The Influence of Mere Social Presence on Message Reception
Stefano Puntoni*
and
Nader T. Tavassoli
Centre for Marketing Working Paper
No. 03-204
Stefano Puntoni is a doctoral candidate in the Marketing PhD Programme, London Business School.
Nader Tavassoli is Associate Professor of Marketing at London Business School.
London Business School, Regent's Park, London NW1 4SA, U.K.
Tel: +44 (0)20 7262-5050 Fax: +44 (0)20 7724-1145
http://www.london.edu/Marketing
Copyright © London Business School 2000
*This research is part of the first author’s dissertation and was partly funded by the
Economic and Social Research Council of the U.K. (R42200134387) and the Centre
of Marketing at London Business School. The authors would like to thank John Lynch
and Marc Vanhuele for their insightful comments and suggestions.
The article presents three experiments that investigate the consequences of social context–seeing an ad
alone or in company of another person–for message reception. We propose that people are intrinsically
motivated to make a good impression on others in social situations and that priming the goal of social
desirability can affect cognitive activity. Using a lexical decision task, two experiments demonstrate
that the mere presence of a person of the opposite gender increases goal accessibility. Participants were
relatively faster at identifying words related to social desirability in the presence of another person, but
they were relatively faster at identifying words neutral with respect to this goal when alone. This
difference in the accessibility of the goal of social desirability had important implications for recall
memory, which mirrored the relative shift in reaction times. A third experiment generalized the
memory findings to an advertising setting.
Keywords: advertising, goal priming, social desirability, automatic processes.
Advertisers have little control over the specific circumstance in which a
consumer comes into contact with their message. Ads can be consumed in isolation or
in the presence of other individuals. People who are not living alone typically view
television commercials in the company of others. People are also not alone when
being exposed to outdoor advertising, to ads in movie theaters, on airplanes, in bars,
or when exposed to online ads in an Internet café. Yet, to our knowledge, no previous
research has examined the influence that the presence of others at the time of
exposure may exert on message reception. We present and empirically test a
framework to explain one mechanism by which the mere presence of another
individual–in the absence of direct interaction–can alter information processing.
Our framework is based on the premise that the presence of another person
raises a concern about the impression that the person may be forming of us.
Specifically, we hypothesize that the presence of another person will prime the goal
of social desirability. The increased accessibility of this goal should affect the
processing of, and memory for, stimuli that have high applicability to the goal. Our
research, therefore, diverges from previous research on social desirability that has
examined overt response biases. Instead, we view social desirability as reflecting a
more pervasive motivational determinant of individual behavior, even if this behavior
is purely cognitive and entirely private. In particular, we focus on the consequences of
social context for memory. Memory is important not only as a phenomenon in itself,
but also as a component in virtually all thinking: It is embedded in complex
behavioral representations such as thoughts, emotions, judgments, intentions, and
choices. Memory is also a key measure of advertising effectiveness offered by leading
testing agencies such as The Burke Institute, Gallup & Robinson, and Millward
Brown.
4
THEORETICAL BACKGROUND
There has been a growing body of consumer research on the influence of
social context. This includes research on embarrassment during purchase (Dahl,
Manchanda, and Argo 2001), emotional contagion on product attitudes (Howard and
Gengler 2001), the impact of public consumption on variety seeking (Ratner and
Khan 2002), gender differences in emotional responding (Fisher and Dubé in press),
and sequential choice in group settings (Ariely and Levav 2000). Research has also
examined the role of advertising in social interactions, showing that consumers
instrumentally use advertising messages to communicate specific ideas or to highlight
group structure (Ritson and Elliott 1999). However, research has neglected to
examine how the presence of other people affects message reception in the first place
(cf. Ritson and Elliott 1999). Paradoxically, research in social psychology mirrors this
under-investigation of the effect of social context. Eagly and Chaiken (1993) argue
that “psychologists’ paradigm for studying attitudes have become decreasingly
attentive to social context over the years … This increasingly psychological focus is
not unusual for social psychology” (p. 628).
Previous research has suggested that the presence of others can affect persons’
cognitions even when these are private and not observed as part of a social interaction
(Huguet et al. 1999). For example, research has shown that women’s self-evaluations
are diminished by the presence of a man during exposure to ads featuring attractive
female models (Henderson-King, Henderson-King, and Hoffman 2001). This finding
demonstrates that advertising and social context can interact to influence cognitions
about the self, an internal focus. We propose that social context can also influence the
processing of ads themselves, an external focus. This is not only an important
theoretical issue but a practical concern as well: Little attention is paid to the social
5
context in which ads are tested vis-à-vis the social context in which they are likely to
be consumed.
Motivational Consequences of Mere Social Presence
We propose that one effect of social context on message reception is
motivational in nature. As demonstrated by the large body of literature on impression
management since Goffman (1959), individuals are motivated to make a good
impression on others. We thus broadly define social desirability as a concern with the
impression others are forming of us. Social desirability is a very general goal and can
manifest itself in a variety of behaviors, such as how we attend at our environment
(Emmons 1989). People scan the social environment for information regarding how
others regard them most, if not all, of the time, even though often at a nonconscious
level (Leary and Kowalski 1990).
Social desirability concerns have been shown to affect overt behaviors such as
eating (for a review, see Herman, Roth, and Polivy 2003). For example, Mori,
Chaiken, and Pliner (1987) found that the presence of a male confederate caused
women to eat less. The authors provide a motivational explanation for this effect.
They argue that due to our culture’s propensity to equate women’s attractiveness with
thinness participants ate less to appear more feminine and thus more socially desirable
in a social context. The presence of others can affect cognitive activity even when
individuals’ responses remain unobserved and there is no opportunity for interaction
(for a review, see Levine, Resnick, and Higgins 1993). For example, the results of the
Implicit Association Test (IAT)–a measure of automatic attitudes–have been found to
be sensitive to social context (Lowery, Hardin, and Sinclair 2001). Participants scored
lower on racial bias on the IAT in the presence of a black experimenter than in the
presence of a white experimenter. The authors offer a motivational explanation for
6
this result, conceptualizing automatic social tuning as a consequence of a concern
with interpersonal relationships (Lowery et al. 2001). We similarly examine how
social context can affect behaviors that are outside of public scrutiny through a
motivational route.
Goal Priming, Goal Accessibility, and Processing Efficiency
Contextual differences tend to cause differences in the accessibility of
knowledge constructs (Higgins and King 1981), where accessibility is moderated by
the applicability of an attended stimulus to stored knowledge (Higgins 1996). Among
the knowledge structures that can be activated by contextual cues are goals and
motives (Bargh 1990). We propose that social context–the mere presence of another
person–can trigger a concern with the impression we make on others. More formally,
H1:
The mere presence of another person will automatically activate the
goal of social desirability.
Of particular importance to the literature on goal priming has been goals’
accessibility, or how readily they come to mind in different situations (Shah 2003).
We therefore test hypothesis 1 by measuring the accessibility of the goal of social
desirability when participants are alone or in the presence of another person. To
measure accessibility of a psychological construct the method most commonly
employed is to assess its processing efficiency, operationalized in terms of response
latencies (Bargh and Chartrand 2000; Fazio 1990). In particular, goal accessibility has
been often operationalized in terms of reaction time during a lexical decision task in
which participants are asked to judge as quickly as possible whether a word is real of
invented (e.g., Mikulincer 1998; Shah 2003). If social context activates the goal of
social desirability then there should be a decrease in response latencies for cues
applicable to the goal due to an increased processing efficiency of these cues.
7
Memory Consequences of Mere Social Presence
An important consequence of the activation of the goal of social desirability is
that social context should affect message reception. The relationship between a to-belearned item and an operating goal is an important determinant of memory for that
item (Wyer and Srull 1989). Such goal relevance has been shown to moderate
information acquisition, especially for novice consumers (Huffman and Houston
1993). Huffman and Houston gave consumers the explicit goal of selecting a product
for a fictitious person who had a specific consumption need. They found that features
that were applicable to the stated goal were acquired more efficiently and more likely
to be recalled than features that were irrelevant to the stated goal. This can be
explained by an activated goal directing attention to goal-applicable information
(Moskowitz 2002) and serving to create associations in memory between the goal,
applicable product features, and available brands (Huffman and Houston 1993). In
other words, recall of goal-applicable information can be improved both from a
heightened level of activation of a concept in memory as well as via connections
between a learned concept and other concepts.
Research on the relationship between involvement and attention has come to
similar conclusions. For example, Celsi and Olson (1988) suggest that the
motivational state of felt involvement–a feeling of personal relevance–affects both the
amount and focus of attention exerted with respect to an advertising message.
Messages that were seen as personally relevant were processed longer and resulted in
the generation of more thoughts, especially of product-related ones.
To summarize, the activation of a goal should result in greater attention paid
to items with high applicability to the goal than to items with low applicability to the
goal. This should affect the probability that these items are committed to long-term
8
memory, the strength of activation of these items in memory, and the associations
between these items and the goal, goal-applicable situational cues, and goal-relevant
knowledge structures. These encoding effects, in turn, should increase the likelihood
that an item is subsequently remembered. If the mere presence of another person
primes the goal of social desirability, our framework predicts that the mere presence
of another person will enhance the encoding of items applicable to the goal of social
desirability but attenuate the encoding of items that are not applicable to this goal.
Social context should not only affect encoding of information but also its
retrieval, even when a person is no longer in the presence of another individual.
Consider free recall. Recall is based on a search for information that is guided by
competing retrieval cues (Burgess and Hitch 1992). One of these retrieval cues is the
association between the to-be-remembered item and the context it was encountered in
(Gillund and Shiffrin 1984). When a presented stimulus is applicable to the goal
primed by social context, this should strengthen context-item associations at
encoding. Similarly, social context should strengthen associations between items
applicable to the goal of social desirability and goal-applicable knowledge structures
(Greenwald and Leavitt 1984). Because the social context at exposure and activated
knowledge structures are part of the memory representation, these are more potent
internal retrieval cues for items with high applicability to the goal of social
desirability than for items with low applicability to the goal. In other words, because
recall is determined by a “competitive filter” that selects the most accessible items
from memory (Burgess and Hitch 1992), items applicable to social desirability benefit
from contextual and self-generated retrieval cues available in memory during a recall
attempt. More formally, encoding and retrieval factors jointly predict that,
9
H2:
The mere presence of another person will enhance memory for items
with
high
applicability to social desirability, relative to items with low
applicability to the goal.
EXPERIMENT 1
The purpose of this study is to assess whether mere social presence increases
the accessibility of the goal of social desirability. Participants were provided the focal
goal of performing a lexical decision task in which words applicable to social
desirability and neutral words were presented along with nonsensical words. After a
filler task, recall and recognition memory for the real words presented earlier were
assessed. We hypothesized that the mere presence of a confederate should prime the
background goal of social desirability. This should increase the level of accessibility
of, and enhance memory for, cues applicable to social desirability relative to cues
neutral with respect to this goal.
Method
Design and participants. Experiment 1 used a mixed 2 (viewing context:
social vs. alone) x 2 (stimulus type: applicable to social desirability vs. neutral)
design. Viewing context was manipulated between-subjects and stimulus type withinsubjects. Sixty-three female students (Mage = 21.2, SD = 3.97) participated in the
experiment for a monetary reward. Data from two participants were eliminated, one
because the fire alarm interrupted the study and one because the participant’s mobile
phone rang during the reaction time task.
Procedure. Participants were randomly assigned to conditions. They were met
in a waiting room by a male experimenter who led them to the room where the study
took place. A male confederate in his mid-twenties, who was already in the room, was
10
introduced to participants in the social condition as a research assistant who was to
observe them performing the task in order to learn about the experimental procedure.
The confederate was seated one meter behind an empty chair reserved for the
participants and positioned in front of a computer screen. Participants in the alone
condition were left alone in the room for the entire duration of the study.
Participants were told to follow the instructions on the screen. During the first
task, participants were asked to judge as quickly and as accurately as possible whether
a series of target words were real or invented using two keys marked in green and red,
respectively. Each word was displayed for two seconds following a one second
fixation mark and reaction time was recorded in milliseconds. Duration of exposure
was therefore held constant across trials and participants, regardless of speed of
response. After 10 practice trials using words that were neutral with respect to social
desirability, 32 words and 32 nonwords (e.g., “hybac”) were displayed in random
order.
After participants completed the lexical decision task the confederate left the
room. The confederate had been instructed by the experimenter in the presence of the
participants to go back to his office once the first part of the study had been
completed. Thus, participants in all conditions completed the remainder of the study
alone. After the confederate left, an unrelated task of about seven minutes was
administered to clear short-term memory. Next, free recall and recognition tests were
administered. For free recall, participants had two minutes to list all the real words
that they could remember having seen during the lexical decision task. Immediately
following, the computer introduced the recognition test. Sixty-four words (the original
32 real words and the same number of matched distractors) were displayed one at a
11
time and participants were asked to indicate if they had seen the word during the
lexical decision task or not.
At the end of the study, participants were probed about the purpose of the
study and those in the social condition were asked about the presence of the
confederate. None of the participants guessed either the purpose of the study or the
role of the confederate.
Stimuli. Of the 32 real words, 16 were applicable to social desirability and 16
were unrelated, or neutral with respect to this goal (e.g., “shelf”). Of the 16 related
words, eight were positively related (e.g., “smart”) and eight were negatively related
(e.g., “ugly”).
To select the stimuli, 17 women (Mage = 28.3, SD = 5.4) from the same pool of
participants partook in a pre-test in return for a pack of candies. They were provided
with 90 words (adjectives, nouns and adverbs) that were four to six letters long and
were asked to imagine each word used in a statement describing a women. They then
indicated how such a statement would reflect on the impression that she would make
on a man who doesn’t know her using a 7-point scale (1 = “very bad impression”; 7 =
“very good impression”).
Three sets of words were created from this pretest. One set contained words
that were negatively related to social desirability based on scoring below the first
quartile; one contained neutral words that scored between the first and the third
quartile; and one contained words that were positively related to social desirability
based on scoring above the third quartile. Stimuli were drawn from these sets (eight
from the first, 16 from the second, and eight from the third). The final selection was
made considering extremity of scores, avoiding words with similar lexical properties,
and balancing across sets the frequency of grammatical elements. Examples of neutral
12
stimuli selected are: “mail”, “twice”, “indoor”, and “retail” (Mneu = 3.96). Examples
of stimuli positively related to social desirability are: “charm”, “beauty”, “cute”, and
“smart” (Mpos = 6.21). Examples of stimuli negatively related to social desirability
are: “ugly”, “acne”, “nasty”, and “inept” (Mneg = 1.61). The paired t-tests among each
of the sets of words were significant (ps < .0001).
Results
Response latencies. Only response latencies for correct responses were
analyzed because latencies for incorrect responses are difficult to interpret in terms of
accessibility (Shah 2003). Furthermore, data were eliminated from six participants
who had exceptionally high error rates (Fazio 1990), based on a cut-off point of 10
mistakes across the 64 experimental trials. The average error rate for these
participants was 23% versus 5% for the participants included in the analysis. In all the
analyses reported below, we therefore used data from 55 participants (29 in the alone
condition and 26 in the social condition).
It is also important to control for heterogeneity across individuals in the highly
variable baseline reaction times (Fazio 1990). To account for heterogeneity in
participants’ speed of response we treated average speed of response as an
unobserved variable normally distributed across the population. We therefore used a
mixed model in which viewing context, stimulus type and the two-way interaction
term are treated as fixed effects whereas heterogeneity in speed of response is
modeled using a random effect:
yijk = µ + rij + αj + βk + (αβ)jk + εijk
where yijk is the response on the k-th stimulus type provided by the i-th
participant in the j-th viewing condition, µ is the grand mean, rij is a normally
distributed random effect corresponding to the i-th participant nested in the j-th
13
viewing condition, αj is the fixed effect for the between-subjects factor (viewing
context), βk is the fixed effect for the within-subjects factor (stimulus type), (αβ)jk
indicates the two-way interaction, and εijk is the error term. We estimated the model
using Generalized Least Squares (Hsiao 2003; Littell, Henry, and Ammerman 1998).
There was a main effect of viewing context, F(1, 53) = 4.07, p < .05. Social
context had an inhibitory effect such that participants in the alone condition were, on
average, faster than participants in the social condition (Malone = 579 ms vs. Msocial =
626 ms). The main effect of stimulus type was non significant. Qualifying these
results, was the interaction effect, F(1, 1642) = 4.07, p < .05. In the alone condition,
words applicable to social desirability (Mapplicable = 588) were processed more slowly
than neutral words (Mneutral = 570). In contrast, in the social condition, words
applicable to social desirability (Mapplicable = 622) were processed faster than neutral
words (Mneutral = 630). These results provide support for hypothesis 1 (see Figure 1a).
--------------------------------------Insert Figure 1 about here
--------------------------------------Memory. We performed a repeated measures ANOVA on the applicable
versus neutral words correctly recalled by the participants. There was a main effect of
stimulus type, F(1, 53) = 42.63, p < .0001. More words applicable to social
desirability were recalled than were neutral words (Mapplicable = 2.69 vs. Mneutral =
1.38). There was no main effect of viewing context. The interaction effect qualified
these results, F(1, 53) = 5.35, p < .05 (see Figure 1b). Participants remembered more
neutral words in the alone condition (for neutral words: Msocial = 1.15 vs. Malone =
1.59) but remembered more words applicable to social desirability in the social
condition (for applicable words: Msocial = 2.96 vs. Malone = 2.45). Hypothesis 2 was,
14
therefore, supported with respect to free recall. In contrast, hypothesis 2 was not
supported based on recognition accuracy (hit rates). In the alone condition hit rates
were 67% for neutral words and 82% for words applicable to social desirability. In the
social condition hit rates were 69% for neutral words and 83% for words applicable to
social desirability. The interaction was not significant, F < 1.
Discussion
The results of experiment 1 support hypothesis 1, that the presence of another
person activates a concern with the impression others may be forming of us. This
conclusion is based on the significant interaction between viewing context and
stimulus type for reaction time. Participants reacted relatively faster to words
applicable to social desirability when in the presence of another person, but they
reacted relatively faster to neutral words with low applicability to social desirability
when alone.
The increased accessibility of the goal of social desirability explains the
finding that participants recalled a greater proportion of words applicable to social
desirability in the social condition, even though the confederate was no longer present
during recall. The effect could be the result of a shift in covert attention towards
words applicable to social desirability at encoding–overt attention was always on the
lexical decision task–as well as based on retrieval. Recall is determined by a
“competitive filter” which selects the most accessible items from memory determined
by the relative strengths of retrieval cues established during encoding (Burgess and
Hitch 1992). Social context and pre-existing knowledge structures applicable to social
desirability are part of the associative memory trace for the words presented during
the lexical decision task. These can act as retrieval cues and increase the likelihood of
15
recalling words applicable to social desirability at the expense of recalling words that
are not applicable to social desirability.
We did not, however, find a parallel effect for recognition accuracy. Models of
recall and recognition generally pose that recall relies largely on a controlled process
of recollection, whereas recognition relies primarily on an automatic signal-detectionlike familiarity judgment and only secondarily on the process of recollection
(Yonelinas 2002). On the one hand, therefore, an effect on recall but not on
recognition can be interpreted as the activated goal having affected recollection–an
associative and meaning-based process–but not familiarity–a fast and mainly
perceptual, signal-detection-like process (Yonelinas 2002). This suggests that social
context affected semantic (associative) processes but not perceptual processes. On the
other hand, recognition judgments may have been unaffected because the focus of
attention was always on the lexical decision task during which exposure durations
were fixed. This reduces the likelihood that perceptual processes that mainly affect
recognition would differ across conditions. There may also have been a ceiling effect
for recognition judgments in the social conditions where hit rates were above 82%.
Finally, the memory tasks assessed episodic memory for common words instead of
memory for new information. This reduces the likelihood that differences in
familiarity would moderate the results. In sum, it is on weak grounds that we conclude
social context plays a lesser role for recognition than for recall. We revisit this issue
in experiment 3.
Experiment 2 was designed to address an unanticipated finding in experiment
1: the inhibitory effect of social context on reaction times. The presence of a
confederate, introduced as a research assistant, resulted in slower reaction times.
Interestingly, it also resulted in a higher error rates (Msocial = 4.11 vs. Malone = 2.93,
16
t[53] = -2.13, p < .05). It is possible that the main effects in the form of slower
reaction times and higher error rates were an outcome of the specific role of the
confederate as a research assistant passively observing the participant. This may have
shifted covert attention to the confederate and away from the task at hand. The cover
story may have also invested the confederate with a sense of authority, an important
determinant of social influence (Milgram 1974). This may have led to an especially
strong activation of the goal of social desirability. We therefore replicated experiment
1 with a less intrusive manipulation of social context, one where the confederate was
a fellow participant and where the confederate was not observing the participant but
attending the target information.
EXPERIMENT 2
Experiment 2 was designed to replicate experiment 1 with one major
difference: the purported role of the confederate in the social condition. Whereas in
experiment 1 the confederate was presented as a passive observer in the persona of a
research assistant, the confederate in experiment 2 was presented as a fellow
participant performing the lexical decision task in parallel to the participant. This
change in the experimental procedure was designed to reduce evaluation
apprehension by preventing the confederate from potentially being perceived as an
authoritative figure and by switching the confederate’s focus of attention from the
participant to the performance at hand. This type of coactivity is known to reduce
social influence (Huguet et al. 1999) and should mitigate the inhibitory effect of
social context on the error rate and on speed of response. In fact, this more
competitive orientation in the social condition should lead to faster performance
during the lexical decision task than in the alone condition because of automatic
social comparison (Gilbert, Giesler, and Morris 1995). We therefore aim to replicate
17
the results of experiment 1 under conditions where the main effect of social context
on speeded responses is designed to be reversed.
Method
Design and participants. The experimental design was the same as in
experiment 1. Participants were 37 female students (Mage = 23.9, SD = 4.54) who
participated in exchange for a monetary reward. One participant was removed from
the sample because she did not follow the experimental instructions. This resulted in
36 participants (19 in the social and 17 in the alone condition).
Procedure. Two participants at a time were met by an experimenter and by a
male confederate who acted as if he was a third participant. Participants were
randomly assigned to the alone condition to a room by themselves or to the social
condition to a room along with the confederate. To discourage direct social interaction
the experimenter asked participants and confederate to avoid talking to each other
during the duration of the study. As in experiment 1, participants first partook in the
reaction time task. In the social condition, the confederate sat next to the participant
and pretended to also perform the lexical decision task using an inactive keyboard. At
the end of the reaction-time task the experimenter entered the room and asked the
confederate to follow him to another room to complete the next task. Participants in
both social and alone condition where thus alone for the remainder of the experiment.
They were first asked to fill out an unrelated questionnaire lasting approximately six
minutes to clear short-term memory before completing the free recall and recognition
tests.
After the completion of the memory measures participants were probed for the
real purpose of the study, debriefed, and thanked. Participants in the social condition
were asked about their thoughts about the other participant (i.e., the confederate).
18
None of the participants guessed the purpose of the study or expressed doubts about
the purported role of confederate.
Results
Response latencies. As in experiment 1, we eliminated data from participants
who made more than 10 mistakes across the 64 experimental trials. This led to the
exclusion of seven participants. The average error rate for these participants was 27%,
as opposed to 4% for the participants included in the analysis. In all the analysis
reported below we therefore used data from 29 participants (16 in the social and 13 in
the alone).
We relied on the same mixed model used in experiment 1. The main effect for
viewing context approached significance, F(1, 27) = 2.96, p < .10. However, contrary
to the inhibitory effect of social context in experiment 1, the direction of the effect
was a facilitating one. Participants were faster in the social condition (Msocial = 599 vs.
Malone = 673) with no differences in the error rates (Msocial = 2.37 vs. Malone = 2.46,
t[27] = .11). As in experiment 1, the main effect of stimulus type was not significant.
Moreover, despite the directional change in the main effect of social context from
experiment 1, the interaction between social context and stimulus type was again
significant and supportive of H1 (see Figure 2a), F(1, 870) = 4.13, p < .05. In the
alone condition, words applicable to the goal of social desirability were again reacted
to more slowly than neutral words (Mapplicable = 684 vs. Mneutral = 661). In contrast, in
the social condition, words applicable to the goal of social desirability were identified
faster than neutral words (Mapplicable = 596 vs. Mneutral = 602).
--------------------------------------Insert Figure 2 about here
---------------------------------------
19
Memory. The pattern of results for free recall was the same as that in
experiment 1. There was no main effect of viewing condition, but the main effect for
stimulus type was significant, F(1, 27) = 22.99, p < .0001. As in experiment 1, words
applicable to social desirability were more likely to be recalled than neutral words
(Mapplicable = 3.59 vs. Mneutral = 1.59). The interaction between viewing context and
stimulus type was again significant and in the hypothesized direction (H2), F(1, 27) =
5.20, p < .05 (see Figure 2b). Participants recalled more neutral words in the alone
condition (Msocial = .94 vs. Malone = 2.38) but more words applicable to social
desirability in the social condition (Msocial = 3.75 vs. Malone = 3.38). As in experiment
1, we observed no differences in recognition accuracy across conditions. In the alone
condition, hit rates were 88% for neutral words and 76% for words applicable to
social desirability. In the social condition, hit rates were 76% for neutral words and
84% for words applicable to social desirability. The interaction was not significant, F
< 1.
Discussion
The results of experiment 2 provide further support for the theoretical
framework. They provide evidence that the mere presence of another person leads to
the automatic activation of the goal of social desirability (H1) and that this has
important implications for memory (H2), as evidenced in free recall. Most
importantly, the interaction effects were robust to the social persona of the
confederate. We observed a directional shift in the main effect of social context on
response latencies due to the change in the confederate’s social persona from passive
research assistant to active participant. Regardless, the results for reaction time and
free recall were the same.
20
Experiment 2 again did not find a significant effect of viewing context for
recognition. This is not surprising given that the design of experiment 2 essentially
replicated that of experiment 1. In experiment 3, we address this issue by assessing
recognition of nonsensical brand names, that are neutral cues per se, but that are
embedded in ads that either have a social desirability positioning or not. This
mitigates semantic (associative) effects on recognition accuracy and allows us to
examine whether goal-priming moderates the amount of attention devoted to an ad
(e.g., Celsi and Olson 1988). Experiment 3 was also designed to test the
generalizability of the results of experiments 1 and 2 in several respects. First, we
examine the effect of social context in an advertising setting rather than in purely
cognitive psychological tasks. We also lengthen time of stimulus exposure, allowing
the focus of attention to shift and perceptual processes to vary. This increases the
likelihood of observing an effect of goal priming on recognition accuracy. Finally, we
rely on male participants and a female confederate.
EXPERIMENT 3
The results of experiments 1 and 2 suggest that message reception, as
measured by recall, is influenced by social context. It is important to consider the task
characteristics, however. Reception was assessed for a series of single words that had
been presented only briefly in a reaction time task. The effect of social context may
have been especially powerful because the time to elaborate was limited and social
context provided one of the few retrieval cues. It is thus unclear if the mere presence
of another person would have the same effect on message reception for complex
stimuli such as advertisements that are consumed under more leisurely conditions.
Experiment 3, therefore, is designed to test the generalizability of the recall findings
in a setting more typical of advertising exposure. We expect that the presence of
21
another person will increase recall memory for advertising cues related to the goal of
social desirability at the expense of cues that are neutral with respect to this goal (H2).
Method
Design and participants. Experiment 3 was again a mixed 2 (viewing context:
social vs. alone) x 2 (stimulus type: applicable to social desirability vs. neutral)
design, with stimulus type being a within-subjects factor. Fifty-five male students
(Mage = 20.7, SD = 2.1) participated for a monetary reward. Data from two participants
were eliminated from the analysis. In one case the participant reported having
previous knowledge of the study’s purpose. In the other case a fire alarm interrupted
the study. We therefore used data from 53 participants (27 in the alone and 26 in the
social condition).
Stimuli. Advertisements contain multiple sources of information including the
brand name, the product category, visuals and verbal copy. Each of these cues can be
a vehicle for information applicable to social desirability. We used print ads and four
to six letters long nonsensical brand names. This enabled us to assess memory for
cues that were part of stimuli applicable to social desirability but that did not possess
any applicability to the goal per se. Furthermore, we controlled for possible
unanticipated relations these nonsense words might have to social desirability by
creating two sets of ads and rotating the brand names across ads that were applicable
to or neutral with respect to social desirability.
Twenty ads (from an initial set of 50 ads) were included in a final online
pretest. The primary purpose of the pre-test was to assess the level of applicability of
the ads to the goal of social desirability. These ads included a brand name, product
category, an image and a slogan. Thirty-six male students participated in the pre-test
in return for a monetary reward (Mage = 23.08, SD = 6.14). Half of the ads were
22
selected to have high applicability and half to have low applicability to social
desirability. Also, half of the stimuli with high applicability endorsed an avoidance
orientation. These ads indicate how the product can help the consumer to avoid
making a bad impression on others. The other half of the ads with high applicability to
the goal endorsed an approach orientation, and show how to make a good impression
on others.
For each of the 20 ads participants rated the (1) complexity of the ad, (2)
attitudes towards the ad, (3) potential embarrassment induced by ad exposure
(“Would you feel uncomfortable in any way if a person of the opposite sex observes
you while you are reading this ad?”), and (4) level of applicability to social
desirability. Order of ad presentation was varied using four versions of the Internet
survey that were counterbalanced across participants to eliminate order effects. Paired
t-tests on perceived level of difficulty showed no differences between ads applicable
to or neutral with respect to social desirability, t(35) = 0.93, or between approach and
avoidance social desirability ads, t(35) = -0.98. Attitudes towards the ad also did not
differ between ads applicable or neutral with respect to social desirability, t(35) = 1.07, or between approach and avoidance social desirability ads, t(35) = -0.12.
However, ads applicable to social desirability were perceived as more potentially
embarrassing than neutral ads, t(35) = -4.13, p < .001. Moreover, ads endorsing an
avoidance focus were perceived as more potentially embarrassing than ads endorsing
an approach focus, t(35) = -2.68, p = .01.
Most importantly, ad applicability to social desirability differed as intended.
For the approach social desirability category 75.6% of responses chose option: “It
suggests how I can make a good impression on others”. For the avoidance social
desirability category 69.4% chose the option: “It suggests how I can avoid making a
23
bad impression on others”. For the neutral ads 95% chose the option: “It contains no
reference to the way I am perceived by others”. A multimomial logit model using
choice between these three options as limited dependent variable demonstrates the
effectiveness of the manipulation. Ad type was highly predictive of choice with the
signs of the coefficients for the dummy variables in the hypothesized direction (βappr =
.88, p < .0001; βavoid = -.80, p < .0001).
In sum, we selected as stimuli for the experiment 20 print ads for fictitious
products and with nonsensical brand names (e.g., “Monec”, “Trez” and “Aleon”).
Half of the ads were unrelated to social desirability (e.g., ad for a toaster) whereas
half were related to social desirability. Of these, five contained messages aimed at
showing how to make a good impression in public (e.g., ad for a perfume) whereas
five contained messages aimed at showing how to avoid a bad impression in public
(e.g., ad for anti-dandruff shampoo). Sample ads are provided in Figure 3.
--------------------------------------Insert Figure 3 about here
--------------------------------------Procedure. The assignment of male participants to conditions followed the
procedure of experiment 1. The cover story presented the confederate, a female in her
mid-twenties, as a research assistant observing participants performing the task in
order to learn about the experimental procedure. To prevent the confederate from
being able to see the keyed responses of the participants an inconspicuous cardboard
screen was placed around the keyboard used by respondents. This screen was present
in both the alone and social condition.
As an additional cover story, participants were told to rate each ad
immediately after it was displayed. They were not alerted to the memory tests.
24
Participants first participated in two practice trials during which they rated photos
depicting neutral mountain views. Next, participants were exposed to the 20 ads in
random order. Each ad was displayed for 15 seconds each. Immediately after
exposure to every ad, participants rated the ad on a scale from one to five (1 = “I
disliked it very much”; 5 = “I liked it very much”). When the sequence of ads was
completed the confederate left the room. All participants completed the remaining of
the study alone in the room.
After a filler task lasting approximately seven minutes, participants had two
minutes to recall the product categories used in the ads. Next, participants were asked
to complete a recognition task including the nonsensical brand names. Half of the
target words were the brand names used for the ads and half were distractors.
At the end of the study participants were probed about the real purpose of the
study and about the role of the confederate. None of the participants guessed the
purpose of the study. Participants were then debriefed and thanked.
Results
Free recall of product category. Data from six participants were eliminated
because they did not follow the instructions and recalled brand names or images
instead of product categories, leaving 47 participants in the analysis of the recall data.
A repeated measures ANCOVA using ad set as a covariate (the brand-product
counterbalancing) revealed no main effect of condition, but a main effect of stimulus
type, F(1, 43) = 80.49, p < .0001 (see Figure 3). Participants recalled more product
categories from ads applicable to social desirability than from neutral ads (Mapplicable =
4.62 and Mneutral = 2.42). A significant interaction between social context and stimulus
type qualifies these results, F(1, 43) = 6.44, p < .05. Memory for product categories in
ads applicable to social desirability was higher in the social condition (Msocial = 4.86
25
vs. Malone = 4.40) whereas memory for the product categories in neutral ads was
higher in the alone condition (Msocial = 2.09 vs. Malone = 2.72)1.
----------------------------------Insert Figure 4 about here
---------------------------------------
Brand name recognition. We did not observe an effect of viewing context on
recognition memory for the nonsensical brand names. In the alone condition the hit
rate for brand names of ads applicable to social desirability was 56% whereas for
brand names of neutral ads was 62%. In the social condition the hit rate for brand
names of ads applicable to social desirability was 60% whereas for brand names of
neutral ads was 59%. The interaction was not significant, χ2 < 1.
Discussion
Experiment 3 mirrors the memory results of experiments 1 and 2 using male
participants and a female confederate and stimuli typical of print ads. We found that,
relative to neutral ads, participants in the social condition were more likely than
participants in the alone condition to recall product categories presented in ads
applicable to social desirability. This provides further evidence on recollection
processes mediating the influence of social context on cognition. As in the previous
two studies, no effect was found for recognition. On the one hand, this suggests that
goal priming influenced cognition by moderating associative and not perceptual
processes. On the other hand, there was a floor effect in the recognition judgments of
the nonsensical brand names and it is difficult to interpret the results. Moreover,
because the brand names were nonsensical, they were not applicable to the goal1
Even though the ads were evaluated similarly in the pretest, there was an unexpected significant main
effect for stimulus type, F(1, 1052) = 49.92, p < .0001. In both the social and the alone conditions, ads
26
priming manipulation. Thus, even though the goal-priming manipulation may have
worked to heighten attention for ads applicable to social desirability as a whole (e.g.,
Moskowitz 2002), it could have drawn attention to ad features applicable to the goal
of social desirability and detracted attention away from the “neutral” brand names
(Huffman and Houston 1993). Therefore, whereas the recall results allow us to
conclude that social context affects recollection processes, we cannot dismiss the
possibility that it affects perceptual processes as well.
GENERAL DISCUSSION
Theoretical contribution
This article presents the results a set of experiments designed to evaluate the
effect of mere social presence on message processing. Three studies demonstrate that
the presence of others affects how we process and remember cues that have high
applicability to the goal of making a good impression on others. The theoretical
framework is based on the notion that people will be intrinsically motivated to make a
good impression on others in social situations and that this concern for social
desirability can affect cognitive activity. Using a lexical decision task and response
latencies as dependent variable, experiments 1 and 2 demonstrated that the presence
of a person of the opposite gender increases the accessibility of the goal of social
desirability. Participants were relatively faster at identifying words applicable to
social desirability in the mere presence of another person, whereas they were
relatively faster at responding to neutral words when alone. This interaction between
viewing context and stimulus type was robust to changes in the main effect of viewing
applicable to social desirability were liked more than were neutral ads (Mapplicable = 3.37, Mneutral = 2.88).
We have no explanation for this result.
27
context, when social context had an inhibitory (as in experiment 1) or a facilitating (as
in experiment 2) effect on speeded performance.
The most important finding of this article concerns the memory consequences
of social context. Three studies consistently demonstrated a shift in recall memory
such that more information applicable to social desirability was recalled when this
information was learned in the mere presence of another person, whereas more
information neutral with respect to social desirability was recalled when the
information was learned by participants who were alone. Because there was no
similar interaction in the recognition data, we must tentatively conclude that this
effect is based on recollection and not familiarity (Lynch and Srull 1982; Yonelinas
2002).
In discussing the response latencies and the recall differences, we referred to
the influence of social context on message reception as automatic. However,
automaticity is not a monolithic concept and it is therefore important to clearly state
what is meant by automaticity in a certain context (Bargh and Chartrand 2000). We
followed the suggestion put forth by Higgins (1996): “if you are concerned with how
a laboratory effect might or might not generalize to a busy real-life social interaction
setting, focus on the efficiency of the process, not its possible unintentionality” (p.
173). Using goal accessibility as the operationalization of goal activation, our
definition of automaticity was an increase in processing efficiency of cues related to
the goal of social desirability experienced by participants in a social context. This
implies neither that participants were unaware of their concern with the impression
the confederate was forming of them nor that such a concern was outside participants’
control. What makes the process investigated automatic in our case is instead its
effect on processing efficiency.
28
Our research contributes to the literature on nonconscious goal priming in two
ways. First, it answers calls for a broadening of scope of automaticity research (e.g.,
Bargh and Ferguson 2000; Bargh 2002) by showing how a very general goal such as
social desirability can be primed using an unobtrusive manipulation. Research on goal
priming has primarily focused on the automaticity of specific concrete goals–such as
memorization or task performance–instead of more abstract goals (Levesque and
Pelletier 2003). We investigated a general goal related to the way individuals construe
their social environment. Evidence of automatic activation of such a goal therefore
broadens our understanding of automatic processes in everyday life.
Second, research on nonconscious goal priming typically relies on priming
manipulations that directly activate a goal (e.g., a scrambled sentence or anagram
task). These manipulations serve as proxy for what happens in a naturalistic setting
and provide limited evidence for the way in which environmental elements
automatically activates the goal (Chartrand and Bargh 2002). These studies, in other
words, are mostly concerned with “the theoretical viability of preconscious goal
activation, not the ecological validity” (Moskowitz 2002, p. 403). The experiments
reported above contribute to the literature on goal priming by showing how social
context, manipulated not via a proxy but using an ecologically common procedure,
can increase the level of activation of the goal of social desirability.
Practical implications
Together the results provide compelling support for the notion that viewing
context is an important determinant of message reception. This has important
implications for the advertising industry, especially in the areas of media planning and
pre-testing, as top-of-mind recall is one of the most commonly used measures of
advertising effectiveness. First, our findings indicate that campaigns that play on the
29
theme of social desirability are more memorable when viewed in a social context.
Second, they suggest that effort should be placed to match pre-testing conditions with
the conditions in which consumers are most likely to encounter the ad. For example,
when pre-testing Super Bowl ads applicable to social desirability it may be more
informative to use a group setting than individual research participants.
Our findings also have methodological implications. The methodological
individualism that characterizes much of social psychology has been the focus of an
active debate within the discipline (e.g., Ritzer and Gindoff 1992). As a result of
methodological individualism, “psychology ordinarily deals with individual processes
and structures, without much systematic consideration of social context” (Eagly and
Chaiken 1993, p. 627). However, as participants interact with the researcher, an
experiment takes on social as well as scientific aspects (Paulus and Nagar 1989, p.
131; Singleton and Straits 1999, pp. 194-195).
Experimental confounds due to a failure to account for the effects of social
context can be either created by the experimenters, research assistants and
confederates or by the participants themselves when several participants complete the
study at the same time. We suspect that social context effects might be the most
common among the various types of priming effects introduced in the data by the
experimental procedure without the experimenters’ awareness (Bargh and Chartrand
2000). Our research thus begins to answer calls for an increased focus on the effect of
social context (e.g., Levine and Moreland 1990) and can provide methodological
insight into the potential threat to external validity posed by a failure to account for
the effect of social context in experimental settings (cf. Fisher and Dubé in press).
Future Research
30
Several areas for future research deserve to be mentioned. The first pertains to
the role of audience member familiarity. This is an important issue because one form
of advertising often consumed socially, television advertising, is also very likely to be
consumed in company of close relationships. Because research on social influence is
often carried out in a laboratory setting with the help of confederates, social
psychologists have rarely investigated the moderating effect of level of acquaintance.
Certain types of people may prime specific goals that we did not consider. For
example, priming the mental representation of father can increase pursuit of a task
goal (Shah 2003), whereas priming the mental representation of good friend can
activate a helping goal (Fitzsimons and Bargh 2003).
Another area of investigation is how social context affects evaluative
processes. The literature on the effect of social context on attitudes has previously
found a role for self-presentation motives, where participants strategically alter their
expressed attitudes to appeal to others (Eagly and Chaiken 1993). An intra-psychic
perspective on the process of persuasion in a social context, however, suggests that
social context could affect attitudes even when these are not subject to public
scrutiny. In experiment 3, we did not find online attitudes to be moderated by social
context. Whereas online judgments are typically not correlated with recall (e.g., Wyer
and Srull 1986) memory-based judgments are largely based on the information
retrieved from memory (Hastie and Park 1986). Therefore, social context may play a
more important role in memory-based attitudes. However, the relationship between
message reception and evaluations is not straightforward and simple quantitative
measures of recall are naïve inasmuch as product attributes vary in their desirability
and attribute valence needs to be considered (Reyes, Thompson, and Bower 1980).
31
Our theoretical framework, therefore, would need to be extended to anticipate the
effect of social context on attitude formation.
Another important issue that needs to be addressed is whether an audience
needs to be present in order to activate a goal of social desirability or if an imagined
audience could trigger the same effect. Evidence supports the contention that priming
a mental representation of others can affect cognition (e.g., Fitzsimons and Bargh
2003). Studies in this area generally prime the mental representation of significant
others but research has shown that priming a mental representation of an individual of
unspecified identity can trigger a concern with social desirability (e.g., Dahl et al.
2001). It is thus possible that advertisements that encourage consumers to imagine
being in a social setting have the same effect as an actual social context or that the
virtual audience in an Internet chat room creates a social presence.
CONCLUSION
Human behavior is inherently social in nature. Previous advertising research
has disregarded the potential influence of social context on motivational forces. This
article presents the first investigation of the effect of the presence of others for goal
priming and message reception. We proposed that a concern with social desirability
becomes automatically activated in social situations. Empirical evidence provides
support for the framework by showing significant consequences of mere social
presence for goal accessibility and free recall. Because memory is a component in
virtually all thinking these findings may have far-reaching implications. Memory is
also one of the key measures of advertising effectiveness and our findings have direct
implications for advertising practice.
32
REFERENCES
Ariely, Dan and Jonathan Levav (2000), "Sequential Choice in Group Settings:
Taking the Road Less Traveled and Less Enjoyed," Journal of Consumer
Research, 27 (December), 279-290.
Bargh, John A. (1990), "Auto-Motives: Preconscious Determinants of Social
Interaction," in Handbook of Motivation and Cognition: Foundations of Social
Behavior, ed. E. Tory Higgins and Richard M. Sorrentino, New York: Guilford,
93-130.
Bargh, John A. and Tanya L. Chartrand (2000), "The Mind in the Middle: A Practical
Guide to Priming and Automaticity Research," in Handbook of Research
Methods in Social and Personality Psychology, ed. Harry T. Reis and Charles
M. Judd. New York: Cambridge University Press, 253-285.
Bargh, John A. and Melissa J. Ferguson (2000), "Beyond Behaviorism: On the
Automaticity of Higher Order Processes," Psychological Bulletin, 126 (6), 925945.
----- (2002) "Loosing Consciousness: Automatic Influences on Consumer Judgments,
Behavior, and Motivation, " Journal of Consumer Research, 29 (September),
280-285.
Burgess, Neil and Graham J. Hitch (1992), "Toward a Network Model of the
Articulatory Loop," Journal of Memory and Language, 31, 429-460.
Celsi, Richard L. and Jerry C. Olson (1988), "The Role of Involvement in Attention
and Comprehension Processes," Journal of Consumer Research, 15
(September), 210-224.
Chartrand, Tanya L. and John A. Bargh (2002), "Nonconscious Motivations: Their
Activation, Operation, and Consequences," in Self and Motivation: Emerging
33
Psychological Perspectives, ed. Abraham Tesser, Diederik A. Stapel, and
Joanna W. Wodd. Washington DC: American Psychological Association, 13-41.
Dahl, Darren W., Rajesh V. Manchanda, and Jennifer J. Argo (2001),
"Embarrassment in Consumer Purchase: The Roles of Social Presence and
Purchase Familiarity," Journal of Consumer Research, 28 (December), 473481.
Eagly, Alice H. and Shelly Chaiken (1993), The Psychology of Attitudes. Orlando,
FL: Hartcourt Brace.
Emmons, Robert A. (1989), "The Personal Striving Approach to Personality," in Goal
Concepts in Personality and Social Psychology, ed. Lawrence A. Pervin.
Hillsdale, NJ: Erlbaum, 87-126.
Fazio, Russell H. (1990), "A Practical Guide to the Use of Response Latencies in
Social Psychological Research," in Review of Personality and Social
Psychology, ed. Ladd Wheeler. Beverly Hills, CA: Sage, 74-97.
Fisher, Robert J. and Laurette Dubé (in press), "Gender Differences in Responses to
Emotional Advertising: A Social Desirability Perspective," Journal of
Consumer Research.
Fitzsimons, Gráinne M. and John A. Bargh (2003), "Thinking of You: Nonconscious
Pursuit of Interpersonal Goals Associated With Relationship Partners," Journal
of Personality and Social Psychology, 84 (1), 148-164.
Gilbert, Daniel T., R. Brian Giesler, and Kathryn A. Morris (1995), "When
Comparisons Arise," Journal of Personality and Social Psychology, 62 (2), 227236.
Gillund, Gary and Richard M. Shiffrin (1984), "A Retrieval Model for Both
Recognition and Recall," Psychological Review, 91 (1), 1-67.
34
Goffman, Erving (1959), The Presentation of Self in Everyday Life. New York:
Doubleday.
Greenwald, Anthony A. and Clark Leavitt (1984), "Audience Involvement in
Advertising: Four Levels," Journal of Consumer Research, 11 (June), 581-592.
Hastie, Reid and Bernadette Park (1986), "The Relationship between Memory and
Judgment Depends on Whether the Judgment Task is Memory-Based or OnLine," Psychological Review, 93 (June), 258-268.
Henderson-King, Donna, Eaaron Henderson-King, and Lisa Hoffman (2001), "Media
Images and Women's Self-Evaluations: Social Context and Importance of
Attractiveness as Moderators," Personality and Social Psychology Bulletin, 27
(11), 1407-1416.
Herman, Peter C., Deborah A. Roth, and Janet Polivy (2003), "Effects of the Presence
of Others on Food Intake: A Normative Interpretation," Psychological Bulletin,
129 (6), 873-886.
Higgins, E. Tory and Gillian King (1981), "Accessibility of Social Constructs:
Information-Processing Consequences of Individual and Contextual
Variability," in Personality, Cognition, and Social Interaction, ed. Nancy
Cantor and John F. Kihlstrom. Hillsdale, NJ: Erlbaum, 69-121.
----- (1996), "Knowledge Activation: Accessibility, Applicability, and Salience," in
Social Psychology: Handbook of Basic Principles, ed. E. Tory Higgins and Arie
W. Kruglanski. New York: Guilford, 133-168.
Howard, Charles and Daniel J. Gengler (2001), "Emotional Contagion Effect on
Product Attitudes," Journal of Consumer Research, 28 (September), 189-201.
Huguet, Pascal, Marie P. Galvaing, Jean M. Monteil, and Florence Dumas (1999),
"Social Presence Effects in the Stroop Task: Further Evidence for an Attentional
35
View of Social Facilitation," Journal of Personality and Social Psychology, 77
(5), 1011-1025.
Huffman, Cynthia and Michael J. Houston (1993), "Goal-oriented Experiences and
the Development of Knowledge," Journal of Consumer Research, 20
(September), 190-207.
Hsiao, Cheng (2003), Analysis of Panel Data, 2nd edition. Cambridge, UK:
Cambridge University Press.
Leary, Mark R. and Robin M. Kowalski (1990), "Impression Management: A
Literature Review and Two-Component Model," Psychological Bulletin, 107
(1), 24-47.
Levesque, Chantal and Luc G. Pelletier (2003), "On the Investigation of Primed and
Chronic Autonomous and Heteronomous Motivational Orientations,"
Personality and Social Psychology Bulletin, 29 (12), 1570-1584.
Levine, John H. and Richard L. Moreland (1990), "Progress in Small Group
Research," Annual Review of Psychology, 41, 585-634.
Levine, John H., Lauren B. Resnick, and Tory E. Higgins (1993), "Social Foundations
of Cognition," Annual Review of Psychology, 44, 585-612.
Littell, R. C., P. R. Henry, and C. B. Ammerman (1998), "Statistical Analysis of
Repeated Measures Data Using SAS Procedures," Journal of Animal Science,
76, 1216-1231.
Lowery, Brian S., Curtis D. Hardin, and Stacey Sinclair (2001), "Social Influence
Effects on Automatic Racial Prejudice," Journal of Personality and Social
Psychology, 81 (5), 842-855.
Lynch, John G. and Thomas K. Srull (1982), "Memory and Attentional Factors in
Consumer Choice: Concepts and Research Methods," Journal of Consumer
36
Research, 9 (June), 18-37.
Mikulincer, Mario (1998), "Attachment Working Models and the Sense of Trust: An
Exploration of Interaction Goals and Affect Regulation," Journal of Personality
and Social Psychology, 74 (5), 1209-1224.
Milgram, Stanley (1974), Obedience to Authority. New York: Harper and Row.
Mori, DeAnna, Patricia Pliner, and Shelly Chaiken (1987), ""Eating Lightly" and the
Self-Presentation of Femininity," Journal of Personality and Social Psychology,
53 (4), 693-702.
Moskowitz, Gordon B. (2002), "Preconscious Effects of Temporary Goals on
Attention," Journal of Experimental Social Psychology, 38, 397-404.
Paulus, Paul B. and Dinesh Nagar (1989), "Environmental Influences on Groups," in
Psychology of Group Influence, ed. Paul B. Paulus. Hillsdale, NJ: Erlbaum.
Ratner, Rebecca K. and Barbara E. Kahn (2002), "The Impact of Public versus Private
Consumption on Variety-Seeking Behavior," Journal of Consumer Research, 29
(September), 246-257.
Reyes, Robert M., William C. Thompson, and Gordon H. Bower (1980), “Judgmental
Biases Resulting from Differing Abilities of Arguments,” Journal of
Personality and Social Psychology, 39 (July), 2-12.
Ritson, Mark and Richard Elliott (1999), "The Social Uses of Advertising: An
Ethnographic Study of Adolescent Audiences," Journal of Consumer Research,
26 (December), 260-277.
Ritzer, George and Pamela Gindoff (1992), "Methodological Relationism," Social
Psychology Quarterly, 55 (2), 128-140.
Shah, James (2003), "Automatic for the People: How Representations of Significant
Others Implicitly Affect Goal Pursuit," Journal of Personality and Social
37
Psychology, 84 (4), 661-681.
Singleton, Royce A. and Bruce C. Straits (1999), Approaches to Social Research.
New York: Oxford University Press.
Yonelinas, Andrew P. (2002), "The Nature of Recollection and Familiarity: A Review
of 30 Years of Research," Journal of Memory and Language, 46, 441-517.
Wyer, Robert S. and Thomas K. Srull (1986), "Human Cognition in Its Social
Context," Psychological Review, 93 (July), 322 359.
----- and ----- (1989), Memory and Cognition in Its Social Context. Hillsdale, NJ:
Erlbaum.
38
FIGURE 1
EXPERIMENT 1: RESPONSE LATENCIES AND FREE RECALL
640
630
630
622
Response Latencies
620
610
600
588
590
580
570
570
560
550
540
Alone
Neutral words
Social
Words related to social desirability
3.5
2.96
3
2.45
Words recalled
2.5
2
1.59
1.5
1.15
1
0.5
0
Neutral words
Words related to social desirability
Alone
Social
39
FIGURE 2
EXPERIMENT 2: RESPONSE LATENCIES AND FREE RECALL
700
684
680
661
Response Latencies
660
640
620
602
596
600
580
560
540
Alone
Neutral words
Social
Words related to social desirability
3.75
4
3.38
3.5
Words recalled
3
2.5
2.38
2
1.5
1
0.94
0.5
0
Neutral words
Words related to social desirability
Alone
Social
40
FIGURE 3
EXAMPLES OF APPROACH SOCIAL DESIRABILITY ADS (A), AVOIDANCE
SOCIAL DESIRABILITY ADS (B) AND NEUTRAL ADS (C)
A
B
C
A
B
C
41
FIGURE 4
EXPERIMENT 3: FREE RECALL OF PRODUCT CATEGORY
6
4.86
Product categories recalled
5
4.4
4
3
2.72
2.09
2
1
0
Neutral ads
Social desirability ads
Alone
Social
42