20
online retail vouchers. Data from three participants were excluded
on the same basis as in Study 1. Seventy-one of the participants
were female and 54 male, and their mean age was 26.2 years
(SD 7.2).
Procedure. Except where mentioned, the procedure was iden-
tical to that of Study 1. The introductory instructions were made
more similar to those used by Griskevicius et al. (2007, p. 91) and
stated that the goal of the research was to “investigate consumer
behavior and decision making.” Participants were told they would
all be asked to read the same standard scenario before filling out
the survey to ensure that all participants were in “the same frame
of mind” when answering the following questions and to thus
reduce extraneous bias in the research. The respondents were
thereby ostensibly “led to believe that everyone was reading the
same scenario and that the nature of the scenario was irrelevant to
the study as long as it served to focus everyone on the same thing.”
Priming. The method used to prime the participants was iden-
tical to that of Study 1. However, the instruction preceding both of
the primes was adapted to the revised cover story: “Please read the
following text. All participants are requested to read the same
standard scenario before completing the rest of the survey. This is
to ensure that everyone is in the same ‘frame of mind’ and to thus
reduce extraneous bias in the study.”
Willingness to pay. Participants were presented with the same
set of goods and services as in Study 1 to ascertain their willing-
ness to pay for conspicuous purchases. The order of the goods and
services was no longer randomized and the first four items in the
set—that is, a new watch, a dinner with friends, a new mobile
phone and a short vacation abroad—corresponded to the four
goods and services of Griskevicius et al. (2007, Study 1) that were
included in Study 1 of the present research. To eliminate potential
order effects, the arrangement of these first four items was re-
versed roughly half way through data collection and the same was
separately done for the remaining six goods and services.
In line with the response format of Griskevicius et al.’s first
experiment (2007, Study 1), participants indicated how much
money they would be willing to spend on the various items using
an 11-point scale for each item. Each point on the scales repre-
sented a specific monetary value and the range of values was
separately predefined for each item on the basis of the frequency
distributions of prices on online retail websites. The middle value
(6) on each scale was set to the respective average price used as the
anchor value in Study 1. The minimum and maximum scale points
were set as round integer values allowing for simple integer
increments between the scale points given the average price as the
middle scale point, and the scale increments between points was
constant for each item. Participants were no longer given the
option to not spend money on individual items.
Results
Willingness to pay. The scale point values (not the monetary
values associated with the scale points) were summed individually
for each participant across the four items taken from Griskevicius
et al.’s research (2007, Study 1) and also across the larger set of 10
items from Study 1 above.
Across the 10-item set, and replicating the findings of Study 1,
gender did not significantly moderate the effect of priming condi-
tion on the amount of money participants indicated they would pay
for goods and services, F(1, 121) 1.27, p.26, p
2.01. In
addition, neither the main effect of prime condition, F(1, 121)
0.02, p.88, p
2.00, nor of gender, F(1, 121) 2.22, p.14,
p
2.018, was significant. Contrast analyses also did not support
the costly signaling interpretation of male luxury good consump-
tion: Men in the experimental condition did not spend more on
conspicuous purchases than those in the control condition, and
indeed the effect was in the wrong direction, t(52) ⫽⫺0.65,
one-tailed p.74. The willingness to pay of women primed with
mating motives also did not significantly differ from those who
were presented with the neutral prime, t(69) 0.97, two-tailed
p.34.
Limiting the statistical analysis to the subset taken from
Griskevicius et al.’s experiment (2007, Study 1) did not change the
pattern of results. There was no significant interaction between
gender and priming condition on expenditure for the reduced set of
goods and services, F(1, 121) 0.77, p.38, p
2.006. Neither
the main effect of prime condition, F(1, 121) 0.00, p.99,
p
2.00, nor of gender, F(1, 121) 2.96, p.09, p
2.024, was
significant. Men primed with the romantic scenario did not have a
significantly higher willingness to pay in comparison to those in
the control condition, t(52) ⫽⫺0.60, one-tailed p.73. The same
result emerged with respect to the effect of priming condition on
expenditure levels for female participants, t(69) 0.65, two-tailed
p.52.
Bayes factors are reported in Table 3. These are again in excess
of four for male participants for both the complete and the reduced
item sets, indicating clear support for the null hypothesis.
Discussion
Studies 1 and 2 failed to replicate the pattern of results of
previous studies, in particular the findings of Study 1 of Griskevi-
cius et al. (2007). Contrary to what was expected, priming male
participants with mating motives did not lead them (in contrast to
female participants) to increase their willingness to pay for pub-
licly visible goods and services. The spending behavior of both
male and female participants in the experimental conditions did
not significantly differ from that in the control conditions.
Study 3
Study 3 is a further attempt to obtain a priming effect, again
making the procedure closer to that of the previously published
studies. Specifically, the study adopts the same design as Griskevi-
cius et al.’s (2007) Study 2, with participants responding not only
to consumption but also to benevolence questions, and using the
same response format that they used. Griskevicius et al. observed
a reliable tendency for romantic primes to increase males’ con-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
MATING MOTIVES AND DECISION MAKING
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
spicuous consumption judgments and to increase females’ con-
spicuous benevolence judgments.
Method
Participants. One hundred and sixty-three UCL students
completed the survey. They were given the option of receiving a
£2 Amazon voucher or partial course credit for filling out a
self-administered online survey. Data from 12 participants were
excluded on the same basis as in the previous studies. This resulted
in a final sample of 151 participants, of whom 87 were female and
64 were male. Their mean age was 27.7 (SD 10.9).
Design. The study used a 2 (Participant Gender: Male vs.
Female) 2 (Priming Condition: Prime vs. Control) 2 (Behav-
ior: Consumption vs. Benevolence) 2 (Conspicuousness: Con-
spicuous vs. Inconspicuous) mixed-factorial design. Gender and
priming condition were both between-subjects factors, and behav-
ior and conspicuousness were both within-subjects factors. Partic-
ipants were randomly assigned to one of the two priming condi-
tions and then all answered questions on spending and helping.
Procedure. Except where specifically mentioned, the proce-
dure was identical to that of Study 2. The text prime was again
used.
Willingness to pay. The items, presented one at a time in
randomized order, were the same goods and services used by
Griskevicius et al. (2007, Study 2). The response format was also
the same as used in that study. Five items were conspicuous goods
and five inconspicuous. The five conspicuous consumption items
were: a new car, a new watch, taking a group of friends out for
dinner, a new mobile phone, and a nice holiday somewhere in
Europe. The five inconspicuous consumption items were: basic
toiletries (e.g., tissues), household medication (e.g., headache med-
ication), a bedroom alarm clock, kitchen staples (e.g., salt), and
household cleaning products (e.g., tile cleaner). Participants indi-
cated how much they would be willing to pay on a 9-point scale:
1(much less than the average person), 5 (about average), and 9
(much more than the average person).
Motivational booster. As in Griskevicius et al.’s (2007) Study
2, a motivational booster was given after the randomly assigned
first block of either consumption or benevolence questions. Par-
ticipants were told they were completing a recall task and that they
were to imagine themselves in the scenario they had read at the
start of the experiment. In the experimental condition participants
were told they had up to 3 min to describe their ideal mate. In the
control condition participants were asked to spend up to 3 min
describing the anticipated concert venue.
Willingness to help. Participants were asked to indicate their
willingness to help in 10 randomized situations. Five were con-
spicuous benevolence situations: volunteering at a homeless shel-
ter, helping to build houses for poor families, teaching underpriv-
ileged youths how to read, mentoring a young person, and
volunteering in a children’s hospital. The other five were incon-
spicuous benevolence situations: spending an afternoon each
weekend picking up rubbish alone in a park, taking much shorter
showers to conserve water, putting money into a stranger’s parking
meter when time had expired, posting a letter someone had
dropped on the way to the post office, and going to the library to
drop off a found library book in the drop box. Responses were
given on the same 9-point scale used for the consumption items
above.
Results
Willingness to pay and help. The conspicuous and incon-
spicuous consumption and benevolence scores were aggregated for
each participant by averaging across each set of five questions. A
full statistical analysis, revealing a pattern of strategic sex-specific
displays similar to that observed by Griskevicius et al. (2007,
Study 2), is provided as supplemental materials. In brief, it dem-
onstrates that our measurement instrument is sensitive to variation
in gender, domain, and conspicuousness. For example participants
rated themselves as more willing than the average person to pay
for acts of benevolence but less willing than the average person to
pay for consumption items; and rated themselves as relatively
more likely to engage in conspicuous compared with inconspicu-
ous activities.
However, whether participants were in the experimental condi-
tion viewing the romantic prime or in the control condition view-
ing the neutral prime did not influence how likely they were to
engage in conspicuous or inconspicuous consumption or benevo-
lence. The relevant means are presented in Table 3 together with
the confidence intervals on the differences. In contradiction to
Griskevicius et al.’s (2007) findings, there was no reliable ten-
dency for romantic primes to increase males’ conspicuous con-
sumption WTP judgments nor to increase females’ conspicuous
benevolence WTP judgments. Table 3 also reports the relevant
Bayes factors that are again in excess of 4 for male participants,
supporting the null hypothesis.
Study 4
This study attempts to see if an increase in conspicuous con-
sumption can also be found using a different cover story and prime
(pictures of attractive opposite sex members). These were adapted
from Study 1 of Griskevicius et al. (2007) that also used picture
primes.
Method
Except where noted this study was identical to Study 3.
Participants. One hundred and fifty-two participants, re-
cruited through the UCL psychology participant panel, completed
the survey. Twelve datasets were excluded on the same basis as
previously. This resulted in a sample of 140 participants of whom
68 were males and 72 female. Their mean age was 22.3 (SD
6.1).
Procedure. A 2 (Gender: Male vs. Female) 2 (Priming
Condition: Prime vs. Control) factorial design was used. The
introductory information simply stated that participants would
have to do “Two brief unrelated studies to provide materials for
future experiments.” The first survey was said to be a “Picture
Preference” task, rating photographs for a future experiment, and
the second survey was said to be a market research questionnaire
on consumer preferences.
Priming. The primes were adapted from Griskevicius et al.
(2007, Study 1). In the experimental condition participants viewed
photos of three attractive opposite sex faces and were told that
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 SHANKS ET AL.
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
each individual was interested in pursuing a relationship with
them. They were then asked to rate each face’s attractiveness on a
scale from 0 (very unattractive)to10(very attractive). Participants
were subsequently asked “Who would you like to go on a first date
with the most?” and were again shown the images. For the chosen
individual participants then spent up to 3 min describing their
perfect date with that person.
Participants in the control condition viewed three photos of
streets and rated how much they liked each on a scale from 0
(dislike extremely)to10(like extremely). They then selected which
street they liked the most out of the three, and spent up to 3 min
describing “the most pleasant weather conditions in which to walk
around and look at the buildings.” This is identical to the control
task used by Griskevicius et al. (2007, Study 1) except that their
participants only viewed one street photo. By requiring a selection
from among three photos, the present control condition more
closely matches the task in the experimental condition.
Conspicuous consumption. Participants were then told they
were starting study two, which required them to indicate how
much they would be willing to pay for various goods and services
in comparison to the average person, again using a 9-point scale in
relation to the average person. As this study was only looking at
the influence of primes on conspicuous consumption, only five
questions were included.
Results
Conspicuous consumption. An aggregate conspicuous con-
sumption score was generated by averaging the five willingness-
to-pay scores. Then a 2 (Gender) 2 (Condition: Prime vs.
Control) ANOVA was conducted. This revealed no main effect of
gender, F(1, 136) .20, p.65, p
2.001, or condition, F(1,
136) .12, p.73, p
2.001, and no interaction, F(1, 136)
0.65, p.42, p
2.005. Planned contrasts found no priming
effect in either male, t(66) 0.30, one-tailed p.38, or female,
t(70) ⫽⫺0.89, two-tailed p.38, participants. Means are listed
in Table 3. The Bayes factor for males is once again in excess of
4.
Study 5
In Study 5 we again used the picture prime procedure but
extended the dependent measure to include risk-taking behavior
(Greitemeyer et al., 2013) as well as inconspicuous consumption
(Griskevicius et al., 2007).
The experiment compared risk-taking behavior in two groups of
male participants who were either exposed to the experimental
prime or a control one. The priming procedure closely matched
that used by Griskevicius et al. (2007, Study 1) and in the present
Study 4. In the experimental condition participants were asked to
rate the attractiveness of women and to imagine a perfect date with
one of them. In the control condition participants rated the com-
petence of male managers and imagined what perfect team-
working with one of them would be like. The effect of these primes
on risk-taking behavior and consumption was measured in three
different questionnaires. One focused on social situations, one on
reckless driving, and one on conspicuous and inconspicuous con-
sumption. The effect of priming on risky driving behavior was
assessed by Greitemeyer et al. (2013, Experiment 3). Greitemeyer
et al. (2013, Experiment 4) included questions on social behavior
in a larger set of risk-propensity questions (the DOSPERT Scale;
Blais & Weber, 2006). Here we assess it as a separate domain.
The major difference between the design of this study and those
of Griskevicius et al. (2007) and Greitemeyer et al. (2013) was that
a more appropriate control prime was used in this experiment: it
involved pictures of people and hence was socially oriented
(though not mating-oriented), while the previous research used
images of streets.
Method
Participants. For this experiment, 85 males were recruited
from the UCL psychology participant panel. Six datasets were
excluded on the same basis as previously. The remaining partici-
pants had a mean age of 22.9 (SD 3.5) with 39 assigned to the
experimental group and 40 to the control group.
Materials and design. The picture primes for the experimental
condition were taken from HotnessRater.com. This website allows
users to rate the attractiveness of pictures of women from 0–10.
All pictures used in this study had at least 1,500 user votes and an
average rating of 9.5. The original source used by Greitemeyer et
al., binichsexy.de, is no longer active. Greitemeyer et al. (2013)
used a similar source and the same criteria of a minimum user vote
and rating to obtain their mating primes. Primes for the control
condition were pictures of smartly dressed men who might come
from an office-type working environment.
Three questionnaires were designed to assess risk-taking behav-
ior in everyday situations as well as consumption. Driving behav-
ior was assessed using the same 10-item questionnaire as Greit-
emeyer et al. (2013), originally designed by Ben-Ari, Florian, and
Mikulincer (1999). Participants were given a scenario, such as
“You are on your way to a weekend vacation. A very slow lorry is
driving just in front of you. A continuous white line separates you
and the other direction of the road. What do you think are the
chances that you will go for an overtake?” They then had to
indicate the likelihood of taking the suggested action on a scale of
1–100.
Social risk-taking was measured in five different scenarios of
possible confrontation. Again, participants were given a scenario
and this time two possible actions. One was the “risky” option,
which included confronting another person, and the other the
“safe” one, which did not. The following is an example scenario:
You are shopping at your local store and in queue behind a rather
attractive woman. As she approaches the till to pay for her groceries
the shop assistant loudly makes an insulting comment towards her.
What do you think the chances are that you A. step forward and tell
him to apologize, B. keep queuing and let her deal with it?
Here, A is the risky option and B the safe one. Participants were
asked to indicate how likely they were to take each of the sug-
gested actions on a scale of 1–100.
Finally, participants’ conspicuous and inconspicuous spending
behavior was examined using tests similar to those used previously
(but with only five items each), with responses made on scales
similar to those used in Study 2 (for conspicuous items) and Study
3 (inconspicuous items).
Procedure. The experiment consisted of a priming phase fol-
lowed by three blocks of questions. During the priming stage,
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
MATING MOTIVES AND DECISION MAKING
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
participants in the experimental condition initially rated the three
pictures of women individually, from 1 (not at all attractive)to10
(extremely attractive). They were then asked to choose which of
the women they would like to take out on a date and spent about
3 min writing about what the perfect date with that woman would
be like. In the control condition participants rated the three pictures
of businessmen individually, from 1 (not at all competent)to10
(extremely competent). They then had to choose which of them
they would like to work with most and write a short text about
what a perfect team-working experience with that person would be
like. After the priming stage, both groups were given the same
three blocks of questions in a randomized order.
Results
Risk-taking. For each participant, scores from the individual
items on the risky driving and social risk-taking questionnaires
were averaged, with responses reverse coded where relevant. High
scores indicate greater risk-taking. Because of a programming
error data for two of the social risk-taking questions were not
correctly recorded for 22 participants and their mean scores are,
therefore, based on the remaining eight questions.
Means and Bayes factors are listed in Table 3. There was no
significant priming effect on mean risky driving scores,
t(77) ⫽⫺0.85, one-tailed p.80, or on mean social risk-taking,
t(77) .59, p.28. Hence, the results from Greitemeyer et al.
(2013) were not replicated. Mating primes did not increase will-
ingness to take risks while driving, which is especially surprising
considering the unusually large effect reported by Greitemeyer et
al. (2013),d
s
1.62. Similarly, mating primes were not found to
increase men’s willingness to engage in risk-taking or confronta-
tion in social situations.
Conspicuous and inconspicuous consumption. There was
no effect of priming condition on participants’ spending behavior
for conspicuous, t(77) .14, one-tailed p.44, or inconspicuous
items, t(77) .08, p.93.
Study 6
Study 5 once again failed to detect any effect of priming mating
motives on males’ consumption behavior, and extended this null
effect to risk-taking. In Study 6 we undertook another evaluation
of priming effects on risk-taking, again following the procedure of
Greitemeyer et al. (2013), and incorporating measures of sexual,
substance, and gambling risk-taking. The present study is a repli-
cation of Greitemeyer et al.’s (2013) Experiments 1 and 2 in which
sexual and gambling risk-taking, respectively, were measured.
Baker and Maner (2008) reported a similar study in which the
primes were attractive (vs. unattractive) opposite sex faces and the
gambling dependent measure was risky choice in a simulated
blackjack game.
Method
Participants. Unlike Study 5, both male and female partici-
pants were included in the sample to allow us to determine whether
our questionnaires are sensitive to gender differences. There were
108 UCL students who completed the study online, with 11
datasets being excluded on the same basis as previously. This
resulted in a final sample of 97 participants of whom 56 were
males and 41 female. Their mean age was 22.0 (SD 5.5).
Approximately half (43/97) the participants were tested in person
via a laptop computer in a quiet room whereas the remainder were
tested online.
Procedure and materials. Except where noted, the procedure
was identical to that of Study 5 with the following changes.
Participants were asked their gender at the start of the experiment
so that appropriate opposite sex pictures could be presented. For
the control group we reverted to the pictures of street scenes to
make the experiment more similar to that of Greitemeyer et al.
(2013).
Presentation of the three risk-taking question sets (sexual risk-
taking, gambling, and substance abuse) was randomized. Four
gambling questions (two from Greitemeyer et al., 2013, plus two
similar new items) were constructed to measure gambling risk-
taking. For example, one question asked participants to choose
between a lottery ticket to win £100 where 1 million people enter
the lottery and 50,000 winners are chosen (the conservative op-
tion), and a ticket to win £500 where 1 million people enter the
lottery and 5,000 winners are chosen (the risky option). The sexual
risk-taking questionnaire incorporated eight items from the same
source as Greitemeyer et al. (2013), such as “If I find someone
attractive, I would agree to sexual intercourse even if it is unpro-
tected” and scored on an 11-point scale from 0 (strongly disagree)
to 10 (strongly agree). A 14-item substance risk-taking question-
naire was constructed asking participants to judge how likely (0
very unlikely,10very likely) they were, for instance, to consume
a Class A drug (e.g., cocaine) in their lifetime.
Results
For each participant, scores from the individual items on the
substance and sexual risk-taking questionnaires were averaged,
with responses reverse coded where relevant. High scores indicate
greater risk-taking. Responses on the gambling questions were
scored as 0 if the conservative option was selected and 1 if the
risky option was selected, and then summed, yielding a score
between 0 and 4. Means scores on each measure for males and
females are reported in Table 3 together with the 95% CIs on the
priming effects, and the ensuing Bayes factors. Mode of testing
(experimenter present vs. online) had no effect on the results.
Sexual risk-taking. A 2 (Gender) 2 (Condition: Prime vs.
Control) ANOVA revealed a significant main effect of gender,
F(1, 93) 7.34, p.01, p
2.072. In the domain of sexual
risk-taking, males were substantially more risk-seeking than fe-
males, as found by Greitemeyer et al. (2013, Experiment 1).
However there was no effect of priming condition, F(1, 93)
1.15, p.29, p
2.011, and no interaction, F1.
Gambling risk-taking. A comparable ANOVA revealed no
main effects and no interaction, F1 in each case. The absence
of a significant gender effect is consistent with the results of
Greitemeyer et al. (2013, Experiment 2).
Substance risk-taking. A comparable ANOVA revealed no
main effects and no interaction, F1 in each case.
Study 6 once again fails to obtain any evidence for priming
effects on measures of risk-taking, in this case in the domains of
sexual, gambling, and substance-abuse behavior. There was no
hint that risk-taking is affected by romantic primes more in males
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
12 SHANKS ET AL.
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
than females, as Greitemeyer et al. (2013) reported for sexual
(Experiment 1) and gambling (Experiment 2) behavior (indeed
Greitemeyer et al.’s Experiment 2 found significantly less risk
taking in females in the mating prime compared with the control
condition). More important, these null effects do not simply reflect
complete insensitivity with our participants using these measures,
as a robust gender effect on sexual risk-taking was observed. There
was no gender effect on the gambling questionnaire. Both of these
patterns align with what Greitemeyer et al. (2013) found. For the
domains common to this study and the experiments of Greitemeyer
et al. (sexual and gambling behavior) the overall mean risk-taking
scores were very comparable.
Studies 7a and 7b
In Study 7a we undertook an even closer replication of Greit-
emeyer et al.’s (2013) Experiment 1 measuring sexual risk-taking.
In Study 7b we did the same for their Experiment 2 measuring
gambling. To test the possibility that participants might respond
differently in laboratory and online conditions, participants in
Study 7a were tested in both contexts.
Method
Participants. Only male participants were included in the
sample. There were 126 participants who completed Study 7a,
with 6 datasets being excluded on the same basis as previously.
This resulted in a final sample of 120 participants with a mean age
of 21.7 (SD 3.7). Thirty-seven UCL participants were tested in
laboratory cubicles (this is larger than Greitemeyer et al.’s sample,
n31), whereas the remainder, recruited via Prolific Academic
(www.prolific.ac) or the UCL participant pool, were tested online.
One hundred and nine participants, all recruited via Prolific Aca-
demic, completed Study 7b, with 3 datasets being excluded. This
resulted in a final sample of 106 participants with a mean age of
27.3 (SD 10.3). Here, all participants were tested online.
Procedure and materials. The procedure was identical to that
of Study 6 with the following changes. In Study 7a participants
completed the sexual risk-taking questionnaire. In Study 7b they
answered four gambling questions, identical to those used by
Greitemeyer et al. (2013, Experiment 2). In both studies, a final set
of questions asked all participants to provide attractiveness ratings
of the three pictures of women and the street scene. Combining
data across both experiments, the former received significantly
higher ratings: M6.58 (SD 1.38) and M3.99 (SD 1.87)
for the faces and street picture, respectively, t(218) 17.40,
two-tailed p.001.
Results
Means scores are reported in Table 3 together with the 95% CIs
on the priming effects and Bayes factors.
Study 7a: Sexual risk-taking. There was no effect of priming
condition. Men primed with mating motives showed no tendency
toward more risky behavior in comparison with those in the
control condition, t(118) 0.07, one-tailed p.47. The subset of
participants tested under laboratory conditions showed the same
absence of a priming effect [M3.30 (SD 1.73) and M3.62
(SD 2.04) for the priming and control groups, respectively],
t(35) ⫽⫺0.52, one-tailed p.70.
Study 7b: Gambling risk-taking. There was again no effect
of priming condition. Men primed with mating motives did not
have a significantly higher willingness to choose risky gambles in
comparison with those in the control condition, t(104) ⫽⫺0.35,
one-tailed p.64.
Study 8
In the final study we examine the reproducibility of the effect of
mating and self-protection primes on another dependent measure
from the domain of decision making, loss aversion, using a method
closely modeled on that of Li et al. (2012, Study 1). Self-protection
primes activate thoughts of fear and the need to protect oneself
against danger, an evolutionary motivation assumed to be gender
nonspecific. Li et al.’s key findings were that, relative to the
neutral control condition, a mating prime rendered males but not
females less averse to losses, while a self-protection prime in-
creased loss aversion, regardless of gender.
Study 8 was preregistered and used a large sample, comprising
(like Study 7a) samples tested in the laboratory and online. Pre-
registration virtually eliminates the possibility of publication bias
and p-hacking (Chambers, Feredoes, Muthukumaraswamy, &
Etchells, 2014), because the study protocol and analysis plan are
specified ahead of time. In a large preregistered experiment,
Gomes and McCullough (2015) failed to reproduce a different
priming effect, of religious primes on prosocial behavior.
Method
Participants. The total sample comprised 670 participants of
both genders, with 20 datasets being excluded on the same basis as
previously. This resulted in a final sample of 650 participants (325
males) with a mean age of 27.9 (SD 9.3). Seventy UCL
participants were tested in laboratory cubicles, whereas the re-
mainder were recruited via Prolific Academic and tested online.
Participants were allocated at random to one of three groups,
control, mating prime, and self-protection prime.
Procedure and materials. The study was preregistered in
detail at https://osf.io/g2nek/. The priming stage was similar to
previous experiments but also included a self-protection text, as
used by Li et al. (2012, Study 3). The test phase included three
measures, presented in a randomized order. Participants responded
to a set of questions, identical to those used by Li et al., designed
to measure loss aversion. Participants were presented with seven
different attribute items (being liked, being respected, providing
for their family, safety from physical danger, safety from conta-
gious disease, dating ability, and romantic relationship stability)
and asked to imagine they were on the 50th percentile on each of
the attributes. They then indicated how much of £1000 they would
pay to improve (gain) or avoid a decrease (loss) on each of the
attributes by 10 and 30%. The order of the 28 questions [7
attributes 2 percentiles (30%, 50%) 2 outcomes (gain/loss)]
was randomized.
Li et al. (2012, Study 1) elicited willingness-to-pay judgments
with an 11-point scale, from $0 to $1,000 in increments of $100.
The mean effect of the mating prime on male participants’ loss
aversion was a reduction of only about $7, which is equivalent to
a change of one point on the 11-point scale for 2 of the 28 items
rated by a participant in the prime condition compared to a par-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
13
MATING MOTIVES AND DECISION MAKING
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
ticipant in the control condition. Given this very small influence,
we deemed it appropriate to replace the 11-point scale with a
continuous scale where participants could enter any monetary
amount between £0 and £1000.
Participants also completed another loss aversion test, taken
from Tom, Fox, Trepel, and Poldrack (2007). In this test, partic-
ipants were presented with 20 gambles, in randomized order. Each
gamble offered a 50% chance of winning an amount shown in
green and a 50% chance of losing an amount shown in red. The
wins ranged from £10-£40 (in increments of £2) and the losses
from £5-£20 (in increments of £1), and half the gambles displayed
greater losses than gains. Participants decided whether to accept or
reject each gamble.
A single gamble similar to those used by Greitemeyer et al. (2013)
and in Studies 6 and 7b was also included in the test. Participants
chose between a conservative option (a 1 in 4 chance of winning £20)
and a riskier one (a 1 in 25 chance of winning £50).
Results
Data collection proceeded in accordance with the preregistration
plan with no noteworthy deviations. A full description of the results is
provided as supplemental material. We computed a loss aversion
measure in the same way as Li et al. (2012), comprising a single score
calculated for each participant by subtracting the amounts paid to
avoid a loss from the amounts paid to achieve a gain, aggregated
across all attributes and both percentile changes. Negative scores thus
imply loss aversion. Mean scores in the mating prime and control
conditions are reported in Table 3 together with the 95% CIs on the
priming effects and Bayes factors.
Participants were loss-averse on average, £9.19, 95% CI
[15.22, 3.17], meaning that they were willing to pay more to
avoid a loss than to acquire a comparable gain. Those tested online
were significantly more loss averse than those tested in the labo-
ratory. However mode of testing did not interact with any other
factors. Combined across this factor, there were no effects of
priming condition. Males primed with mating motives were no less
averse to losses than those in the control condition, t(216) 0.02,
one-tailed p.49, and the Bayes factor in support of the null is
nearly 10. The effect for females was also nonsignificant,
t(215) ⫽⫺0.06, two-tailed p.95. Also contrary to the predic-
tions, the self-protection prime did not make either males, t(214)
0.16, one-tailed p.44, or females, t(214) ⫽⫺0.83, one-tailed
p.80, more averse to losses.
General Discussion
The studies reported here can be readily summarized: They have
failed to detect any effects of mating primes on risk-taking, ex-
penditure on publicly consumed goods and services, or loss aver-
sion. Indeed, as indicated by the Bayes factor analyses, their results
strongly support the null hypothesis of no effect. Together with the
asymmetric funnel plot shown in Figure 2, which implies the
existence either of p-hacking in previously published studies or
selective publication of results (or both), our results suggest the
real possibility that romantic primes have no meaningful effect on
decision-making behaviors.
Although the major findings comprise null results, our experi-
ments were able to replicate other anticipated effects unrelated to
priming, including quite subtle ones. For example we confirmed in
Study 3 that males and females differ in their judgments concern-
ing consumption and benevolence, depending on whether the
behaviors in question are conspicuous or inconspicuous, mirroring
the pattern reported by Griskevicius et al. (2007, Study 2); in Study
6, we confirmed that males were substantially more risk-seeking
than females but only in the domain of sexual risk-taking and not
gambling, exactly as found by Greitemeyer et al. (2013); and in
Study 8 participants were significantly loss averse.
There is no such thing as an “exact” replication (Stroebe & Strack,
2014) and hence it must be acknowledged that the published studies
(notwithstanding the evidence for p-hacking and/or publication bias)
may have obtained genuine effects and that undetected moderator
variables explain why the present studies failed to obtain priming.
Some of the experiments reported here differed in important ways
from those on which they were modeled (although others were closer
replications and even these failed to obtain evidence of reliable
romantic priming). As Stroebe and Strack (2014) point out, what is
crucial is not so much exact surface replication but rather identical
operationalization of the theoretically relevant variables. In the pres-
ent case, the crucial factors are the activation of romantic motives and
the appropriate assessment of consumption, risk-taking, and other
measures. For instance testing the same participant population as an
original study but with different pictures of attractive individuals is
likely to be a closer and more valid replication than using the same
pictures but in participants from a different culture with different
views of attractiveness. In the former but not the latter, a theoretically
relevant factor—activation of romantic motives—will be reproduced.
Is it possible that the present studies failed to achieve these
requirements? It seems unlikely. Published studies have used a
range of priming methods with no hint that these methods differ in
efficacy. For instance, both picture and text primes have been used,
with and without participants writing about their thoughts. Indeed
published studies have found priming effects with minimal induc-
tions such as that used by Kim and Zauberman (2013, Study 2) in
which participants rated the attractiveness of each of seven pho-
tographs, shown for 7 s each, a task that presumably took little
more than 1 min. Our participants come from a very similar
language and cultural population as those previously studied
(themselves quite varied), and the priming materials surely acti-
vated romantic concepts effectively in our participants (as explic-
itly confirmed in Studies 7a and 7b and in the manipulation check
reported in the supplemental materials). Our measures of con-
sumption, risk-taking and other decisions used similar or identical
methods to those used previously.
One potential moderator variable is study format (laboratory or
online) with the majority of published studies using laboratory data
collection and the majority of our studies using online collection.
Is it likely that this difference is crucial? Some of the previous
studies (Chan, 2015;Kim & Zauberman, 2013, Study 5; Sundie et
al., 2011, Study 2) were conducted online (as were other studies
within this general field, e.g., Wang & Griskevicius, 2014) and
obtained priming effects similar to those found in their companion
laboratory experiments. At the same time, two of the experiments
reported here (Studies 7a and 8) specifically asked whether prim-
ing effects differed between the two formats and obtained no
evidence of this, consistent with a growing literature from a range
of domains demonstrating a similar conclusion (Germine et al.,
2012;R. A. Klein et al., 2014). Indeed in their meta-analysis of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
14 SHANKS ET AL.
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
religious priming studies, Shariff et al. (2015) tested whether the
results varied between online and laboratory studies and found no
evidence that they did. This is certainly something to be explored
further in future research, but the available evidence provides no
strong support for this being an important moderating factor. One
of the clear advantages of online experiments is that direct human
interaction between researcher and participant is limited thus leav-
ing less room for experimenters to bias participants’ responses.
A second potential moderator is whether participants are primed
with or have a bias toward short- versus long-term mating motives.
This refers to the fact that individuals differ in their preference for
uncommitted, low-investment, unrestricted sexual strategies in
contrast to committed, high-investment, restricted, ones. As well
as this intrinsic preference, participants can be presented with
inductions designed to trigger either short- or long-term mating
motives, for example via texts that emphasize either the low- or
high-commitment interests of the protagonists in a romantic en-
counter. Some research (Sundie et al., 2011) has suggested that
romantic priming of male conspicuous consumption is limited to
those following a low-investment strategy and to short-term mat-
ing primes. Although this moderator should continue to be evalu-
ated in future research, it is unlikely that it is an important factor,
for several reasons: (a) the evidence reported by Sundie et al.
(2011) is inconclusive: they found significant priming in partici-
pants presented with a short-term prime but not in those presented
with a long-term prime, and likewise they found significant prim-
ing in unrestricted but not in restricted males; however in neither
case did they report that these factors significantly moderated the
degree of priming; (b) other studies (Greitemeyer et al., 2013,
Experiments 3 and 4; Griskevicius et al., 2007, Study 2, see
Footnote 1, p. 90) found no difference between short- and long-
term mating primes, while Griskevicius et al. (2007, Study 3)
found no effect of male participants’ preference for pursuing a
short-term compared to a long-term mating strategy on the extent
to which consumption behavior was primed; and (c) the effects
obtained in published studies on other aspects of decision-making
(see Table 1) did not depend on (and indeed took no account of)
short- versus long-term mating motives.
The conclusion we drew in the Introduction from the funnel plot
asymmetry revealed in Figure 2 was that the published studies do
not license any firm statement about the true magnitude of the
effect of mating primes on decision making. Despite the fact that
43 contrasts on data from 3,252 participants yielded 42 significant
priming effects and a meta-analytic effect size of d0.57, those
studies do not allow the true effect size to be determined unam-
biguously because of the correlation they manifest between effect
size and sample size. However, in light of the results of the
additional data we have collected, this conclusion can now be
considerably strengthened: We infer that the true effect size for
those 43 contrasts is very close to zero. Egger’s test yields an
estimated intercept of d⫽⫺0.06 [0.27, 0.14], and for the main
experimental tests in Studies 1–8 the meta-analytic effect size is
d0.00 [0.12, 0.11]. The individual effect sizes
2
contributing
to the latter are illustrated in the open triangles in Figure 2 and in
the forest plot in Figure 3. Strikingly, the two datasets (previously
published studies represented by the black circles in Figure 2, and
the main results across Studies 1–8 reported here, represented by
the open triangles) fall into complete alignment. The previous
studies show evidence of publication bias and the correction for
this yields an estimated effect size close to zero, similar to that of
the studies reported here (for which a publication bias correction is
not required).
Our conclusion may appear surprising but it is consistent with
meta-analyses elsewhere in decision-making research. For exam-
ple, Nieuwenstein et al. (2015) described a meta-analysis of stud-
ies on the ‘unconscious thought’ effect and concluded that, despite
many positive findings (e.g., Dijksterhuis, Bos, Nordgren, & van
Baaren, 2006), the true effect size after correcting for publication
bias is negligible. Renkewitz, Fuchs, and Fiedler (2011) reana-
lyzed a meta-analysis by Dato-on and Dahlstrom (2003) on prim-
ing effects in decision making and found evidence of publication
bias, leading them to conclude that Dato-on and Dahlstrom had
overestimated the true effect size. And Carter and McCullough
(2014) reported a meta-analysis which implied that—after correct-
ing for publication bias—the tendency for acts of self-control to
cause “depletion” of a common resource has an effect size that is
no greater than zero. In all of these cases substantial bodies of
evidence are severely compromised by publication bias.
Although the present results cast some doubt on the claim that
aspects of decision-making behavior can be primed by the activation
of mating motives, there are other phenomena which, though super-
ficially similar, are not at all challenged by the null results obtained
here. For example, it has been demonstrated that the presence of an
opposite sex observer may alter an individual’s willingness to take
risks (e.g., Ronay & von Hippel, 2010). Such effects may be mediated
by direct physiological arousal and indeed Ariely and Loewenstein
(2006) showed that sexual arousal rendered males more willing to
take risks. The concept of behavior priming (Molden, 2014) does not
need to be invoked to explain such effects.
Behavior priming is both theoretically and empirically question-
able (Newell & Shanks, 2014a,2014b). It is theoretically prob-
lematic because the claim that subtle cues can unconsciously
activate the “mental representations of social targets, events, or
situations that then influences subsequent evaluations, judgments,
or actions” (Molden, 2014, p. 4) runs counter to well-established
theories in which behavior is understood in relation to the con-
scious, cognitive appraisal of cues and situations (Baumeister,
Masicampo, & Vohs, 2011;Lovibond & Shanks, 2002;Newell &
Shanks, 2014b). It is empirically problematic because—consistent
with the findings reported here—many of the most influential
demonstrations of behavior priming have proved to be very diffi-
cult to replicate (e.g., Gomes & McCullough, 2015;R. A. Klein et
al., 2014;Pashler et al., 2012;Shanks et al., 2013).
As noted earlier in the article, research on the effects of mating
primes on decision-making behaviors is connected to a larger body
of work in which (a) these primes have been paired with other
dependent measures such as creativity and aggression, and (b)
decision-making behaviors have been paired with other types of
primes (see Kenrick & Griskevicius, 2013, for a review). Our
findings of course must not be overgeneralized, but future research
should take seriously the possibility that publication bias and/or
p-hacking are present in these associated domains too and should
2
The dependent variables in the meta-analysis of Studies 1–8 are not all
statistically independent: Within-study measures come from the same
participants. We include all those data points for illustrative purposes. The
results do not change substantially if all nonindependent data points from
a study are collated into a single composite effect size.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
15
MATING MOTIVES AND DECISION MAKING
Fn2
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
take steps to minimize that risk. Preregistration represents an
important new method that future studies could profitably use to
enable firmer conclusions about these priming effects to be
reached. This refers to the practice of specifying the study protocol
and analytic strategy ahead of data collection. Preregistration, with
its attendant minimization of the possibility of publication bias and
p-hacking (Chambers et al., 2014), offers the possibility of placing
romantic priming of decision-making behaviors (and related forms
of priming) on a firmer footing than has been achieved to date.
References
Ariely, D., & Loewenstein, G. (2006). The heat of the moment: The effect
of sexual arousal on sexual decision making. Journal of Behavioral
Decision Making, 19, 87–98. http://dx.doi.org/10.1002/bdm.501
Asendorpf, J. B., Conner, M., De Fruyt, F., De Houwer, J., Denissen,
J. J. A., Fiedler, K.,...Wicherts, J. M. (2013). Recommendations for
increasing replicability in psychology. European Journal of Personality,
27, 108–119. http://dx.doi.org/10.1002/per.1919
Baker, M. D., Jr., & Maner, J. K. (2008). Risk-taking as a situationally
sensitive male mating strategy. Evolution and Human Behavior, 29,
391–395. http://dx.doi.org/10.1016/j.evolhumbehav.2008.06.001
Baker, M. D., Jr., & Maner, J. K. (2009). Male risk-taking as a context-
sensitive signaling device. Journal of Experimental Social Psychology,
45, 1136–1139. http://dx.doi.org/10.1016/j.jesp.2009.06.006
Bakker, M., van Dijk, A., & Wicherts, J. M. (2012). The rules of the game
called psychological science. Perspectives on Psychological Science, 7,
543–554. http://dx.doi.org/10.1177/1745691612459060
Baumeister, R. F., Masicampo, E. J., & Vohs, K. D. (2011). Do conscious
thoughts cause behavior? Annual Review of Psychology, 62, 331–361.
http://dx.doi.org/10.1146/annurev.psych.093008.131126
Ben-Ari, O. T., Florian, V., & Mikulincer, M. (1999). The impact of
mortality salience on reckless driving: A test of terror management
mechanisms. Journal of Personality and Social Psychology, 76, 35–45.
http://dx.doi.org/10.1037/0022-3514.76.1.35
Blais, A. R., & Weber, E. U. (2006). A Domain-Specific Risk-Taking
(DOSPERT) scale for adult populations. Judgment and Decision Mak-
ing, 1, 33–47.
Brener, N. D., Billy, J. O. G., & Grady, W. R. (2003). Assessment of
factors affecting the validity of self-reported health-risk behavior among
adolescents: Evidence from the scientific literature. Journal of Adoles-
cent Health, 33, 436457. http://dx.doi.org/10.1016/S1054-
139X(03)00052-1
Carter, E. C., & McCullough, M. E. (2014). Publication bias and the
limited strength model of self-control: Has the evidence for ego deple-
tion been overestimated? Frontiers in Psychology, 5, 823. http://dx.doi
.org/10.3389/fpsyg.2014.00823
Chambers, C. D., Feredoes, E., Muthukumaraswamy, S. D., & Etchells,
P. J. (2014). Instead of “playing the game” it is time to change the rules:
Registered Reports at AIMS Neuroscience and beyond. AIMS Neurosci-
ence, 1, 4–17.
Chan, E. Y. (2015). Physically-attractive males increase men’s financial
risk-taking. Evolution and Human Behavior, 36, 407–413. http://dx.doi
.org/10.1016/j.evolhumbehav.2015.03.005
Dato-on, M. C., & Dahlstrom, R. (2003). A meta-analytic investigation of
contrast effects in decision making. Psychology & Marketing, 20, 707–
731. http://dx.doi.org/10.1002/mar.10093
Dijksterhuis, A., Bos, M. W., Nordgren, L. F., & van Baaren, R. B. (2006).
On making the right choice: The deliberation-without-attention effect.
Science, 311, 1005–1007. http://dx.doi.org/10.1126/science.1121629
Durgin, F. H., Baird, J. A., Greenburg, M., Russell, R., Shaughnessy, K.,
& Waymouth, S. (2009). Who is being deceived? The experimental
demands of wearing a backpack. Psychonomic Bulletin & Review, 16,
964–969. http://dx.doi.org/10.3758/PBR.16.5.964
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in
meta-analysis detected by a simple, graphical test. British Medical
Journal, 315, 629634. http://dx.doi.org/10.1136/bmj.315.7109.629
Ferguson, C. J., & Brannick, M. T. (2012). Publication bias in psycholog-
ical science: Prevalence, methods for identifying and controlling, and
implications for the use of meta-analyses. Psychological Methods, 17,
120–128. http://dx.doi.org/10.1037/a0024445
Festjens, A., Bruyneel, S., & Dewitte, S. (2014). What a feeling! Touching
sexually laden stimuli makes women seek rewards. Journal of Consumer
Psychology, 24, 387–393. http://dx.doi.org/10.1016/j.jcps.2013.10.001
Flore, P. C., & Wicherts, J. M. (2015). Does stereotype threat influence
performance of girls in stereotyped domains? A meta-analysis. Journal
of School Psychology, 53, 25–44. http://dx.doi.org/10.1016/j.jsp.2014
.10.002
Germine, L., Nakayama, K., Duchaine, B. C., Chabris, C. F., Chatterjee,
G., & Wilmer, J. B. (2012). Is the Web as good as the lab? Comparable
performance from Web and lab in cognitive/perceptual experiments.
Psychonomic Bulletin & Review, 19, 847–857. http://dx.doi.org/10
.3758/s13423-012-0296-9
Gigerenzer, G. (1991). How to make cognitive illusions disappear: Beyond
“heuristics and biases.” European Review of Social Psychology, 2,
83–115. http://dx.doi.org/10.1080/14792779143000033
Gomes, C. M., & McCullough, M. E. (2015). The effects of implicit
religious primes on dictator game allocations: A preregistered replica-
tion experiment. Journal of Experimental Psychology: General. [Ad-
vance online publication.] http://dx.doi.org/10.1037/xge0000027
Greitemeyer, T., Kastenmüller, A., & Fischer, P. (2013). Romantic motives
and risk-taking: An evolutionary approach. Journal of Risk Research,
16, 19–38. http://dx.doi.org/10.1080/13669877.2012.713388
Griskevicius, V., Tybur, J. M., Sundie, J. M., Cialdini, R. B., Miller, G. F.,
& Kenrick, D. T. (2007). Blatant benevolence and conspicuous con-
sumption: When romantic motives elicit strategic costly signals. Journal
of Personality and Social Psychology, 93, 85–102. http://dx.doi.org/10
.1037/0022-3514.93.1.85
Hill, S. E., & Durante, K. M. (2011). Courtship, competition, and the
pursuit of attractiveness: Mating goals facilitate health-related risk tak-
ing and strategic risk suppression in women. Personality and Social
Psychology Bulletin, 37, 383–394. http://dx.doi.org/10.1177/
0146167210395603
Janssens, K., Pandelaere, M., Van den Bergh, B., Millet, K., Lens, I., &
Roe, K. (2011). Can buy me love: Mate attraction goals lead to percep-
tual readiness for status products. Journal of Experimental Social Psy-
chology, 47, 254–258. http://dx.doi.org/10.1016/j.jesp.2010.08.009
Jeffreys, H. (1961). Theory of probability (3rd ed.). Oxford, United King-
dom: Oxford University Press.
John, L. K., Loewenstein, G., & Prelec, D. (2012). Measuring the preva-
lence of questionable research practices with incentives for truth telling.
Psychological Science, 23, 524–532. http://dx.doi.org/10.1177/
0956797611430953
Kenrick, D. T., & Griskevicius, V. (2013). The rational animal: How
evolution made us smarter than we think. New York, NY: Basic Books.
Kim, B. K., & Zauberman, G. (2013). Can Victoria’s Secret change the
future? A subjective time perception account of sexual-cue effects on
impatience. Journal of Experimental Psychology: General, 142, 328
335. http://dx.doi.org/10.1037/a0028954
King, J., McClelland, A., & Furnham, A. (2015). Sex really does sell: The
recall of sexual and non-sexual television advertisements in sexual and
non-sexual programmes. Applied Cognitive Psychology, 29, 210–216.
http://dx.doi.org/10.1002/acp.3095
Klein, O., Doyen, S., Leys, C., Magalhães de Saldanha da Gama, P. A.,
Miller, S., Questienne, L., & Cleeremans, A. (2012). Low hopes, high
expectations: Expectancy effects and the replicability of behavioral
experiments. Perspectives on Psychological Science, 7, 572–584. http://
dx.doi.org/10.1177/1745691612463704
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
16 SHANKS ET AL.
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
Klein, R. A., Ratliff, K. A., Vianello, M., Adams, R. B., Jr., Bahník, Š.,
Bernstein, M. J.,...Nosek, B. A. (2014). Investigating variation in
replicability: A “Many Labs” replication project. Social Psychology, 45,
142–152. http://dx.doi.org/10.1027/1864-9335/a000178
Larrick, R. P. (2004). Debiasing. In D. J. Koehler & N. Harvey (Eds.),
Blackwell handbook of judgment and decision making (pp. 316–337).
Oxford: Blackwell. http://dx.doi.org/10.1002/9780470752937.ch16
Li, Y. J. (2012). The functionality of risk-taking: Mating motivation,
relationship status, and sex differences. Dissertation, Arizona State
University.
Li, Y. J., Kenrick, D. T., Griskevicius, V., & Neuberg, S. L. (2012).
Economic decision biases and fundamental motivations: How mating
and self-protection alter loss aversion. Journal of Personality and Social
Psychology, 102, 550–561. http://dx.doi.org/10.1037/a0025844
Love, J., Selker, R., Marsman, M., Jamil, T., Verhagen, A. J., Ly, A.,...
Wagenmakers, E.-J. (2015). JASP (Version 0.6.5) [Computer software].
Lovibond, P. F., & Shanks, D. R. (2002). The role of awareness in
Pavlovian conditioning: Empirical evidence and theoretical implica-
tions. Journal of Experimental Psychology: Animal Behavior Processes,
28, 3–26. http://dx.doi.org/10.1037/0097-7403.28.1.3
McAlvanah, P. (2009). Are people more risk-taking in the presence of the
opposite sex? Journal of Economic Psychology, 30, 136–146. http://dx
.doi.org/10.1016/j.joep.2008.10.002
Miller, G. (2000). The mating mind: How sexual choice shaped the
evolution of human nature. New York, NY: Anchor Books.
Molden, D. C. (2014). Understanding priming effects in social psychology:
What is “social priming” and how does it occur? Social Cognition,
32(Suppl.), 1–11. http://dx.doi.org/10.1521/soco.2014.32.supp.1
Newell, B. R., & Shanks, D. R. (2014a). Prime numbers: Anchoring and its
implications for theories of behavior priming. Social Cognition,
32(Suppl.), 88–108. http://dx.doi.org/10.1521/soco.2014.32.supp.88
Newell, B. R., & Shanks, D. R. (2014b). Unconscious influences on
decision making: A critical review. Behavioral and Brain Sciences, 37,
1–19. http://dx.doi.org/10.1017/S0140525X12003214
Nieuwenstein, M. R., Wierenga, T., Morey, R. D., Wicherts, J. M., Blom,
T. N., Wagenmakers, E.-J., & van Rijn, H. (2015). On making the right
choice: A meta-analysis and large-scale replication attempt of the un-
conscious thought advantage. Judgment and Decision Making, 10, 1–17.
Orne, M. T. (1962). On the social psychology of the psychological exper-
iment: With particular reference to demand charcteristics and their
implications. American Psychologist, 17, 776–783. http://dx.doi.org/10
.1037/h0043424
Pashler, H., Coburn, N., & Harris, C. R. (2012). Priming of social distance?
Failure to replicate effects on social and food judgments. PLoS ONE, 7,
e42510. http://dx.doi.org/10.1371/journal.pone.0042510
Reichert, T. (2002). Sex in advertising research: A review of content,
effects, and functions of sexual information in consumer advertising.
Annual Review of Sex Research, 13, 241–273.
Renkewitz, F., Fuchs, H. M., & Fiedler, S. (2011). Is there evidence of
publication biases in JDM research? Judgment and Decision Making, 6,
870881.
Ronay, R., & von Hippel, W. (2010). The presence of an attractive woman
elevates testosterone and physical risk taking in young men. Social
Psychological & Personality Science, 1, 57–64. http://dx.doi.org/10
.1177/1948550609352807
Roney, J. R. (2003). Effects of visual exposure to the opposite sex:
Cognitive aspects of mate attraction in human males. Personality and
Social Psychology Bulletin, 29, 393–404. http://dx.doi.org/10.1177/
0146167202250221
Rosenthal, R., & Rubin, D. B. (1978). Interpersonal expectancy effects:
The first 345 studies. Behavioral and Brain Sciences, 1, 377–386.
http://dx.doi.org/10.1017/S0140525X00075506
Rouder, J. N., Speckman, P. L., Sun, D., Morey, R. D., & Iverson, G.
(2009). Bayesian ttests for accepting and rejecting the null hypothesis.
Psychonomic Bulletin & Review, 16, 225–237. http://dx.doi.org/10
.3758/PBR.16.2.225
Shanks, D. R., Newell, B. R., Lee, E. H., Balakrishnan, D., Ekelund, L.,
Cenac, Z.,...Moore, C. (2013). Priming intelligent behavior: An
elusive phenomenon. PLoS ONE, 8, e56515. http://dx.doi.org/10.1371/
journal.pone.0056515
Shariff, A. F., Willard, A. K., Andersen, T., & Norenzayan, A. (2015).
Religious priming: A meta-analysis with a focus on prosociality. Per-
sonality and Social Psychology Review. [Advance online publication.]
http://dx.doi.org/10.1177/1088868314568811
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive
psychology: Undisclosed flexibility in data collection and analysis al-
lows presenting anything as significant. Psychological Science, 22,
1359–1366. http://dx.doi.org/10.1177/0956797611417632
Simons, D. J. (2014). The value of direct replication. Perspectives on
Psychological Science, 9, 7680. http://dx.doi.org/10.1177/
1745691613514755
Stanley, T. D., & Doucouliagos, H. (2014). Meta-regression approxima-
tions to reduce publication selection bias. Research Synthesis Methods,
5, 60–78. http://dx.doi.org/10.1002/jrsm.1095
Sterne, J. A. C., Sutton, A. J., Ioannidis, J. P. A., Terrin, N., Jones, D. R.,
Lau, J.,...Higgins, J. P. T. (2011). Recommendations for examining
and interpreting funnel plot asymmetry in meta-analyses of randomised
controlled trials. British Medical Journal, 343, d4002. http://dx.doi.org/
10.1136/bmj.d4002
Stroebe, W., & Strack, F. (2014). The alleged crisis and the illusion of
exact replication. Perspectives on Psychological Science, 9, 59–71.
http://dx.doi.org/10.1177/1745691613514450
Sundie, J. M., Kenrick, D. T., Griskevicius, V., Tybur, J. M., Vohs, K. D.,
& Beal, D. J. (2011). Peacocks, Porsches, and Thorstein Veblen: Con-
spicuous consumption as a sexual signaling system. Journal of Person-
ality and Social Psychology, 100, 664680. http://dx.doi.org/10.1037/
a0021669
Tom, S. M., Fox, C. R., Trepel, C., & Poldrack, R. A. (2007). The neural
basis of loss aversion in decision-making under risk. Science, 315,
515–518. http://dx.doi.org/10.1126/science.1134239
Van den Bergh, B., & Dewitte, S. (2006). Digit ratio (2D:4D) moderates
the impact of sexual cues on men’s decisions in ultimatum games.
Proceedings Biological Sciences, 273, 2091–2095. http://dx.doi.org/10
.1098/rspb.2006.3550
Van den Bergh, B., Dewitte, S., & Warlop, L. (2008). Bikinis instigate
generalized impatience in intertemporal choice. Journal of Consumer
Research, 35, 85–97. http://dx.doi.org/10.1086/525505
Wang, Y., & Griskevicius, V. (2014). Conspicuous consumption, relation-
ships, and rivals: Women’s luxury products as signals to other women.
Journal of Consumer Research, 40, 834854. http://dx.doi.org/10.1086/
673256
Wetzels, R., Matzke, D., Lee, M. D., Rouder, J. N., Iverson, G. J., &
Wagenmakers, E. J. (2011). Statistical evidence in experimental psy-
chology: An empirical comparison using 855 t tests. Perspectives on
Psychological Science, 6, 291–298. http://dx.doi.org/10.1177/
1745691611406923
Wilson, M., & Daly, M. (2004). Do pretty women inspire men to discount
the future? Proceedings of the Royal Society B: Biological Sciences, 271
(Suppl. 4), S177–S179. http://dx.doi.org/10.1098/rsbl.2003.0134
Yu, E. C., Sprenger, A. M., Thomas, R. P., & Dougherty, M. R. (2014).
When decision heuristics and science collide. Psychonomic Bulletin &
Review, 21, 268–282. http://dx.doi.org/10.3758/s13423-013-0495-z
Received July 20, 2015
Revision received September 4, 2015
Accepted September 5, 2015
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
17
MATING MOTIVES AND DECISION MAKING
tapraid5/zfr-xge/zfr-xge/zfr00615/zfr2698d15z
xppws S1 10/8/15 4:52 Art: 2015-1187
APA NLM
... religious priming on decisions in an economic game despite running 455 participants across two critical between-subjects conditions. Similarly, Shanks et al. (2015), across 9 separate experiments (N ¼ 1,325), failed to find evidence consistent with previous work showing that priming mating motives affects people's spending behavior. Shanks et al. (2015) specifically had at least 80% power to detect a between-subjects effect as small as d s ¼ .14. ...
... Similarly, Shanks et al. (2015), across 9 separate experiments (N ¼ 1,325), failed to find evidence consistent with previous work showing that priming mating motives affects people's spending behavior. Shanks et al. (2015) specifically had at least 80% power to detect a between-subjects effect as small as d s ¼ .14. This highly powered failure to replicate in tandem with demonstrable evidence of publication bias indicates that the reported effect-priming mating motives influences consumer behavior-is most likely a Type I error. ...
Article
Failures to replicate high-profile priming effects have raised questions about the reliability of so-called “social priming” phenomena. However, not only are many of the relevant studies not particularly social in nature, but other robust priming effects that are clearly social in nature do not count as social priming. Most importantly, the focus on the supposedly social aspect of the work has obscured factors that help to account for the relative reliability of priming effects. Here, we examine the construct of social priming, describe some simple demonstrations on the role of experimental design in priming reproducibility, and discuss future avenues for building a better understanding of priming. We conclude that the term “social priming” should be laid to rest, and that it is time to move past arguments about the reliability of specific effects and shift our energy to building theories that help us better understand the mechanisms underlying priming effects.
... However, as our main aim here is to test whether the interaction between mating motives, participants' sex, and sociosexuality predicts the relative importance of products' green features, and cheapness, we conducted a power analysis in G*Power. Although previous literature suggests a relatively large interactive effect (i.e., f = 0.40; Griskevicius et al., 2007) between mating motives, sociosexual orientation and sex on related outcomes such as financial generosity, recent work suggests that published effect sizes of mating motive primes on consumer behavior are likely inflated due to publication bias (Shanks et al., 2015;Shanks & Vadillo, 2019). We thus estimated a small effect size (f = 0.14). ...
... Future developments should explore whether personal income serves as an anchor that people use to evaluate how expensive a product is, and whether this affects their consumer preferences. Lastly, and taking in consideration the recent discussions on the effect sizes associated to priming mating motives on consumer behavior (Shanks et al., 2015;Shanks & Vadillo, 2019), future studies should better validate priming manipulations such as that used in Study 2, perhaps by including manipulation checks. ...
Article
Full-text available
Conspicuous conservation refers to pro-environmental activities that are intended as signals of some attractive quality of the actor. As some of these qualities are desirable in romantic partners, people may purchase green products or services to impress potential romantic partners. We propose that conspicuous conservation communicates generosity – a trait that is especially valued in long-term romantic partners. Two studies tested whether people’s sustainable product preferences influence how they are perceived as romantic partners (Study 1), and whether actual product preferences are aligned with these perceptions (Study 2). Results from Study 1 suggest that people presented as having purchased green products are perceived as more generous and more attractive as long-term – but also short-term – romantic partners. Results from Study 2 suggest that individuals primed to think about a romantic context are no more likely to prefer sustainable products, suggesting an actor-observer discrepancy that potentially adds to the honesty of the conspicuous conservation signal. The potential communicative value of conspicuous conservation is discussed in relation to the literature on costly signaling, sexual selection, and green marketing.
... These patterns are consistent with studies employing small samples and finding non-significant results being harder to publish. Another meta-analysis [15,16] similarly found clear evidence of publication bias in 43 independent measures of another form of priming in which risk-taking, gambling and other potentially harmful behaviours are claimed to be increased by primes which activate evolutionary 'mating' motives (young male syndrome). In the demonstrable presence of reporting and publication biases [17], exploitation of 'researcher degrees of freedom' (RDF) [18][19][20], and inadequate power [21,22] in behavioural research, residual evidence for priming effects must be regarded as weak and requiring confirmation in large-scale, pre-registered studies that can exclude these and other sources of bias as an alternative explanation of the observed effects. ...
Article
Full-text available
Research on goal priming asks whether the subtle activation of an achievement goal can improve task performance. Studies in this domain employ a range of priming methods, such as surreptitiously displaying a photograph of an athlete winning a race, and a range of dependent variables including measures of creativity and workplace performance. Chen, Latham, Piccolo and Itzchakov (Chen et al. 2021 J. Appl. Psychol. 70 , 216–253) recently undertook a meta-analysis of this research and reported positive overall effects in both laboratory and field studies, with field studies yielding a moderate-to-large effect that was significantly larger than that obtained in laboratory experiments. We highlight a number of issues with Chen et al .'s selection of field studies and then report a new meta-analysis ( k = 13, N = 683) that corrects these. The new meta-analysis reveals suggestive evidence of publication bias and low power in goal priming field studies. We conclude that the available evidence falls short of demonstrating goal priming effects in the workplace, and offer proposals for how future research can provide stronger tests.
... They also suggest that measuring participants' perception of stimuli might be particularly important to find relevant hypothesized effects. Careful experimental design is essential, given that recent research was not able to replicate the effects of mating-related experimental primes on decision-making and risk-taking behavior (e.g., Shanks et al. 2015). Measuring participants' perceptions as mediating variables allows researchers to establish whether the experimental stimuli had their desired effects and whether the shifted perceptions in turn influence attitudes and behaviors. ...
Article
Full-text available
Objectives Individual differences in socio-political attitudes can reflect mating interests, and attitudes can also shift in response to mating market cues, including mating competitor quality. In four experiments, we tested whether competitors’ attractiveness (Experiments 1F&1M) and income (Experiments 2F&2M) would influence socio-political attitudes (participants’ self-reported attitudes towards promiscuity and sexual liberalism, traditional gender roles, and the minimum wage and healthcare). Methods We collected data from American participants online through Amazon’s Mechanical Turk (total N = 787). In all experiments, each participant was randomly assigned to one of four experimental treatments in a between-subjects design (three levels of mating competitor quality and a control group), and to one of five stimuli within each treatment. Results Overall, the experimental treatments largely did not predict participants’ socio-political attitudes. The fifteen unique experimental stimuli, however, did significantly affect participants’ perception of their competitors’ quality. That perception, in turn, affected some socio-political attitudes. Namely, individuals who perceived their competitors to be of high mate-value were more supportive of traditional gender roles and, only for men in Experiment 2M, more opposed to promiscuity and sexual liberalism than individuals who perceived competitors to be of low mate-value. These results only applied to sexually unrestricted, but not restricted, women. Perceived mating competition did not affect attitudes towards the minimum wage and healthcare. Conclusions Experimental cues of mating competition shifted participants’ perceptions of their competitors’ mating quality and these perceptions in turn shifted some socio-political attitudes. We interpret these results considering broader arguments about plasticity in socio-political attitudes.
Article
Two recent meta-analyses on inattentional blindness (Kreitz, Pugnaghi, & Memmert, 2020; Nobre et al., 2020) concluded that objects can be processed implicitly even when attention is directed elsewhere. However, signs of publication bias are evident in both of these meta-analyses. Here, we employed multiple tools to correct for publication bias in the data aggregated in those meta-analyses. Analyses using the Precision-Effect Test (PET) and robust Bayesian meta-analysis (RoBMA) suggest that the estimates in the original meta-analyses were inflated, together with strong evidence of publication bias. Indeed, the data are consistent with no overall implicit effects. We suggest that more evidence, particularly from well-powered pre-registered experiments, is needed before solid conclusions can be drawn regarding implicit processing during inattentional blindness.
Preprint
To assess the replicability of social priming findings we reviewed the extant close replication attempts in the field. In total, we found 65 close replications, that replicated 46 unique findings. Ninety-four percent of the replications had effect sizes smaller than the effect they replicated, only 18% of the replications reported a significant p-value in the original direction, and the 95% confidence interval of the replication effects included the original effects only 26% of the time. The strongest predictor of replication success was whether or not the replication team included at least one of the authors of the original paper. Twelve of the 16 replications with at least one original author produced a significant effect in the original direction and the meta-analytic average of these studies suggest a significant priming effect (d = 0.33, 95% CI[0.26; 0.65]). In stark contrast, none of the 49 replications by independent research teams produced a significant effect in the original direction and the meta-analytic average was virtually zero (d = 0.001, 95% CI[-0.03; 0.03]). We argue that these results have shifted the burden of proof back onto advocates of social priming. Successful replications from independent research teams will likely be required to convince sceptics that social priming exists at all.
Article
Science is facing a reproducibility crisis, and there is a concern about the reproducibility of the research findings, especially in psychology. The research findings in leading journals of consumer research have been found to be unreproducible. Questionable research practices (QRPs) such as p-hacking and HARKing may be involved in this reproducibility crisis. It has been pointed out that QRPs have been employed in consumer research. In this paper, we discuss the problem of reproducibility and its related issues from the perspective of consumer research. First, we indicate that an implicit research practice in consumer research encompasses QRPs. Next, we discuss how the reproducibility crisis has been perceived in consumer research. Finally, we propose the research practices required for consumer research to deal with QRPs. Focusing on the pre-registration, we introduce credible research practices, such as open data, and accurate and transparent reporting of statistics and experimental materials.
Article
Studies of unconscious mental processes often compare a performance measure (e.g., some assessment of perception or memory) with a measure of awareness (e.g., a verbal report or forced-choice response) of the critical cue or contingency taken either concurrently or separately. The resulting patterns of bivariate data across participants lend themselves to several analytic approaches for inferring the existence of unconscious mental processes, but it is rare for researchers to consider the underlying generative processes that might cause these patterns. We show that bivariate data are generally insufficient to discriminate single-process models, with a unitary latent process determining both performance and awareness, from dual-process models, comprising distinct latent processes for performance and awareness. Future research attempting to isolate and investigate unconscious processes will need to employ richer types of data and analyses.
Article
Full-text available
Determining the reasons why people attend events, as well characterizing visitors’ attitudes and factors influencing their choices of particular events, are of key importance for event managers and destination management organizations. Accordingly, this study aimed to identify the factors influencing visitors’ behavioural intention towards attending events. Theoretically, this study offers a reconceptualized model positing optimum stimulation and needs for variety as antecedents of motivation, going beyond the more classical approaches of focusing on attitude and behavioural intention. The proposed model was empirically tested in the context of a mountain sport event hosted in Norway, one in which participants were deemed to engage in both exploratory and non-exploratory behaviours. The results show a valid relationship between visitors’ optimum stimulation level and their variety-seeking tendencies; these factors, in turn, influence motivation and have an indirect effect upon visitors’ attitudes and behavioural intentions towards event participation.
Preprint
Full-text available
Data is the cornerstone of the modern academic industry, like a constant production line of consumable goods packaged with a veneer of statistical techniques. Biomedical and psychological sciences invert the traditional logic of the physical sciences where hypotheses were tested against data rather than data against hypotheses. This is likely to reflect the immaturity of the new paradigms that barely cope with the data output of a precocious enterprise. However, the current state leads to distortions of the practice of science and entrenchment of its dysfunctional politics reflected in the science itself. With the debate now on the reproduction of data and statistical sleights of hand accompanying many if not most studies, what is lost in the debate is the preeminent role of good theoretical discovery. This is partly as a result of the structure of funding, the nature of reporting in biomedical fields based on the symbiotic relationship of high ranking institutions and journals. Open access could fill a gap in traditional publishing literature which has entrenched a culture of highly restrictive practices at a time when revolutionary science is required. OA is not in danger of lowering the standards of science as its critics claim, but because it is ‘open’ to new and radical ideas that would never see the light of day otherwise, it may well provide a rejuvenating energy. OA is a self-regulatory response to the consequences of the distorted and inhibitory contemporary practice of science.
Article
Full-text available
The data includes measures collected for the two experiments reported in “False-Positive Psychology” [1] where listening to a randomly assigned song made people feel younger (Study 1) or actually be younger (Study 2). These data are useful because they illustrate inflations of false positive rates due to flexibility in data collection, analysis, and reporting of results. Data are useful for educational purposes.
Article
Full-text available
There has been increasing criticism of the way psychologists conduct and analyze studies. These critiques as well as failures to replicate several high-profile studies have been used as justification to proclaim a "replication crisis" in psychology. Psychologists are encouraged to conduct more "exact" replications of published studies to assess the reproducibility of psychological research. This article argues that the alleged "crisis of replicability" is primarily due to an epistemological misunderstanding that emphasizes the phenomenon instead of its underlying mechanisms. As a consequence, a replicated phenomenon may not serve as a rigorous test of a theoretical hypothesis because identical operationalizations of variables in studies conducted at different times and with different subject populations might test different theoretical constructs. Therefore, we propose that for meaningful replications, attempts at reinstating the original circumstances are not sufficient. Instead, replicators must ascertain that conditions are realized that reflect the theoretical variable(s) manipulated (and/or measured) in the original study. © The Author(s) 2013.
Article
Shariff and Norenzayan (2007) discovered that people allocate more money to anonymous strangers in a dictator game following a scrambled sentence task that involved words with religious meanings. We conducted a direct replication of key elements of Shariff and Norenzayan's (2007) Experiment 2, with some additional changes. Specifically, we (a) collected data from a much larger sample of participants (N = 650); (b) added a second religious priming condition that attempted to prime thoughts of religion less conspicuously; (c) modified the wording of some of their task explanations to avoid deceiving our participants; (d) added a more explicit awareness probe; (e) reduced prime-probe time; and (f) performed statistical analyses that are more appropriate for non-normal data. We did not find a statistically significant effect for religious priming. Additional tests for possible between-subjects moderators of the religious priming effect also yielded nonsignificant results. A small-scale meta-analysis, which included all known studies investigating the effect of religious priming on dictator game offers, suggested that the mean effect size is not different from zero, although the wide confidence intervals indicate that conclusions regarding this effect should be drawn with caution. Finally, we found some evidence of small-study effects: Studies with larger samples tended to produce smaller effects (a pattern consistent with publication bias). Overall, these results suggest that the effects of religious priming on dictator game allocations might be either not reliable or else quite sensitive to differences in methods or in the populations in which the effect has been examined. (PsycINFO Database Record (c) 2015 APA, all rights reserved).
Article
Reproducibility is the cornerstone of science. If an effect is reliable, any competent researcher should be able to obtain it when using the same procedures with adequate statistical power. Two of the articles in this special section question the value of direct replication by other laboratories. In this commentary, I discuss the problematic implications of some of their assumptions and argue that direct replication by multiple laboratories is the only way to verify the reliability of an effect. © The Author(s) 2013.
Article
Prior research has examined how sexual opposite-sex stimuli impact people's choices and behaviors. However, it is largely unknown whether sexual same-sex stimuli also do so. This research reports an intriguing phenomenon: men who see attractive males take greater financial risks than those who do not. An evolution-based account is proffered and tested across four experiments. In evolutionary history, men have faced greater intrasexual competition in attracting women as a mating partner. Thus, when the average heterosexual man sees males who are more physically-attractive than he is, he is motivated to increase his desirability as a mating partner to women, prompting him to accrue money, and taking financial risks helps him to do so. This research concludes by discussing the implications of the present findings for men today who are constantly bombarded by not only sexual opposite- but also same-sex others, such as images that are commonly used in advertising.
Article
Past research shows that luxury products can function to boost self-esteem, express identity, and signal status. We propose that luxury products also have important signaling functions in relationships. Whereas men use conspicuous luxury products to attract mates, women use such products to deter female rivals. Drawing on both evolutionary and cultural perspectives, five experiments investigated how women's luxury products function as a signaling system directed at other women who pose threats to their romantic relationships. Findings showed that activating a motive to guard one's mate triggered women to seek and display lavish possessions. Additional studies revealed that women use pricey possessions to signal that their romantic partner is especially devoted to them. In turn, flaunting designer handbags and shoes was effective at deterring other women from poaching a relationship partner. This research identifies a novel function of conspicuous consumption, revealing that luxury products and brands play important roles in relationships.
Article
Although the effect of stereotype threat concerning women and mathematics has been subject to various systematic reviews, none of them have been performed on the sub-population of children and adolescents. In this meta-analysis we estimated the effects of stereotype threat on performance of girls on math, science and spatial skills (MSSS) tests. Moreover, we studied publication bias and four moderators: test difficulty, presence of boys, gender equality within countries, and the type of control group that was used in the studies. We selected study samples when the study included girls, samples had a mean age below 18years, the design was (quasi-)experimental, the stereotype threat manipulation was administered between-subjects, and the dependent variable was a MSSS test related to a gender stereotype favoring boys. To analyze the 47 effect sizes, we used random effects and mixed effects models. The estimated mean effect size equaled -0.22 and significantly differed from 0. None of the moderator variables was significant; however, there were several signs for the presence of publication bias. We conclude that publication bias might seriously distort the literature on the effects of stereotype threat among schoolgirls. We propose a large replication study to provide a less biased effect size estimate. Copyright © 2014 Society for the Study of School Psychology. Published by Elsevier Ltd. All rights reserved.