Appetite 56 (2011) 747–752

Contents lists available at ScienceDirect

Appetite
journal homepage: www.elsevier.com/locate/appet

Research report

Social matching of food intake and the need for social acceptance
Eric Robinson *, Talia Tobias, Lucy Shaw, Elyse Freeman, Suzanne Higgs
School of Psychology, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK

A R T I C L E I N F O A B S T R A C T

Article history: In a social eating context individuals tend to match their food intake to that of eating companions,
Received 14 December 2010 regardless of whether the other person is eating a large amount. However, relatively little is known about
Received in revised form 1 March 2011 the underlying processes of food intake matching. Findings from behavioural mimicry research suggest
Accepted 3 March 2011
that individuals may copy how those around them act in order to facilitate social interactions and
Available online 10 March 2011
ingratiate themselves. The present paper reports two studies which were designed to examine whether
ingratiation strategies may in part explain social matching of food intake in young females. In Study 1,
Keywords:
female dyads completed a problem solving task together whilst having access to chocolate M&M’s. We
Social modeling
Food intake
hypothesised that the extent to which individuals have a need to be socially accepted (trait self esteem)
Self-esteem and are competent in social interactions (trait empathy) would predict the degree of matching. In Study 2
Social acceptance we directly manipulated the desire to ingratiate by priming social acceptance in half of participants prior
to eating popcorn in the presence of a high eating confederate. In Study 1, both self esteem and empathy
were associated with degree of matching within female dyads. In Study 2, priming social acceptance
reduced the matching effect in females. These findings suggest that desire for social acceptance may be
an underlying cause of social matching of food intake.
ß 2011 Elsevier Ltd. All rights reserved.

Introduction absence of obvious satiety signals, individuals look outwards to

social cues to establish appropriate amounts to eat. However,
The amount of food consumed in a social eating context can be matching has been shown to occur even when there are strong
strongly influenced by the amount eaten by fellow dining signals regarding nutrient status, for example when participants
companions, such that individuals tend to eat more when others have been starved for 24 h before being exposed to a low eating
eat more and less when others eat less (Herman, Roth, & Polivy, confederate (Goldman, Herman, & Polivy, 1991). This suggests that
2003). Social matching of food intake is a pervasive and well people may match the intake of others around them not only to
recognised influence on eating behaviour. The effect has been resolve of uncertainty about what is appropriate to eat in the
replicated in studies that have employed a confederate as the absence of strong hunger-related signals but also for other reasons
eating companion as well as natural dyads (Herman, Koenig- (Herman et al., 2003).
Nobert, Peterson, & Polivy, 2005; Rosenthal & McSweeney, 1979) One reason why people adhere to social norms and match the
and it occurs for consumption of different foods (Hermans, Larsen, behaviour of others is to gain social approval. Humans have a strong
Herman, & Engels, 2009; Pliner & Mann, 2004; Rosenthal & Marx, desire to be liked by others (Baumeister & Leary, 1995) and there is
1979). As many meals are eaten in a social context, matching may evidence that mimicking the behaviour of others is a strategy by
have relevance to current global obesity concerns because such which this is achieved (Chartrand & Bargh, 1999; Chartrand & van
effects could theoretically have long term effects on energy intake. Baaren, 2009). Behavioural mimicry has been shown to facilitate
In a similar vein, recent findings have also started to tie down links social interaction, possibly because mimicry of behaviour increases
between weight gain and the behaviour of those around us, by perceived similarity (Chartrand & van Baaren, 2009). For example,
showing that obesity tends to spread across social networks Lakin and Chartrand (2003) informed participants it was important
through family members and friends (Christakis & Fowler, 2007). to get along well with a confederate and this resulted in participants
Although matching effects in relation to eating behaviour have showing a greater degree of behavioural mimicry (in the form of face
been well documented, the underlying mechanisms are less clear. touching). Hence it is possible that individuals match their food
Herman et al. (2003) suggest a normative model, whereby in the intake to a companion to ingratiate themselves with that person and
ease social interaction (Hermans, Engels, Larsen, & Herman, 2009;
Salvy, Jarrin, Paluch, Irfan, & Pliner, 2007). Interestingly, Hermans,
* Corresponding author. Engels et al. (2009) reported that participants only showed a
E-mail address: [email protected] (E. Robinson). matching effect when exposed to a socially ‘cold’ confederate and

0195-6663/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.appet.2011.03.001
748 E. Robinson et al. / Appetite 56 (2011) 747–752

not when exposed to a friendly confederate. The authors explained recruited in exchange for course credit (mean age = 19.8 yrs old,
their findings by suggesting that exposure to a socially ‘cold’ s.d = 1.4). BMI was within the healthy range; 23.0 (s.d = 3.4).
confederate may have caused participants to try and ingratiate Participants gave informed signed consent and the study protocol
themselves with their eating companion through matching. Perhaps was approved by University of Birmingham Research Ethics
what is more interesting is that when in the company of the ‘warm’ Committee.
confederate there was no matching which may have been because
participants no longer felt a need to ingratiate themselves with the Measures
confederate, although such an interpretation is speculative. To measure trait self esteem participants completed the
If social matching of food intake is influenced by the need for Rosenberg self esteem scale (1965).
social approval then we would expect differences in the degree to The scale consists of 10 items answered on a 4 point likert
which individuals feel the need to for affiliation to be related to response format; 1 = strongly agree, 4 = strongly disagree (Cron-
matching. In support, it has been reported that priming the need for bach’s a = 0.88). To measure trait empathy participants completed
affiliation increases behavioural mimicry (Lakin & Chartrand, 2003). the Interpersonal Reactivity Index (IRI; Davis, 1980). The scale
In addition, Caudill and Kong (2001) report that individuals with a consists of 28 items answered on a 7 point likert response format
high need for social acceptance show a greater degree of matching (1 = not at all true, 7 = totally true), with questions across four
when drinking alcohol with a confederate. An individual trait linked factors; ‘perspective taking’ (a = 0.71–0.75), ‘emotional distress’
to need for affiliation is empathy. Empathic individuals are thought (a = 0.75–0.77), ‘empathic concern’ (Cronbach’s a = 0.68–0.73)
use behavioural mimicry to increase liking and rapport (Chartrand & and ‘fantasy’ (Cronbach’s a = 0.78–0.79), which tap into the global
Bargh, 1999; Chartrand & van Baaren, 2009) and this helps them concept of empathy (Davis, 1980). The IRI has previously been
develop social bonds more easily than non empathic individuals shown to be strongly correlated with other measures of trait
(Davis, 1980, 1983). Thus, we might expect a link between empathy empathy (Davis, 1983). The two questionnaires were administered
and mimicking of eating behaviour because individuals high in after the social eating task. This was to prevent participants
empathy will use matching to ‘smooth’ social interactions in eating becoming aware that the study examined social interaction (the
situations. Another trait that has been linked to need for affiliation is empathy questionnaire includes questions concerning social
self-esteem (Leary & Baumeister, 2000; Pyszczynski, Greenberg, functioning). Weight and height were measured at the end of all
Solomon, Arndt, & Schimel, 2004). Leary and Baumeister (2000) sessions, using electronic digital scales and a stadiometer to
suggest that because self esteem is determined by perceptions of calculate BMI; body mass index (kg/m2).
belongingness, individuals with low self esteem should exhibit a
high need to affiliate with others (see also Baumeister & Leary, 1995). Procedure
In a social eating paradigm, if ingratiation is important, individuals Participants signed up for sessions online and were unaware of
with lower self esteem may be more likely to match their partner’s the identity of their dyad partner. Sessions took place between 9
intake in an attempt to facilitate social acceptance. am and 11 am or 2 pm and 5 pm on weekdays. Both participants
The present paper presents two studies designed to investigate were met in the reception of the lab facilities by the experimenter
whether ingratiation strategies explain the extent to which and accompanied to a room with a desk and two chairs that were
individuals match their intake to that of an eating companion. placed equal distances apart from each other to create a
In Study 1, we examined whether self esteem and empathy scores comfortable environment. Participants were asked to read an
predicted matching of food intake in natural dyads that had access information sheet about the study in which they were told that
to snack food whilst completing a problem solving task. Examining they would be completing a social interaction task and that they
whether self esteem and empathy were associated with matching should inform the experimenter of any food allergies, as snack food
of food intake enabled a preliminary test of our social ingratiation would be provided to make for a more naturalistic and comfortable
hypothesis by observing whether a need to ingratiate (as measured experience. After reading the information sheet and completing
by trait self esteem) and natural tendency to ingratiate (as demographics, participants completed a set of 6 mood ratings ‘how
measured by trait empathy) were associated with matching of food tired are you?’ as a cover for the aims of the study (5 point tick box
intake. In Study 2 we directly manipulated the need for social scale, anchors; ‘Not at all’’ and ‘‘Extremely’’).
ingratiation by priming social acceptance in half of our participants Participants were then informed that they would be playing a
prior to exposure to a high eating confederate. As the need for fun game together (‘hangman’). It was explained that this is a paper
social acceptance is thought to be a basic human need, temporary and pencil word guessing game, in which one player tries to work
feelings of social acceptance should saturate the need for social out a word by guessing individual letters one at a time. If a
acceptance and reduce the desire to ingratiate (Leary and participant had not played hangman before the experimenter
Baumeister, 2000; Carvallo & Pelham, 2006). We hypothesised provided a sheet which explained the rules of the game. Each
that if matching of food intake is influenced by the need for participant in the dyad was given a set word for hangman ‘Kanye
affiliation and ingratiation, then removing this need (by increasing West’ or ‘David Schwimmer’ and they were then informed that the
feelings of acceptance) should reduce matching. Across both experimenter would return in ten minutes. Two small bowls of
studies we decided to test only female participants as our sample M&M’s were left (one in front of each participant) in reaching
was taken from a largely female population (undergraduate distance only to that individual. Each bowl of M&M’s was pre-
psychology students). In addition, since there are effects of gender weighed and contained 30 M&M’s in each, so that the bowl was
on social eating (Salvy et al., 2007) we decided to test only females close to being full (chocolate M&Ms; 491 kcal per 100 g). Before
to avoid gender effects. leaving the experimenter casually told participants if they felt like
eating any they should feel free to do so.
Study 1 The experimenter returned after 10 min and removed the
hangman materials and bowls from the room (intake of each
Method participant was calculated by weighing the remaining M&M’s in
their bowl). Participants were then given the same mood ratings as
Participants described earlier and were asked to complete the empathy and self
One hundred and twenty two female participants (under- esteem questionnaires. They were then asked separately what they
graduate students from the University of Birmingham) were believed the purpose of the experiment was.
 E. Robinson et al. / Appetite 56 (2011) 747–752 749

Analysis In dyads that contained at least one individual with low self
esteem a strong and significant matching effect was observed
To examine overall intradyadic similarity (the degree of (r = 0.81). Dyads that consisted of two individuals categorised as
matching of intake within dyads) intraclass correlation coefficient being high in self esteem exhibited no significant matching
(ICC’s) were used. ICC’s were computed using a one way random (r = 0.20). The difference between the degree of matching between
model. To examine whether dyadic self esteem and empathy were these two groups was significantly different (Z = 3.04, p < 0.05).
associated with matching, for each dyad, the self esteem and Similarly, in dyads containing at least one individual with high
empathy scores were averaged (dyadic self esteem and dyadic empathy a strong and significant matching effect was observed
empathy) and entered into regression analysis to predict the degree (r = 0.76). However, in dyads that contained only individuals
of matching. Degree of matching was measured by the difference in categorised as being low in empathy, no significant matching effect
grams of M&Ms eaten by dyad partners. For example, if one dyad was observed (r = 0.14). The difference between the degree of
partner ate 5 g of M&Ms and another ate 8 g, there degree of matching between these two groups was significantly different
matching score would be 3. The smaller the degree of matching (Z = 2.69, p < 0.05). See Table 1.
score, the greater the similarity between the intake of dyad partners. As BMI has been shown to be of importance in the matching
Although the regression analysis is useful in determining food intake (Hermans, Larsen, Herman, & Engels, 2008), one
whether dyadic self esteem and empathy are associated with explanation of these findings could be differences in BMI. For
matching of food intake, it does not directly examine whether low example, it may be that BMI similarity between dyad partners may
and high scorers on trait self esteem or empathy match their differ in the low vs. high self esteem or empathy groups. However,
partner’s intake, or whether only high scorers are responsible for no evidence of this was found in the data set. For the overall
the matching effect observed in the overall sample. Therefore, we sample, BMI difference between dyad partners was 2.9 (s.d = 2.7).
used median splits of self esteem and empathy scores (score of 70 A one way ANOVA indicated no significant differences in dyad
on the empathy questionnaire, score of 20 on the self esteem partner BMI similarity across the four groups [F(3,98) = 0.79,
questionnaire) to determine ‘low’ and ‘high’ scorers on each p = 0.51). In addition, a small number of participants were
measure and divided dyads into 4 groups: dyads including at least classified as being overweight (BMI > 24.9). We examined the
one high scorer on empathy, dyads including only low scorers on effect of removing these participants, as well as removing dyads in
empathy, dyads including at least one low scorer on self esteem which one partner was overweight and one was of healthy weight
and dyads including only high scorers on self esteem. If need for (BMI = 18.5–24.9) but this did not affect the pattern of results.
ingratiation plays a role in matching, we hypothesised that the
matching effect would be smaller in dyads consisting of two Discussion
individuals with high self esteem (as there would be little desire to
ingratiate) and dyads consisting only of individuals with low As expected, a strong matching effect was observed in the
empathy (as neither would use matching to facilitate the social overall sample. In line with our hypotheses, both trait self esteem
interaction). and empathy were associated with a stronger matching effect,
with lower dyadic self esteem scores associated with a greater
Results degree of matching and higher dyadic empathy scores also
associated with a stronger matching effect. Intra-dyad correlations
One participant indicated that they had guessed the aims of the suggested that the matching effect in the overall sample may have
experiment so the data for that dyad was excluded from the final been strongly influenced by dyads which contained at least one
analysis. On average, participants consumed 14.4 g of M&M’s individual that was either low in self esteem or high in empathy.
(approximately 7 M&M’s). Thirty-six participants did not consume No significant matching effects were observed for dyads consisting
any M&M’s. Of these thirty-six, eleven dyads contained two non- only of individuals high in self esteem or low in empathy. However,
eaters. Responses on the empathy questionnaire ranged from 28 to only a small number of dyads contained two participants with high
196, mean = 69.1 (s.d = 11.1). Responses on the self esteem self esteem (n = 18) or two participants with low empathy (n = 16).
questionnaire ranged from 10 to 40 (mean = 19.2, s.d = 5.2). There Thus, some caution should be taken when interpreting the
was no significant association between self esteem and empathy difference in matching observed between those and dyads
scores (r = 0.05, p = 0.56). containing participants with low self esteem (n = 42) or high
empathy (n = 44), due to an imbalance of group size.
Overall matching effect These results provide some support for our hypothesis that
The overall degree of matching within samples was high, with social ingratiation attempts may be an important factor behind
an intradyadic correlation of M&M intake = 0.59 (df = 60, p < 0.05). social matching of food intake. However, as these results were
correlational it is not possible to infer causality. A more stringent
Self esteem & empathy test of the hypothesis that social matching may occur because of a
The regression model significantly predicted the degree of need to ingratiate oneself with others would be to directly
matching between dyad partners (adjusted R2 = 0.12, p < 0.05). manipulate the need to ingratiate. Thus, building on the findings of
Both dyadic self esteem (b = 0.27, p < 0.05) and dyadic empathy Study 1, in Study 2 we primed social acceptance by having half our
(b = 0.25, p < 0.05) uniquely predicted the degree of matching. participants complete a cross word containing words associated
The variance inflation factor for the two predictors was low; 1.0, with feelings of social acceptance before being exposed to a high
suggesting no problem of multi-collinearity in the data set. eating confederate. We hypothesised that this priming effect

Table 1
Intra-dyad intake correlations as a function of high and low for self esteem and empathy.

Dyad’s by self esteem Dyad’s by empathy

*
1 partner = low scorer in self esteem r = 0.81, (df = 42, p < 0.05) 1 partners = high scorer in empathy r = 0.76, (df = 44, p < 0.05)*
Both partners = high scorers in self esteem r = 0.20, (df = 18, p = 0.19) Both partners = low scorers in empathy r = 0.14, (df = 16, p = 0.39)
*
Indicates significant at p < 0.05. Intraclass correlations reported.
750 E. Robinson et al. / Appetite 56 (2011) 747–752

would temporarily reduce the need to ingratiate and result in with which they were to test the other with. The confederate and
reduced matching of food intake. participant took it in turns to guess the words. Before leaving, the
experimenter placed two bowls of popcorn; 30 pieces of popcorn in
Study 2 each (Butterkist toffee popcorn, 415 kcal per 100 g), one next to the
participant and one next to the confederate and informed the pair
Method that they could eat during the games if they felt like it. The
confederate was instructed to eat at a steady rate of 4 pieces of
Participants popcorn per game, resulting in a total of 16 pieces of popcorn being
Fifty female participants from the University of Birmingham consumed by the confederate. The experimenter returned after
were recruited in exchange for course credit (mean age = 19.1 - 10 min. If the games lasted less than 10 min the confederate was
years, s.d = 1.0). BMI was within healthy range; 21.1 (s.d = 2.6). The instructed to let the participant lead any conversation and to reply
study was advertised as research examining mood and social to any questions and ask the participant a similar question back.
problem solving. Participants gave informed signed consent and On completion of the game, the experimenter removed the
the study protocol was approved by University of Birmingham bowls of popcorn and asked the confederate to leave the room so
Research Ethics Committee. that the subsequent questionnaires could be completed alone. To
corroborate the cover story participants then completed the same
Experimental conditions mood questionnaire described earlier. They also completed the
Participants were assigned to one of two experimental cognitive restraint scale of the Three Factor Eating Questionnaire
conditions. To prime feelings of social acceptance we used a (Stunkard & Messick, 1985). Finally, participants were asked to
priming manipulation as reported by Carvallo and Pelham (2006). guess what the aims of the study were, before weight and height
In the ‘primed’ condition participants were asked to complete a were measured to calculate BMI (see Study 1). Intake was
word search task that contained 10 words related to social measured by weighing and then counting the remaining pieces
acceptance (i.e. accepted, belonging, approved). In the neutral of popcorn in the separate bowls.
condition participants were asked to complete the word search
task, but containing 10 neutral words that were of similar length Analysis
(i.e. location, beauty, table).
t-tests were used to examine whether the two experimental
Confederate groups were matched for hunger at the start of the session, BMI and
One female undergraduate research student acted as the cognitive restraint and whether the groups differed in popcorn
confederate in all experimental sessions. She was selected due eaten. To further examine the degree of matching between
to her similarity to our sampled population; aged 20, BMI = 19.5 participant intake and confederate intake, a matching index was
(height 175 cm). The confederate was instructed to behave used, whereby a score was generated for each participant by
similarly for all participants (see procedure for instructions subtracting 16 (amount of popcorn eaten by confederate) from the
concerning conversational style during sessions). The confederate number of pieces of popcorn they ate. A score of 0 indicates that
was instructed to eat 16 pieces of popcorn during the experiment. participants matched the confederate perfectly, a minus score
This amount was chosen based on a pilot study with a separate indicates that participants consumed less than the confederate and
group of participants that found 16 pieces to be more than a plus score indicates participants consumed more than the
participants imagined eating in that setting. The confederate was confederate.
unaware of which experimental condition each participant was
assigned to and intake checks indicated that she ate 16 pieces of Results
popcorn in all sessions.
Three participants guessed the aims of the study and were
Procedure removed from data analyses. On average, participants consumed
Sessions took place between 9 am and 11 am and 2 pm and 5 pm 11.3 g (s.d = 7.9) of popcorn (approximately 14 pieces). Four
on weekdays. It was ensured that as participants arrived for the participants did not eat any of the popcorn.
study, the confederate was already waiting in the meeting room.
Shortly after, both the confederate and the participant were Intake and matching by experimental condition
greeted by the experimenter and were taken to the laboratory. The Both groups were matched for baseline hunger [t(45) = 0.8,
confederate and participant were seated at opposite ends of a small p = 0.4], BMI [t(45) = 0.3, p = 0.97] and restraint [t(45) = 0.6,
table before being asked to complete demographic questionnaires p = 0.58]. See Table 2. Participants in the neutral condition
and a dummy mood questionnaire, to corroborate the cover story. consumed 14.0 g of popcorn (s.d = 7.1), whereas participants in
Mood and appetite items (happy, hungry, tired, anxious and alert) the primed condition consumed significantly less; 9.1 g of popcorn
were rated using a 10 cm visual analogue line rating scale (V.A.S) (s.d = 8.1) [t(45) = 2.2, p < 0.05]. In line with the intake data,
with ‘‘Not at all’’ and ‘‘Extremely’’ as end anchors and the question participants in the primed condition exhibited a significantly
‘‘How... do you feel right now?’’ (centred above the line scale). The
experimenter then returned and instructed participants that for
the next part of the experiment they were each required to Table 2
complete a word search task in separate rooms. The experimenter Participant characteristics and food intake by experimental condition.
then handed both the participant and confederate separate word
Primed condition Neutral condition
search tasks before instructing the confederate to follow her to (n = 25) (n = 22)
another room. Participants were left alone for 5 min to complete
BMI 21.1 (2.7) 21.1 (2.4)
the experimental manipulation tasks then the experimenter Hunger (0–10 cm scale) 4.8 (2.4) 4.2 (2.8)
returned with the confederate and explained to both the Restraint (0–21 scale) 7.8 (5.3) 7.0 (5.4)
confederate and participant that they would be required to Degree of matching to confederate 5.32 (8.4)* 0.59 (9.6)*
complete 4 games of ‘hangman’ together. The experimenter *
Indicates significant difference between primed and neutral condition
presented both the confederate and participant with two words, (standard deviations in brackets).
 E. Robinson et al. / Appetite 56 (2011) 747–752 751

weaker matching effect than participants in the control condition We provide the first evidence that individual differences (self
[t(45) = 2.2, p < 0.05]. See Table 2. Although our groups did not esteem and empathy) moderate social matching. Identification of
differ in baseline hunger, BMI or restraint, further analysis was individual differences in matching of eating behaviour is of
undertaken to examine whether these factors moderated degree of interest because up to now the matching effect has appeared
matching across our two groups but no evidence was found (data impervious to attempts to find individual difference moderators
not reported). (Herman et al., 2005). Herman et al. (2005) found no evidence that
individual differences in self-monitoring or extraversion affected
Discussion matching.
We in no way believe that the need for social acceptance is the
After priming feelings of social acceptance, participants in only underlying cause of matching effects in social eating. Roth,
Study 2 exhibited a smaller matching effect than participants in Herman, Polivy, & Pilner (2001) report a study in which intake
the neutral condition. These results suggest that ingratiation matching was observed due to knowledge of what a remote
attempts may be an underlying process behind social matching confederate had eaten and so need for social affiliation cannot
of food intake. One weakness of Study 2 is that we did not have explain matching in that context. Furthermore, Hermans et al.
a manipulation check. Unfortunately, the nature of the experi- (2008) recently reported data that suggests individuals are more
ment meant that introducing a manipulation check in between inclined to model the intake of a confederate of similar body
the priming manipulation and eating task may have influenced weight. Both of these findings clearly underline that social norms
participant behaviour or made the aims of the study transparent. are important in explaining matching of food intake (Herman et al.,
Furthermore, as priming effects are short term in their duration, 2005). In addition, the importance of personality factors in
a manipulation check after the 10 min eating period may not explaining variance in food intake matching is unclear since we
have detected difference between the groups. Therefore, we found that self esteem and empathy only accounted for a small
cannot be entirely sure that the manipulation directly affected amount of variance in matching.
the need to affiliate and subsequently influenced matching. A further issue is the extent to which situational factors may
Although it should be noted we were using a method to prime influence the relationship between social ingratiation and match-
social acceptance that has been shown to be effective in previous ing of food intake. Presumably, social ingratiation is of particular
studies (Carvallo & Pelham, 2006). However, we were able to relevance when eating with strangers, but whether ingratiation
probe participants to see if any were aware of the study aims and concerns are great enough amongst familiar others to be a cause of
no participants picked up on the priming manipulation matching is unclear. If ingratiation attempts are important then we
(although 3 participants were removed from analyses as they might expect to see stronger matching effects in unfamiliar versus
mentioned the study may have been about social interactions familiar groups. Salvy et al. (2007) found that dyads composed of
and eating). unfamiliar participants did not produce a statistically significant
Another issue is that matching of eating was assessed during a greater matching effect than dyads composed of ‘friends’.
game which some participants may have played competitively However, as the authors recognised, one suggestion why this
(even though it was introduced as a ‘fun game to play together’). difference was not significant is because no measure of the degree
Further research testing the effect social ingratiation has on of interpersonal bond or ‘closeness’ between friends is reported in
matching in more natural settings is advised. The same person the study.
acted as the confederate for all participants in Study 2. The Many studies of social matching have used single sex, usually
confederate was matched for age and BMI to the participants. To all female, samples. However, Salvy et al. (2007) examined sex
ensure that the results are not specific to this confederate it would differences and found no evidence of intake matching in male
be necessary to replicate using more than one confederate. We did dyads. Additionally, Hermans, Herman, Larsen, and Engels
not include a control condition in which participants ate alone, so (2010) reported that sated males did not show a matching
although our social acceptance prime significantly reduced the effect. This pattern of results may be due to women possessing
matching effect, we cannot comment on whether it removed it greater interests in facilitating positive social bonds than men
completely. It would therefore be interesting in future studies to although this remains to be investigated (Eagly, 1978; see also
examine differences between a control eating alone condition and Salvy et al., 2007). Another interesting question is the extent to
a primed condition. which social matching of food intake is occurring unconsciously.
Much of the literature that links behavioural mimicry to
General discussion ingratiation suggests such processes are automatic and uncon-
scious, whereby the action of one person triggers a close to
It has been speculated that ingratiation attempts may underlie automatic mimicked response by the other person (Chartrand &
food intake matching effects (Herman et al., 2003; Salvy et al., van Baaren, 2009). Whether this applies to social matching of
2007). The present studies examined this proposition. In Study 1, a food intake is an intriguing question. Does the motor movement
measurement of desire for ingratiation (trait self esteem) was of a nearby individual reaching for food trigger a similar
associated with food intake matching in female dyads, as was trait response in their eating partner? Studies to date suggest that
empathy. In Study 2, we found that reducing the need to ingratiate participants’ are not aware that their intake was influenced by
with others, by priming feelings of social acceptance, reduced the the intake of their eating partner (Vartanian, Herman, &
degree of intake matching to a high eating confederate. The results Wansink, 2008).
of Study 1 and Study 2 are in line with other results suggesting In conclusion, the present studies provide preliminary evidence
behavioural mimicry occurs to facilitate social interactions and for the role of social acceptance concerns in the matching of food
ingratiation (Chartrand & van Baaren, 2009; Lakin & Chartrand, intake amongst female eating partners. Study 1 suggested that
2003). Our present findings also are in line with those of Hermans, personality variables associated with social ingratiation were
Engels et al. (2009) who reported that no matching effect was related to matching of intake. Study 2 suggested that matching of
observed when participants were exposed to a socially warm food intake can be reduced by priming feelings of social
participant. It is possible that participants in the socially warm acceptance. Further work examining underlying mechanisms
condition did not feel a need for ingratiation because this had and whether social acceptance concerns moderate matching of
already been achieved. intake amongst familiar others, as well as strangers, is advised.
752 E. Robinson et al. / Appetite 56 (2011) 747–752

References Hermans, R. C. J., Engels, R. C. M. E. , Larsen, J. K., & Herman, P. C. (2009). Modeling of
palatable food intake. The influence of quality of social interaction. Appetite, 52,
Baumeister, R. F., & Leary, M. R. (1995). The need to belong. Desire for interpersonal 801–804.
attachments as a fundamental human motivation. Psychological Bulletin, 117, 497– Hermans, R. C. J., Larsen, J. K., Herman, P. C., & Engels, R. C. M. E. (2009). Effects of social
529. modeling on young women’s nutrient-dense food intake. Appetite, 53, 135–138.
Carvallo, M., & Pelham, B. W. (2006). When fiends become friends. The need to belong Hermans, R. C. J., Larsen, J. K., Herman, P. C., & Engels, C. M. E. (2008). Modeling of
and perceptions of personal and group discrimination. Journal of Personality & palatable food intake in female young adults. Effects of perceived body size.
Social Psychology, 90, 94–108. Appetite, 51, 512–518.
Caudill, B. D., & Kong, F. H. (2001). Social approval and facilitation in predicting Lakin, J. L., & Chartrand, T. L. (2003). Using nonconscious behavioral mimicry to create
modeling effects in alcohol consumption. Journal of Substance Abuse, 13, 425– affiliation and rapport. Psychological Science, 14, 334–339.
441. Leary, M. R., & Baumeister, B. F. (2000). The nature and function of self esteem.
Chartrand, T. L., & Bargh, J. A. (1999). The Chameleon effect. The perception-behavior Sociometer theory. Advances in Experimental Social Psychology, 32, 1–62.
link and social interaction. Journal of Personality and Social Psychology, 76, 893–910. Pliner, P., & Mann, N. (2004). Influence of social norms and palatability on amount
Chartrand, T. L., & van Baaren, R. (2009). Human mimicry. Advances in Experimental consumed and food choice. Appetite, 42, 227–237.
Social Psychology, 41, 219–274. Pyszczynski, T., Greenberg, J., Solomon, S., Arndt, J., & Schimel, J. (2004). Why do people
Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network need self-esteem? A theoretical and empirical review. Psychological Bulletin, 130,
over 32 years. New England Journal of Medicine, 357, 370–379. 435–468.
Davis, M. H. (1980). A multidimensional approach to individual differences in empathy. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
JSAS Catalog of Selected Documents in Psychology, 10, 85. University Press.
Davis, M. H. (1983). Measuring individual differences in empathy. Evidence for a Rosenthal, B., & McSweeney, F. K. (1979). Modeling influences on eating behaviour.
multidimensional approach. Journal of Personality & Social Psychology, 44, 113– Addictive Behaviors, 4, 205–214.
126. Rosenthal, B., & Marx, R. D. (1979). Modeling influences on the eating behavior of
Eagly, A. H. (1978). Sex differences in influence ability. Psychological Bulletin, 85, 86– successful and unsuccessful dieters and untreated normal weight individuals.
116. Addictive Behaviors, 4, 215–221.
Goldman, S. J., Herman, C. P., & Polivy, J. (1991). Is the effect of a social model attenuated Roth, D., Herman, C. P., Polivy, J., & Pilner, P. (2001). Self-presentational conflict in social
by hunger? Appetite, 17, 129–140. eating situations. A normative perspective. Appetite, 36, 165–171.
Herman, C. P., Roth, D. A., & Polivy, J. (2003). Effects of the presence of others on food Salvy, S. J., Jarrin, D., Paluch, R., Irfan, N., & Pliner, P. (2007). Effects of social influence on
intake. A normative interpretation. Psychological Bulletin, 129, 873–886. eating in couples, friends and strangers. Appetite, 49, 92–99.
Herman, C. P., Koenig-Nobert, S., Peterson, J. B., & Polivy, J. (2005). Matching effects on Stunkard, A. J., & Messick, S. (1985). The three factor eating questionnaire to measure.
eating. Do individual differences make a difference? Appetite, 45, 108–109. Dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29,
Hermans, R. C. J., Herman, P. C., Larsen, J. K., Herman, R. C., & Engels, M. E. (2010). Social 71–83.
modeling effects on snack intake among young men. The role of hunger. Appetite, Vartanian, L. R., Herman, P. C., & Wansink, B. (2008). Are we aware of the external
54, 378–383. factors that influence our food intake? Healthy Psychology, 27, 533–538.