Wednesday, October 18, 2017

Psychological Wellbeing of Students at the Lahore University of Management Sciences

Use of Social Network Analysis to Provide Better Mental Health Support to Students at the Lahore University of Management Sciences (LUMS)

Background: The Lahore University of Management Sciences (LUMS) is the premier research university in Pakistan. It was founded in 1984 by the benefaction of the business industrial community. LUMS established the business school in 1986 with advice from the Harvard Business School and $10 million grant from USAID and since then, it has established an undergraduate liberal arts school, an engineering school and a law school. LUMS also launched a program, under the name of the National Outreach Program, to provide financial aid to students coming from remote underdeveloped regions of the country and to those who can otherwise not support their degree due to financial constraints. In order to diversify its student body, it collaborated with the Afghan government to provide scholarships to Afghan students to come and study at LUMS[1]. Therefore, LUMS hosts a diverse set of student body (4000) ranging from different cultures and ethnicities, social setting and conservative to liberal backgrounds.

LUMS, where I studied from 2011-2015, is the country’s most expensive, selective and progressive university with high academic standards and integrity. The university has been known for its conducive and ‘healthy’ environment, however, over the past few years the incidence of mental health issues among the students has been growing. This is quite alarming as this year alone LUMS witnessed two suicides along with one attempt to suicide and 5 drop outs[2]. Students have also begun to come out in the open on social media platforms about their growing anxiety, depression and stress issues. During my time, I had personally witnessed the metal health of fellow students deteriorate and had heard of a few suicide attempts. Although a few of these reported issues have been attributed to family or medical emergencies and sexual harassment, some have been related to drug abuse, sexual orientation and depression which has been traced back to the stressful environment and the inability of students to adjust[3].

LUMS attempted to address this issue by taking a few immediate steps in 2015: First, the institute an anti-narcotics raid was launched to have a ‘clean university’. The institute had an intolerable policy towards drugs. Night wardens were employed so that they could keep tabs on the consumption of these illegal items in the dormitories[4]. Students who were suspected of consuming drugs had to get urine tests conducted. Second, the university got a counsellor on board, whom the students could visit regularly for their mental, emotional and psychological well-being. Third, the university instituted a sexual harassment policy with strict actions taken against harassers[5]. 

While it is pertinent to have a regulated school environment with code of conduct requirements and a psychological support mechanism in place, it is also important to nurture student support networks to help sustain challenges through these years. Prevailing literature on this topic establishes a direct relation of student support networks with healthy, depression –free school years. LUMS itself has been keen on fostering friendships among new students by dividing them into groups during the orientation week to engage in lots of activities.

Primary Question: What is the correlation between weak social support networks and dropouts/ students with psychological issues? How can Social Network Analysis be used to identify these students and to foster stronger connections with their peers?

Secondary Questions: What is the network of these students? Do birds of feather flock together? Are connections made on the basis of same school/degree, regional associations or extra curriculars?
Are there certain students who can leverage their connections to develop stronger connections?

Hypothesis: Students who have a bigger or stronger network within LUMS are more likely to adjust and cope up with the challenging environment and less likely to fall prey to depression and psychological issues. These students complete all years of their degree program.

Core courses, group projects, extra-curriculars and similar backgrounds (belonging to same region or high school) are known to foster friendships among students at LUMS.

Data Collection and Methodology: Ideally I would want to conduct surveys with students in their freshman year first and then with the same students in their last year to assess how their network has evolved and which factors have attributed to it. As that is not realistically possible due to the long time lag I would want to conduct a survey among students in the second semester of their sophomore year. This is a good population as students in their second year have had sufficient time and exposure to this new diverse environment and can articulate if their friendships from the orientation week sustained. Moreover, sophomores would be useful in testing the hypothesis as they would account for the role of co-curricular activities, group projects in evolving their network.

In order to maintain confidentiality, students would be ascribed with aliases. In my experience students would willingly fill out this survey as they are concerned with the emotional and mental well-being of their peers, and most of them have gone through a similar ‘rough’ patch while some passing with flying colors and others struggling in the quagmire of depression.

Some sample questions of the survey would be:

- Where are you from?
- Which high school did you attend?
- What is the strength of friendships/connections on the basis of talk/meet frequently, less frequently, rarely, never
- How many friendships have you sustained since your first year? Who? At what level?  (For 4th year students)
- How many friendships have you developed from your group projects? Who? At what level? 
- How many friendships have you developed by being force enrolled in core courses? Who? At what level? 
- How many elective courses do you take in a year with group projects? None, 1,  > 2, > 4, > 6
- How many friendships have you developed from your extra-curricular activities? Who? At what level?
- Do you identify yourself as minority of religion, cast, ethnicity, sexual orientation, financial status?
- What is your spoken proficiency of the English language? Native, professional, conversational, basic, none
- Have you experienced feelings of loneliness, depression, stress, culture shock? Yes No
- Do you talk to your peers/friends or do your peers/friends help you out on the above issues? Often, sometimes, rarely
- How frequently do you visit the counsellor? Regularly, sometimes, rarely, never?
- How helpful is it? Very, little, useless

I would first analyze the data to look at the friendship networks at various levels by applying some whole network measures such as density and average distance to give an idea of how dense the overall network is and what is the level of interconnectedness among various friendship groups. After this I would analyze the data on a subgroup level by Girvan Newman to find out the more closely connected friendship groups. Applying some centrality measures on these subgroups would give an idea of who are the bridges and eigenvectors in these networks.

I would also be interested in noting how students are connected to other students via certain extra-curricular activities/clubs and students connected to other students via group project courses, therefore, developing 1 to 1 mode matrices would be helpful in this case. Which students are emerging as informal leaders in those? It would be interesting to note if students who were emerging as leaders or influencers were also on the student council. Then I would find the overlap of friendships and group projects and core courses to assess the importance of these tools in fostering closer ties.

I would also look into how cliques are formed-  are they on the basis of high schools, core courses, ethnicity, language? Do students having feelings of loneliness or depression forge deeper connections with one another? Application of node centrality measures will allow me to assess if there is a direct correlation between students having fewer connections and having feelings of depression, culture shock, loneliness. If so then do these students share attributes of ethnicity, cast, language, etc.? I would also like to determine if these students visit the counsellor, and those who visit the counsellor frequently how many connections/friendships are they fostering. This would explain the effectiveness of the school counsellor.

To directly test the hypothesis, I would find out if students with a high degree and eigenvectors have feelings of isolation, depression, culture shock. If this does not hold true, then students are less likely to drop out.

I would pay close attention to students who have identified themselves as minority. By looking into their ego network I would get an estimate of their number of connections and the strength of their ties. If these students have many friends or strong ties, then this would support the hypothesis. However, it would be interesting to notice if these students have formed clusters on the basis of their minority, if they share negative feelings and if any one of them has a high eigenvector. This would make them ‘negative influencers’ and translate into more negativity being spread.

Recommendations & Conclusion: SNA would be helpful in leading towards an organizational change in LUMS. Students who emerge as eigenvectors or influencers should be nominated on the student council and should be asked to hold weekly sessions where other students can come up to discuss problems. If courses with group projects are helping to foster relationships, then the design of some popular courses should be changed to include group projects. Similarly, if students with core courses have greater/stronger friendships then more of the degree programs should have at least a few core courses. The analysis would also demonstrate the effectiveness of the counsellor and if there is a need to replace him/her or give him/her training on how to be an effective therapist. Similarly, the effectiveness of extra-curricular activities would also demonstrate if it is necessary for the students to be a part of at least one club.   If language comes out to be a strong barrier, then LUMS should also hold regular English classes after school hours, as the medium of instruction is in English and students coming from remote areas are not fluent in the language at all.



[1] https://lums.edu.pk/
[2] This information is not publicly available but is revealed by the LUMS admin
[3] This information is not publicly available but is revealed by the LUMS admin
[4] http://dailytimes.com.pk/features/27-Dec-16/to-lums-youve-got-it-all-wrong?utm_campaign=shareaholic&utm_medium=facebook&utm_source=socialnetwork
[5] https://osa.lums.edu.pk/content/student-code-conduct

1 comment:

Christopher Tunnard said...

Interesting and worthwhile project, especially after the rather confrontational approach taken by the school (e.g. urine tests.) Your SNA approach is certainly extensive, although in a shotgun kind of way, as you suggest a range of survey Qs and analytic tools without linking many of them to the recommendations (which make good sense, by the way.) Also, some of that are linked are difficult to understand, e.g. "I would find out if students with a high degree and eigenvectors have feelings of isolation, depression, culture shock. If this does not hold true, then students are less likely to drop out."

My other major point is that you need to define what you mean by strong and weak support. A clearer definition would help you refine your net Qs and your analysis.