Ndaesha Dryden-Mason

True Crime Fanatic

Blog 9: A Serial Killer and Dark Networks

The first article that I found that appied SNA to crime really piqued my interests.  Bichler, Lim, and Larkin (2012) sought to determine if SNA could offer a complementary theory and empirical methods to crime pattern theory  for linking people based on shared activities.  They believed that SNA could be applied to identifying relative suspects in a criminal case.  In order to test this hypothesis,  the researchers used the case of the Green River Killer.  The Green River Killer, Gary Ridgeway, was active in the 1980s and 1990s in the Seattle and Tacoma areas of Washignton.  He is linked to the murder of 49 different women, but authorities believe that he killed upwards of 90 women.  He was finally caught in 2001 and sentenced to life in prison in 2003.

Gary Ridgeway

Bichler et al (2012) collected data from 3 different sources.  The first source was a journalistic account written by two journalist, Smith and Guillen, that contained interviews with victims’ families, law enforcement, and witness information.  Smith and Guillen supplemented this information with official information such as court records.  The second source was a book written by the lead detective in the case.  It included first had accounts about the investigation and confirmed information from Smith and Guillen.  The final source were the court transcripts from Ridgeway’s trial.

The sample population for the study included victims, suspects, witnesses, body finders, locations, and other persons of interest.  The final sample included 88 people who were victims, had investigatory involvement, suspects, or family/associates. The sample also included 58 different locations.  The nodes in the study were the people and the links were connections to geographic locations where either the bodies were found or last seen locations.

Network observations were taken every 6 months for five of the six analysis phases.  The final phases consisted of 30 months of analysis.  The researchers observed the density of the network to determine overall network cohesion.  Betweenness was also analyzed to identify possible suspects.  Finally, the Jaccard Coefficient was analyzed to determine if the network changed every six months of the police investigations.

Analysis of the network revealed that the network increases in size overtime which means that the new information that was gathered changed the actual charcteristics of the network.  In addition, developments in the investigation up to the 18 month mark contributed to indetifying people which changed the network structure.  After 18 months, the networks stabilized.  Even though the structure of the network stabilized, the new information changed the actor level centrality scores.  This leads to the most important finding which is related to the betweeness centrality measure.  At first on suspect was identified as the main suspect due to his high level of betweenness centrality.  However, after the 18 month mark, Ridgeway became the most prominent figure in the network because his betweenness centrality measured 2.4 times higher than the original main suspect.

The methods used in this study this study could be used in cold cases such as the Zodiac Killer to narrow down the number of potential suspects and determine persons of interests who may have information that could help the investigation.

The second article looked at the dark networks that were responsible for the July 7,  2005 and July 21, 2005 London Bombings.  Burcher and Whelan (2015) sought to determine how the limitations of SNA on dark networks impacted analysis of the network in a crime intelligence context.  In order to do this, greater emphasis was placed on the issue of fuzzy boundaries in the analysis of small dynamic networks.

Data was collected from emails, phones calls, text messages, and face to face meetings between the perpetrators of the bombings.  The information was gathered from open sources such as the BBC, CNN, The Guardian, and government reports that came out after the bombings.

The sample consisted of 12 individuals from the July 7 bombing group and the July 25 bombing group.  The networks were analyzed separately, by group, and then combined to determine if there were links between the two groups.  The networks were analyzed based on the degree and betweenness centralities in order to aid in the identification of fuzzy boaundaries for small group dark networks.  The nodes in the network were the individuals and the links were the interactions between them.

The results revealed that one individual, Ibrahim, was the central individual when the two networks were combined . However, when the networks were separated, each network had a different individual who was centa  The results also revealed that the way in which the network boundaries are defined influence the findings of an analysis.

References

Bichler, G., Lim, S., & Larin, E. (2017). Tactical social network analysis: Using affiliation networks to aid serial homicide investigation. Homicide Studies, 21(2), 133-158

Bucher, M. & Whelan, C. (2015) Social network analysis and small ‘dark’ group networks: An analysis of the London bombers and the problem of ‘fuzzy’ boundaries, Global Crime, 16(2), 104-122

 

 

Blog 9: Team Cohesion and Peer Influence

Article 1

Warner, Matthews, and Dixon (2012), sought to more fully understand the relationship between team cohesion and team performance.  In order to achieve this understanding, the study used SNA to look at the structural cohesiveness of two different women’s college basketball teams.

At four different points during the basketball season, the team members completed online-surveys.  The researchers measured cohesivness based on friendship, trust, advice, and sport-specific measure of  individual efficacy based on an individual’s basketball related knowledge and/or ability (Warner, et al., 2012).  Structural cohesion was measured using density to  determine the proportion of the number of ties that existed between individuals in relation to the number of  the maximum possible ties in the network.  In the network, the nodes consists of the individual players that make up the teams. The links were the connections between each player to other players on the team.

  

The results of this study revealed how networks changed and evolved over time.  The study also revealed that the application of SNA to team cohesion allows researchers to examine a team network at the individual level to determine the role that they play on the team. This information could be used by a coach or team manager to determine the possible obstacles that are hindering the team from becoming a cohesive unit. This information identify interpersonal conflict and figure out strategies to mitigate it.

Article 2

Delay et al. (2016), examined how social-emotional learning (SEL) intervention could be associated with peer socialization on academic performance.  The goal of SEL programs are to improve the social and emotional skills of children which can lead to close social bonds which will ultimately lead to increased chances of positive peer influence.

The sample for this study consisted of 631 fifth graders from six different elementary schools. There were 14 intervention classrooms that received relationship building intervention (RBI) and 8 control classrooms. Before and after the intervention, the students nominated their friends and teachers completed assessments of  each student’s writing and math performance. The use of SNA in this study was used to determine how interventions affect social processes.

The results of the study revealed that RBI was associated with more significantly diverse friend choices when compared to control classrooms.  The results also revealed that peer influence was observed in RBI classrooms in relation to improved writing and math performance.

References

Warner, S., Bowers, M.T., & Dixon, M.A. (2012) Team dynamics: A social network perspective. Journal of Sport Management, 26, 53-66

DeLay, D., Zhang, L. , Hanish, L., Miller, C., Fabes, R., Martin, C., Kochel, K., & Updegraff, K. (2016). Peer influence on academic performance: A social network analysis of social-emotional intervention effects. Prevention Science, 17(8), 903-913

Blog 8: Identification of influential spreaders in online social networks using interaction weighted K-core decomposition method

As we all know, online social networks are a key component of everyday life.  More and more of our interactions with others are taking place online through social media sites such as Instagram and Twitter. Al-garadi, Varathan, and Ravana (2017), sought to improve the existing K-core method for online social networks by using a link-weighing method based on interactions among individuals.  The purpose of the study was to identify the influential spreaders in online social networks more accurately  in comparison to other methods such as degree centrality and PageRank.

 

K-core decomposition

The study sample consisted of two large online social networks from Twitter.  The dataset for the first network came before, during , and after the announcement of the discovery of a new particle. The first  dataset was made up of three data of the same user IDs: the social network, the retweet network, and the mention network. The social network accounted for the social structure while the retweet and mention networks were used to weigh the social network (Al-garadi et al., 2017).  The second dataset consisted of 121,807,378 tweets that were posted by 14,599,240 different users.  The second dataset was used to create an undirected, unweighted social network.

The nodes in this network were the individual Twitter user IDs. The links were the connections (follows) between the individual user IDs that were then weighted using the retweet and mention networks.

In order to determine which method was the most accurate, the study calculated the imprecision functions of the other methods used to determine influential spreaders in a network. Imprecision function values that were close to 0 are indicative of high diffusion efficiency given that the users selected are nearly those that provide the most information dissemination (Al-garadi et al., 2017). The study found that using a weighed K-core decomposition method was the most accurate in identifying the most influential spreaders in the network. This method can be used in the future to propose a new model for the spread of information in online social networks as well as lead to the creation of new nerve ration of information spreading models (Al-garadi et al., 2017).  In my own field of criminal justice, this method could be used to identify the influential spreaders within a criminal network in order to target them to hinder the spread of information through the criminal network.

Refernces

Al-garadi, M.A., Varathan, K.D., Ravana, S.D. (2017. Identification of influential spreaders in online social networks using interaction weighted k-core decomposition method. Physica A: Statistical Mechanics and its Applications, 468, 278–288

Blog 7: Changing Networks

Habermas and Castells present two very interesting perspectives related to the field of social network analysis. According to Habermas, the public sphere is the connection between public life and civil society emerging as a neutral space when individuals can discuss their concerns freely and democratically to form public opinion (Habermas, 1974). Ideally, the public sphere is available to everyone and it is an integral part of democracy. Success is dependent upon rational-critical debate which is accessible by all members of the public. In the past, not all members of the public were a part of the public sphere. At many points in history, certain ethic, racial, and gender groups were not allowed to be a part of the public sphere.

 

Unfortunately, the public sphere is being threatened by the media. Through a shift in reporting, the media now has an influence over the discourse. Instead of the goal of the media being the dissemination of information, it is now focused on how much commerce it can generate through it’s reporting. 

According to Castells, network society has led to decentralized networks that are highly efficient because they are better at managing complexity.  Increases in technology have made networks increasingly available to more and more people. As a result, space and time are becoming more and more irrelevant. The sharing of information with in these decentralized networks is driven by micro-electronic devices such as smartphones.  

 

These new forms of social organizations have and will have a major impact of the work around us. In the field of education, classes at the primary, secondary, and post-secondary levels are now offered online.  I took a course last semester online where the professor lived in Oregon. Due to the time and geographic differences, we were able to schedule times that we could meet with her online to discuss any questions or concerns that we had during the course. I have even had professors who live here in the Richmond area who are willing to schedule video chats if you are not able to physically meet them during their on-campus office hours. Education has become more available to people no matter where they are. In the field of healthcare, technology has made it easy and convenient to speak with healthcare providers. I listen to a number of podcasts who are sponsored by a company called Talkspace. Talkspace is an online counseling and therapy company that allows individuals to talk with licensed counselors by phone, text, or video chats. Many primary care practices are shifting to online applications to communicate with patients. For example, instead of having to wait for my blood test results through the mail, I can now view them online once they have been processed. In addition, if I need to schedule an appointment or have a specific question for my doctor, I can just use the online application instead of calling the office.  Finally, the new forms of social organizations have had an impact on the economy.  In this Information Age, both public sector and private sector companies have moved jobs that were traditionally in the office to the field. More and more  people are able to work remotely instead of working in an office. Teleworking costs less for the government and businesses. This means that companies and the government could hire more individuals because they would be saving money on rented office space, and each new individual who enters the work force adds to the economy.

I believe that these changes within the network are both good and bad. They are good in the sense that more people are able to obtain and disseminate information to a larger number of individuals. Unfortunately, not everyone will be able to benefit from this.  Low-income countries and individuals will be left behind until they are able to obtain advanced technology and have an equal voice. I think the changes are good for those who are able to actively benefit from them but bad for those who are not.  The individuals who are not will continue to be left behind with very little voice.

Reference

Habermas, J. (1974). The public sphere: An encyclopedia article (1964). New German Critique, (3), 49-55

Blog 6: Friendship and Risky Behavior

The main focus for my research over the course of my studies in the Criminal Justice program are the factors that compell an adolescent to participate in deviant behavior.  An adolescent’s environment  is one of the factors that affects an adolescent’s behavior. Therefore, my question for this research project is do adolescent peers have an effect on adolescent behavior?

For the purposes of this project I will be studying a network of student at a single school in order to investigate my research question.  I will be using existing data from the first wave of the National Longitudinal Study of Adolescent to Adult Health (Add Health).  The first wave of the study took place during the 1994-1995 school year.  The study consisted of a nationally representative  sample of students in grades 7 through 12 in the United States.  As a part of the survey process for the first wave, participants were asked to identify five close female friends and five close male friends from a school roster. In addition, participants were also asked questions about alcohol and drug use as well as sexual activity.  Due to the number of  participants in the study and the  sheer amount of data that is available, I will focus my analysis on a single school. I believe that the key to analyzing peer influence is looking at the in-degrees and out-degrees of a node to determine if there are reciprocal relationships between the individual and the other individuals that they consider to be close friends.

High School hallway showing student lockers

There is an abundance of research that has looked at the influence peers have on adolscent behavior.   Many studies have used the dataset from Add Health to study the influence of peers on adolescents. Keon and Goodson (2015) used to data from two schools that participated in the study to analyze friendship networks and risky behavior of participants. Their main focus was the degree centrality of each individual node and that node’s response to the following question: Did you  engage in sexual activity and alcohol consumption simultaneously?  Fujimoto, Unger, and Valente (2012), used SNA to also look at how peers influence behavior through the measurement of affiliation- based peer influence.  The study specifically analyzed networks within school based team sports. Kobus and Henry (2009) looked at the position of individual nodes in a network consisting of adolescents  in sixth through eighth grade as well as interactions with other individual nodes in the network to  predict the use of cigarettes, marijuana, and alcohol in adolescents.

To build on the existing research, I want to also analyze the existing data to determine if there is an individual  or individuals who are the main source of the deviant behavior. This indentification could be helpful in the formulation and implementation of programs that are aimed at stopping and preventing deviant behavior in adolescents.

The data collection method used in Fu (2005) is useful to my own research. Fu used responses generated from participants to develop the social network much like what I will do for my own research.

Blog 5: Centrality Measures

Centrality measures are used to describe the role that a particular node plays in a network. Degree centrality is a measure that accounts for the number of neighbors that a single node has. In a directed network, the focus would be in the number of incoming links or in-degrees. The higher the number of degree centrality, the more active that node is within the network. Closeness centrality measures the sum of the length of the shortest paths between a single node and all other nodes  in the network. A node with a high measure for closeness centrality has to ability to pass information through the network at a higher volume and faster than other nodes in the network. The measure of centrality in a graph based on the shortest paths between nodes is betweenness centrality. A node with a high measure of betweenness centrality has a big influence within the network. Nodes with a high measure of betweenness centrality  control the information that is passed to other nodes on the network.  Finally, eigenvector centrality  is the measure of the influence of a node within a network. A node is more is considered important if it is linked to other important nodes in a network.Image result for node centralities

Grassi et al. used the betweenness centrality measure to determine who the leaders in a criminal network.  The researchers acknowledges that nodes with high measures were critical in social networks because they can effectively manage the flows of resources and information in a network.  While other studies have focused on other measures of node centrality to determine leaders, this particular study sought to use betweenness centrality because it distinguishes strategically positioned criminals in a criminal network. The study use a dual projection approach which looked at individual and meetings.  Criminal leaders were identified based on their individual participation in meetings.

Image result for criminal networks

Criminal network visualization

Saidi et al. also used node centralities to determine the leaders with in cyber terrorist communities. The closeness centrality measure was used the determine who the players were with in the network who had control over the information that was passed through the network. Degree centrality was used to identify the players within the network who would be the key players and influencers within the network. Betweenness centrality was used to identify the players within the network who had a strong influence on the transfer of resources and information passing through the network. Eigenvector centrality was used to identify the leader of the cyber terrorist network.

Image result for cyber terrorism

In both studies, centrality measures were used to identify the key members in a network and the influence they have within a network. This is important for law enforcement agencies in the development of tactics to dismantle criminal networks.

Blog 3: Social Capital

Based on the readings for this week, the definition of social capital varies immensely.  The numerous views of social capital make it difficult to nail down a concrete definition.  Basically, social capital can be defined as those resources essential in social relations that promote  collective action.  We can think of social capital as links, understandings, and shared values in our society that allow individuals and groups to work together and trust one another. Kadushin (2003), states that at both the social capital investment and individual social capital investment levels are the two main consequences.  While social capital investment can result in greater social capital, individual social capital increases well-being (Kadushin, 2003).

Image result for social capital

Many theories look at the positive effects of community participation such as the transfer of information, access to resources, and job networking (Kadushin, 2003; Lin, 2001). Portes and Landolt (2002) assert that many of the theories on social capital only take into account the positive effects of community participation without considering the possible negative implications.  As cited in Smith’s study, even though the black job seekers had the personal connections with individuals to aid in the job hiring process, many of these connections were not passing information to their employers (Smith, 2005).  Many individual’s feared feared providing a referral knowing that if the other individual did not perform well, it would be a direct reflection on them.  As a result, access to resources did not result in mobilization (Smith, 2005).

Image result for community engagement

 

According to Lin, the internet  is contributing to a rise in social capital.  As a result more people are connecting through online discussion boards, forums, and clubs (Lin, 2001). Putnam did not believe that cyber technologies would have the same effect as in person interactions have and believed that these new technologies were contributing to the erosion of social capital (Putnam, 1995).   Putnam also attributed the erosion of social capital to women joining the labor force, decreases in marriage, increases in the divorce rate, and mobility (Putnam, 1995).   Putnam felt that the decline of community networks, such as bowling leagues, represents a loss of social capital.

 

Image result for bowling

 

Contrary to Putnam, I believe that technology has contributed to the rise in social capital to an extent.  For example, social media allows us to connect with others and aids with the efficient transfer of information.  However, technology can make people less individualized  and more anonymous which can lead to a relaxation  of social norms.  People tend to hide behind their social media facades and feel empowered to say whatever they feel, even if it is not socially acceptable.

For my personal research, social capital could be used to determine the likelihood of an individual participating in deviant behavior.  Research has shown that individuals that are more involved in school  and community activities, work, and are married are less likely to engage in deviant behavior. Is an individual with a higher level of social capital less likely to engage in deviant behavior? SNA can help us to determine which individuals engage in a high rate of deviant behavior and their links within the network.

Blog 3: Small World Theory

The small network theory asserts that actors in a network belong to distinct groups within the network, but these actors have to ability to reach across the network in an efficient manner (Robbins,  2015). The small world theory emphasizes the balance between closure and connectivity.  The concept of the six degrees on separation did not take into consideration how peoples’ abilities to connect varies from person to person. For example, there are people who find talking to others and making connections very easy. On the other had, there are others who struggle with making social connection and who are not willing to venture socially outside of their group of friends.

Image result for no new friends meme

Big data presents a challenge to social science research.  According to Granovetter (1973), the Achilles heel of modern sociological theory is that it does not relate micro-level interactions to macro-level patterns. The small world theory looks at Big Data from the small group level  and then applies to the full network phenomena such as the diffusion of information.

Image result for small world theory

Small World Graph

In the small world theory, emphasis is placed on weak ties. Weak ties are the vehicles to transmit new information and innovation across social networks (Robbins, 2015).  In addition, weak ties help with the integrations of social systems (Kadushin, 2012). Many of our friend groups can be characterized by homophily. We are drawn to people who are like us. Weak ties allow individuals to obtain information that they may not have come to acquire if they only interacted with individuals within their close friend group. People with few weak ties are not going to see outside of their close circle (Kadushin, 2012).  This is important when it comes to finding a new job or learning about new educational opportunities. For example, I was able to be hired for a consistent babysitting job during my first two years of undergrad as a result of my weak ties. One of my sister’s friends knew a professor in DC who needed to have her daughter watched when she was teaching classes and her husband, who was also a professor, was not home.

I personally believe that weak ties are more important than strong ties because they remove us from our “bubble”.  When we connect with weak ties in the network, we are introduced to a number of things such as new ways of thinking and other perspectives of the world around us. Weak ties play a role in our understanding of our world and help to present us with new opportunities for individual growth.

Image result for weak ties

In 1994, three college students wrote to the Jon Stewart Show stating that they believed that they could connect Kevin Bacon to almost any actor in Hollywood with two or three links (Barabasi, 2003).  Two students from UVA saw the episode and took on the Kevin Bacon game as a computer science project. Using the already establish IMDb.com database, they created The Oracle of Bacon Website which by simply inputting the names of two actors into the search, one would be shown the shortest path connecting the two actors.

Blog 2: SNA Data Structures

The social network analysis perspective described in the reading by Keim, can be applied to a diverse number of topics. This reinforces what we learned through last week’s readings that social networks are a part of every aspect of society. 

According to Keim, even though groups consists of a defined, specific number of individuals, the interactions of individuals on a daily basis leads complexity with in the social network and individuals not being involved with just one social group (Keim 2011). For example, I have a group of individuals that I interact with when I am at work, but I have a different and very distinct group that I interact with outside of work.  This contributes to the complexity of social networks within our society.

Social network data differs from traditional social science data because it seeks to derive social structure empirically and not through inferable classifications ( Haythornthwaite, 1996).  The content and the pattern of relationships are examined in the social network approach in order to determine how and what resources flow from one individual to another. This is important for the development and implementation of methods to disseminate important information to a large number individuals as quickly and efficiently as possible. 

 

  

 

Relational data refers to the relevant network relationship among nodes (Robbins, 2015). This data consists of information that  is specific to at minimum pairs, but can also be specific to a greater number.  Descriptive analysis looks at  the similarities an differences between networks.  Descriptive analysis is used to find trends in relational data.

relational model diagram 

Relational data and descriptive analysis are two important factors to paint a full picture of how the social network is constructed and functions.  Predictive analytics is the use of data to identify the likelihood of future outcomes on historical data.  Predictive analytics is used in many fields such as marketing, health, and criminal justice.

 

 

Linked: Blog 1

In the book Linked: How Everything is Connected to Everything Else and What it Means for Business, Science and Everyday Life, Barabasi uses various real life stories to show the developments in the understanding of how social networks are formed and function in all aspects of our daily lives. Prior to starting the readings for this course, I only thought about networks in terms of how they applied to the internet. This book caused me to think of networks in every part of daily life.  It presents the idea that the old understanding of networks formed and functioned do not explain how networks, such as Erdos’ model, do not explain how these networks are dominant in our lives. The newer models that are based on network growth which does explain how networks dominate our lives.

The most interesting discussion that was introduced to me in the book is how digital networks function on the internet. Barabasi discussed how many people view the internet as an equal facet of society where everyone’s voice is heard (Barabasi, 2003). This would lead one to believe that everyone’s voice would have an equal chance of being “heard” online, but this is not the case. Some have louder voices than others due to the sheer volume of content that is available on the internet. The more incoming links that a webpage has, the more visible it is to others (Barabasi, 2003).  Sites such as Google and Yahoo are apart of a rare group of nodes that are referred to by Barabasi as hubs. One will find links back to these hubs on various sites throughout the internet which is why they are visited so frequently.

Related image

Barabasi’s explanation of how digital networks functioned helped me to connect the dots on how social networks work in everyday society. I believe that we all know someone who seems to know everybody. Barabasi explains how these individuals, who are referred to as connectors, are rare within the social network, but I compare them to the major websites on the internet. Barabasi.

In the chapter titled “The Fourth Link: Small Worlds”, Barabasi references The Strength of Weak Ties by Mark Granovetter. According to Granovetter, most people find jobs through acquaintances than through their close friend group. This is because your close friends are more likely to have the same personal connections that you do and access to the same information (Barabasi, 2013). Even though this seems to be the opposite of what we would conventionally believe to be true, it does make a lot of sense.

Image result for social networks

I agree with Barabasi that we can find networks everywhere. The evidence that he presented from the world of business, to the internet, to the world of economics demonstrate that this is true. I also agree that there is constant growth in the various networks due to the advances in technology that we see on a daily basis

« Older posts
Privacy Statement