Mapping Social Networks
“Invisible threads are the strongest ties.” Friedrich Nietzsche
Social mapping and the study of network theory offers marketers and organizational leaders a deeper understanding of how people are likely to behave based on their interactions with each other. This understanding gives context and meaning to individuals’ personal data during a specific time period.
In creating a social map of the Ballarat Wellness Study participants, I intend to identify the key roles of participants within the social structure of this study and recognize how these roles can be utilized for the benefit of their entire network’s long-term health. Although most social maps are a complex set of data points and algorithms generated by a computer program, my work is far more simple form; quilted together between Excel spreadsheets and Photoshop layers.
The social mapping I have created is a visual analysis of data that shows how subjects in this study have interacted with each other. Individuals are represented as nodes connected by a set of social interactions or links. Each node is an aggregated data set of each subject’s motivation level, social and online engagement, as well as their mental and/or physical health improvement throughout the program. The links between each node are based on specific data collected about the positive, supportive, or competitive interactions of every subject.
Data collected from weekly surveys, BallaratWellness.com comments, Fitbit Forum posts, the initial and final health surveys, and other communication channels was organized and aggregated into a basic spreadsheet saved HERE.
Creating individual Nodes: Assigning Scores & Aggregating Data
The first worksheet is a compilation of the total online, social, and motivation scores of each study subject. Scores of each criteria are not concise measurements, but are based on the effort involved to do each task. Higher points were given to tasks that required greater effort to complete.
There is a challenge in comparing the difficulty of completing a social task to an online task or either to a subject’s motivation level, therefore these scores are not meant to compare against each other. One subject’s motivation score can be compared to another’s motivation score, but not compared to their own social engagement score.
Online Engagement Score:
Online scores are points given to subjects for their interaction with the online program through BallaratWellness.com, Fitbit.com, or by completing the weekly surveys.
Each week, subjects were emailed the a link to a 5-to-8 question survey on SurveyMonkey.com. The survey questions were also pasted into the body of the email. Subjects could submit their surveys through SurveyMonkey.com or reply to the email with their answers in the body. Since each subject’s email address was included in the email’s BCC field, subjects could reply to the email with their questions and have it only be received by email@example.com.
The number of points went up by 1 every week, to reflect the decrease in motivation to complete a survey each week. Subjects also received 10 points for completing the final 36-question health survey in Week 8. To complete the health survey subjects had to log onto a different site and take the long survey on a computer that could only open on Internet Explorer 8 or earlier. The current version of Internet Explorer is 10, and most Macs and Apple devices from the past 2 years are not able to access IE 8. This meant that some subjects had to fill out this deeply personal health survey at work or on a friend’s computer if they did not have an older device or PC.
Furthermore, subjects were given a point for each time they commented on BallaratWellness.com or the Fitbit Forum. One point was also given to each subject if they became Fitbit “Friends” with my account during the program. Being friends gave me access to view their activity points and food log.
Five points were given to subjects who submitted their official Fitbit activity report at least once in a 4 week period. Subjects could not self report their activity reports, but had to sent me a report the was either downloaded from the site or emailed from Fitbit.
Each week, subjects were asked to rate their motivation level on a scale of 1-to-10. The question was asked in a way that reflected the subject’s physical effort as well as their emotional motivation.
The exact question was asked as follows:
Currently, how motivated are you to improve one or more of your health behaviors? Select one number.
1 = pulling the covers over my head
5 = making the effort, but not excited about it
10 = ready to climb Mt. Everest with a smile!
Subjects were given the exact number of points that they listed as their responses each week.
In each weekly survey, subjects were asked if they had spoken to anyone about their Wellness Goals in the previous week. If the subject answered that they had, they received a point. It did not matter to how many people the subject socially interacted with, what communication channel they used, or if the conversation was positive.
Most people responded to this survey question with an account of exactly who they spoke with and how they communicated with them (for more on this, visit the Communication Channels & Messaging Distribution page). With this detailed data, I was able to conclude the average number of people each subject who had socially interacted with others that week.
Subjects were given 5 points if the average number of people they interacted with that week was 2. If the average number of people subjects interacted with was 3, each subject who had answered that they had some kind of social interaction was given 10 points.
At the end of Week 3, many of the subjects planned a lunch break wellness session, where I presented to them some initial findings via Skype. Subjects were given 10 points for attending this session or emailing me for more information about my presentation if they could not make the session themselves.
Finally, subjects were asked to list the first and last names of people who had been health allies, Fitbit competitors, or all-around positive influences throughout the study in the Week 7 survey.
The exact question read as:
Please name the people in this program who you consider to be your health allies, Fitbit competitors, or all- around positive influences in the last 7 weeks of this study? Include FIRST and LAST NAME if you know them. **All names will be kept confidential, I am measuring the social connectivity of participants in the study.**
Subjects were given 2 points every time another subject of the opposite sex listed them. Subjects received 4 points for every time another subject of the same sex listed them. This is based on Christakis and Fowler’s (2007) study on the spread of obesity in social networks. They found that the likelihood of someone becoming obese increased by 71% if their same sex, mutual friend was obese. This probability actually increases to 100% for men . Gender may not have any affect on all of the different wellness goals subjects had in this study. I wanted to see, however, if same sex links had any noticeable affect on subjects physical and/or mental health.
Designing the Ballarat Wellness Social Map
Using the data collected through this process for each individual I was able to create a node and link for each subject and corresponding interaction. As shown in the KEY, each subject’s Motivation, Social, and Online Engagement scores are reflected by the size of the black, pink, and grey circles respectively. The circle size matches each score in size of pixels across the diameter. For example, if a subject has a Motivation score of 30, the number of pixels across the diameter of the circle is also 30. The smaller circles are on top, showing the relative size of each compared to the other subjects around them.
Subjects are connected to each other by mint green or blue line. A mint green line indicates that the two subjects are the same sex, while the blue line indicated the subjects are opposite sex. There are no values associated with the size or length of each line. I simply tried to fit each connection onto the page in a readable, 2-D format which caused me to make adjustments to the size and length of each line.
Radiating behind certain subjects are bright green, purple, or yellow circles. These circles have no data represented by their size. I choose to have the health improvements graphically radiate from the subjects because these improvements manifest subtly in social interactions. We radiate wellness when we are healthy, whereas the motivation, social and online engagement scores represent specific action, emotion, and participation in the program.
The bright green radiating circle appears behind all of the subjects who had improved only their physical health during the study, according to the initial and final health surveys. With the bright green radiating circle, the subject’s mental health may have declined or been unchanged.
The purple radiating circle appears behind all of the subjects who had improved only their mental health during the study, according to the initial and final health surveys. With the purple radiating circle, the subject’s physical health may have declined or been unchanged.
Finally, the yellow radiating circle appears behind all of the subjects who had improved both their mental and physical health during the study, according to the initial and final health surveys. Again, the size of each radiating circle does not reflect the extent of how much the improvement is made.
Immediately, the eye catches a large pink node in the center of the map. This subject has socialized with a considerable amount of people in the program and has increased their physical health to some degree. They were not, however, able to improve their mental health and may have had their mental health decline.
Like this central node, others who have improved their physical health and not their mental health also have multiple links to other nodes, although not to same extent. There is no discernible trend with nodes that radiate the bright green of increased physical health, most of them have an average motivation and online engagement score, except that they all have at least one link to another node. None of the nodes towards the outside, without any links have a green hallo.
At the same time, many more of the outer unconnected nodes have achieved some kind of improvement to their mental health with no physical health improvement. Although many of the nodes who are connected to others have also improved only their mental health, it is striking how many “loners” there are who have improved their mental health. Maybe improving their physical health wasn’t a focus, or they felt uncomfortable joining an exercise group with their co-workers. There is very little data to make any assertions. Many subjects who had not socialized with others in the program, but who may or may not have improved their mental health, had explained that they worked in different buildings and departments from other subjects in the program or that they had support from other sources such as Weight Watchers, personal friends, or fitness groups outside of the program.
Nodes who have improved both their mental and physical health in some way radiate a sunny yellow glow. Some of these nodes are disconnected, while others are a part of social clusters. There are others who are only linked to one node, but the node they are linked to is at the center of a larger social cluster.
Why is it that some of these nodes with the golden auras have hardly any motivation scores, below average online engagement, and considerably fewer links than other nodes who did not achieve the same improvements?
As expected, there were several clusters of nodes linked by the same gender. A greater number of participants in the program were women, which affects the concentration of same-sex links. With this in mind, many of subjects who socialized with others both of the same and opposite gender had improved their mental and physical health. However, this may be an indication of more socialization in general.
Attitudes, Behaviors, and Social Norms
As mentioned on the Communication Channels & Messaging Distribution page, we are a social species that has evolved to physically, emotionally, and mentally benefit from social well being. We seek out competition with others as a means to relate with those competitors, just as we seek emotional support and inspiration through subtle interactions each day.
For this reason creating small but far reaching, long-term behavior pivots in a social network are far more important than creating larger changes in individuals’ personal beliefs and attitudes. As many dieters attending holiday parties can attest, unless a personal belief, attitude, or resolution is extremely important to them a person will forego their own opinions and exhibit the behaviors reflected by those around them.
In Behavioural Conflict: Why Understanding People and Their Motivations will Prove Decisive in Future Conflict (2011), Andrew Mackay and Steve Tatham explore how the social behaviors of specific populations can be changed to avoid or quickly resolve military conflict. Mackay and Tatham write:
A social norm is defined as a behavioural expectation or cue within a society or group… The stronger the norms in a group of society, the more highly will people conform. (p. 171)
Mackay and Tatham use an example of how a Harvard psychologist in 2009 found that a popular radio soap opera in Rwanda promoting conflict resolution between the Hutu and Tutsi impacted social behaviors while having little effect on the population’s personal beliefs. This research concluded that because listener’s perceptions of social norms changed, so did their behaviors in respect to intermarriage, open dissent, trust, empathy, corporation and trauma healing even though each individual’s personal opinions were quite the opposite (p. 172).
Once behavior changes are made to reflect new social norms, individual attitudes are likely
develop over time. Mackay and Tathm note that in Britain drivers’ attitudes about wearing seat belts were largely negative until after they began wearing them. Understanding this concept of how change occurs is crucial to supporting a long-term, scalable wellness program. Exactly how to follow this concept is debatable and currently being tested in a multitude of ways from adding a tax on tobacco products to selling smaller soft drink sizes in New York City.
Making dramatic changes to social norms do not have to be institutionalized to be effective and can actually hinder change by causing controversy, although having the support and cooperation from large institutions is likely increase efficiency. To creating long-term behavior change Managers, Leaders, and Organizers of programs must focus on messaging to, gathering data on, and interacting with groups of people and the roles individuals play within those groups more than the individuals themselves.
Developers of health apps and activity monitory devices are eager to point out that highly detailed data about individuals’ online engagement combined with their personal health stats can play an important role in effective behavior change. Especially when empowering and motivation individuals to be mindful of their wellness. The value of this information is lost, however, if users and developers alike are unable to place this data in a social context.
To create effective, long-term behavior change the organizers of programs need to:
- Identify and quantify the social network of the population they are working with so they can further understand how individuals, groups, social clusters, sub-cultures, or tribes of people within this population affect each other.
- Understand how the social norms and common personal beliefs contradict or support the desired behaviors needed to achieve specific goals in various common settings. The most impactful social norms that contradict or support the desired behavior should then be analyzed to identify subtle changes in these norms within the smaller social clusters.
- Effectively communicate and engage with individuals by messaging to their social clusters. Empower those who belong to multiple clusters to share information and support between the clusters.
All of this requires time, a progressive data capture strategy and analysis, as well as a deep understanding of network theory. In many ways, the work I have done in this study is just beginning to creating a truly effective wellness program. The most common way to collect data throughout this study was to rely on responses from the weekly surveys. However, in a commercial program, weekly surveys may annoy users and generate a lower participation rate.
Integrating Google Analytics into the site is the easiest way to capture real-time data on the general trends and actions of users on the site. For this study, I hosted BallaratWellness.com on a WordPress site because it was easy to build a visually engaging site with little coding knowledge. A drawback to hosting sites on WordPress is that their sites do not allow for Google Analytics integration. WordPress offers it’s own “Site Stats” tools, but they are extremely vague compared to Google Analytics.
CrazyEgg.com is a new company that offers websites user analysis tools complementary to Google Analytics through heat-mapping technology. The service is free, and shows where users aren’t going on the site as opposed to Google Analytics’ metrics of where users are going. Again, due to my lack of coding knowledge I wasn’t able to hack CrazyEgg’s tools into my WordPress site.
Google Analytics and CrazyEgg don’t give any data about how users are interacting on the site or with each other offline. One way to track this is to create a Facebook group or Twitter account that uses hashtags to track online sharing. Another way is to encourage people to join wellness groups such as Weight Watchers or a tennis club and manually track their participation through a partnership with the program or an online poll.
By using data collected in weekly surveys applied to social mapping, I may not have created the clearest picture of my study’s social network, but I was able to determine different social clusters within the program and identify the roles subjects played within the study and their smaller clusters.
Communication Roles within a Network
In a social network, individuals, organizations, and entire societies can be identified as a single node. Links between nodes may depict an array of relations, including the communication type, work flow, proximity, and cognitive tie to name a few. The links may also denote the direction and strength of the relation.
With so many different kinds of social networks, nodes, and interactions studied, a social map can quickly become quite complex. To understand the behavioral implications on the individuals of a given network structure, researchers apply the concepts of communication networks.
Communication networks are patterns that appear in all types of communication within a social or organizational network. These patterns vary widely depending on the communication type and context of the network; some are ideal for specific circumstances within a network and poorly suited in others. For example a centralized communication network, where orders are directed from the top down in an organization, is more effective when accomplishing simple tasks like putting out a fire or short term weight loss. Decentralized communication networks, where everyone participates in communication with multiple nodes, are more effective when accomplishing complex tasks or tasks with moral implications, like passing legislation on gun control or preventing the risk of Type II Diabetes.
The City of Ballarat has multiple communication networks for different departments of their organization. Garry Davis, Executive Manager of Organisation Services and Development explained that most of the full-time employees are encouraged to have multi-directional communication with each other and their superiors through email, office exchanges, meetings, etc. Other part-time employees, many of whom do road construction for the city, have a single-direction, centralized communication with their superiors through pre-shift announcements or by having notes stapled onto their paychecks.
My 8-week program was promoted to many employees in announcements through email, office conversation, and notes stapled onto paychecks. Once study subjects showed interest in the program, they were communicated in both multi-directional formats such as the Fitbit Forum, limited interaction such as the weekly surveys, and single-direction formats such as email announcements throughout the program.
Segmenting participants in an organization is the most effective way to manage communication in a decentralized communication network because it increases individuals’ participation. Working with a group of employees who are already broken into departments within an organization, it is easy to create smaller groups for people to work with each other and socialize their wellness goals. However, when making a scalable online wellness program, there are no pre-set groups to work with. Many health apps and websites like Earndit.com or Fitbit.com make it easy to find “people like you” to communicate with about specific goals. Earndit.com lets the users compare their stats to other users by city, gender, and contacts users have already made on the site. Fitbit.com allows users to join or create their own forum groups based on their interests and goals. Users are then able to invite others to the forum, make the forum private, or open the forum up to all Fitbit users.
Increasing individual engagement by segmenting the communication network into smaller groups will only be effective if those smaller groups are in communication with each other. To ensure this, we must address the roles of individuals in a communication network.
As Lunenburg (2011) writes, there are four basic communication roles that individual nodes play in the varying degrees of centralized and decentralized communication networks within an organization:
- Gate Keepers: Individuals who have control over the flow communication.
- Liaisons: Individuals who connect two or more cliques.
- Bridges: Individuals who connect one clique to another.
- Isolates: Individuals who have limited communication with the network.
Each kind communication role can be used to further develop the acceptance of social norms that cultivate healthy behaviors. The way these roles form and interact with each other directly create the patterns of communication that affect behavior. These patterns can then be broken into small segments, such as the different departments within an organization or a group of co-workers who sit near each other.
I have identified the smaller groups within the Ballarat Social Map as “cliques”. Based on my research, it is unknown how these cliques formed, how long these cliques have lasted, or the value of importance these cliques play for each individual.
Figure 1 shows how the different roles of individuals may appear in a more structured organization with preset departments and groups. My social map, however, reflects the natural interactions of participants in the program so the roles are not as clearly defined. However, the preset roles of employees and departments at the City of Ballarat have certainly affected my social map in some ways. Many of the cliques are made of people who work in the same department, commute to work together, or sit next to each other at the office.
For instance, the Gatekeeper I have identified in Figure 2 represented by the large pink node in a central position, worked with me on the study to distribute the Fitbits and make sure everyone was receiving my emails. This person was not a Gatekeeper in the traditional sense that others would not have been able to communicate without this role, but a great many of the subjects relied on this person when they had trouble logging into Fitbit, filling out the health surveys and so on.
The Gatekeepers in a social network are the easiest to spot as they are either in positions of power or related to leaders in an organization. Their role is messaging distribution, they determine what messages are the most important to distribute at the right moments and how to transmit this message clearly. Marketers know Gatekeeprs as Brand Ambassadors or people with the highest social distribution. Sites like Klout.com offer ways to score people based on their distribution networks, engagement, and interaction with others in a social network. Gatekeepers will likely have the largest distribution channels, using these channels often.
When directing the messaging within a network, it is important to empower the Gatekeepers at the center of clique with a clear idea of how they can incorporate positive messaging throughout their network. Although Gatekeepers have the largest audience they are not always the most important spreading social norms and behaviors.
Bridges and Liaisons may not be the most active in a social network, but are often the most important in a communication network. Instead of leading a conversation in one particular social group, Bridges and Liaisons take part in an array of conversations in multiple cliques. Although they have a smaller messaging distribution, their communication is crucial to the spread of behaviors because they share the social norms from one cluster to the next. This role is especially important for cliques that may otherwise be isolated, like the different departments in the City of Ballarat.
Sometimes, an individual can be both a Gatekeeper and Liaison. As Liaisons belong to multiple cliques, they will eventually receive conflicting messages from the different cliques to sort through. This makes the Liaison act as a discerning communication source. If multiple individuals from different clusters agree on a certain social norm or behavior, such as walking a mile every day before work, that message is more likely to be be picked up and spread by liaisons to other social clusters. Liaisons are often marketed to because they provide a much more trusted, personal recommendation of a brand than a celebrity endorsement even though the celebrity’s message will reach more people.
Because Liaisons have such and important, trusted role in a network, organization leaders must ensure that Liaisons are positive and engaged in their cliques. For instance, a recent Gallup analysis found that wellbeing levels among employee team members were connected to and dependent on the wellbeing of others on the team. While the wellbeing connection within teams was much stronger than among non-team members, these levels of wellbeing can be easily transmitter to different teams through people in Liaison roles.
Looking back on my study, I would have focused more on the social network of the subjects to understand their social interactions. With this knowledge, I may have been able to identify and target the Bridges and Liaisons of cliques with direct messaging instead of relying on my the two Gatekeepers who helped the most. This may have lead to higher participation with the program, better mental and physical health improvements, and less isolates. I can’t help but wonder if the burden of this program kept the central Gatekeeper from achieving an improved mental health score.
As an organization or program leader, finding ways to engage all members of the network in a dialogue about the new social norm you are trying to create or strengthen. As mentioned on the Communication Channels & Messaging Distribution page, simply encouraging people to socialize with their networks has shown to be extremely beneficial. Researchers at Gallup found that employees who have strong friendships with their coworkers are 7 times more likely to be “engaged in their jobs, are better at engaging customers, produce higher quality work, have higher wellbeing, and are less likely to get injured on the job”.
Social norms, after all, are the original crowd sourcing technology. The more people engaged in dialogue by sharing what works and what doesn’t, or showing how it can fit into their daily lives, a more effective and beneficial norm will emerge.
Christakis, N. A., & Fowler, J.H. (2007). The Spread of Obseity in a Large Social Network over 32 Years. The New England Journal of Medicine, 357 (4). Retrieved from http://www.nejm.org
Mackay, A., & Tatham, S. (2011). Behavioural Conflict: Why Understanding People and Their Motivations Will Prove Decisive in Future Conflict. Essex UK: Military Studies Press.
Zimmerman, J. (2012). Cigarette Taxes Worked. Why Not Soda? NewsWorks. Retrieved from http://www.newsworks.org/index.php/speak-easy-archive/item/47484-cigarette-taxes-worked-why-not-soda
Katz, N., Lazer, D., Arrow, H., Contractor, Noshir. (2004). Network Theory and Small Groups. Small Group Research, 35 (3). Retrieved from http://www.hks.harvard.edu/davidlazer/files/papers/Lazer_Katz_Small_Group.pdf
Lunenburg, F. (2011). Network Patterns and Analysis: Underused Sources to Improve Communication Effectiveness. National Forum of Education Administration and Supervision Journal, 28 (4). Retrieved from http://www.nationalforum.com/Electronic%20Journal%20Volumes/Lunenburg,%20Fred%20C%20Network%20Patterns%20and%20Analysis%20NFEASJ%20V28%20N4%202011.pdf
Robinson, J. (2012). Wellbeing is Contagious (for Better or Worse). Gallup Business Journal. Retrieved from http://businessjournal.gallup.com/content/158732/wellbeing-contagious-better-worse.aspx?utm_source=WWW&utm_medium=csm&utm_campaign=syndication
Rath, T., & Harter, J. (2012). Your Friends and Your Social Wellbeing. Gallup Business Journal. Retrieved from http://businessjournal.gallup.com/content/127043/Friends-Social-Wellbeing.aspx#2
Related Articles: Applications of Network Theory
This Document is a Master’s Project which has been prepared at the request of and in connection with the University of San Francisco Sport Management Program. Neither this Master’s Project nor any of the information contained therein may be reproduced or disclosed to any person under any circumstances without the express written permission of Mae Schultz.