The influence of approaching behavior on the development of likeability in social anxiety.
waiting room there was a hidden camera that records the behavior of the participant.
Permission to view this video among observers was asked after the experiment.
The examiner had a short conversation with the confederate. In the conversation, the
participant was made to believe that the confederate wrote an appointment in his agenda.
However, instead of making an appointment he was noting his first impression (FI) of the
When the examiner left, the confederate had strict instructions about approaching the
participant. He was instructed to make eye contact twice. After this, he had instructions to ask
one question: “have you been participating in another study before?”. The confederate was
instructed to be reserved, but empathic toward the participant.
After the waiting room situation, the confederate and participant filled in the DFI, and
a second impression rate (SI). Subsequently, another social task would start that was not the
focus of the current study.
The video observers watched the recorded video tape and wrote down the first
impression of the participant. After watching the entire video tape, they had to fill in several
questionnaires: the Desire of Future Interaction (DFI), Judgments of Similarity (JoS) and the
Second Impression (SI). After filling in those questionnaires, they could watch the tapes as
many times as wanted to fill in the Scale of Approaching Behavior (SAB) and Self-Disclosure
and Reciprocity rating Scale (SRS). The JoS and SRS are scored, but did not belong to the
focus of the study. Therefore, those scales are excluded for further analysis. All scores are
filled in on a rating form, enclosed in appendix 6.
Five men participated in this study as confederate. Three men were approached by the local
drama club (‘Alles is Drama’). Two other actors were students of the faculty of psychology.
All the confederates were trained to follow a protocol of consistent behavior during social
interaction. The confederates had strict instructions which they had to follow. During the
conversation they were instructed to give a maximum of three answers per question and to
leave the initiative to the participant. Because of privacy, the names of the confederates are
changed in the results of the current study.
The examiners of the study of the social task are all females. Two of them are students of the
master of Mental Health Sciences. The third examiner is a student of the research master of
2.7 Video observers
Three master students of Mental Health Sciences have rated the video-tapes. The video
observers have had a training to observe the videos and to score the questionnaires. One video
observer had only scored the DFI and JoS. The two other observers scored those scales and
additionally the SAB, SRS, and FC. The scales DFI, SAB, and FC are used for the current
study. The JoS and SRS are scored, but did not belong to the focus of the study. Therefore,
those scales are excluded for further analysis.
2.8 Statistical analyses
All tests will be made by the use of the statistical program SPSS 12.0. The following tests will
Firstly, a Cronbachs Alpha test will be done to measure the reliability of the used scales.
Secondly, t-tests are done to see the difference between the HSA and LSA groups in social
anxiety, depressive symptoms, the likeability, DFI, and approaching behavior. Thirdly,
correlation tests will be used for the relation between approaching behavior and likeability.
Fourthly and finally, repeated measures ANOVA will be used for factors influencing the
development of likeability: social anxiety and approaching behavior.
3.1 Descriptive data
As expected the HSA group has higher levels of social anxiety than the LSA group, t(47) = -
8.94, p < .001. In addition, HSA reported significant more depressive symptoms than LSA,
t(47) = -4.47, p < .001. The average age of the HSA group is 19.55 (SD = 1.6); of the LSA
group it is 19.13 (SD = 1.4).
3.2 Reliability of the scales
An overview of the internal consistency and the inter rater reliability (I.C.C) of the different
scales used in this study are given below in table 1.
Internal consistency and inter rater reliability (I.C.C.) of the scales used in the current study
Alpha F.W. Alpha F.C. Alpha C. I.C.C
DFI .983 .949 .953 .811
SAB .923 .900 .955
Functioning of Confederate .916 .900 .744
First Impression .682
Second Impression .836
DFI: Desire for Future Interaction F.W.: video observer 1
SAB: Scale of Approaching Behavior F.C.: video observer 2
ICC: inter rater reliability C.: confederate
The DFI showed to be highly reliable. Possibly, the DFI can be easily used in this unstructured
task because of the short duration of the conversation. Because of having less information, it
might be easier to score high or low rates of the participant.
The scales developed for this study, the SAB (Cappendijk, 2009) and Functioning of
Confederate (Weijtens, 2009), both show a high internal consistency and inter rater reliability
The rate of the first impression and the second impression are both highly subjective.
However, there is found a high I.C.C. and, therefore, can be used for the results of the current
3.3 Influence of one of the confederate
There is found a difference between the five confederates on the scale of functioning of
confederates, F(4, 311) = 3.56, p = .013. This can be seen in figure 5. One of the confederates,
Fred, has scored significant lower on the scale of functioning of confederates than the other
confederates. In addition, this confederate has seen more participants of the HSA group than
the other confederates, χ²(4, N = 43) = .05, p < .05. Moreover, the participants were
significant scored lower on the likeability than other participants when faced by confederate
Fred, F(1, 55) = .02, p < .05. This can be caused by the higher level of participants with social
anxiety than other confederates or by the way this confederate responded to the participants.
Therefore, all analyses were done with and without the confederate. No significant difference
is found, or otherwise described.
Confederates on the scale of functioning of confederates
3.4 Differences between the HSA and LSA group on the likeability
A significant difference is found between HSA and LSA group on the likeability, measured
with the DFI, t(40) = 3.24, p < .01. The HSA group is less liked than the LSA group. To test
whether the increase in likeability differs between the LSA and the HSA group, a repeated
measures ANOVA was done with likeability (First Impression, Second Impression) as within
subject variable and group (HSA and LSA) as between subject variable. No main effect of
likeability was found, F(1, 40) = 1.37, p > .05. That is, there is no significant development of
likeability. The main effect of group is found significant, F(1, 40) = 11.1, p < .01. This means,
the HSA group received lower likeability ratings than the LSA group. No interaction effect
was found, F(1, 40) = 1.99, p > .05. That is, no difference was found in the increase in
likeability between the groups. The analysis without confederate Fred resulted in the same
conclusions. The development of likeability from the first to the second impression can be
seen in figure 6.
Difference over time at the first and second impression.
Left: with all confederates; Right: without confederate Fred
3.5 Differences between the HSA and LSA group on approaching behavior
Participants in the HSA group showed lower levels of approaching behavior than participants
in the LSA group, t(37) = 3.59, p < .01.
3.6 Correlations between likeability and approaching behavior
A high correlation is found between approaching behavior and likeability (SAB and DFI),
r(49) = .873, p < .01. Higher levels of approaching behavior related to higher likeability on the
DFI scale. A high correlation is also found between approaching behavior and the first
impression, r(47) = .602, p < .01. This means that the first impression relates to the following
approaching behavior of the participant. The correlation between approaching behavior and
the second impression was high, r(47) = .839, p < .01. This means that approaching behavior
relates to how the participants were liked after the social interaction. The difference between
the first and the second impression in relation with approaching behavior can be seen in
scatter plots, figure 7.
Scatter plots: first impression and second impression in relation with approaching behavior
3.7 The influence of approaching behavior on the likeability
To test whether the increase in likeability is moderated by approaching behavior in LSA and
the HSA group, a repeated measures ANOVA was done with likeability (First Impression,
Second Impression) as within subject variable, group (HSA and LSA) as between subject
variable, and approaching behavior (SAB) as a covariate. A borderline significant main effect
of likeability was found, F(1, 35) = 1.86, p = .18. That is, there is a tendency of development
of likeability by showing approaching behavior. The main effect of group was not significant,
F(1, 35) = 1.88, p > .05. That is, the effect of behavior on the development of likeability is
equal across both the groups. No interaction effect was found, F(1, 35) = 0.63, p > .05. That
is, no difference was found in the increase in likeability between the groups. Approaching
behavior as a covariate is found significant, F(1, 35) = 19.91, p < .01. That is, the
development of likeability can be explained by the level of approaching behavior. The
analyses without confederate Fred resulted in no significant results, but have the same
There is a correlation between the development of likeability and approaching
behavior in the HSA group, r(20) = .697, p < .01, and in the LSA group, r(16) = .523, p < .05.
That is, the development of likeability relates with approaching behavior in both HSA and
In conclusion, the HSA group has a lower level of likeability than the LSA group. Across
the two groups the level of likeability increased from before to after the waiting room
situation. However, this increase was only borderline significant. The development of
likeability was moderated by the level of approaching behavior, but not by the level social
anxiety (i.e., the HSA and LSA group). That is, the effect of approaching behavior on the
development of likeability is equal across both the groups.
4.1 Summary of the results
As described in the introduction, people with social anxiety are less liked by others compared
to people without social anxiety. Being less liked by others might well lead to an increase of
social anxiety complaints, less performance in social situations, more avoidance, and an
amplification of the social anxiety. In addition, having good contact with others is necessary
for good development on personal level, interpersonal level as on career level. A low
likeability might lead to maintaining the cognitive circles by people with SAD, like the
confirmation of the basic assumption (“I am stupid”, or “I am boring”), physical reactions,
self-focused attention, and safety behaviors. This will consequently lead to more impairment.
Thus, being less liked by others is an important issue in social anxiety and, thereby, seems to
be an important maintenance factor in SAD.
According to the hypotheses, as written in the introduction, the following results are found.
First, as was expected, HSA group was rated lower on the likeability than the LSA group.
Second, the HSA group showed lower levels of approaching behavior than the LSA
Third, it is found that the higher the level of approaching behavior the more the
increase of likeability from first to second impression. However, the change in likeability from
first to second impression was only borderline significant. In addition, it was hypothesized
that the effect of approaching behavior in the development of likeability would be reduced in
the HSA group compared by the LSA group. However, it is found in the study that the relation
between approaching behavior and the development of likeability is equal across the two
This study provides evidence for increasing the likeability by showing approaching behavior.
It is seen in the current study that the HSA group has lower levels at the first impression than
the LSA group. In addition, the HSA group shows lower levels of approaching behavior than
the LSA group. This approaching behavior has a strong influence as a moderating factor in the
development of likeability: high levels of approaching behavior lead to a higher likeability,
low levels of approaching behavior lead to a lower likeability. However, the development of
likeability is only borderline significant, what means that the level of likeability is hard to
increase during the conversation. This development of likeability is equal across both HSA
and LSA group, what means that this is independent of the levels of social anxiety. However,
the level of social anxiety indirectly corresponds with the development of likeability, because
of the low levels of approaching behavior showed by the HSA group. In addition, a low
likeability is suggested as a maintaining factor on the social anxiety. However, this is not
investigated yet. It is important to know for people with SAD that, although, the first
impression is important, approaching behavior has an influence on the development of
likeability. With a high likeability, an individual with SAD will learn that the basic
assumption of being stupid or boring will not have to be correct, and, thereby, the level of
impairment might decrease as a consequence of a change in the cognitive vicious circles.
4.2 Links with earlier described theories
As written in the introduction, it has been seen that people with social anxiety are being less
liked, warm, relaxed, attractive, friendly, interested, and less likeable than people without
social anxiety (Meleshko & Alden, 1993; Pilkonis, 1977, Voncken et al., 2008). Thus,
individuals with social anxiety seem to receive more negative reactions by others. This will
probably lead to more social rejection, and this is also the core fear of socially anxious
individuals. This increases their urge to show safety behaviors what leads to more
inappropriate social behavior. All this leads to the confirmation of the basic assumption and
core fear of not being liked by others (Alden, 2001). The current study corresponds with this
theory of a low likeability as a consequence of certain behavior. It is seen that the participants
of the study, the group with high social anxiety, are less liked by others than the control group.
They are scored significant lower at the first and second impression than the controls
participants. Even though there is made use of a subjective scale to measure the likeability,
there is found a high inter rater reliability for these impression assessments. Moreover, the
HSA group show lower levels of approaching behavior than the LSA group, what corresponds
with the idea that people with SAD show other behavior than the people without SAD, for
instance safety behavior and avoidance behavior. In addition, Collins and Miller (1994)
described three effects of self-disclosure: people who self-discloses 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. As described in the introduction, self-disclosure can be seen as a form of
approaching behavior. A person approaches by showing certain types of behavior, such as
self-disclosure. In line with Collins and Miller (1994), we have seen in this study that the
likeability increases with the level of approaching behavior, and so also self-disclosure. Thus,
people with SAD show more avoidance behavior among with less self-disclosure than people
without SAD, what leads to a lower likeability.
Furthermore, the theory of Alden (2001) is in line with the current study. Alden (2001)
describes an effect of behavior on the likeability. This interpersonal model describes people
with SAD that are in a vicious circle of a negative social expectation, leading to avoidance
behavior and, therefore, a lower level of likeability. Based on the findings in this study and the
theory of Alden (2001), it can be concluded that due to their avoidance behavior, a person
with SAD has actually reason to believe that others do not like them. Therefore, they find
evidence for their basic assumption of being less nice or their main fear of being rejected.
When people with SAD would approach more and avoid less, they probably will be liked
more, what they will notice and so develop less anxiety in the interaction with others.
Moreover, Thompson and Rapee (2002) and Papsdorf and Alden (1998) also describe
a deficit in the behavior of people with SAD. Thompson and Rapee (2002) suggest that safety
behaviors maintain social anxiety. Papsdorf and Alden (1998) suggest low rates of selfdisclosure
leading to a low level of likeability.
In conclusion, according to previous studies and the current study there are found
deficits in the behavior of people with SAD. Deficits are found in the approaching behavior,
self-disclosure, avoidance behavior, and safety behaviors. These deficits lead to a lower
likeability, what, in turn, might lead to the amplification of the anxiety and more impairment.
4.3 Clinical relevance
The conclusions of this study lead to changes in the clinical settings for treatments for people
with social anxiety. Therapy nowadays focuses on the cognitive circles of social anxiety. This
lead to improvements in the complaints, but still a lot of impairments remains. The knowledge
of the interpersonal perspective and the current study can be useful to improve current
therapy. As seen in this study, the development of likeability is independent of social anxiety.
However, the HSA group has lower levels of approaching behavior than the LSA group.