The influence of approaching behavior on the development of likeability in social anxiety.
Case 2 Persistence of the vicious circles of SAD according to the cognitive model
Lunch break at work; Maria wants to join the conversation and has the feeling that she should
participate. Entering the social situation her basic assumptions activate. Maria has the
assumption that she is socially an inadequate person. Therefore, she perceives the social situation
as threatening. Negative automatic thoughts come running up in her mind, such as “they think I
must be stupid” and “I am boring”. Automatic responses to threat can evoke physical symptoms,
such as trembling, blushing, warmth, and stumbling. Maria starts to blush and to stutter, what
leads to a high self-consciousness: she perceives herself as a social object in the situation. Those
anxiety symptoms have an impact on the self-presentation, more negative thoughts, and more
anxiety. Maria starts to show inappropriate behaviors: safety behaviors. She sits still and says
nothing as she fears that others think her remarks or questions are stupid or inappropriate.
Furthermore, Maria thinks she shows inadequate behavior and thinks she will therefore be liked
less by others and, will be negatively evaluated and socially rejected by them. Because of her high
self-consciousness, Maria does not start looking for evidence if their assumptions are correct by
noticing on how others respond to her. She remains in the circle of anxiety, safety behavior, and
Cognitive behavioral therapy includes exposure techniques and cognitive interventions
(Keijsers, van Minnen, & Hoogduin, 1999). It suggests cognitive impairments in persons with
high levels of social anxiety and sets schemes and assumptions central in the therapy (Beck
and Emery, 1985). Hofmann and Bogels (2006) describe 40 to 50% with little or no
improvement after usual used cognitive behavioral therapy (CBT). Several patients could not
complete the therapy or remain to have high levels of anxiety after therapy. Feske and
Chambless (1995) studied by meta-analysis over 21 studies the effectiveness of CBT and
exposure in vivo. They concluded that CBT is not more effective than a therapy alone with
exposure in vivo (Scholing & Emmelkamp, 1995). In addition to CBT, Westenberg (2009)
studies the effectiveness of other treatments and concludes that only 40 to 60% of patients
with SAD show a treatment response after both pharmacologic and non-pharmacologic
Concluding, CBT therapy leads to some improvement of social anxiety, but because of
the severity of the disorder, further improvement on the therapy is still needed. The disorder
knows a high impairment on the daily life and high levels of anxiety. Although treatment for
SAD has been developed, it is still one of the most common mental disorders (Hofmann &
1.5 The interpersonal perspective
The cognitive behavioral theory describes the effects of the anxiety on the person with SAD,
but does not describe the interpersonal impact it has. The interpersonal perspective, as
described below, does describe this impact and the influences within. This new view on social
anxiety is developed additionally to the cognitive theory.
1.5.1 Behavioral deficits
The reactions of others depend for a great part on how others see us. And how they see us is
influenced by the interaction of the social situation. It is suggested that behavior in a social
situation has a great influence on this interaction, but the evidence is mixed whether people
with SAD really do show impairments in behavioral skills (Rapee, 1995; Papsdorf & Alden,
1998). Some studies have not found any differences in social or performance related skills
between people with high levels of social anxiety and control participants (Clark & Arkowitz,
1975; Pilkonis, 1977b; Rapee & Lim, 1992). Other studies did find differences between
people with high and low levels of social anxiety on global measures of performance, but not
on specific social skills, such as eye-contact (Beidel, Turner, & Dancu, 1985). Rapee, Gaston
and Abbott (2009) focused by use of an enhanced treatment with several additional treatment
techniques on the behavioral aspects instead of the cognitive aspects (e.g., performance
feedback, attention retraining). They found significant greater effects than CBT on diagnoses,
diagnostic severity, and anxiety during a speech. This means that, in contrast with other
studies, that behavior does have an impact on SAD.
1.5.2 Approaching behavior
Behavior in social situations is an important aspect in the interpersonal perspective on SAD. It
can be divided in four types: approaching behavior, avoidance behavior, safety behaviors (as
described in paragraph 1.4), and self-disclosure. All behavioral aspects have its influence on
the interaction between people and the likeability within.
Approaching behavior is the behavior to approach to the interaction partner and to show the
willingness to start and remain a conversation. It contains aspects such as smiling, open
questions, eye-contact, high quantity of questions, fluency of speech, long duration of
responds, and an open posture (Papsdorf & Alden, 1998; and master theses by van der Kloet,
2008 and Reichgelt, 2007).
Avoidance behavior contains aspects such as: avoiding eye-contact, doing other
activities during the social situation, short responds, closed posture, and low quantity of
questions. Like written above, persons with high levels of social anxiety have the basic
assumption that they will be rejected and/or criticized. Avoidance behavior can be seen as the
absence of approaching behavior. The difference with safety behaviors is that in safety
behavior the reason why the behavior is used is defined. That is, avoidance behavior is merely
the absence of certain types of behavior; safety behavior is inappropriate behavior that a
person with social anxiety uses (erroneously) in order to avoid being less liked by others.
Thompson and Rapee (2002) studied the behavioral features of avoidance behavior and
capability of behavior by individuals with high (N=26) and low levels (N=24) of social
anxiety. The participants were observed in an unstructured and a structured social interaction
situation with a confederate. The finding of the study is that the performance of the group of
high social anxiety (HSA) was scored lower than the group with low social anxiety (LSA) in
both situations, but the difference was much greater in the unstructured than in the structured
social situation. Because of the difference of performance between the structured and the
unstructured task, Thompson and Rapee (2002) suggest that HSA group performs lower than
the LSA group as a result of avoidance rather than incapability. In unstructured situations a
person with SAD has more opportunity to show safety behaviors and avoidance behavior what
lead to a worse performance. Other studies (Clark & Wells, 1995; Rapee & Heimberg, 1997;
Thompson & Rapee, 1997) support this vision. However, when this would be the only
problem, there would be more positive therapy results on behavioral therapies, such as social
skills training. There are other studies (Rapee & Lim, 1992; Clark & Arkowitz, 1975) that fail
to show a difference between HSA and LSA, probably because of the use of a structured task
instead of an unstructured task (Thompson & Rapee, 2002). However, as suggested by
Thompson and Rapee (2002), more research is needed to see if the main problem is a skill
deficit or if there are other problems.
The last aspect of approaching behavior is self-disclosure. Self-disclosure is telling
personal information from oneself to another, such as the own interests, personal ideas, and
intimate thoughts (Collins & Miller, 1994). It can also be seen as part of approaching behavior
in this study: giving personal information to the interaction partner is a way to approach in
contact. Self-disclosure has also an influence on the likeability and interaction. Collins &
Miller (1994) studied the relation between self-disclosure and likeability by the interaction
partners in undergraduate women (N=84). Their results suggest that persons who interact in
intimate disclosure situations tend to be liked more than people who disclose at lower levels.
In addition, people disclosure more to whom they like and people like others as a result of
having disclosed to them. This can be seen in case 3.
Case 3 Self-disclosure and reciprocity
Thomas participates in three conversations with three different people. Firstly, Thomas is having a
conversation with Peter. He self-discloses to Peter, and, therefore, Peter likes Thomas. As a result
of this, Peter discloses to Thomas and, therefore, will be liked by Thomas.
Secondly, Thomas has a conversation with Sarah. However, when Thomas self-discloses to Sarah,
she does not self-discloses back to Thomas, Thomas will like Sarah less because of the absence
self-disclosure. Thirdly, the final conversation is with Chris. Thomas self-discloses to Chris and
Chris self-discloses back to Thomas. However, when Chris started to talk, he keeps on talking
without listening or paying attention to Thomas.
As can be seen in case 3, the interaction must be two-sided to increase the levels of likeability.
According to the first example, when Thomas is self-disclosing to Peter, there is a good
interaction with an increasing level of likeability. However, in the case of Sarah and Chris,
there is no reciprocity, leading to lower levels of likeability. It is possible that people with low
levels of social anxiety disclose more easily and, therefore, are more liked by others (Papsdorf
& Alden, 1998). Additionally, reciprocity is essential in the process of self-disclosure.
Resuming, three effects are found by Collins and Miller (1994): people who self-disclose tend
to be liked more than others, people disclose more to the people they like more, and people
like others as a result of disclosure.
1.5.3 Interpersonal model of Alden (2001)
People with high levels of social anxiety have negative beliefs of on how they appear to
others. Is it possible that, therefore, people with high levels of social anxiety tend to have less
approaching behavior and show less self-disclosure? Meleshko and Alden (1993) have found
self-protective strategies in people with high levels of social anxiety, which can lead to
negative interpersonal reactions that maintain the implications in the interpersonal
relationships. They are tending to turn themselves internally instead of externally. Moreover,
they will not find any proof that others are willing to give a positive response (Clark et al.,
2006). They tend to reduce the risk of negative evaluation by avoiding approaching behavior
and avoiding self-disclosure. However, this has a reversed effect. The interpersonal model of
Alden is a vicious circle of erroneous schemes and thoughts by people with social anxiety.
The person with social anxiety gets a reaction of others that confirms the schemes of
inadequacy. Both interaction partners self-disclose less and, thereby, show less approaching
behavior (Alden, 2001). This theory can be summarized as in figure 3 and case 4.
Case 4 Interpersonal model of Alden (2001)
Matthew has social anxiety; he is afraid that others will not like him. He expects that others will
reject him, and therefore he remains quiet. Jessica, the interaction partner of Matthew, receives
this message as negatively and interprets this message as: he does not like me. This thought will
turn into the behavior of saying nothing by Jessica. The behavior of Jessica will be interpreted
again by Matthew as a rejection: ‘you see, she does not like me, because she says nothing’.
‘they do not like
Message to others
‘he/she does not
Interpersonal model of Alden (2001)
1.6 Current theory
As indicated in Du Toit & Stein (2003) and Thompson and Rapee (2002), more investigation
is needed on the behavioral impairments of people with high levels of social anxiety in
interaction with other people. The current study focuses more on those behavioral features in
According to the described theories, there is a vicious circle in social anxiety beginning by
avoidance of contact; the respondent will not go into conversation. This leads to the
confirmation of the basic assumption (being an inadequate person or being stupid), which
leads to more avoidance of contact and reduced levels of likeability, as can be seen in figure 3.
Thus it seems that in social anxiety not only cognitive processes, but approaching behavior
has a large influence as well on the development of likeability. However remaining questions
are: Why are people with high levels of social anxiety less liked by others? What is the
influence of approaching behavior on the development of likeability? In this study these
questions will be studied by observing an unstructured social interaction of participants with
high (HSA) and low levels of social anxiety (LSA).
The following hypotheses are assumed for the results of the study.
The first hypothesis is that HSA group will be less liked than LSA group.
The second hypothesis is that the HSA group has lower levels of approaching behavior
than the LSA group.
The third hypothesis is that the increase of likeability is moderated by approaching
behavior. That is, the higher the levels of approaching behavior the more the increase of
likeability. In addition, it is expected that social anxiety affects this moderation effect of
approaching behavior in the development of likeability. That is, the HSA group will have a
lower increase of likeability than the LSA group after showing approaching behavior.
The hypotheses are resumed on the next page in figure 4.
(SI, DFI, JoS)
Current study: development of likeability by social anxiety, moderated by approaching behavior
In this study we compared people with high levels of social anxiety (HSA) and low levels of
social anxiety (LSA) in an unstructured task by the design of an observational study. The
social anxiety was measured by assessment and questionnaires, and was checked by the
second assessment point. In addition, the level of approaching behavior was rated and a rate of
likeability was given at the beginning and at the end of a conversation.
118 first year students of psychology of Maastricht University were selected based on their
social anxiety scores on the Social Interaction Anxiety Scale (SIAS, see below). From this
group, 58 females were selected with high and low ratings (top and bottom 15%) ratings on
the SIAS of which 29 females participated in the low socially anxious group (LSA) and 29 in
the high socially anxious group (HSA).
Of the 58 participants there were 55 video tapes available of the waiting room situation. The
other tapes were lost, due to erroneous recording. Of these 55 video tapes, 49 were valid to
rate. That is, two participants were out of sight, two participants were unclear recorded and
could not been seen if they were smiling or not, one participant was entirely not on the tape,
and another participant was not recorded for the last two minutes. Those participants were
excluded from the analysis.
The following scales are used during the assessment point for the participants, confederates
and for the video observers.
2.3.1 The Social Interaction Anxiety Scale (SIAS; Mattick & Clarke, 1997)
The SIAS is used to diagnose the social anxiety. It assesses anxiety for general social
interaction. The scale has a high level of internal consistency, validness, and test-retest
reliability (Mattick & Clarke, 1997). It is used at the first and second assessment point. Ten
participants changed in diagnose over these assessment points and were excluded for further
analysis. The scale is enclosed in appendix 2.
2.3.2 The Beck Depression Inventory (BDI; Beck et al., 1988)
Meta-analysis of the BDI shows a high internal consistency, validness, and high correlations
with other depression scales (Beck et al, 1988). It is used in this study to retrieve a complete
view of psychopathological symptoms of the participants. The scale is enclosed in appendix 3.
2.3.3 The Desire for Future Interaction (DFI; Coyne, 1976)
The Desire for Future Interaction scale (DFI) rates the degree in which the rater wants to meet
the other person. According to Papsdorf & Alden (1998) it generally measures the degree of
liking or rejecting the other person. In the current study this scale is used to measure the level
of likeability. There is found a high internal consistency. The scale consists of eight items
from a scale from 1 to 5. In the current study, the DFI is used by the participant, confederate,
and three video observers. The scale is part of the rating form for the video observers and is
enclosed in appendix 6.
2.3.4 First Impression (FI) and Second Impression (SI)
For measuring the likeability of the participant there is given a score from a 1 to 10 scale. The
scale is an impression score of the participant and measures the likeability of the participant
by the raters. The FI and SI are rated by the confederate and the two video observers. The
scale is part of the rating form for the video observers and is enclosed in appendix 6.
2.3.5 Scale of Approaching Behavior (SAB; Cappendijk, 2009)
For measuring the levels of approaching behavior of the participant a new scale is developed
for this study. This scale is based on previous developed scales by two graduated master
students of Mental Health Sciences (van der Kloet, 2008; Reichgelt, 2007). Those scales rated
several types of behaviors in social interaction. The developed scale is focused on
approaching behavior. Those scales, and the developed scale, are enclosed in appendices 4, 5
and 6. It is developed by pointing out the several types of approaching behavior by watching
the videotapes of the participants with high and low levels of social anxiety. The scale is
tested by watching the tapes and comparing the results afterwards by five raters (namely, two
supervisors, one graduated master student of Mental Health Sciences, and two master students
of Mental Health Sciences who will be the video observers and raters in this study). Finally,
there are nineteen items developed to score the levels of approaching behavior on a scale from
1 to 5, such as the duration of the conversation, the place of seat taken by the participant, the
times and types of smiling, the type of questions, and the posture.
2.3.6 Scale of Functioning of the Confederate (FC; Weijtens, 2009)
The scale Functioning of the Confederate (FC) is developed to measure the reactions, posture,
and interaction of the confederate. The confederate had strict instruction about approaching
the participant, such as making eye contact twice, asking one question, and to be reserved, but
empathic toward the participant. In this scale, a rate can be given of the degree the confederate
keeps to the given instructions. This scale is rated by two video observers and is enclosed in
The participants were assessed at two time points with 2 weeks in between. At one of these
assessments, the participants were assessed during a waiting room situation and a social task.
The waiting room situation is the object of this study. At the other assessment, the participants
conducted a computer task and filled in some questionnaires. This assessment point is not
relevant for this study. The order of the assessments was randomized.
The exact instructions for the confederate and examiner can be found in appendix 1. The
experimenter explained the purpose and procedure of the social task when the participant
arrived. The participant then conducted a computer task and filled in several questionnaires
(appendices 2 and 3). Then, she was asked to wait for a few minutes in a waiting room. When
the experimenter arrived with the participant in the waiting room, the confederate already had
taken place in the room. The participant could choose where to sit between four chairs. In the