Social Influence Moderates Placebo Response 1&...2016/02/05 · Social Influence Moderates Placebo Response Draft&underreview,&last&updated2.5.16& 3& Although we are told, “a rose

Social Influence Moderates Placebo Response

Draft under review, last updated 2.5.16

1

Transforming Water: Social Influence Moderates Psychological, Physiological, and

Functional Response to a Faux Product

Alia J. Crum1*, Damon J. Phillips2, E. Tory Higgins2

1Stanford University, 450 Serra Mall, Stanford, CA 94305

2Columbia University, 116th St & Broadway, New York, NY 10027

Correspondence to:

Dr. Alia J. Crum Bldg. 420, Jordan Hall, Rm 244 Stanford, CA 94305-2130 ph: 970 987 9182 email: [email protected]



2

Abstract

This paper investigates how social influence can alter physiological, psychological, and

functional responses to a faux product and how such responses influence the ultimate

endorsement of the product. Participants consumed a product, “AquaCharge Energy Water,”

falsely-labeled as containing 200 mg of caffeine but which was actually plain spring water in one

of three conditions: a no social influence condition, a disconfirming social influence condition,

and a confirming social influence condition. Results demonstrated that the effect of the product

labeling on physiological, psychological and functional alertness was moderated by social

influence: participants experienced more subjective, physiological and functional alertness after

consuming the product in the confirming social influence condition and experienced less

subjective, physiological and functional alertness after consuming the product in the

disconfirming social influence condition. Moreover, participants in the confirming social

influence condition more strongly endorsed the product, an effect that was mediated by changes

in blood pressure. These results suggest that social influence can alter subjective, physiological

and functional responses to a faux product, in this case transforming the effects of plain water.



3

Although we are told, “a rose by any other name would smell as sweet,” decades of

psychological research establishes that the sensory properties of an object or substance are not

the sole determinant of how we experience it. This is especially true for shaping experiences of

food and beverage consumption, in which case simply altering the name or label of a food can

literally make the product taste sweeter. For example, people enjoy the taste of Coke more when

it is consumed in a brand name cup (McClure et al., 2004), children enjoy the taste of french

fries, milk, and carrots more when they believe them to be from McDonald’s (Robinson,

Borzekowski, Matheson, & Kraemer, 2007), and beer infused with vinegar is enjoyed more if it

is labeled as having a “special ingredient” than when the actual ingredient is unveiled (Lee,

Frederick, & Ariely, 2006).

Inspired by the neurological and physiological effects found in clinical trials and placebo

research (Finniss, Kaptchuk, Miller, & Benedetti, 2010), a growing body of research suggests

that, in addition to influencing perception and behavior, mindsets and expectations can also alter

more “objective” outcomes such as sensory experience and physiological processing. For

example, increasing the perceived cost of wine results in heightened activity in areas of the brain

related to pleasure and reward when consuming the exact same wine. (Plassmann, O’Doherty,

Shiv & Rangel, 2008). Additionally, participants who thought they were drinking an indulgent,

high-calorie milkshake showed steeper declines in ghrelin, a hunger-inducing hormone, than

when they thought that the same shake was a sensible, low-calorie milkshake (Crum, Corbin,

Brownell, & Salovey, 2011).

Such effects do not exist in a vacuum. Instead, social forces dynamically inform them.

Social forces supply new information, confirm or correct old information, and serve as a

foundational source for our mindsets, beliefs, and expectations (Asch, 1948; Robert B. Cialdini,



4

2005; Hardin & Higgins, 1996; Muchnik, Aral, & Taylor, 2013; Salganik, Dodds, & Watts,

2006). Individuals often turn to others for information about what foods are good or bad, what

diets work or fail, and what medications are most effective. Decades of research support the idea

that social forces can change subjective emotions and preferences (R. B. Cialdini & Goldstein,

2004; Colloca & Benedetti, 2009; Moscovici, Lage, & Naffrechoux, 1969; Salganik et al., 2006;

Schachter & Singer, 1962; Sherif, 1937). And more recent research has accumulated to suggest

that social information (e.g., about other’s attractiveness) can increase or decrease the

neurological activity in the orbitofrontal cortex and the nucleus accumbens, suggesting that

social influence can produce changes in brain areas corresponding to hedonic experience (Zaki,

Schirmer, & Mitchell, 2011).

Can social influence literally “get under the skin” and affect physiological experiences of

a product? Literature from medical and placebo research suggests that this may indeed be

possible. Research on mass psychogenic illness (MPI) demonstrates that the collective

occurrence of self-reported physical symptoms in the absence of an identifiable pathogen is a

relatively common occurrence (Mazzoni, Foan, Hyland, & Kirsch, 2010). Studies exploring

MPI empirically have shown that the belief that one has inhaled a substance described as an

environmental toxin, when in fact it is just odorless ambient air, can evoke the experience of

corresponding symptoms (e.g., headaches, nausea, and drowsiness). Most importantly,

researchers have documented that these symptoms are increased (up to an 11x increase in one

study) when in the presence of a confederate participant displaying the expected symptoms

(Broderick, Kaplan-Liss, & Bass, 2011; Lorber, Mazzoni, & Kirsch, 2007; Mazzoni et al., 2010).

Placebo analgesia research has demonstrated that pain relief can be conditioned by witnessing

another person receive pain relief (Colloca & Benedetti, 2009). These studies suggest that social



5

influence may affect physiological or visceral experience. However, these studies are limited to a

clinical context and have only been reflected in self-reported measures of physiological

experiences such as pain or nausea. Thus, an important question remains unanswered: can

social influence moderate the physiological effects of a product or experience as indicated by

objective measures such as cardiovascular, immune or neuroendocrine response?

Previous research has demonstrated that beliefs and expectations about a product or

experience induced by labels or beliefs can affect physiological and psychological responses to

that product. A separate body of research suggests that social influence can affect psychological

responses and possibly physiological response although no study to our knowledge has directly

tested the later. No studies, to our knowledge, have explored how the psychological,

physiological and functional effects of social influence may be responsible for the degree to

which a person ultimately endorses a product. In other words, it remains to be understood

whether the effects of social influence on product endorsement are mediated through changes in

psychological response, physiological response or both.

To test these questions we created a product, “AquaCharge Energy Water.” The bottle’s

label indicated that it contained 200 mg of performance and energy-enhancing caffeine. In

reality, it was plain spring water. After participants consumed the product we compared

measures of subjective arousal, physiological arousal and cognitive function known to respond to

the effects of caffeine across three conditions: a no social influence condition, in which the

participant consumed the water in a room alone, a disconfirming social influence condition, in

which the participant consumed the water alongside a confederate who denied the effect of the

product, and a confirming social influence condition, where the confederate participant endorsed

the effect of the product. We predicted that, in line with research on the effects of placebo



6

caffeine, all participants would experience increases in subjective, physiological and functional

response. Additionally, we predicted that in line with research on social influence these effects

would be moderated by social influence such that they would increase in the presence of an

affirming confederate and decrease in the presence of a denying confederate. We also

investigated whether subjective and physiological responses to the product would mediate

endorsement of the product.

Participants and Design

We recruited 98 females (24% Black, 40% White, 36% Asian) (mean age=24 years;

SD=5) from a university study pool. The decision to have at least 30 participants in each

condition was predetermined based on published research on MPI, placebo caffeine, and social

influence (Mazzoni et al., 2010; Boothby et al., 2010; Harrell & Juliano, 2009). Participants

were screened at the onset and excluded from the study if they were a smoker, had high blood

pressure, or reported intake of alcohol, antihistamines, caffeine, or anything besides water two

hours prior to experiment. Participants received $20 for their participation, with the opportunity

to earn a $100 gift card reward if they completed the one-week follow-up questionnaire.

Participants were randomly assigned to: 1) no social influence condition (N = 35); 2) confirming

social influence condition (N= 32); or 3) disconfirming social influence condition (N = 31). To

heighten the effect of social influence (Echterhoff, Higgins, & Levine, 2009) and reduce

variability from gender effects on arousal (Hartley, Lovallo, & Whitsett, 2004) participants were

matched to the same-race and same-sex confederates.

Procedure

Participants were asked to not drink caffeine or consume anything except water for two

hours prior to the onset of the 1-hour study, which took place during the morning hours of 9am-



7

12pm. After arriving at the laboratory, participants filled out two short questionnaires assesing

their caffeine exposure and expectancies as well as their subjective arousal levels. Participants

were then directed into a room in which the confederate participant was present (conditions 1

and 3) or into an empty room with no other participant present (condition 2). The participant

and (for conditions 1 and 3) the confederate were then connected to a Noninvasive Blood

Pressure System (NIBPH100D) to record systolic blood pressure (SBP) responses following

which they completed a Stroop task. Following these baseline measurements, participants were

given an iPad and asked to view and rate the product’s website (Figure 1). The purpose of this

was to add to the credibility of the study guise (beta-testing the marketing for a new product) and

ensure that the participants paid attention to the placebo/mindset intervention (that the water is

infused with caffeine and thus will increase their arousal, energy and reaction time when

consumed).

INSERT FIGURE 1

After completing the website-rating form, participants were given an 8 oz bottle of

AquaCharge (Figure 2) and asked to consume the product in its entirety within two minutes.

Participants were then told to wait “a few minutes for the energy water to take effect.” Three

minutes after consuming the water, the confederate participant spoke. In the disconfirming

confederate condition, the confederate stated: “Hmm. I don’t really feel any change. I’m

definitely NOT feeling charged up. How about you?” In the confirming confederate condition,

the confederate participant stated: “Wow! This is really something. This is really waking me up!

How about you?” After the confederate spoke, the experimenter entered the room and reminded

participants to stay quiet during the testing phase. For the next five minutes, participants sat

quietly and completed a few questions asking about the product label, taste of the product. These



8

were filler questions designed to keep them focused on the product and maintain alignment with

the study guise.

INSERT FIGURE 2

Ten minutes after consuming the product, the participant completed the subjective

alertness questionnaire, the Stroop task and a “Product Endorsement Survey.” One week later

participants were contacted by email and asked to fill out a short survey asking them how many

people they had told about the product and for what reasons.

Measures. Subjective, physiological and functional tasks were chosen based on research

supporting the effects of caffeine on subjective arousal (Childs & de Wit, 2006; Smith, 2002),

systolic blood pressure (Hartley et al., 2004; Karatzis et al., 2005) and cognitive interference

(Dixit, Goyal, Thawani, & Vaney, 2012; Kenemans, Wieleman, Zeegers, & Verbaten, 1999).

Subjective Alertness: Participants’ subjective alertness was measured by a scale

developed and validated by Kirsch & Weixel (Kirsch & Weixel, 1988) for the purposes of

studying the effects of caffeine. The scale includes fifteen adjectives along a 5-point Likert

scale, which factor in three subscales indicating alertness, relaxation, and tension. In the current

sample, Cronbach’s alpha was adequate (alertness subscale alpha = .79; tense subscale alpha =

.73; relaxation subscale alpha = .86). There were no significant effects of time or condition for

the tense or relaxation subscales.

Physiological Alertness: Participants’ physiological arousal was measured using a

Noninvasive Blood Pressure System (NIBPH100D) that recorded systolic blood pressure (SBP)

responses. To equate in time with the subjective responses we analyzed blood pressure at

immediately before consumption and 5 minutes after consumption. Due to technical issues with

the blood pressure machine, data from several participants was unreadable, thus blood pressure



9

was analyzed for 79 participants (N = 22 Disconfirming Confederate; N=25 No Confederate;

N=32 Confirming Confederate).

Functional Alertness (Cognitive Interference): Because cognitive interference has shown

to be affected by caffeine intake, participants’ cognitive interference was measured using the

Stroop color-naming task (Stroop, 1935; Macleod, 1991). Participants were instructed to

indicate the font color of a letter string by pressing one of four appropriate color-coded keys as

quickly as possible. On incongruent trials a color word appeared in a font color different from its

semantic meaning (e.g., “BLUE” in red font), whereas confirming trails displayed a color word

that matched its font color (e.g., “BLUE” in blue font). Control trials consisted of a string of

“@”s in one of the four font colors. Participants completed 20 practice and 90 experimental

trials. Stroop interference scores were computed as the difference in response latencies (in

milliseconds) between incongruent and congruent trials, with higher scores indicating greater

cognitive depletion. Incorrect responses and latencies above 2000 ms and below 200 ms were

recoded as missing data.

Product Endorsement: To assess product endorsement, participants were asked a variety

of questions pertaining to their endorsement of the AquaCharge product. Should Cost was

measured by asking people, “How much would you pay for AquaCharge Energy Water in a

store?” They were given several benchmarks to assist them in making their decision including

Gatorade ($1.50), Poland Springs Water ($1.75) and Five Hour Energy ($2.10). Probability of

Buying the product was measured by asking participants to rate “What is the likelihood you

would BUY AquaCharge in a store,” on a Likert scale ranging from 1= Not at all to 7 =

Extremely. Ambassador Initiative was measured by taking the mean rating of participants’

likelihood to: “Become a College Ambassador for AquaCharge”, “Boost my resume by working



10

for AquaCharge”, and “Help make AquaCharge available on or near your campus,” (1= Not at

all, 4 = A little, 7 = Extremely). One week following their participation, participants were

contacted by email and asked to fill out a short survey asking them how many people they had

had told about the product and the reasons for telling them (e.g., because “I love the product and

want people to know about it”; “I am just a talker and tell everyone things”; “I didn’t like the

product and wanted them to know I didn’t like it”; “I thought it was an interesting idea” or “I

thought they would be interested in it”). Number of People Told was calculated as the total

number told either because they checked “I love the product and want people to know about it”

or “I thought they would be interested in the product.” They then chose whether they wanted to

be entered to win either a $100 gift card or an $80 gift card and a case of 30 bottles of

AquaCharge, serving as a behavioral measure of product endorsement. Chose AquaCharge over

Cash was calculated as the total number of participants choosing this second option (0 = no, 1=

yes) per condition. Data was analyzed for the 79 participants who completed this follow-up

survey (N = 21 disconfirming confederate; N=12 no confederate; N=19 confirming confederate).

Complete means and standard deviations for each dependent variable are detailed in Table 2.

Results

To test whether the social influence condition moderated changes in subjective,

physiological, and functional response, a hierarchical regression model predicting post

consumption subjective alertness, physiological alertness (SBP), and functional response was

conducted with baseline levels entered in Step 1 and social influence condition (disconfirming

confederate=1, no confederate=2, confirming confederate=3) entered in Step 2. These results

indicate a significant effect of social influence condition on subjective alertness [β=.14,

t(96)=2.28 p=.025] as well as change in physiological alertness as measured by systolic blood



11

pressure [β=.15, t(76)=2.07 p=.04] (Table 1). Furthermore, there was a significant effect of

social influence condition on functional alertness as indicated by change on the Stroop task from

baseline to post-consumption [β=-.22, t(92)=-2.29 p=.024].

INSERT TABLE 1 AND FIGURE 3

Linear regression also tested the effect of condition on product endorsement variables

(likelihood of purchasing, should cost, ambassador initiative, number of people told about

product, and choice of AquaCharge over cash). These results indicate a significant effect of

social influence condition on product endorsement as indicated by what participants thought the

product should cost [β=.27, t(96)=2.77 p=.007], likelihood to buy AquaCharge [β=.31, t(97)=

3.25 p=.002], and marginally significant effects on willingness to be an ambassador for the

product [β=.19, t(97)=1.92 p=.057], the number of people told about the product at a 1 week

follow-up [β=.24, t(50)= 1.76 p=.08] and the extent to which participants chose the option to

win AquaCharge over cash [β=.25, t(50)=1.84 p=.07] (Figure 4).

INSERT FIGURE 4 AND TABLE 2

To test whether the effect of social influence on product endorsement was mediated

through differences in physiological alertness, subjective alertness, or functional alertness, we

employed a bootstrapping method (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002;

Preacher & Hayes, 2004) with n=5000 bootstrap resamples in which subjective alertness and

SBP were included as potential mediators in the model, and baseline alertness and pre

consumption SBP were included as covariates. Because the product endorsement variables hung

reasonably well together (alpha=.67) we created an aggregate index by taking the mean of the

standardized variables. The analyses revealed with 95% confidence that the effect of social

influence on aggregate product endorsement was significantly mediated by changes in SBP



12

[point estimate=.04; 95% CI .0020, .1271] and not by changes in subjective alertness [point

estimate=.04 ; 95% CI .-.0005, .1580] or changes in functional response [point estimate=.002 ;

95% CI .-.0172, .0050]. This suggests that endorsement of the product was largely driven by the

changes in systolic blood pressure evoked by social influence rather than by changes in self-

reported arousal or reductions in cognitive interference.

Discussion

When participants entered our lab they thought they were drinking the latest in a line of

enriched bottled waters. They believed that they were drinking water infused with caffeine, when

in fact it was just plain water. Our results demonstrated that the effect of this belief on

physiological, psychological and functional alertness was moderated by social influence:

participants experienced more subjective, physiological and functional alertness after consuming

the product in the confirming social influence condition and experienced less subjective,

physiological and functional alertness after consuming the product in the disconfirming social

influence condition. Furthermore, the social influence manipulation proved to have effects on

product endorsement. Participants in the confirming social influence condition were willing to

pay more for the product, more likely to buy it, more likely to choose the product over cash,

more interested in being an ambassador for the product and more likely to tell their friends about

the product. Finally, mediation analyses indicated that the effect of social influence condition on

product endorsement was mediated by changes in physiological alertness as measured by systolic

blood pressure and not changes in subjective alertness or functional alertness.

These results inform and extend the literature on social influence and product

consumption in two important ways. First, researchers have traditionally struggled to distinguish

the effects of social influence on true changes in preference versus mere public compliance (R.



13

B. Cialdini & Goldstein, 2004; Moscovici et al., 1969; Sherif, 1937). Recent research has

informed this debate by employing neurological techniques that suggest that social influence can

produce changes in brain areas corresponding to hedonic experience (Zaki et al., 2011). Our

results extend this research and suggest that social influence can also produce physiological and

functional responses, such as differences in systolic blood pressure and cognitive interference to

a faux stimulant.

Second, our research extends the existing body of research on social influence by

furthering our understanding of mechanisms through which social influence might influence

one’s ultimate endorsement of a product. Intriguingly, our mediation results suggest that

endorsements of a product or substance may not simply be driven by social desirability or

subjective reports of effectiveness. Instead, they seem to be most strongly driven by

physiological differences (in this case systolic blood pressure) induced by the social-

psychological nature of the experience.

These findings are particularly interesting to take into account when considering the

increasing popularity of “aquaceuticals” and other products that contain active ingredients which

claim to have effects that are not scientifically supported. If these products are actually

producing the proclaimed effects—even if the cause of the change is driven primarily by the

psychosocial context as opposed to the ingredients themselves—does that make the purchase of

those products and their growing popularity more legitimate? Assessing the legitimacy of

purchasing such products is far out of the scope of this paper, of course; at the very least, these

results remind us that to understand the true effects of any product or substance we must work to

account for the psycho-social setting in which it is consumed.



14

When reflecting on the legitimacy of marketing claims, it is important to note that the

differences in blood pressure in our study did not result from an increase from baseline levels;

instead, overall blood pressure dropped during the study procedure and this drop was

exacerbated in the incongruent social influence condition but eliminated in the congruent social

influence condition. This decline in blood pressure is to be expected given the prolonged

sedentary behavior throughout the course of the experiment and is a common occurrence in

similar experimental paradigms. However, future research aimed at understanding the effects of

the expectation alone would benefit from including a control condition in which participants

consumed the same bottled water without the belief that it was caffeinated. Of course, research

has already established that placebo caffeine can increase blood pressure response as compared

to decaffeinated control (Kirsch & Weixel, 1988) and our primary aim was to test the moderating

role of social influence on physiological responses and the differences in decline are indicative of

the influence of social influence on physiological arousal.

Although the results of this study provide important data on the impact of social influence

on product experience and endorsement, several questions remain. First, how did the presence of

a confederate participant influence physiological or functional responding? Although it is

possible that there was an influence from the mere presence of the confederate (Zajonc, 1965),

the same presence of a confederate either increased or decreased the effects depending on what

the confederate said about the product experience. A more likely mechanism is that what the

confederate said about the product experience altered the response expectancies of the

participant, which in-turn influenced the participant’s physiological and functional responding.

In the current study we did not measure the momentary change in expectations because we were

concerned that asking participants about their expectations after the social encounter would lead



15

them to question the study guise. Thus, more research is needed to thoroughly understand the

precise mechanisms (i.e. increased expectancy or other avenues) through which social influence

can affect physiological and functional responding.

What about the form through which social information is communicated? In today’s

world, social influence is often transmuted through technology (via social media, email, etc.).

Does in-person social influence have the same effect as technologically-transmuted social

influence? Hunter and colleagues (2014) demonstrated the social modeling of a placebo

analgesic is effective both when presented in person as well as when participants view a

recording of a participant (Hunter, Siess, & Colloca, 2014). However, more work is needed to

understand the context in which social influence is likely to have the most potent effects. One

question that remains is the degree to which social information is merely providing additional

information about the benefits of the product or if there is something inherently unique about the

human connection or what is sometimes called “shared reality” (Hardin & Higgins, 1996).

Third, what is the relative weight of impact from social influence and objective qualities

of a product, substance, or experience? We intentionally chose water as our product, rather than

something like decaf coffee, in an effort to isolate the effect of social influence from

conditioning or other effects that might result from the smell of decaf coffee (Benedetti, 2008).

Future research is needed to understand how social influence interacts with sensory properties to

engineer the ultimate impact of a product, medication, or experience.

Finally, what are the boundaries of these effects? In the current study we chose to study

women only and to match the race and gender of the confederate with the participant to reduce

variability and maximize perceived similarity between the participant and the confederate.

According to Bandura’s Social Learning Theory, male subjects are more prone to imitate male



16

models and female subjects more prone to imitate female models (Bandura, Ross, & Ross, 1961;

Bussey & Bandura, 1984). Research exploring social modeling on placebo responding has

shown mixed effects of gender, with some studies demonstrating greater effects with matched

confederates (Lorber et al., 2007; Mazzoni et al., 2010), and some demonstrating greater effects

with male models (Świder & Bąbel, 2013) and some showing no gender differences in

responding or matching affect (Broderick et al., 2011). The present study is limited in its ability

to inform this debate and or make claims outside of female-female interactions.

Though much remains to be explored, our results are intriguing in demonstrating that

social influence—working hand in hand with the psychological construction of sensory input—

can alter physiological and functional response to a product, in this case changing the experience

of, and attitudes toward, simple water.

Acknowledgments

We thank Jennifer Aaker, Ted Kaptchuk, Irving Kirsch, Stephen Kosslyn, Hazel Markus, Ann

McGill, Lee Ross and Jamil Zaki for their comments on this manuscript and Nana Amoh, Hayley

Blunden, Ellen Hada, and Yael Warach for their assistance in conducting this experiment



17

References:

Asch, S. E. (1948). The doctrine of suggestion, prestige and imitation in social psychology. Psychol Rev, 55(5), 250-‐276.

Bandura, A., Ross, D., & Ross, S. A. (1961). Transmission of aggression through imitation of aggressive models. The Journal of Abnormal and Social Psychology, 63(3), 575-‐582. doi: 10.1037/h0045925

Benedetti, F. (2008). Mechanisms of placebo and placebo-‐related effects across diseases and treatments. Annu Rev Pharmacol Toxicol, 48, 33-‐60. doi: 10.1146/annurev.pharmtox.48.113006.094711

Broderick, J. E., Kaplan-‐Liss, E., & Bass, E. (2011). Experimental induction of psychogenic illness in the context of a medical event and media exposure. American journal of disaster medicine, 6(3), 163.

Bussey, K., & Bandura, A. (1984). Influence of gender constancy and social power on sex-‐linked modeling. J Pers Soc Psychol, 47(6), 1292-‐1302. doi: 10.1037/0022-‐3514.47.6.1292

Childs, E., & de Wit, H. (2006). Subjective, behavioral, and physiological effects of acute caffeine in light, nondependent caffeine users. Psychopharmacology (Berl), 185(4), 514-‐523. doi: 10.1007/s00213-‐006-‐0341-‐3

Cialdini, R. B. (2005). Basic Social Influence Is Underestimated. Psychological Inquiry, 16(4), 158-‐161. doi: 10.2307/20447283

Cialdini, R. B., & Goldstein, N. J. (2004). Social influence: compliance and conformity. Annu Rev Psychol, 55, 591-‐621. doi: 10.1146/annurev.psych.55.090902.142015

Colloca, L., & Benedetti, F. (2009). Placebo analgesia induced by social observational learning. Pain, 144(1-‐2), 28-‐34. doi: 10.1016/j.pain.2009.01.033

Crum, A. J., Corbin, W. R., Brownell, K. D., & Salovey, P. (2011). Mind over milkshakes: mindsets, not just nutrients, determine ghrelin response. Health Psychol, 30(4), 424-‐429; discussion 430-‐421. doi: 10.1037/a0023467

Dixit, A., Goyal, A., Thawani, R., & Vaney, N. (2012). Effect of caffeine on information processing: evidence from stroop task. Indian J Psychol Med, 34(3), 218-‐222. doi: 10.4103/0253-‐7176.106013

Echterhoff, G., Higgins, E. T., & Levine, J. M. (2009). Shared Reality: Experiencing Commonality With Others' Inner States About the World. Perspectives on Psychological Science, 4(5), 496-‐521. doi: 10.1111/j.1745-‐6924.2009.01161.x

Finniss, D. G., Kaptchuk, T. J., Miller, F., & Benedetti, F. (2010). Biological, clinical, and ethical advances of placebo effects. Lancet, 375(9715), 686-‐695. doi: 10.1016/s0140-‐6736(09)61706-‐2

Hardin, C. D., & Higgins, E. T. (1996). Shared reality: How social verification makes the subjective objective. In R. M. S. E. T. Higgins (Ed.), Handbook of motivation and cognition, Vol. 3: The interpersonal context (pp. 28-‐84). New York, NY, US: Guilford Press.

Hartley, T. R., Lovallo, W. R., & Whitsett, T. L. (2004). Cardiovascular effects of caffeine in men and women. Am J Cardiol, 93(8), 1022-‐1026. doi: 10.1016/j.amjcard.2003.12.057



18

Hunter, T., Siess, F., & Colloca, L. (2014). Socially induced placebo analgesia: A comparison of a pre‐recorded versus live face‐to‐face observation. European Journal of Pain, 18(7), 914-‐922.

Karatzis, E., Papaioannou, T. G., Aznaouridis, K., Karatzi, K., Stamatelopoulos, K., Zampelas, A., . . . Mavrikakis, M. (2005). Acute effects of caffeine on blood pressure and wave reflections in healthy subjects: should we consider monitoring central blood pressure? Int J Cardiol, 98(3), 425-‐430. doi: 10.1016/j.ijcard.2003.11.013

Kenemans, J. L., Wieleman, J. S., Zeegers, M., & Verbaten, M. N. (1999). Caffeine and stroop interference. Pharmacol Biochem Behav, 63(4), 589-‐598.

Kirsch, I., & Weixel, L. J. (1988). Double-‐blind versus deceptive administration of a placebo. Behav Neurosci, 102(2), 319-‐323.

Lee, L., Frederick, S., & Ariely, D. (2006). Try It, You'll Like It The Influence of Expectation, Consumption, and Revelation on Preferences for Beer. Psychol Sci, 17(12), 1054-‐1058.

Lorber, W., Mazzoni, G., & Kirsch, I. (2007). Illness by suggestion: expectancy, modeling, and gender in the production of psychosomatic symptoms. Annals of Behavioral Medicine, 33(1), 112-‐116.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychol Methods, 7(1), 83-‐104.

Mazzoni, G., Foan, L., Hyland, M. E., & Kirsch, I. (2010). The effects of observation and gender on psychogenic symptoms. Health Psychology, 29(2), 181.

McClure, S. M., Li, J., Tomlin, D., Cypert, K. S., Montague, L. M., & Montague, P. R. (2004). Neural correlates of behavioral preference for culturally familiar drinks. Neuron, 44(2), 379-‐387.

Moscovici, S., Lage, E., & Naffrechoux, M. (1969). Influence of a consistent minority on the responses of a majority in a color perception task. Sociometry, 32(4), 365-‐380.

Muchnik, L., Aral, S., & Taylor, S. J. (2013). Social Influence Bias: A Randomized Experiment. Science, 341(6146), 647-‐651. doi: 10.2307/23491296

Preacher, K., & Hayes, A. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36(4), 717-‐731. doi: 10.3758/BF03206553

Robinson, T. N., Borzekowski, D. L., Matheson, D. M., & Kraemer, H. C. (2007). Effects of fast food branding on young children's taste preferences. Archives of Pediatrics & Adolescent Medicine, 161(8), 792-‐797.

Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). Experimental study of inequality and unpredictability in an artificial cultural market. Science, 311(5762), 854-‐856. doi: 10.1126/science.1121066

Schachter, S., & Singer, J. (1962). Cognitive, social, and physiological determinants of emotional state. Psychol Rev, 69(5), 379.

Sherif, M. (1937). An Experimental Approach to the Study of Attitudes. Sociometry, 1(1/2), 90-‐98. doi: 10.2307/2785261

Smith, A. (2002). Effects of caffeine on human behavior. Food Chem Toxicol, 40(9), 1243-‐1255.



19

Świder, K., & Bąbel, P. (2013). The effect of the sex of a model on nocebo hyperalgesia induced by social observational learning. Pain, 154(8), 1312-‐1317.

Zaki, J., Schirmer, J., & Mitchell, J. P. (2011). Social influence modulates the neural computation of value. Psychol Sci, 22(7), 894-‐900. doi: 10.1177/0956797611411057



20

Table 1. Stepwise regression results for changes in subjective, physiological and functional alertness. Subjective

Alertness Systolic

Blood Pressure Cognitive

Interference β R2 ∆R2 β R2 ∆R2 β R2 ∆R2 Step 1 .34** .34** .42** .42** .07** .07** Baseline .58** .65** .27** Step 2 .37** .03** .45* .03* .12* .05* Baseline .59** .64** .27** Condition .19* .18* -.22*



21

Table 2. Means and (standard deviations) for outcome variables as a function of condition

Condition Disconfirming Confederate

No Confederate

Confirming Confederate

Subjective Alertness Baseline 3.38(.62) 3.34(.53) 3.31(.73) Post Consumption 3.63(.69) 3.78(.52) 3.86(.53) Systolic Blood Pressure Baseline 126.0(18.8) 130.9(25.8) 131.1(25.9) Post Consumption 115.0(32.1) 120.3(24.8) 130.6(27.6) Cognitive Interference Baseline 89.64(94.35) 81.61(130.02) 91.56(78.78) Post Consumption 100.09(94.78) 59.70(68.67) 54.13(81.65) Should Cost 1.57(.41) 1.77(.41) 1.88 (.45) Probability of Buying 2.97(1.77) 3.59(2.12) 4.50(1.74) Ambassador Initiative 2.34(1.68) 2.36(1.78) 3.19(1.79) Chose ACW over Cash* 1.10(.30) 1.08(.29) 1.32(.48) # of People Told* .05(.21) .08(.28) .63(1.67)

Note: * indicates measure taken 1 week after experiment.



22

Fig. 1. Screenshot of the AquaCharge website “The Buzz” page.



23

Fig. 2. Label affixed to 8oz plastic bottles of spring water.



24

Fig. 3. Effect of condition on subjective alertness (A) physiological alertness (SBP) (B), and

functional alertness (cognitive interference) (C).

(A) (B)

(C)

Note. Error bars reflect standard +/- 1 SE.



25

Fig. 4. Effect of condition on product endorsement. (A) (B)

(C) (D)

(E)

Note. Error bars reflect standard +/- 1 SE.

Documents

Social Influence Moderates Placebo Response 1&...2016/02/05 · Social Influence Moderates Placebo Response Draft&underreview,&last&updated2.5.16& 3& Although we are told, “a rose