What Type of Social Media Users are More Likely To Have Social Anxiety Disorder?

It is very common nowadays for people with mobile smart phone as well as on social media. For all these advantages of the technologies, People are no longer require to see each other physically, yet still able to obtain latest or even more information about the people in your your social circle on your mobile, it either via Social Media or Instant Messenger Application. Hence, few year down the line, there is a phenomenon that people rather to communicate digitally (CMC - Computer Mediated Communication) than face to face interaction (FTF - Unmediated Face-To-Face).

I was wondering is this behavior changes simply because of the conveniences of the technology and people has become more Introvert or Social Anxiety now compare to 10 years ago? Or the rise of the Social Media attracted the group of people who are already Introvert or Social Anxiety. Interacting face-to-face are basically too complex for them, where they might have forgotten how to communicate in person, not able to withdraw or edit the sentence already said; not sensitive enough to read the room how other people feel or even the body language of the person they are talking too. Regardless what’s the root cause or the factor of it, it is also interesting to find out which segment of the population on social media might have Social Anxiety Disorder.

The objective for this blog is to understand which group of Malaysia University student are more vulnerable to Social Anxiety Disorder. A survey had been conducted by using google form and distributed to several locations with high number of university students. Total 93 surveys are collected with 69 are students and 24 are no students.

A very comprehensive paper about the use of Internet and Social Anxiety was published by Shiri Prizant-Passal et al. [1]. In the paper, the author performed meta-analysis on approximately 23 similar papers that published by researchers. The scope of Shiri’s studies are focused on the following four aspects, which is (1) Perceptions of CMC: comfort online, (2) Problematic internet use and social anxiety, (3) Developmental aspects, (4) Behavioral preference for CMC: time online.

From all the interesting aspects above, the item (4) behavioral preference for CMC: time online particularly aroused my interest. According to the Shiri’s paper, he mentioned there number of studies or researches about the social media and social anxiety disorder are quite limited, this is why the data collected for the meta-analysis is mainly about the number time online and not the time on social media. Hence for my study in this blog is focusing on the number of time on Social Media against Potential of Social Anxiety Disorder to specific group of people.

Data Collection

The survey had been executed approximately 1 months by using google form and total 93 feedback collected. The survey form consists of basic personal information like age range, gender and Marital Status and Number of hour spend on social media per day. Then follow with another 6 questions in Likert-Scale method for behavioral questions as below.

  1. How do you feel being the center of attention
  2. Working while being observed
  3. Talking face to face with someone you don’t know very well
  4. Expressing disagreement or disapproval to someone you don’t know very well
  5. Are you extremely conscious of your actions when in social settings because you fear they might offend someone or you could be rejected?
  6. Do you feel anxious or panicky before social situations?

Candidates must select the answer from 1 (Very Comfortable) to 5 (very Uncomfortable) for all the 6 Liebowitz Social Anxiety Scale (LSAS-SR) questions above.

Data Processing

The data collected from the survey are quite clean, hence it does not required much cleaning. However, the existing data structure from the survey is a little bit restricted for the further analysis and minor changes as below. Figure 1.0 above shows the count of the answers for each questions. For example more people selected Neither Agree nor Disagree (3) and Uncomfortable (4) in Question 1. For Question 2, majority people feels Uncomfortable (4) and Very Uncomfortable (5). First change is to sum up the total scores for all the questions for each survey collected.

The 2nd change is to categories the score to None, Mild and High Social Anxiety. By referring to the Figure 2.0 above. Total score is 6 if the candidate selected 1 (Very Comfortable) for all the 6 questions (6 x 1); Or total score is 30 if candidate selected all 5 (Very Uncomfortable) for all the 6 questions (6 x 5). Converted the score for each individual to categorical data like None, Mild and High. The benchmarks for those categories are set as Mild at 18 and High at 24. The benchmarks was decided according to the total scores and also by referring the distribution graph below. Among the total 93 surveys, the minimum total score is 13, maximum total score is 30 and mean is 20.83 (blue). From the graph below, it shows that the area less than 18 and more than 24 quite similar to the lower quartile and higher quartile. Hence, following benchmarks are being used

  • None: 0 to 17 scores
  • Mild: 18 to 23 scores
  • High: 24 and above

Due to the mean is smaller than median (on the left of the median), the bell curve is slightly skew to left by 0.172.

Who They Are?

4 pie charts below (Figure 4)show high-levels type of data of the collected 93 surveys. Due to the large population of the participants are students (74%), where majority also single (78%), the diagram below also shown up to 59% are female. The chart on the most right indicates the number of hours spent on social media. 46% (43) of candidates spent more than 5 hours per day and only 11% (10) spent up to 1 hour per day.

In this blog will only focus on Student sample. By comparing to Figure 4.0 (All candidates) above, there are more Female (42%) and Single (87%) candidates in Student Sample. Besides, the number of house on social media are higher (54%)when only student sample data.

Age Range

In this section, we are going to explore more by age. There are 2 diagrams in Figure 6 below, the left diagram is the bar chart that showing the number of participants for 3 different age group, which is ≥20, 21–30, and 31–40. Due to the participants are mainly student, where most of them are around the age between 21–30. The left diagram is the box-plot diagram. In order to make it more understandable, few items added to the diagram. The yellow box at the bottom is the one with total 93 participants and the white boxes are only the 69 students.

The yellow line on the most left (18 scores) indicates the start of mild social anxiety range, the most right (24 scores) indicates the start of high social anxiety range. The blue (20.828) is the mean of the total 93 samples. By simply looking at the yellow box, it matches with the Figure 3 bell shape distribution chart above, where the mean before yellow box’s median and the right whisker is longer.

According to the box-plot in Figure 6 above, the students under age 31–40 shows that they are less social anxiety as the maximum value is still lower than the high social anxiety benchmark (24) and close to 75% of the participants are below the sample mean 20.828. Students between age 21–30 are those group of people with highest social anxiety disorder where close of 75% are mild and high. However, the entire group of student age 20 and below are mild and high, but luckily the less than 25% are in the high. From the box-plot we got to know the younger students will be higher chance to have social anxiety disorder.


Refer to the Figure 8.0 below, it does not have very obvious different about the skewness for all gender and candidate type, even the median line is approximately same and closes to the mean point.

Male student generally are less social anxiety than the Female student. As the min and max point of Male’s box plot are fall between Female’s. However, by comparing the whiskers between male and female, you can see Female student has longer lower whisker then Male student; and Male student has longer upper whisker compare to Female student. Which is mean although Male students are less social anxiety then Female, but in general majority of them are tend to be more anxiety then Female (Majority of Female student are less social anxiety).

P/S: there is one outlier in Male Student’s box-plot

Marital Status

Marital Status box-plot below (Figure 9) is kind of align with the interpretation above regarding age, where we assumed younger student are majority single.

Hence it is not surprised to see the Married Student are less Social Anxiety than Students who are Single. For Married Students, only less than 50% are fall between the mild zone, where for the Single Student are more than 75% are fall between Mild and High Zone.

From the information above, we can make say that Single Female students around age 21–30 have high chances for Social Anxiety Disorder. In the next section, we will be looking deep into the correlation between the social anxiety disorder and the number of hour on social media.

Correlation between Hours on Social Media and Social Anxiety Disorder

Linear regression and Spearmen‘s Correlation (Confident Level: 0.95) will be used to determine if the number of hours on social media and social anxiety disorder are correlated? Figure 10 below shows 2 linear regression plot for students and none-students. From the linear equation y=0.145x + 20.257 (student) and y=0.903x + 18.013 (none students). According to the Student Regression line on the left, it shows that even the students only spend less than 1 hours on social media, the social anxiety score is 20.257 which is fall in the mild zone and yet the slope of the line is not significant. However for None- student the slope of the line is more obvious. Spearmen’s Correlation is being used to determine the correlation between hours on social media and social anxiety disorder. The rho for all participants (rho:0.1624), all students (rho:0.0806) and None Student (rho: 0.3691) are all consider weak correlation.


According to the result from the surveys, we can conclude that the correlation between the number of hours spend on social media and social anxiety disorder is weak and not significant. However, the student population segment who might have high chance to have Social Anxiety Disorder are Single Female Student between age 21–30.

R Code and Files

For those you are interested on this project can find my R code on github and dataset on Kaggle.


[1] Shiri Prizant-Passal, Tomer Shechner, Idan M. Aderka — Social anxiety and internet use — A meta-analysis: What do we know? What are we missing?,
Computers in Human Behavior, Volume 62, 2016, Pages 221–229, ISSN 0747–5632, https://doi.org/10.1016/j.chb.2016.04.003.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store