Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: Psychol Addict Behav. 2019 Oct 17;33(8):721–729. doi: 10.1037/adb0000519

Experiencing Cigarette Craving with a Friend: A Shared Reality Analysis

John D Dimoff 1, Michael A Sayette 1, John M Levine 1
PMCID: PMC6889042  NIHMSID: NIHMS1050262  PMID: 31621342

Abstract

Cigarette craving predicts relapse to smoking, which remains the leading cause of preventable death in the United States. Understanding why individuals smoke has important clinical implications and is a research priority. Non-laboratory studies reveal that social factors, such as the presence of other people, are associated with self-reported craving, yet laboratory smoking research has largely ignored these factors by testing participants in isolation. In this study, a shared reality framework was used to evaluate social processes that may change when smokers experience craving while in the presence of a smoking friend vs. in social isolation. Sixty pairs of smoking friends (n = 120) arrived together at the laboratory following a required 5-hr of smoking abstinence. One pre-selected (target) participant then underwent an in vivo smoking cue-exposure craving induction either with their friend present or in another room, completing an unrelated task. Target participants who were together with their friend while craving experienced a greater sense of similarity and felt closer to their friend than did those who were alone. Further, in the together condition, shared Duchenne smiles (using the Facial Action Coding System) were associated with targets’ ratings of perceived similarity to their friend. Though social context did not influence affect or urge to smoke, urge was associated with targets’ ratings of similarity in the together, but not the alone condition. Results highlight the potential social utility of craving (satisfying epistemic and relational goals) and highlight the need for increased laboratory research on smoking that includes a social context.

Keywords: cigarette, smoking, social, shared reality, facial expressions

Smoking remains the single largest cause of premature death in the United States (U.S. Department of Health and Human Services, 2014). While smoking rates have declined since the first Surgeon General’s report on smoking was published in 1964, this trend has slowed in recent years and nearly 40 million Americans continue to smoke (Centers for Disease Control and Prevention, 2016). Neurobiological and socioeconomic factors are most often considered when investigating the stubborn persistence of smoking (Watkins, Koob, & Markou, 2000; Peretti-Watel, Seror, Constance, & Beck, 2009). However, social contextual factors (e.g., having friends who smoke) also have been identified as a powerful determinant of smoking initiation, maintenance, and cessation failure (Dimoff & Sayette, 2017; Poland et al., 2006).

For decades, researchers have recognized that smoking is influenced by social context (Glad & Adesso, 1976; Hatsukami, Morgan, Pickens, & Champagne, 1990; Piper et al., 2004; Simons-Morton & Farhat, 2010), yet few experimental smoking studies have been designed to test the effect of social contextual factors (Dimoff & Sayette, 2017). While many smoking-relevant social relationships warrant investigation in the laboratory, the present study focused on nonromantic same-sex friendships, which are common in the real-world and correlated with important smoking outcomes (e.g., reduced likelihood of quitting smoking; Hitchman, Fong, Zanna, Thrasher, & Laux, 2014). Due to the lack of experimental research on social-contextual factors in smoking, it is unclear what mechanisms determine how friends influence one another’s smoking behavior.

While there has been little experimental research examining smoking in social settings, there is even less work examining craving for cigarettes in a social context. Some studies have used smoking confederates as craving cues for participants (e.g., Niaura, Abrams, Demuth, Pinto, & Monti, 1989), but the focus was on the individual participant, and the social aspect of a shared craving was not examined (Sayette, 2016). This is unfortunate, as craving often occurs in social contexts, and studying craving while smokers are in isolation precludes assessment of a host of social factors that may contribute to the perpetuation of a smoking habit. Like others (Baker, Morse, & Sherman, 1987), we view craving as affective in nature (Sayette et al., 2003a). Affect plays a pivotal role in motivating drug use (Baker, Piper, McCarthy, Majeskie, & Fiore, 2004) and varies depending on whether it is experienced with another person or in isolation (Fairbairn et al., 2015; Knobloch, & Metts, 2013; Pliner & Cappell, 1974). Sharing the affective state of craving may influence smoking through uniquely “social” processes.

Chief among these processes is shared reality, which occurs when two or more people perceive that they have common inner states (e.g., beliefs, feelings) about some referent (Hardin & Higgins, 1996). Shared reality is sought because of its utility in satisfying two fundamental human motives—the “epistemic” need to achieve a valid and reliable understanding of the world and the “relational” need to affiliate with and feel connected to other people (Echterhoff, Higgins, & Levine, 2009). Although shared reality has certain similarities to other constructs, such as common ground, perspective taking, empathy, mood contagion, embodied synchrony, and informational social influence, it provides a novel and comprehensive framework for analyzing the causes and consequences of social sharing (see Echterhoff et al., 2009; Echterhoff & Higgins, 2017, 2018a).

Much of the empirical work on shared reality has utilized a communication paradigm in which people transmit (or, more accurately, believe they transmit) information about a target person to an audience that ostensibly has an opinion about the target (Echterhoff & Higgins, 2017). However, verbal communication is not the only way in which people provide information about their inner states. Other cues to a person’s inner state about a referent include facial expressions and gestures (Echterhoff et al., 2009), as well as eye gaze and pointing (Echterhoff & Higgins, 2018b).

Shared reality theory offers an ideal framework to analyze cigarette craving, which is a central feature of nicotine use disorder (American Psychiatric Association, 2013). Because craving can evoke powerful facial reactions in smokers (see Sayette et al., 2003a), experiencing a discrete craving state with a friend who is perceived to be craving as well (hereafter referred to as shared craving) is likely to increase smokers’ sense of shared reality about their inner state. Facial expression analysis is an especially useful tool for assessing shared craving, as it offers an unobtrusive, continuous, and objective index of smokers’ affect. Moreover, it allows examination of shared expressions across partners, which has been linked to feelings of closeness and bonding (Sayette et al., 2012). To the extent that shared craving elicits the perception of shared inner states, this shared reality will deserve consideration as a possible motive for continued smoking over and above such widely held motives as reducing nicotine withdrawal.

To our knowledge, no studies have tested the impact of the social context (e.g., the presence vs. absence of another person) during a cigarette craving induction task or any other task involving anticipation of consumptive behaviors (e.g., the marshmallow test; Mischel, Shoda, & Rodriguez, 1989; see also Watts, Duncan, & Quan, 2018). The present study was thus the first to examine craving in a social context.

In this experiment, we studied pairs of nonromantic same-sex smoking friends. Each pair was comprised of one participant who had completed a prior study and another participant who had not participated in that study. Participants who had completed the parent study experienced a peak-provoked craving (combining smoking abstinence and smoking cue-exposure; see Sayette & Tiffany, 2013) either together with their friend or in isolation. (Note that in the latter condition, it was made explicit to each participant that the friend without prior research involvement would be completing a task unrelated to smoking in a separate room.) Our first goal was to assess the impact of being with a friend vs. being alone during an in vivo smoking cue-exposure on shared reality (perceived similarity and closeness), self-report craving (urge, affect), and facial reactions (Duchenne smiles, which are characterized by raised mouth corners and cheeks and are associated with felt enjoyment; Ekman, 1989). We predicted that, compared to individuals who were by themselves (Alone), individuals who were with a smoking friend (Together) would 1) report stronger perceptions of similarity and closeness toward their friend, 2) report higher urges to smoke and more intense affect, consistent with the idea that affective experiences tend to be amplified when shared (Boothby, Clark, & Bargh, 2014), and 3) display more Duchenne smiles.

Our second goal was to assess the effect of urge to smoke on shared reality in the Together and Alone conditions. Because it is not enough for people merely to be together to achieve shared reality, we predicted that urge to smoke would moderate the predicted effect of social context on shared reality measures, such that the differential impact of the Together and Alone conditions on perceived similarity and closeness would be greater when participants experienced higher (vs. lower) urge to smoke.

Lastly, we aimed to test whether dyadic smiling behavior in the Together condition was associated with participants’ sense of shared reality. In line with the proposition that synchronous nonverbal expressions are used to communicate commonality of experience (Echterhoff et al., 2009), we predicted that participants would report greater perceptions of similarity and closeness to their friend when they engaged in more (vs. less) co-occurring Duchenne smiling.

By using a “social” paradigm, we sought to expand the scope of craving-related processes to include social experiences associated with the sharing of a craving state. Although craving has been studied in socially isolated individuals, in natural environments it often occurs in social contexts. Failure to address the social aspects of shared craving precludes a comprehensive analysis of the construct (Sayette, 2016). In addition, we sought to expand the scope of research on shared reality, which has thus far primarily focused on shared cognitive states (e.g., attitudes and memories; Echterhoff, Higgins, Kopietz, & Groll, 2008), by examining shared affective states.

Method

Participants

The present study included 120 participants aged 18-55. This sample size was chosen because it gave us sufficient power to detect meaningful (i.e., medium-to-large-sized) effects of social context, while operating within budgetary and logistical constraints. Sixty participants (half female) were recruited from a prior study (Sayette et al., in press) in which they participated alone (described herein as targets). The remaining 60 participants (half female) were smoking friends of these targets, who had not participated in the prior study (described herein as friends). The purpose of the prior study was to test the impact of olfactory cues on craving reduction in response to an in vivo smoking cue and to examine whether individual difference factors, such as working memory, personality, and motivation to quit smoking, moderated the craving-reducing effects of olfactory cues.

Targets were recruited via telephone calls and e-mails informing them that they were eligible to participate in an optional follow-up to the study they previously had completed. Interested targets were told that the current study would require them to participate with a same-sex nonromantic friend with whom they regularly smoke and that they would need to provide this friend with our contact information so that the friend could initiate contact with the lab.

All participants (i.e., targets and friends) were screened over the phone to ensure that they were appropriate for the present study based on their current smoking patterns. To qualify, participants were required to be between the ages of 18–55 and to smoke an average of 10–30 cigarettes/day for at least 12 continuous months. Participants were excluded if they reported any of the following: a health condition that contraindicated nicotine administration, difficulties with reading suggestive of illiteracy, or intent to quit smoking within the next 30 days (see Wertz & Sayette, 2001). Participants were informed that they would be paid $30 and reimbursed for travel expenses for successful completion of the study.

Procedures

Informed consent was obtained from all participants, using procedures approved by the University of Pittsburgh’s Institutional Review Board. The study included two experimental conditions: one in which targets completed the smoking cue-exposure task with their friend (Together, n = 30), and one in which targets completed this task alone (Alone, n = 30). All target-friend dyads, regardless of condition, were instructed to travel to the laboratory together, and compliance was assessed via self-report. Targets and friends stayed in separate rooms before and after smoking cue-exposure. Targets in both the Together and Alone conditions completed the smoking cue-exposure task, whereas friends completed the task in the Together but not in the Alone condition. [In the Alone condition, while targets completed the smoking cue-exposure task, friends stayed in a separate room and completed a packet of questionnaires about health behaviors unrelated to smoking (e.g., food and alcohol consumption).] The first two hypotheses were tested using self-report and facial-expressive data only from targets.1 The third hypothesis, involving dyadic Duchenne smiling, was tested using data from both targets and friends.

Baseline period.

Sessions began between 3:00 p.m. and 7:00 p.m. Consistent with prior work (see Sayette & Dimoff, 2016), participants were required to bring their preferred brand of cigarettes and lighter to the experimental session and to abstain from nicotine for at least 5-hr prior to their appointment. Our decision to use 5-hr rather than a more typical (i.e., longer) nicotine deprivation interval was based on our desire to induce powerful but not overwhelming cigarette craving, as extreme levels of withdrawal or negative affect might override the impact of social factors.

Upon arriving at the laboratory, targets and friends were placed in separate rooms and asked to verbally report the degree to which they were craving a cigarette using a single-item, 0–100 scale with 0 = “no urge to smoke at all” and 100 = “the most intense urge to smoke that I have ever felt.” They then completed the Positive and Negative Affective Schedule (PANAS; Watson, Clark, & Tellegen, 1988) and the Nicotine Dependence Syndrome Scale (NDSS; Shiffman, Waters, & Hickcox, 2004). In addition, to account for possible baseline differences in the characteristics of friendships, participants completed a 5-item inventory assessing 1) how long they had known each other (in years), 2) how close they felt toward one another (0 = “not at all close,” 10 = “very close”), 3) how often they smoked together (number of times per week), and whether they were 4) living or 5) working together (yes/no).

Participants’ breath carbon monoxide (CO) readings were collected and recorded, with CO readings ≤ 20 ppm considered abstinent.2 (This cutoff point was higher than those used in some previous studies because participants in the present study were required to abstain from smoking for a shorter interval than is typically required.) Once abstinence was confirmed, targets were randomly assigned to either the Together or Alone condition, with the constraint that approximately equal numbers of males and females were assigned to each condition. Because merely separating individuals does not guarantee that they will feel that they are engaged in different activities (Boothby, Smith, Clark, & Bargh, 2016), participants assigned to the Alone condition were told that their friend would be completing a series of tasks unrelated to smoking. Following assignment to condition, targets provided the experimenter with their cigarette pack and lighter, which were placed in another room until the cue-exposure procedure.

Peak-provoked craving induction.

In the Alone condition, targets were moved to the experimental room after baseline assessment, while friends completed questionnaires in a separate room. A tray containing an opaque cover was placed on the desk at which targets were seated, and they were asked not to touch the tray until instructed to do so. Once the experimenter left the room, targets were asked via intercom to pick up the cover, revealing their cigarettes and lighter as well as an ashtray. Targets were instructed to remove a cigarette from their pack and to light it without putting it in their mouth by holding it in the flame for several seconds until the tobacco started to burn. Targets were informed that they could not smoke the cigarette until told that they could do so. They next were told to put down their lighter, to hold their cigarette comfortably in their dominant hand, and to stare at it for 20-sec without placing it in their mouth.

The smiling behavior of targets during these 20-sec was recorded using a digital video camera and later coded using the Facial Action Coding System (FACS), which is the most comprehensive system for coding observable facial muscle movements, referred to as action units (AUs; Ekman et al., 2002). FACS is a continuous (frame by frame) and unobtrusive coding tool with excellent psychometric properties (Sayette, Cohn, Wertz, Perrott, & Parrott, 2001). For the present study, a FACS-certified coder blind to experimental condition coded facial muscle movements associated with Duchenne smiles, which have been linked to craving in studies in which individual smokers underwent a comparable craving induction (Sayette et al, 2003b; Sayette & Hufford, 1995). Moreover, when Duchenne smiles occur in a social context they also are thought to facilitate interpersonal communication (Fridlund, 1997).

When 20-sec had elapsed, targets were told to set, but not extinguish, their cigarette in the ashtray on the tray in front of them. They then were asked to complete a questionnaire assessing their urge to smoke, using the same scale used at baseline, and their affect, using Carter and Tiffany’s (2001) two-item scale, which consists of separate positive and negative affect items and from which a single affect score was calculated ranging from extremely negative (−10) to extremely positive (+10).

Targets next completed questionnaires assessing perceptions of similarity and closeness toward their friend during the cue-exposure task. The similarity inventory, designed to assess the epistemic effects of the social context manipulation, was comprised of three 10-point items: “My friend is feeling the same way Fm feeling,” “My friend can relate to my experiences in this study,” “My friend wants to smoke as badly as I do.” Responses to these items were averaged to create a composite score ranging from 0 = “strongly disagree” to 10 = “strongly agree.” The closeness inventory, designed to assess relational effects of the social context manipulation, was comprised of two 7-point items: a measure of perceived closeness adapted from Orehek, Forest, and Wingrove (2018) and the Inclusion of Other in the Self Scale (IOS) (Aron, Aron, & Smollan, 1992). Responses to these items were averaged to create a composite score ranging from 1 = “not close at all” to 7 = “extremely close”.

Targets and friends then were debriefed in separate rooms, paid $30, reimbursed for travel, and told where they could smoke outside the laboratory if they wished to do so before leaving.

The same procedure was used in the Together condition, with the following exceptions: targets and friends were moved to the experimental room after baseline assessment, and each was given a tray containing the materials noted above. In the experimental room, targets and friends sat at a 90-degree angle from one another, with a small divider between them so that they could see one another’s face but not one another’s written responses. In addition, they were asked not to talk about their cravings during the cue-exposure task. In this condition, both targets’ and friends’ smiling behavior was recorded. Lastly, targets and friends were moved to separate rooms before being given the questionnaires assessing similarity and closeness.

Data Coding and Analysis

Facial coding.

The Observer XT software system (Version 10.5, Noldus Information Technology, Wageningen, The Netherlands) was used to code digital video footage of participants’ facial expressions recorded during smoking cue exposure. A FACS-certified coder blind to experimental condition assessed positive affect by measuring the duration of participants’ Duchenne smiles—defined by the combination of action unit (AU) 6 (“cheek raiser”) and AU 12 (“lip corner puller”; Ekman, 1989). In addition, shared Duchenne smiles (i.e., those occurring simultaneously between targets and friends) were coded in the Together group as an index of positive affect at the dyad level. Consistent with prior work (e.g., Sayette et al., 2012), “trace” levels of AUs were not coded (i.e., those at an “A” intensity; see Ekman et al., 2002). The total duration of Duchenne smiles was calculated in seconds. A second FACS-certified coder coded smiles of a randomly selected subset of 24 participants (20% of the total sample). Intercoder agreement (Cohen, 1960) was high (κs = .91).

Analytic approach.

Analyses were conducted using IBM SPSS Statistics 25. Initial tests were conducted to assess whether baseline measures varied by condition or correlated with outcome measures. Baseline measures meeting either of these criteria were used as covariates in remaining analyses. As noted, self-report data were analyzed for the 60 target participants. Facial expression data were analyzed for all targets and for the 30 friends in the Together condition. ANOVA was used to test for the main effect of social context (Together vs. Alone) on all outcome measures (perceived similarity, closeness, urge, affect, Duchenne smiles). Linear regression was used to test for the moderating effect of urge on perceived similarity and for associations between co-occurring Duchenne smiles and target-reported perceptions of similarity and closeness.

Results

Baseline Measures

Fifty-eight percent of targets identified themselves as African American, 37% as Caucasian, and 5% as more than one race. On average, targets were 43.5 years old and reported having smoked 15.1 cigarettes per day for the past 14.6 years. [Note that friends reported equivelant ages (average = 42.8 years old) and smoking patterns (15.2 cigarettes per day for 14.0 years).]

Preliminary Analyses

Random assignment.

Table 1 presents the characteristics of target participants by social context condition. Participants in the two conditions did not differ significantly on demographic characteristics (age, sex, race), smoking patterns (number of cigarettes per day, years smoking at current rate, nicotine dependence as assessed via NDSS), friendship characteristics (years known, baseline closeness, number of times smoked together per week, whether they live and/or work together), and baseline measures (hours since last cigarette, CO reading, baseline urge to smoke, and baseline affect as assessed via PANAS).

Table 1.

Target participant characteristics by social context condition.

Together (n = 30) Alone (n = 30)
Demographics
 Age 44.05 (9.98) 42.65 (10.06)
 % Female 53.3% 46.7%
 % African American 53.3% 63.3%
 % Caucasian 40.0% 33.3%
 % More than one race 6.7% 3.3%
Smoking patterns
 Cigarettes per day 15.83 (5.53) 14.45 (4.98)
 Years Smoking 15.43 (9.60) 13.77 (10.65)
 Dependence (NDSS) −.14 (1.11) .13 (0.89)
Friendship characteristics
 Years known 11.48 (9.11) 8.97 (8.07)
 Baseline closeness 8.75 (1.39) 8.07 (1.95)
 Smoke together per week 5.35 (7.23) 5.30 (5.26)
 % Live together 6.7% 16.7%
 % Work together 13.3% 13.3%
Baseline measures
 Hours since last cigarette 5.96 (2.27) 7.71 (5.14)
 CO reading 9.67 (5.83) 8.97 (5.33)
 Baseline urge to smoke 65.37 (30.19) 69.03 (28.58)
 Baseline affect (PANAS) 17.77 (9.20) 18.03 (11.40)
Open in a new tab

Note: Participants randomly assigned to the Together and Alone conditions did not differ on any of the above characteristics (all ps > . 10).

Tests of normality.

Negative skew was detected for similarity, closeness, and urge, while positive skew was detected for Duchenne smiles. Similarity, closeness, and urge required squaring to reduce negative skew, while Duchenne smiles required a cube root transformation to reduce positive skew. Once transformed, these measures tested negative for skewness and heteroscedasticity.

Items used to create composite scores.

As shown in Table 2, the three items comprising the similarity inventory were all positively and significantly correlated with one another. In addition, the two items comprising the closeness inventory were positively and significantly correlated, r = .83, p < .001, while positive and negative affect were negatively and significantly correlated, r = −.43, p < .01.

Table 2.

Correlations among items on the similarity inventory.

Similar 1 Similar 2
Similar 2 .77**
Similar 3 .71** .53**
Open in a new tab

Similar 1 = My friend is feeling the same way I’m feeling.

Similar 2 = My friend can relate to my experiences in this study.

Similar 3 = My friend wants to smoke as badly as I do.

**

= Correlation is significant at the .01 level.

Associations Among Baseline and Outcome Measures

Because this study used a novel set of measures for which associations were mostly unknown, covariates were not identified a priori. Post-hoc analyses revealed that each of the four outcome measures (similarity, closeness, urge, affect) was significantly associated with at least three baseline measures (see Table 3). For this reason, two analyses were run for each outcome: one that included the outcome’s associated baseline measures as covariates and one that did not. There were no differences between these analyses with respect to the statistical significance of the relationships between social context, on the one hand, and self-report and smiling measures, on the other hand. Therefore, to avoid the possibility of overfitting, findings are reported from analyses that did not use covariates.

Table 3.

Correlations among baseline and outcome measures across social context conditions.

Similar Close Urge Affect D-smiles
Demographics
 Age −.32* .09 −.28* .26* −.21
 Sex .07 .37** .01 −.06 −.13
 Race −.14 −.20 .04 .11 −.21
Smoking patterns
 Cigarettes per day .03 −.12 −.04 −.05 .21
 Years Smoking −.03 −.13 −.12 −.02 −.26*
 Dependence (NDSS) .32* −.09 .39** −.30* .12
Friendship characteristics
 Years known −.03 .43** −.01 .15 −.02
 Baseline closeness .10 .56** −.01 .00 .17
 Smoke together per week .17 .07 .05 −.12 .04
 % Live together −.08 .08 .03 −.06 −.05
 % Work together −.05 −.08 −.09 −.14 −.06
Baseline measures
 Hours since last cigarette .11 .14 .09 −.17 −.22
 CO reading .16 −.22 −.10 −.10 .21
 Baseline urge to smoke .31* .11 .81** −.33* .25
 Baseline affect (PANAS) −.23 .31* −.15 .46** .08
Open in a new tab
**

= Correlation is significant at the .01 level.

*

= Correlation is significant at the .05 level.

Similar = Similarity composite

Close = Closeness composite

Urge = Urge to smoke

Affect = Craving-related affect

D-smiles = Duchenne smiles

The association of similarity and closeness did not reach significance [r(58) = .19; p = .15], suggesting that these measures captured distinct aspects of shared reality. Furthermore, there was a significant negative association between urge and (positive) affect across social context conditions, r(58) = −.35; p < .01. Consistent with our intention to create a benign (vs. aversive) craving state, however, inspection of the affect scores indicated that participants generally felt neutral to positive during the study, with 76.7% of participants reporting affect scores ranging from neutral (zero) to maximally positive (+10).

Tests of Hypotheses

Shared reality processes.

Consistent with our first hypothesis, there were significant effects of social context on both similarity, F(1, 58) = 4.81, p < .04, d = .58, and closeness, F(1, 58) = 6.69, p < .02, d = .68. As shown in Table 4, targets in the Together condition had a stronger sense of similarity and felt closer to their friend than did targets in the Alone condition.

Table 4.

Mean (SD) responses on outcome measures.

Together Alone
Shared reality processes
 Similarity* 23.81 (7.03) 18.81 (9.93)
 Closeness* 5.38 (1.21) 3.84 (2.35)
Inner states
 Urge to smoke 74.16 (29.50) 78.93 (23.52)
 Craving-related affect 2.89 (4.29) 2.41 (5.91)
Facial expressionsa
 Duchenne smiles** 3.72 (4.23) 0.87 (2.07)
 Duchenne smiles (dyad)b 2.24 (3.27) —
Open in a new tab
**

= Difference is significant at the .01 level.

*

= Difference is significant at the .05 level.

a

Reported in number of seconds.

b

Dyadic smiles were only assessed in the Together condition.

Note: Values were calculated using non-transformed data.

Inner states.

Contrary to our second hypothesis, there was no significant effect of social context on either urge to smoke or craving-related affect (ps > .60).

Facial expressions.

Consistent with our third hypothesis, there was a significant effect of social context on Duchenne smiling during the craving induction, F(1, 58) = 14.48, p < .001, d = .98. Targets in the Together condition displayed significantly longer Duchenne smiles than did targets in the Alone condition.

In the Together condition, 57% of dyads displayed at least one shared Duchenne smile during the craving induction, with shared smiles accounting for an average of 2.34-sec. (SD = 3.27) of the 20-sec induction. More than three-quarters (76.7%) of all Duchenne smiles displayed in the Together condition were shared, with sharing occurring on average 1.90-sec (SD = .95-sec) after smiling was initiated by one member of the dyad. Moreover, shared Duchenne smiles appeared across the entire 20-sec craving induction interval (i.e., they were initiated at 13 of 20 possible 1-sec bins), suggesting that these smiles were not stimulated by a single momentary stimulus during the craving period.

Moderating effect of urge.

As we predicted, there was a significant moderating effect of urge on the effect of social context on similarity, β = .42, t(59) = 2.56, p < .02, 95% CI [.09, .75]. Specifically, there was a robust association between urge and similarity (r = .74, p < .001) in the Together condition during smoking cue-exposure, but these measures were unrelated in the Alone condition (r = .081, p = .67). Inspection of scatterplots revealed a wide range of scores for urge and similarity across experimental conditions, suggesting that the observed moderation effect was unlikely due to unusually high or low scores reported by only a few targets. To further explore the association between urge and similarity, a median split on urge (Mdn = 83.50) revealed that in the Together condition, targets with higher urges reported a stronger sense of similarity (M = 28.14, SD = 1.99) than did targets with lower urges (M = 20.01, SD = 7.70), F(1, 28) = 17.63, p < .001, d = 1.53. In contrast, the parallel comparison in the Alone condition did not reach significance (higher urge M = 20.06, SD = 8.94; lower urge M = 17.37, SD = 11.13; p = .60). Contrary to our prediction, urge did not moderate the impact of social context on perceived closeness (p > .55).

Relationships between dyadic smiling and shared reality measures.

As hypothesized, there was a significant positive association between shared Duchenne smiling and perceptions of similarity, β = .36, t(29) = 2.06, p < .05, 95% CI [.02, .72]. Targets who engaged in more shared Duchenne smiling with their friend—the proposed building block of shared reality in this study—had a stronger sense of similarity. Contrary to our hypothesis, shared Duchenne smiles were not associated with targets’ ratings of closeness (p > .75).

Discussion

The idea that social context may affect an individual’s attitudes about smoking is not new (Mettlin, 1976), and many smokers spend time with other smokers as part of their daily routines (Conklin, Salkeld, Perkins, & Robin, 2013; Dunbar, Scharf, Kirchner, & Shiffman, 2010). Laboratory studies have, however, evaluated craving exclusively among smokers tested alone; thus, it is unclear how social context affects the experience of craving. If craving in the presence of another person satisfies epistemic and relational goals—by reinforcing the subjective validity of one’s perception of craving and promoting interpersonal closeness, respectively—it would suggest that shared reality is a potentially important motive for continued smoking.

Our findings provided support for a shared reality explanation of craving. We found that participants who underwent an in vivo smoking cue-exposure craving induction with a friend experienced a greater sense of shared reality, as evidenced by greater perceived similarity and closeness to the friend, than did those who were alone. This occurred in spite of the fact that the Together and Alone conditions differed minimally (i.e., in both conditions, participants traveled to the laboratory together and spent equal time apart completing questionnaires prior to cue exposure). Moreover, both urge to smoke and shared Duchenne smiling were positively associated with perceived similarity for targets in the presence of a friend during the craving induction.

Several pieces of evidence suggest that the social nature of the together experience was critical to the friend’s impact during craving. First, self-reported urge to smoke was associated with perceptions of similarity when participants were together during craving—but not when they were alone. Moreover, in the Together condition, as urge increased so too did the perception that the friend was having a similar experience. While our correlational data preclude causal inferences, they suggest that the association between urge and similarity was dependent upon the presence of another smoker. The absence of an association between urge and similarity in the Alone condition argues against a less interesting possibility that increased urge to smoke merely enhances perceptions of similarity toward smoking friends regardless of social context.

Second, there was a main effect of social context on Duchenne smiling. This reinforces the social-affective aspect of the Together condition used in the present study, distinguishing it from the Alone condition in which the participant did not interact with his or her friend during craving. More importantly, among participants in the Together condition, dyad-level (but not individual-level) Duchenne smiles expressed during craving predicted subsequent ratings of similarity. While the temporal order of these two assessments is consistent with the view that sharing smiles increased perceptions of similarity, an alternate interpretation—that perceived similarity stimulated the expression of shared Duchenne smiles—cannot be ruled out. Nevertheless, our data are consistent with the proposition that facial expression provides a social signal that enables interaction partners to infer that they are sharing feelings, in this case desire to smoke (Echterhoff et al., 2009).

Although shared Duchenne smiles predicted perceptions of similarity, they did not predict perceptions of closeness. The near-zero correlation between shared Duchenne smiles and closeness ratings suggests that the absence of a significant association was not due to limited power. Nor does this null finding appear to result from ceiling effects, based on inspection of scatterplots. Perhaps our peak-provoked craving manipulation caused friends, who knew one another well, to interpret shared Duchenne smiles as signaling shared attention to the smoking cue (“I’m also really feeling the urge to smoke.”) rather than the expression of warmth. Though speculative, this interpretation is in line with evidence that the meaning and function of facial expressions depends upon the context in which they occur (Barrett, 2014).

It is interesting that social context did not influence self-reported urge or affect during cue-exposure. Inspection of means and frequencies suggests that these null findings cannot be attributed to ceiling effects or limitations in power. One possible explanation for our findings is that the 20-sec craving induction may not have been sufficient to elicit social context effects on urge or affect (cf. Boothby et al., 2016). It should be noted, however, that extending the craving interval without permitting consumption raises methodological challenges, as even brief delays (e.g., 15-sec extensions) before smoking while holding the lit cigarette can increase negative affect and thereby narrow the range of emotions experienced during cue-exposure (Sayette et al., 2003b).

Another possible explanation for the null effect of social context on urge is that some participants may not have abstained from nicotine for the full 5-hr interval, and therefore been less reactive to the study’s craving induction. This explanation cannot be ruled out because we did not collect a pre-abstinence CO value to compare with the post-abstinence value. This limitation notwithstanding, it should be noted that CO values in this study were comparable to those collected in another study that used a 5-hr abstinence period (Sayette & Dimoff, 2016), which included pre- and post-abstinence values. Moreover, urge ratings in the present study were generally high (Mdn = 83.50), which compares favorably to rating obtained in our prior work when both CO rating were collected.

Implications and Caveats

Theoretical implications.

While much research focuses on the role of nicotine and smoking behavior in maintaining smoking, a comprehensive theoretical analysis of smoking maintenance requires consideration of the social context in which smoking and craving occur (Dimoff & Sayette, 2017). Little theoretical attention has been directed toward the processes underlying social contextual effects. From a shared reality perspective, the present study suggests that a smoking habit may be maintained, at least in part, because it routinely creates moments of smoking anticipation that are shared with friends in ways that satisfy epistemic and relational goals. That is, when smokers experience a common anticipation of smoking a cigarette, they may perceive their anticipatory affect as valid (i.e., it is “correct” to crave under these circumstances) and their relationship as close.

In addition to enriching our understanding of the factors that underlie smoking maintenance, this study contributes to the literature on shared reality. It does so by providing evidence supporting the proposition that “processes involved in empathizing and mood contagion ... might serve as precursor mechanisms in the unfolding of a full-blown shared reality” (Echterhoff et al., 2009, p. 511).

Clinical implications.

As noted earlier, the social context manipulation had no effect on our measure of urge, which served as this study’s most proximal measure of cigarette use risk. Such data, in tandem with the shared reality findings, leaves open the possibility that social factors might maintain tobacco use not by increasing the intensity of craving, but by facilitating the social utility of craving. Indeed, to the degree that shared cravings provide smoking friends with anticipatory experiences that satisfy epistemic and relational needs, the loss of shared cravings could be considered a “hidden” barrier for smoking cessation that cannot be accounted for by models that focus only on craving intensity as a predictor of tobacco use. Further research across a range of abstinence intervals would be useful to comprehensively evaluate the impact of a friend on urge intensity.

More generally, our results accord with recent work suggesting that many harmful behaviors are “rational” when viewed as means to goals (Kopetz & Orehek, 2015). The present data suggest that smoking—despite its harmful health consequences—may persist, at least in part, because it allows smokers to share craving states in ways that validate their understanding of the world and cause them to feel closer to one another. Certainly, efforts to enhance shared reality by themselves are unlikely to account for smoking, but they may prove to be a worthy addition to an expanding set of relevant factors for maintaining the smoking habit.

Future Directions

The present data suggest several directions for future research. First, the inclusion of other conditions would shed light on the role that shared reality plays in craving. As one example, including a condition in which smoking-deprived participants and friends are together but not exposed to an in vivo smoking cue would shed light on whether the observed differences between the Together and Alone conditions in the present study are specific to sharing (vs. not sharing) a cigarette craving. As another example, it would be informative to assess how smoking-deprived participants would react if a friend in the same room was experiencing an inner state different from theirs (e.g., nicotine satiation, hunger) or if a friend in a different room was experiencing the same state.

Second, this study observed effects of social context on daily smokers, which makes them all the more interesting because of the long-held assumption that the smoking behavior of such smokers is driven more by pharmacological than environmental factors (Miller, Frederiksen, & Hosford, 1979). Future research should explore the impact of social context on smokers at different stages of dependence, including those motivated to seek cessation treatment (Sayette & Dimoff, 2016). By so doing, future studies would be positioned to identify factors that moderate smokers’ sensitivity to their social context.

Third, future research examining other kinds of dyads, including strangers and romantic partners, would be useful. It may be, for example, that strangers are less sensitive than friends to features of their social context whereas romantic partners are more sensitive.

Conclusions

The present study is the first to test the impact of the social context during a peak-provoked craving state. In addition, the study introduced a conceptual framework (shared reality theory) and measures that have not been used previously to assess craving-related responses. These measures (e.g., shared smiling) are novel to the craving literature in that they assess social processes and cannot be utilized in traditional (i.e., asocial) craving paradigms. Moreover, this is the first study to evaluate a behavioral mechanism potentially underlying the construction of shared reality (mutual smiling behavior). Our findings suggest that a comprehensive analysis of smoking behavior requires systematic attention to social processes, such as shared reality, as well as neurobiological and pharmacological processes.

Acknowledgments

This research was supported in part by the National Cancer Institute (R01 CA184779) and a dissertation award to John D. Dimoff from the American Psychological Association. We thank Ed Orehek, Eric Donny, Michael Pogue-Geile, and Michele Levine for assistance.

Footnotes

Data in the article were presented in a doctoral dissertation completed by the first author.

1

Focusing on targets allowed us to hold constant participants’ recent involvement with our research program.

2

Consistent with prior work, participants whose CO readings exceeded 20 ppm were permitted to wait up to 1-hr in the laboratory to reach acceptable levels. Six participants required extra time to reach ≤ 20 ppm, with no participant requiring more than 30-min. In each of these cases, the participant and his or her partner were informed of the situation and required to wait in their respective rooms.

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: American Psychiatric Association. [Google Scholar]
  2. Aron A, Aron EN, & Smollan D (1992). Inclusion of Other in the Self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596–612. [Google Scholar]
  3. Baker TB, Morse E, & Sherman JE (1987). The motivation to use drugs: A psychobiological analysis of urges In Nebraska symposium on motivation. University of Nebraska Press. [PubMed] [Google Scholar]
  4. Baker TB, Piper ME, McCarthy DE, Majeskie MR, & Fiore MC (2004). Addiction motivation reformulated: an affective processing model of negative reinforcement. Psychological Review, 111(1), 33. [DOI] [PubMed] [Google Scholar]
  5. Barrett LF (2014, February 28). What faces can’t tell us. The New York Times, Retrieved from http://www.nytimes.com. [Google Scholar]
  6. Boothby EJ, Clark MS, & Bargh JA (2014). Shared experiences are amplified. Psychological Science, 25(12), 2209–2216. [DOI] [PubMed] [Google Scholar]
  7. Boothby EJ, Smith LK, Clark MS, & Bargh JA (2016). Psychological distance moderates the amplification of shared experience. Personality and Social Psychology Bulletin, 42(10), 1431–1444. [DOI] [PubMed] [Google Scholar]
  8. Bradford DE, Curtin JJ, & Piper ME (2015). Anticipation of smoking sufficiently dampens stress reactivity in nicotine-deprived smokers. Journal of Abnormal Psychology, 124(1), 128–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carter BL, & Tiffany ST (2001). The cue-availability paradigm: the effects of cigarette availability on cue reactivity in smokers. Experimental and Clinical Psychopharmacology, 9(2), 183–190. [DOI] [PubMed] [Google Scholar]
  10. Centers for Disease Control and Prevention. (2016). Cigarette smoking among adults—United States, 2005–2015. Morbidity and Mortality Weekly Report, 65(44), 1205–11. [DOI] [PubMed] [Google Scholar]
  11. Cohen J (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46. [Google Scholar]
  12. Conklin CA, Salkeld RP, Perkins KA, & Robin N (2013). Do people serve as cues to smoke? Nicotine & Tobacco Research, 75(12), 2081–2087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dimoff JD, & Sayette MA (2017). The case for investigating social context in laboratory studies of smoking. Addiction, 112(3), 388–395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dunbar MS, Scharf D, Kirchner T, & Shiffman S (2010). Do smokers crave cigarettes in some smoking situations more than others? Situational correlates of craving when smoking. Nicotine & Tobacco Research, 12(3), 226–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Echterhoff G, & Higgins ET (2018b). Shared reality: Construct and mechanisms. Current Opinion in Psychology, 23, iv–vii. [DOI] [PubMed] [Google Scholar]
  16. Echterhoff G, & Higgins ET (Eds.) (2018a). Shared reality [Special issue]. Current Opinion in Psychology, 23. [DOI] [PubMed] [Google Scholar]
  17. Echterhoff G, Higgins ET, & Levine JM (2009). Shared reality experiencing commonality with others’ inner states about the world. Perspectives on Psychological Science, 4(5), 496–521. [DOI] [PubMed] [Google Scholar]
  18. Echterhoff G, Higgins ET, Kopietz R, & Groll S (2008). How communication goals determine when audience tuning biases memory. Journal of Experimental Psychology: General, 137 (1), 3–21. [DOI] [PubMed] [Google Scholar]
  19. Ekman P (1989). The argument and evidence about universals in facial expressions. Handbook of social psychophysiology, 143–164. [Google Scholar]
  20. Ekman P, Friesen WV, & Hager JC (2002). FACS investigator’s guide. A Human Face, 96. [Google Scholar]
  21. Fairbairn CE, Sayette MA, Wright AG, Levine JM, Cohn JF, & Creswell KG (2015). Extraversion and the rewarding effects of alcohol in a social context. Journal of Abnormal Psychology, 124(3), 660–673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fridlund AJ (1997). The new ethology of human facial expressions In Russell JA & Fernández-Dols JM (Eds.), Studies in emotion and social interaction, 2nd series. The psychology of facial expression (pp. 103–129). [Google Scholar]
  23. Glad W, & Adesso VJ (1976). The relative importance of socially induced tension and behavioral contagion for smoking behavior. Journal of Abnormal Psychology, 85(1), 119–121. [DOI] [PubMed] [Google Scholar]
  24. Hardin CD, & Higgins ET (1996). Shared reality: How social verification makes the subjective objective In Sorrentino RM & Higgins ET (Eds.), Handbook of motivation and cognition. Handbook of motivation and cognition, Vol. 3. The interpersonal context (pp. 28–84). New York, NY, US: Guilford Press. [Google Scholar]
  25. Hatsukami DK, Morgan SF, Pickens RW, & Champagne SE (1990). Situational factors in cigarette smoking. Addictive Behaviors, 15(1), 1–12. [DOI] [PubMed] [Google Scholar]
  26. Hitchman SC, Fong GT, Zanna MP, Thrasher JF, & Laux FL (2014). The relation between number of smoking friends, and quit intentions, attempts, and success: Findings from the International Tobacco Control (ITC) Four Country Survey. Psychology of Addictive Behaviors, 28(4), 1144–1152. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Knobloch LK, & Metts S (2013). Emotion in relationships In Simpson JA, & Campbell L (Eds.). The Oxford handbook of close relationships (pp. 514–534). New York: Oxford University Press. [Google Scholar]
  28. Kopetz C, & Orehek E (2015). When the end justifies the means: Self-defeating behaviors as “rational” and “successful” self-regulation. Current Directions in Psychological Science, 24(5), 386–391. [Google Scholar]
  29. Mettlin C (1976). Peer and other influences on smoking behavior. Journal of School Health, 46(9), 529–536. [DOI] [PubMed] [Google Scholar]
  30. Miller PM, Frederiksen LW, & Hosford RL (1979). Social interaction and smoking topography in heavy and light smokers. Addictive Behaviors, 4(2), 147–153. [DOI] [PubMed] [Google Scholar]
  31. Mischel W, Shoda Y, & Rodriguez MI (1989). Delay of gratification in children. Science, 244(4907), 933–938. [DOI] [PubMed] [Google Scholar]
  32. Niaura R, Abrams D, Demuth B, Pinto R, & Monti P (1989). Responses to smoking-related stimuli and early relapse to smoking. Addictive Behaviors, 14(4), 419–428. [DOI] [PubMed] [Google Scholar]
  33. Orehek E, Forest AL, & Wingrove S (2018). People as means to multiple goals: Implications for interpersonal relationships. Personality and Social Psychology Bulletin, 44 (10), 1487–1501. [DOI] [PubMed] [Google Scholar]
  34. Peretti-Watel P, Seror V, Constance J, & Beck F (2009). Poverty as a smoking trap. International Journal of Drug Policy, 20(3), 230–236. [DOI] [PubMed] [Google Scholar]
  35. Piper ME, Piasecki TM, Federman EB, Bolt DM, Smith SS, Fiore MC, & Baker TB (2004). A multiple motives approach to tobacco dependence: the Wisconsin Inventory of Smoking Dependence Motives (WISDM-68). Journal of Consulting and Clinical Psychology, 72(2), 139–154. [DOI] [PubMed] [Google Scholar]
  36. Pliner P, & Cappell H (1974). Modification of affective consequences of alcohol: A comparison of social and solitary drinking. Journal of Abnormal Psychology, 83(4), 418–425. [DOI] [PubMed] [Google Scholar]
  37. Poland B, Frohlich K, Haines RJ, Mykhalovskiy E, Rock M, & Sparks R (2006). The social context of smoking: the next frontier in tobacco control?. Tobacco Control, 15(1), 59–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sayette MA (2016). The role of craving in substance use disorders: Theoretical and methodological issues. Annual Review of Clinical Psychology, 12, 407–433. [DOI] [PubMed] [Google Scholar]
  39. Sayette MA, & Dimoff JD (2016). In search of anticipatory cigarette cravings: The impact of perceived smoking opportunity and motivation to seek treatment. Psychology of Addictive Behaviors, 30(3), 277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sayette MA, & Hufford MR (1994). Effects of cue exposure and deprivation on cognitive resources in smokers. Journal of Abnormal Psychology, 103(4), 812–818. [DOI] [PubMed] [Google Scholar]
  41. Sayette MA, & Hufford MR (1995). Urge and affect: A facial coding analysis of smokers. Experimental and Clinical Psychopharmacology, 3(4), 417–423. [Google Scholar]
  42. Sayette MA, & Tiffany ST (2013). Peak provoked craving: an alternative to smoking cue-reactivity. Addiction, 108(6), 1019–1025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sayette MA, Cohn JF, Wertz JM, Perrott MA, & Parrott DJ (2001). A psychometric evaluation of the facial action coding system for assessing spontaneous expression. Journal of Nonverbal Behavior, 25(3), 167–185. [Google Scholar]
  44. Sayette MA, Creswell KG, Dimoff JD, Fairbairn CE, Cohn JF, Heckman BW, … & Moreland RL (2012). Alcohol and group formation: A multimodal investigation of the effects of alcohol on emotion and social bonding. Psychological Science, 23(8), 869–878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sayette MA, Wertz JM, Martin CS, Cohn JF, Perrott MA, & Hobel J (2003b). Effects of smoking opportunity on cue-elicited urge: A facial coding analysis. Experimental and Clinical Psychopharmacology, 11(3), 218–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sayette MA, Martin CS, Hull JG, Wertz JM, & Perrott MA (2003a). The effects of nicotine deprivation on craving response covariation in smokers. Journal of Abnormal Psychology, 112, 110–118. [PMC free article] [PubMed] [Google Scholar]
  47. Shiffman S, Waters AJ, & Hickcox M (2004). The nicotine dependence syndrome scale: a multidimensional measure of nicotine dependence. Nicotine & Tobacco Research, 6(2), 327–348. [DOI] [PubMed] [Google Scholar]
  48. Simons-Morton B, & Farhat T (2010). Recent findings on peer group influences on adolescent smoking. Journal of Primary Prevention, 31(4), 191–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. US Department of Health and Human Services (USDHHS). (2014). The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta, GA: USDHHS. [Google Scholar]
  50. Watkins SS, Koob GF, & Markou A (2000). Neural mechanisms underlying nicotine addiction: acute positive reinforcement and withdrawal. Nicotine & Tobacco Research, 2(1), 19–37. [DOI] [PubMed] [Google Scholar]
  51. Watson D, Clark LA, & Tellegen A (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070. [DOI] [PubMed] [Google Scholar]
  52. Watts TW, Duncan GJ, & Quan H (2018). Revisiting the marshmallow test: A conceptual replication investigating links between early delay of gratification and later outcomes. Psychological Science. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wertz JM, & Sayette MA (2001). A review of the effects of perceived drug use opportunity on self-reported urge. Experimental and Clinical Psychopharmacology, 9(1), 3–13. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES