Blog 10 – Recidivism and Psychopathy

Because this is cool… check out recidivism in the NFL

Ok, back to the real business…

First, after reviewing other’s articles over the past few weeks and diving into my own, it seems obvious that students and most tend to relate and group themselves with those that they find similar to themselves (although this can likely also be latent or implicit), I wanted to find an article to read that, perhaps, touched on this phenomenon.

Work by Fortuin, van Geel, and Vedder (2015), which examined adolescent’s tendency to group themselves with those similar to them and how that relates to internalizing or externalizing problems, seemed to focus right at this observation. They sought to better understand the influence of internalizing and externalizing problem behavior upon the selection, socialization, withdrawal, and avoidance of peers. They presented dualing hypotheses regarding the extent to which different types of socialization patterns would present with those of varying degrees of internalizing or externalizing problem behaviors. That is, over the course of a school year they recorded three instances of each student’s (N = 542) friends that they liked and their internalizing and externalizing problems through survey methods. In this case, the students served as nodes with their peer nominations to other students as the edges.

Using longitudinal social network analyses, Fortuin and colleagues (2015) found that adolescences tend to prefer peers with similar externalizing problems, while no significant selection effect was found. Over time groups tended to become more similar in externalizing problems, but not internalized problems. Interestingly, and through the use of some fancy software, they were able to control for obvious correlates of gender and ethnicity. Using SNA allowed them to further examine how relationships transpire over time and what particular attributes, such as externalizing or internalizing behavior, influence or change within such groups of students. In other words, they employed SNA to examine the forming and existing of groups over time, while examining how those groups changed in respect to their internalizing or externalizing problem behaviors.

Next up, and again after reading other’s reviews, I thought it would be interesting to see the use of networks that are derived from experience sampling techniques (more frequent data collection. Work by Weerman, Wilcox, and Sullivan (2018) examined short-term changes in peer relationships, deviant behavior, and routine activities to better understand selection, socialization, and situational peer influence. Using high-school students (N = 155), they collected survey data to collect their peer relationships, activities, and offending behaviors over five time points over no more than a two-week span. This was in an effort to answer four main research questions displayed below, but in general involve: how volatile are peer relations, substance use, and delinquent behavior? How structural network effects influence iterative peer changes? Do those who socialize adopt the delinquent behavior of their peers? and is delinquent behavior a product of changes in socializing and substance use?

Results indicate very “volatile” networks that exhibited changes in delinquent offending that were related to situational changes in unstructured socializing, alcohol use, and marijuana use. In that, they concluded that “long-term peer influence processes like socialization may be less important in the short run, while situational peer effcts might be more salient.” Ultimately, they demonstrated that adolescent changes in peer relations is volatile and changes very quickly and drastically.

Ultimately, the use of social network analysis here was similar to the first article, in that they viewed peer relationships and how another aspect of them, in this case delinquency and substance use, was influential and influenced therein.

Interestingly, both articles presented here used the R package: RSiena. More information about it can be found HERE. In essence, it is a statistical package used for longitudinal network analysis that focus on nodal relation changes over time.

 

Blog 9: Friendships

Drawing from my psychology background and great interest to learn more about childrens’ learning paths and experience, I focused my search for these two articles upon developmental aspects of a child’s growth, their friendships, and the use of social network analysis.

First, the work by Shin and Ryan (2014) investigated early adolescent friendship selection and social influence in relation to academic motivation, engagement, and achievement. Data was collected in classrooms using surveys at two time points, one in Fall and one in Spring, among approximately 575 (at each time point) 6th graders. Using a stochastic actor-based social network model to estimate friendship selection and influence, Shin and Ryan (2014) used students as nodes and the strength of their friendships to others as the edges or connections therein. This method, which examines changes in friendship networks and behaviors (attributes), while allowing other attributes such as gender, race, and other behaviors to be examined. Although no explicit research questions for outlined, they did offer hypotheses (which can be inferred into RQs…): They “expected that both selection and influence processes occur among friends in all aspects of academic adjustment.” and that academic self-efficacy likely serves as peer influence effects (p. 4).

Therefore, despite having very solidified research questions, they seemed to after being more descriptive than anything. That said, they found network density to decline over time, meaning adolescents dont just nominate anyone as their friend, and a positive reciprocity parameter, whereby they tend to reciprocate friendship nominations (dyad nominations). Therefore, they tended to nominate each other in friend dyads, keep the network closed, and form peer group structures in friendship networks. In terms of other attributes, they found that those with high levels of effortful behavior and GPAs were more often nominated. Interestingly, contrary to my own beliefs, those with higher values of self-efficacy tended to be nominated less. Overall, they found that “selection effects were not as pervasive as influence effects in explaining similarity among friends across the school year” (p. 8). Furthermore, students tend to select friends that are similar in GPA and confidence, while GPA is influenced by friends over time. Being friends with someone else with higher achievement tends to influence those with less initial achievement, although their confidence is not influence. Students dont seek out other that behave like them, yet tend to select those with similar GPA and similar confidence, while their behavior tends to become similar over time. Collectively, the use of SNA here allowed for a dynamic examination of how selection and particular influences mitigate friendships and what the means over time. Understanding what influences selection of friendships and how those relationships change over time is vital to understanding how students change and adjust to their surroundings over time, especially when considering their academic motivation, which is vital to continued sustainment of effort and focus.

Next, work by Parker and colleagues (2014) examined adolescences’ (N = 1,972, grade 10 among 16 schools) friendship groups to examine how hope and well-being influenced are related to such groups. The data was collected through survey methods, whereby measures of hope and well-being were collected alongside a form to fill out both male and female friends. In this case, the nodes in this SNA were the students themselves and their listed relationships were their undirected edges between students. Again, these researched did not provide explicit research questions, however, they did offer hypotheses: 1. Students from the same group of friends will resemble each other in hope and well-being, 2. Average hope in friendship groups will be associated with group well-being over and above individual level hope. To form groups of friends, they used “community detection algorithms to assign individuals to particular friendship groups. To assess if average group hope was significantly related to their subjective well-being over and above their own level of hope, they used multilevel structural equation modeling (MSEM) to both account for level effects and measurement error. 211 friendship groups were formed (~13 per school) and nodal metrics were generated that suggested a good amount of reciprocal friendships and variable centrality therein. In terms of the MSEM, they portrayed the intraclass correlations (ICC) (proportion of variance explained by the grouping effect–in this case ‘friendship group’ was the grouping variable). In that, they found that the ICC for hope was .241 and for well-being it was .293, suggesting about 25-30% of the variance was explained by their group membership, supporting hypothesis 1. Using the MSEM models, they found that individual hope was significantly related to social well-being, while average hope was not associated with emotional well-being, which was consistent with their second hypothesis. Said another way, using the contextual effect model (MSEM), demonstrated that individual subjective well-being was associated with group hope beyond individual levels of hope. Therefore, this research suggests a relationship between individual well-being and the hope of the friendship group, which implies targeting an individual hope and therefore their friendship groups hope, can be a power means to collectively increase hope and well-being. In this case, the use of SNA provided an empirical way to derive groups of students, while also providing the footing to assess a dynamic multilevel model that could pick out the particular attributes between individuals and group level variance in terms of hope and well-being.

Together, these studyies, though not exactly simple, demonstrate the utility of social network analysis to better understand how friendships matter and how other attributes can be dynamically associated therein. As we all likely know anecdotally, friendships and groups of friends can be a lifeline and provide ample means of support in difficult times. Gaining a better understanding of how these groups form, what influences them, and what they mean in terms of other attributes can provide targeting to influential interventions that can help support student success.

Parker, P. D., Ciarrochi, J., Heaven, P., Marshall, S., Sahdra, B., & Kiuru, N. (2014). Hope, friends, and subjective well-being  A social network approach to peer group contextual effects. Child Development, 86(2), 642–650. https://doi.org/10.1111/cdev.12308

Shin, H., & Ryan, A. M. (2014). Early adolescent friendships and academic adjustment: Examining selection and influence processes with longitudinal social network analysis. Developmental Psychology, 50(11), 2462–2472. https://doi.org/10.1037/a0037922

K-core & Creativity Research – Blog 8

In a not so long ago time I dabbled in creativity research, at least through the lens of how children become creative, believe themselves to be creative, and exhibit creative thought, behavior, and ideas. All of that said, the use of social network analysis in educational research is sparse…

Nevertheless, I found an article by Zhang, Zhang, Yu, and Zhao (2015) that caught my eye… and used k-core to help explain the spatial structure of creativity research over a ten year period between 1992 and 2011. They argue that because of the proliferation of creativity research being generated from various disparate fields, it can become difficult to understand where the field as a whole needs to move in the future. Using keywords to express themes and trends in the literature, they used social network analysis to examine the cooccurrence between words to “deeply explore research hotspots, research opportunities, as well as cutting-edge evolution in creativity field” (Zhang et al., 2015, p. 1024). Therefore, their research question, though not explicitly stated, was: Collectively, what are the common themes of research, future research opportunities, and hotspots identifiable from past research and trends in keyword among the extant literature?

The data was collected directly from Web of Science and keywords were extracted and summed using NoteExpress 2.0. The primary keywords were then identified to other keywords in the same paper. In other words, a paper typically has three keywords and these keywords can, if looked at broadly, tell trends of research by their simple connections to each other. Furthermore, using a method to judge the distance between measured objects, multi-dimensional scaling (MDS), they also figured the position of each research topic and schools of thought by employing MDS.

This figure demonstrates the co-occurrence of each work, similar to a correlation matrix.

They used k-core to derive groups of connected keyword paradigms to “pick out the sub-groups according to each keyword degree in the keywords networks” (p. 1029). This resulted in 15 sub-groups, whose node degree ranged from 45 to 22, and in which 89.6% were above 40.

Results indicated that out of 163 top keywords from 4,575 papers on creativity research, keywords can be best described into five main topics: creativity applications into particular areas, pathology of creative generation, individual level creativity, organizational level creativity, and theoretical and methodological studies of creativity research.

Using their MDS analysis, they identified the spatial structure of creativity research. Results there indicated

With the main top keyword of “Creativity” in the center, each quadrant represents key research paradigms found.

From here, they went on to further analyze the five key areas of research and what those fields are up to during the past ten years of research. They then explain the salient areas of future research suggested therein. That is, future research should be focused on the process of creativity and specific sub-processes of the complex creative formation.

Ultimately, this form of analysis, namely k-core, can provide researchers with a metric to determine groups of centrally connected relations in social networks. This further aids researchers in seeing who is important, integral to the network, and important to collectively speak about and examine.


Zhang, W., Zhang, Q., Yu, B., & Zhao, L. (2015). Knowledge map of creativity research based on keywords network and co-word analysis, 1992–2011. Quality & Quantity, 49(3), 1023–1038. https://doi.org/10.1007/s11135-014-0032-9

Blog 7 – Theories of Networks

The construct of a “public sphere” was presented by Habermas as the “nexus between public life and civil society,” whereby the sphere exists outside organized and institutional influence. Although disconnected from government or organized institutional influence, the public sphere is an integral component to a democratic society, whereby citizens can discuss, agree, or disagree with each other to illustrate their collective being, opinions, and ideals without influence from those that have been given the authority to govern. This process, though seemingly haphazard, is actually well organized, or needs to be well organized, in universally agreed upon terms that establish some type of platform for which these discussions can exist. In other words, debate needs to be freely accessible and organized in such a manner as to not oppress or limit anyone’s speech. Nevertheless, with the growth of public organizations rivaling those of organized government, Habermas worries that the public sphere of discourse is actually in decline, as these public organizations now harness and retain sufficient amounts of influence and power. In terms of what characteristics define the public sphere, there are, as those that came a long time before us, many purposely ambiguous and clear directives of free speech that, in some part, define the public sphere. In other words, the Constitution regards free speech likely in some way as a public sphere, whereby any speech can exist, be argued, and be freely expressed. Although this connection does not explicitly separate governmental control, it does in many ways limit governmental powers to limit speech of citizens and permits them the allowance therein.

Although the Constitution was written with no knowledge of or how our “network society” would evolve, it certainly has. That is, the network society now largely exists apart from face-to-face relationships and exists across both place and time, as they are now almost completely irrelevant (as Castells explained). In other words, online communication and social media allow us to transcend space and time because we can instantaneously connect with anyone at virtually any time across the globe. This has reshaped how we view the impact of social organizations, such as Facebook, political parties, conceptual movements that are important to people, and anyone that tends to favor or side with anyone else. The ease of connections has permitted the social organization to have much more influence in both positive and negative ways to how society “operates” as we know it. Barely a week goes by that we are not surprised at what is socially popular, has gained traction in the news, and what people are quickly becoming both experts at and very passionate about. This connectedness of the network society has come with enormous influential power, even reaching and influencing the highest components of organized government.

This has flipped the script completely in terms of how many have fought for a ‘public sphere’ to operate and exist outside the reach of governmental influence, as Habermas lamented, however, it has come with a surprising amount of power and influence that is shaping the landscape of our own society on a daily basis. Therefore, it is safe to say our society is becoming far more fluid, far more influenced by social movements that gain traction purely by ease of connection, and far more dramatic in how volatile things can get and in what time frame. Civil wars are often the culmination of decades of oppression that has gained steam and found a breaking point, as Castells explained, but what access to such information and connection has provided is a time shift. This time shift has enabled mini-civil wars to reach their boiling point much faster, which can be both gratifying and far less organized.

As can be expected, these rapid changes can have very dramatic influence to what we often rely upon in terms of socialized or organized functions like education and health care. Speaking towards education, we consistently see the institution itself flexing, bending, and being shaped by popularized movements that have provided ample progress to issues haunting education over the past thirty years, like equality, teacher pay, and the conditions of our schools. Despite these obvious goods, there also comes some things that tend to be more problematic, like keeping the collective voice well understood or standardized and keeping the organization in front of movements that move at the speed of light. For example, having access to bathrooms that relate one way or another to your gender/sex. When this first became a huge issue in social media and then the news, followed by ample social movements and organizational opinions, school systems had little time to collectively determine their route, opinion themselves, and their directives. Without time to become unified with a route forward they were marked as the problem themselves. Therefore, the speed at which the network society moves can be problematic, yet actually have fairly good intentions.

I try to be a fairly pragmatic person that looks towards making things better and that the future will likely be much better than what we have today. However, it is difficult to grasp our digital future as a society. It is far too volatile and variable to see very clearly. I do, however, believe it is our nature to seek the better, a society that is better today than yesterday, and to make things as good as we can. Therefore, I believe the future is promising, will help more tomorrow than today, and eventually enable more better to happen to all.

This is fascinating: https://www.ebscohost.com/uploads/imported/thisTopic-dbTopic-1248.pdf

In short, it is a research overview of how social movement emerge through 4 stages: emergence, coalescence, bureaucratization, and decline.

Blog 6 – Research Plan

In general, my work tends to involve both student motivation at large (Bae & DeBusk-Lane, 2018) and writing motivation more specifically (Ekholm, Zumbrunn, & DeBusk-Lane, 2017; Zumbrunn, Ekholm, Stringer, McKnight, & DeBusk-Lane, 2017). In short, my personal passion involves using person-centered approaches, namely Latent Transition, Profile, and Class Analyses to assess latent groups of students’ motivation. In a good bit of my work in writing, I focus on latent profiles of writing self-efficacy that focus on the cognitive components of what a student goes through and forecasts themselves having the ability to execute: writing mechanics, self-regulation, and being capable of generating ideas to write (Bruning, Dempsey, Kauffman, McKim, & Zumbrunn, 2013). Over the past 40 years of research, writing self-efficacy has proven to be a dominant theme predictive of future writing performance, which is in-line with many other domains in education (Pajares, 2002, Klassen, 2003, Klassen & Usher, 2010). That being said, there is utility in understanding students’ writing self-efficacy, as it can be an early indicator of doubt, difficulty, and the foundation that can undermine future success.

Zumbrunn et al., 2017

Derived from Albert Bandura’s Social Cognitive Theory, self-efficacy is commonly procured or ‘sourced’ from four main areas and experiences: mastery experiences (past performance), social persuasions, vicarious experiences, and emotional and physiological states (Usher & Pajares, 2008).

Here, I am most interested in social persuasions and vicarious experiences that may or may not serve to bolster efficacious beliefs. To be more specific, and to generally get a grasp of both the method (social network analysis) and the collective interaction between student-to-student relations, I am only interested in ‘academic self-efficacy,’ not specifically writing. I do this for simplicity and to not overcomplicate this relationship. To further explain, social persuasions are commonly the feedback we receive from others that encompass the immediate and applicable environment. In this case, the feedback, relations, and interactions with others can serve as social persuasions. Additionally, vicarious experiences are commonly understood to be self-appraisals of one’s own capabilities in relation to others. The people for which surround an individual within the situational context influence how self-efficay beliefs are judged and determined. Therefore, I’d expect inter-classroom relations to influence this source of efficacy.

Therefore, to make the theoretical and causal leap, I’d hypothesize that those with more robust networks to also receive more self-supporting sources of self-efficacy. Conversely, I’d expect those that do not have a robust network to lack applicable sources of self-efficacy and therefore report less academic self-efficacy. Further, because there are fairly well established social norms associated with particular domains (like females tending to have higher english, writing self-efficacy and males having higher math and science efficacy), I’d expect females from my target sample to exhibit more ‘academic’ self-efficacy, as the class is in the School of Education and is primarily composed of pre-service teachers (who are commonly and dominantly female) (Pajares, 2003; Klassen & Usher, 2010). This is likely paired with more robust networks form females, as they are commonly reported to be more social creatures (Weisberg, DeYoung, & Hirsh, 2011).

Therefore, my question would be: Are those with more diverse and robust social networks also believe they are more academically capable?  Also, do women tend to be more social and report higher academic self-efficacy than men? If so, does the type of network change (those they typically socialize with, those they collaborate with, those they speak to often, ect.) influence this relationship?

This network, as I foresee it, would be a directed graph based upon students’ responses to questions about to who they regularly interact and relate with. That is, the nodes will be students, differentiated by either self-efficacy or gender (both), and the edges will be directed, pointing towards those they express interaction with.

I plan to use one of three of the School of Education’s Developmental classes for pre-service teachers. I have access through a few of their professors. Though they tend to be mostly female, I have access and they can provide preliminary results that may be interesting and suggestive to future work.

So, to clarify, I plan to get the class to answer a fairly short survey that includes their gender, academic self-efficacy (How confident are you that you can get an A in this course?), and, of course, the way in which they relate. As mentioned, this will include those they most regularly collaborate with, interact with, discuss class topics with, have spoken to at all, ect. I think picking three of these types of relations would be adequate enough for this project.

To my knowledge, there is not specific social network research in this area to date. However, based upon the aforementioned theoretical basis, I’d expect particular trends to exist. Therefore, this work would likely further validate gender norms in this gender specific arena and further establish another lens to depict and explain both vicarious experiences and social persuasions in terms of their sourcing of efficacious beliefs.

Nevertheless, experience sampling and other more intensive ways to collect data, such as Fu’s (2005) contact diary, may provide more dynamic and changing views on these relationships between those who students interact with and their efficacy. Using the more detailed approach, at least from my vantage, would offer for detail to pick out the peculiarities of self-efficacy, gender, and academic self-efficacy (especially if I collected the full scale–which I dont plan to do, as it is long and I’d rather respect the class’ time–perhaps for a full study).  I do, however, believe both networks would offer valuable perspectives and either may benefit the researcher depending on time, availability of participants, etc. Even further, networks such as Carrasco, Hogan, Wellman, and Miller (2006), that incorporate the “spatial distribution o social activities” along with the social network, can also offer a nuanced and detailed account of what is actually going on in a network.

 

References:

Bae, C. L., & DeBusk-Lane, M. (2018). Motivation belief profiles in science: Links to classroom goal structures and achievement. Learning and Individual Differences, 67, 91–104. https://doi.org/10.1016/j.lindif.2018.08.003

Ekholm, E., Zumbrunn, S., & DeBusk-Lane, M. (2018). Clarifying an Elusive Construct: a Systematic Review of Writing Attitudes. Educational Psychology Review, 30(3), 827–856. https://doi.org/10.1007/s10648-017-9423-5

Klassen, R. (2002). Writing in early adolescence: A review of the role of self-efficacy beliefs. Educational Psychology Review, 14(2), 173–203.

Klassen, R. M., & Usher, E. L. (2010). Self-efficacy in educational settings: Recent research and emerging directions. In T. C. Urdan & S. A. Karabenick (Eds.), Advances in Motivation and Achievement (Vol. 16, pp. 1–33). Emerald Group Publishing Limited. https://doi.org/10.1108/S0749-7423(2010)000016A004

Pajares, F. (2003). Self-Efficacy Beliefs, Motivation, and Achievement in Writing: A Review of the Literature. Reading & Writing Quarterly: Overcoming Learning Difficulties, 19(2), 139–158. https://doi.org/10.1080/10573560308222

Usher, E. L., & Pajares, F. (2008). Sources of Self-Efficacy in School: Critical Review of the Literature and Future Directions. Review of Educational Research, 78(4), 751–796. https://doi.org/10.3102/0034654308321456

Weisberg, Y. J., DeYoung, C. G., & Hirsh, J. B. (2011). Gender Differences in Personality across the Ten Aspects of the Big Five. Frontiers in Psychology, 2. https://doi.org/10.3389/fpsyg.2011.00178

Zumbrunn, S., Ekholm, E., Stringer, J. K., McKnight, K., & DeBusk-Lane, M. (2017). Student Experiences With Writing: Taking the Temperature of the Classroom. The Reading Teacher, 70(6), 667–677. https://doi.org/10.1002/trtr.1574

Blog 5 – Centrality and Student Motivation

In general, node centrality refers to how influential or significant a node is to the overall network (Yang, Keller, & Zheng, 2017). Further, node centrality is often commonly associated with four types of centrality: degree, betweenness, closeness, and eigenvector centrality.

Briefly, degree is simply how many connections a node has to other nodes, betweenness is a measure to capture how often a particular node serves as a bridge between other nodes, closeness captures how close a node is to any other node in the network, and eigenvector is a measure that seeks to capture how influential a node is to the overall network.

Ultimately, however, each metric of centrality can describe a nuanced difference upon a network, dependent upon the interest of the researcher. In other words, much like statistical descriptive metrics, sometimes certain metrics are more useful to describe the network than others.

Two fairly recent articles display a use of centrality measures well. First, work from Liu, Chen, and Tai (2017), who examined how elementary students collaborated to create multimedia stories, while also assessing their engagement in the process. They sought to investigate how teams or collaborations were formed and whether allowing students to freely form teams also enhanced their engagement. As they described,

“Centrality is also a prominent indicator to show the activeness or the prestige of entites in a network as it indicates the extent to which an individual interacts with other members (Wasserman & Faust, 1994).” (Liu et al., 2017, p. 116).

In doing so, they built upon previous findings that found an inverse relationship between in-degree centralization and the out-degree centralization in relation to student prestige and influence (Lopez et al., 2014). Related, they also were interested in gender’s association to centrality, as previous works had suggested that female students were more likely to be central (Giri et al., 2014). Furthermore, a lack of conclusive evidence between a student’s knowledge level and their centrality was also in question, as previous work has yet to fully understand how high betweenness (“positions between two strongly connected groups”) and social centrality relate (Liu et al., 2017).

Primarily using in and out-degree to representing how and individual invited peers or received invitations from peers, Lui and colleagues (2017)

With girls on the left and boys on the right clearly showing no obvious preferential centrality (but clearly defined by their gender), they found that gender did not align with preferential popularity (in-degree) or activeness (out-degree). Furthermore, they concluded that students’ knowledge had a significant impact on their structural positions in the social network, however, those with less proficiency were more active (out-degree) although their knowledge level did not directly impact their popularity (in-degree). Also, free teaming was significantly related to positive engagement. Interestingly, however, those who were more central in the out-degree structure were also more engaged.

Similarly, work by Brewe, Zwolak Potvin, Williams, and Kramer (2016) sought to examine a marked decline in self-efficacy among physics students by assessing their relationships among their peers. They used three types of centrality to examine changes in self-efficacy: In-degree, out-degree, and PageRank. Although we have familiarity with in and out-degree, PageRank centrality, a derivative to eigenvector centrality, takes link direction and weight into account in determining how influential any one node is. That is, it is not squarely based upon how many connections a node’s connections have, but also their in and out-degrees and the weight of those degrees.

Nevertheless, Brewe and colleagues (2016) found that while controlling for time 1 self-efficacy, PageRank centrality predicted time 2 (end of course) self-efficacy. Interestingly, and more substantively, PageRank centrality was also significantly related to students’ mastery experiences (a source of self-efficacy beliefs). they also found in-degree to be associated with verbal persuasion (another source), and out-degree to be related to both verbal and vicarious learning experiences (sources). Together, they further supported the logic that self-beliefs are very difficult to pin down, understand what promulgates them, and how they develop. Nevertheless, the relationships people have and how their network is structured is clearly predictive of their future self-efficacy. This suggests that students’ social interactions play an important role in further supporting students’ self-beliefs, which are well known and well supported to be a primary, if not the most influential, component of what motivates students to learn, succeed, and keep pushing despite troubles.

 

As can be seen from these two studies, node centrality can tell us a lot of the social positioning, influence, and relationships present in a network or classroom. As outlined in Lui and colleagues’ (2017) work, mean based descriptives about a sample and the interactions often found between students in aggregate can only tell us but so much about who is influential, what their influences means to others, and what it may mean to other, obviously well known and examined parameters of student motivation. In all, it seems social network analysis can serve as an intermediatary between qualitative observation that is often aggregated into frequencies for analyses and more traditional survey methods that rely upon mean based variable centered analyses. The power of social interaction and social environmental influences are, surprisingly, now making a prominent and robust place in student motivational research. That is, just in recent years, scholars in self-efficacy have called to further acknowledge the social, cultural, and interactive power of such relationships (Pajares, 2003; Usher, 2015, Bandura, 2018). They are vital to better understanding motivation.

 

 

 

Social Capital in Networks – Blog 4

 

“social capital” refers to features of social organization such as networks, norms, and social trust that facilitate coordination and cooperation for mutual benefit.

By and large, Robert Putnam’s bestseller, Bowling Along: The Collapse and Revival of American Community, argues that although Americans have increased in wealth and prosperity, their sense of community has declined (2000). His analogy to bowling, likely synonymous to face-to-face communication and interactions, was to say that although something is large, popular, and frutiful, it can be pervasive, lack cohesive congruence, and continue along upon the same footing that established it. In terms of face-to-face communication, or even voice communication (which is perhaps the better aligned), the advent of mobile phones, land lines, and accessibility to each, perpetuated the rise of communication in the American society. Although, similar to Putnam’s analogy, the continued rise of communicative ability has, as we’re all likely aware, led to a preponderance of less genuine, visceral, and authentic communication (texts, emojis, and emails).

Outlined by Portes and Landolt (2002), social capital’s initial footing by Coleman has been extensively expanded well beyond its initial scope. This is to say that social capital has expanded to ‘collective’ capital, fully causal in positive outcomes and benefits, and only positive in effect (Portest & Landolt, 2002). Nevertheless, “sociability, in every sense, cuts both ways” (Portest & Landol, 2002, p. 4). That is, there is a consistent give and take associated with being more social, whereby you have to give up freedom and individual directives in collaborative or communitive relationships. That being said, the notion that all networks or social capital harnessing collectives result in a net positive has to be false. There are plentiful networks that function well enough to be categorically negative or incongruent to society. Examples may include gang networks, drug networks, or perhaps oppressive regimes. Although, constrained within the view of those networks, I’m sure their net capital is, in fact, positive.

To this point, perspective, is everything. The size, shape, and members of any given network and how you define such a network, likely matters to how you may define its net overall social capital effect. There are, as Smith (2005) explained, far more variables, contextual components, and motives propelling the collective outcome of social capital. In his qualitative study, an individual’s reputation, relationship strength, and those seeking job help’s reputation were determinants upon whether they received aid, help, or an outreach of social capital (Smith, 2005). To this end, social network analysis could provide answers to help better understand what is going on here. Examples may include: In what ways does a person’s connections, the quantity of them and quality of them, matter in receiving help or social capital? Does having more distant friends or relationships enable more access to capital? Does being close friends with more people support receiving more social capital?

Outside of gaining jobs or receiving capital to escape poverty (Smith, 2005), the use and nature of social capital can be found in other areas. For example, in academic discussions the extension of capital between studies is the footing of the scientific method. That is, the transfer of knowledge (often) openly extended, is iteratively used by others to further knowledge, and so forth. Never minding the obvious economic railroads evident in publisher’s agendas, the scientific method and perpetual iterative nature of science can be seen as movement of social capital. In other words, science is predicated upon a collectively agreed upon notion of norms, practices, and methods that help facilitate the coordination and cooperation of many for the continuance of knowledge gain. Although, it is worth saying, this doesn’t always work as seamlessly and as congruently as we’d all wish… as personal directives, egos, and the like get in the way of true or pure collaboration.

Nevertheless, the definition provided at the top here is central to how researchers collectively work together in a collaborative effort to gain, progress, and expand human knowledge–social capital. Therefore, in my own trade: research, the concept of social capital is central and vital. This is also evident in institutional (meaning researchers) push for open science, or a purer translation of social capital ingrained within knowledge.

Also, one of my favorite websites for science news: https://www.sciencedaily.com/terms/scientific_method.htm

Check it out and the related articles surrounding building knowledge!

Weak Ties & SWT Blog 3

In essence, small world theory suggests that the natural clustering of relationships is fairly limited on an individual level, yet highly connected through overlap of ajoining clusters. This, as described by Milgram (1967), permits any two people arbitrarily selected to be connected through a series of efficient connections. In Milgram’s (1967) study, the average number of connections it took to send a package from one person to another arbitrary person was 6, however, he also found some other interesting features of the connections. The “pattern of contact” tended to be associated with gender, through mostly friends and acquantences, likely to, at some point, have similar pathways through key people/nodes, and appeared to be limited more by social ties than physical distance (Milgram, 1967).

Striving to take a more detailed approach to describing the structural issues of social networks, Granovetter (1973) specifically assessed the influence of weak ties therein. In this, he explained that weak ties often exist as local bridges (shortest path between two nodes) and that such local bridges among a larger network is the most effecient means to spread information (as it is the shortest path). Therefore, the importance of local bridges, or weak ties, resides in their ability to efficiently move information through a path of least ties. In other words, if information is moved through a dense network, it has a good chance of being only locally dispersed, as those in the immediate cluster have likely quickly heard the information many times and lost the motivation to spread (Granovetter, 1973). Furethermore, the more weak ties a network has the more cohesive in information spread and allowance for innovation, as ideas have a greater probability of moving about equally without a dense network to silo (Granovetter, 1973). Nevertheless, although strong ties tend to directly impact us in our day-to-day lives and provide ample diversity of information, weak ties even further allow outside information to permiate our localized network clusters (Kadushin, 2012).

Saying that either weak or strong ties are more important is a complicated and situational issue. In other words, in some situations weak ties become of greater value and importance, such as spreading salicious information (if the objective is to spread the information), whereas strong ties may be more “important” in generating group cohension of a particular egocentric network (this may be useful or “important” for marketing) (Kadushin, 2012). Therefore, it is difficult to position either as more important than the other. They both serve particular and useful functions that likely bring both bad and good outcomes in their operation.

17 million social ties in 55 countries were used to document the influence of weak ties on job attainment:

More people get jobs where their weak ties work. However, this is not because weak ties are more helpful than strong ties – it is because they are more numerous.

Keven Bacon…

Putting this all together, the “Keven Bacon: Six degree of seperation” is a viable example popularized. That is, it posits that everyone (in Hollywood) is connected with Kevin Bacon  through 6 degrees of seperation. In this sence, through the use of (likely) weak ties (if Hollywood actors have their own egocentric networks… I’m not really sure here), every actor can likely find a route of connections to Kevin. For what it’s worth, this likely helped popularize social network analysis to some degree and even promulgated a National Geographic Channel show titled “The Human Family Tree” that served to educate.

 

Value of SNs – Blog 2

Anton Chekhov (one of the greatest writers of all time) once said,

“Knowledge is of no value unless you put it into practice,”

which exemplifies and typifies value of social networks. What is to value to me may very well not be of value to you. So why then are social networks of value to anyone? Or are they purely applicable to all? What can social network data/analyses tell us, how can it push us forward, how can we, as Chekhov suggested, use it?

The value of social networks is layden in the simplicity of what it is–relationships. As clearly articulated in Kadushin (2012), a social network is “…a set of relations between objects…” that permit the extensions of information, knowledge, and culture (Haythornthwaite, 1996). To this end, diffusion of such provides a means to describe value in terms of social capital (Kadushin, 2012). This social capital ensures people have access to resources, information, help, and the extension of values, culture, and ideals. Through the active involvement in networks, social capital is built, extended, and perpetuated outwardly.

To assess how this works, researcher have collectively pronounced and described what precisely the social network perspective is, how it can be used to describe behavior, and how it can be beneficial in understanding our society (Keim, 2011). That said, a ‘social network perspective’ does not draw limits to particular groups, but instead allows the full scope of interaction to be modeled (Keim, 2011). Further, such a broad scope dispells the reliance on an individual while squarely focusing on the interactions between all players within a given society/sample to identify salient patterns or structures of networks (Keim, 2011). Network structures, or patterns, are, as explained by Keim (2011), the mending of both individual and societal level determinants to help explain the relationship between such micro and macro levels. In other words, the social structure at large may constrain individual behavior, establish particular norms, and provide secure platforms for interests to exist, while individuals also have the ability to change the social structure itself through their own actions and relations with others (Keim, 2011). This, bidirectional relationship between social structure and individual relations permits and predicates the study of social networks.

With a greater reliance on relationships, social network data is inherently different and uniquely arranged. That is, relationships can exist in direction, strength, level, and time. Ajoining all of these dimensions requires data itself to be structured in a far different way than is traditionally found in either long or wide datasets that are conventionally arranged in rectangular arrays. To show network data appropriately requires the individual relations between the players in the network to be quantified. Therefore, network data is often depicted as a square array, whereby rows and collumns are, essentially, the same thing, which enables relational quantities to be recorded (adjacency matrix). This, differentiated from more traditional and conventional data arrays, permits the perspective to shift towards seeking an understanding of relationships between all the rows and all the collumns. Ultimately, however, a focus on relationships between the individuals, nodes, or actors in a relational dataset diminishes the ability to clearly assess the inherent properties therein. In other words, causality is difficult to discern within a relational dataset, when methological design and reliance on individual level attributes are not often primary in analysis (Yang, Keller, & Zheng, 2017). Nevertheless, to fully understand and interpret the relations it is imperative to describe the sample descriptively. In general, as the actors can be just about anything, this include descriptive accounts, averages, minimums, maximums, and distributions of important variable criteria to best describe who the actors are and why they may relate the way they do. As described by Yang and colleagues, such individual level attributes, especially in relationship cross-sections, are context and time dependent. That is, and falling squarely in-line with the notion of a structural-relational model described earlier between individuals and the societal structure, such attributes mean less when focusing on relations at any given moment or between peopel in particular contexts (Yang et al., 2017). Either way, the dynamic interplay here makes prediction inherently difficult from both a theoretical explanatory sense and non-independence statistical sense.

(I found this video particularly useful in better understanding the structural-relational model and the network perspective)

In sum, social networks are dynamic relationships between the societal structure and individual level relationships therein. They are categorically different than traditional conventional data arrays, yet allow for a focus on relations between actors, as opposed to traditional inferential statistics associated with common actor attributes. This situation begs for contextual descriptives, yet complicates predictive analyses because relational data is highly contextual and clearly non-independent (you know, an assumption of regression… ).

 

Everything is Connected – Blog 1

One of the most fascinating things about research, about our lives, and our existence is how prevalent patterns are. We seek patterns, search for anomalies, and, through the scientific method, seek to adjust our view based upon differences in patterns we notice.

“We are all connected; to each other, biologically. To the earth, chemically. To the rest of the universe atomically. We are not figuratively, but literally stardust.” –Neil DeGrasse Tyson

Drawing from Christakis and Fowler’s (2011) book, networks are useful, positively oriented, groups of people who are connected by some facet of life. This can be broad, general, or intricately defined, however, this got me thinking–in what ways can networks be used to figure things out, figure out the world, or answer questions–how can we use them to conduct research.

In searching the internet, I came upon this, which is a perfect example of how networks can show us, or better yet, predict outcomes. In short, they used data that described how internationally connected the players were on each World Cup team, as international players were likely better players than those simply on domestic teams. Using an algorithm to compute a weighted form of centrality based upon how many international connections each player had, each team computed a weighted centrality score. Sure enough, a positive linear relationship existed between teams who were more central (had a greater amount of international players with more international connections) and succeeded in the World Cup.


Although we all have structural relational networks and they exist everywhere, how we define them, as Christakis and Fowler (2011) explain, and what attributes of interest we seek to examine, matters in how we can interpret them. In other words, the “ties of interest” are vital to viewing, understanding, and using social network analyses (Christakis & Fowler, 2011, p. 28). Nevertheless, central to this book, at least from my reading, was that networks are complicated, yet simple in that we are all connected by only six degree on average, flexible, always changing, greatly dependent upon what metric is used to describe the ties of interest, innate (genetically.. we are built for them), and inextricably useful in understanding how connections between people interact, disperse, and influence (such as the Three Degrees Rule in how phenomenon spreads). This is exemplified therein by their own research and examples of other situations and research. Most notably, to describe the complicated nature of networks and the propensity for dispersion to occure, the authors describe laughing fits, hysteria, and even colds that spread in online environments that illicit very human like behavior. Further, the authors describe how natural selection likely shaped our reliance and strive to establish social networks, or at least who we allow to participate in our own social networks. In sum, networks provide advantages above and beyond what any one of us could do alone. Christakis and Fowler (2011) explain these tendencies have continually progressed through time into fairly naturally occurring divisions of people into towns, villages, and cities across the globe. Such that, in order to coexist and cooperate, particular innate tendencies are ignored, such as selfishness. Also, genetics and heritability are explained as very powerful in explaining the natural variation in many arenas such as obesity, good looks, politics, happiness, smoking, and how ept we are to introduce our friends. Ultimately, Christakis and Fowler (2011) describe social networks as forming what they refer to as a “human super-organism.” By this, they explain together, through networks, humans excel, function at far greater abilities, and can replicate, cooperate, and ultimately project positive and desirable outcomes ubiquitously. We are simply stronger together than apart.

Interesting, Christakis and Fowler (2011) say, “Embedded in social networks and influenced by others to whom we are tied, we necessarily lose some of our individuality,” which seems incongruent to Simmel’s statement we reflected upon earlier. In my opinion, both have their points.

Ultimately, we all have a greater impact on each other than we notice or can even begin to fathom. As some put it, “surround yourself with good people” means far more now than it use to.


This conversation is especially relevant in our current accelerated advance of social networks. We can, as never before, connect with others. Although, it seems digital social relationships do not fully mimic personal relations, I do, however, wonder how much hyperdyadic spread is either lost or gained in particular areas. For example, I wonder if obesity has the same spread as, say, political commentary/afflictions seemingly do within a digital landscape.

Based upon a cursory review of social media related social network research, there appears to be a great deal of activity related to understanding how social media use interplays with various elements. Interestingly, I found a few videos that may be of interest. First, by James Surowiecki, a foray into crowds and groups in social media.

Secondly, a video that seems to pull at a lot of heart-strings in fairly natural ways, is Cal Newport’s take on releasing your reliance on social media.

This brings up a few really important questions that we, as active and very busy students, likely deal with. What is the worth of social media participation? How much of our time should be spent online in a social media environment as opposed to face-to-face communication? If we completely withdraw from social media, can we still survive socially or academically? I, personally, do not believe social media is that much a part of our lives and livelihood–it can be reduced, but that is, obviously, a very personal decision.

I’ll leave you with a short task (…if you’ve read this far…). This will take less than 5 minutes.

Step 1:Go to https://www.library.vcu.edu

Step 2: Type in “social media, social network analysis, yourareaofresearchhere.” For example, for me I would enter: social media, social network analysis, education

Step 3: Find a relevant article to your area of research and annotate (with hypothesi.is) the citation.

Cheers! #SNA #SNA18