Examining social networks in workplace safety and health: Why and how?

Transcription

1 National Institute for Occupational Safety and Health Examining social networks in workplace safety and health: Why and how? Jessica MK Streit, PhD(cand) Research Psychologist Work Wellness and Disability Prevention Institute Webinar Sept 28, 2018

2 Agenda Social network fundamentals Definitions and terms relevant for social networks Methods for obtaining social network data Uses of social network data in workplace/workforce research Approaches and programs for analyzing social network data Examples Published literature In-process studies at NIOSH Full network: scientific impact (Streit & Ward) Ego: Social support systems for multi-role workers (Streit & Rosen)

3 Human interconnectedness Who have you seen since waking up this morning? Who have you called, texted, or ed today? Who are your favorite coworkers? Who are your closest friends? Who do you talk to about important matters?

4 Human interconnectedness Who did you see since waking up this morning? Are the individuals you identified... Who are your favorite coworkers?...the same? Who shares meals with you?...different? Who are your closest friends? Who do you talk to about important matters?

5 istockphoto Human interconnectedness Families Friendships Workplaces Schools Self Dependents Points of service Restaurants Hospitals Public places Parks Shopping centers Roadways...and on and on and on...

6 istockphoto Human interconnectedness In-person Snail mail Technology Phone Text Social media Facebook Instagram YouTube Twitter Pinterest LinkedIn Other websites and apps

7 Human interconnectedness Families Friendships Workplaces Schools Self Dependents Public places Parks Shopping centers Roadways Points of service Restaurants Hospitals...and on and on and on... In-person Snail mail Technology Phone Text Social media Facebook Instagram YouTube Twitter Pinterest istockphoto LinkedIn Other websites and apps

8 Traditional research istockphoto

9 Traditional research istockphoto

10 Traditional research istockphoto

11 Social network research istockphoto

12 Social network research: Terms to know Actor or Node: each member in the network -Person -Organization Path: aka tie; connection between actors or nodes Dyad: pairs of actors or nodes connected by a path Network: a group of dyads with specified bounds istockphoto

13 Social networks: Connections and roles istockphoto

14 Social networks: Connections and roles istockphoto

15 Social networks: Connections and roles istockphoto

16 Social networks: Connections and roles istockphoto

17 Social networks: Connections and roles istockphoto

18 Social networks: Common dyads Category Co-occurrences Social relations Interactions Flows Varieties Co-membership Co-participation Physical distance Attribute similarities Kinship Affective (likes, dislikes) Perceptual (who knows whom) Transactions Activities (does with) Information Goods Illness or disease Adapted from Borgatti, Everett, & Johnson (2013) Table 3.3

19 Social network measures Size: how many people (in network, known/related to, etc.) Centrality: nearness to center of network; measure of importance/influence Degree: # links from actor-to-actor Betweenness: how often an actor is a connector for other actors Closeness: average distance from all other actors Centralization: hierarchical structure of network Cohesion: how many connections in unbroken path? Density: # connections / total links possible Reciprocity: # bidirectional (v. unidirectional) relationships Transitivity: # triads (A B, B C, A C; unidirectional or bidirectional) Clusters: clumpiness or dense pockets of activity

20 Social network research designs Whole network Bound set Data from all actors in set Values assigned to all dyads Pros Complete network info More analysis options Cons Costly (time, $) Data quality Personal network Focal actors: egos Ties: alters Egos describe alters Pros Boundaries not required Less costly Richer data Cons Localized networks

21 Social network data collection approaches Primary: ask questions; observe behavior Most common: ask about ties with others Closed-ended (roster) Reduced recall bias Increased workload with large networks Open-ended (write-in) Increased recall bias Reduced workload Secondary: existing data (records, databases) Easier to collect Limited relationships to be studied

22 Social network data collection techniques Techniques Survey Egocentric Sequenced Census Two-mode Archival Description Ask questions about behaviors and relationships Ask respondents to name and describe people in their networks Use snowball approach (nominations) to grow study sample Survey/interview [almost] all members of priority population or target community Aka jointness or duality data; record instances where people participate/attend same event Logs, diaries, telephone records, bibliometrics

23 Social network data uses Exposure Diffusion of innovations Information and ideas Practices and behaviors Recruitment Interventions ID opinion leaders and key players ID groups ID leaders within groups Create leader-learner pairings Changing networks Investigating network and member attributes

24 Social network data roles For statistical analyses Role Social capital paradigm Social homogeneity paradigm Explanatory (IV) Outcome (DV) Predictive: how do network traits explain outcome of interest? Acquisition: how to egos come to have their networks? Influence: how are ego behavior and attitudes affected by alters? Homophily: how similar are egos and alters on socially relevant attributes? Adapted from Borgatti, Everett, & Johnson (2013) Table 15.3

25 Social network data analyses Packages and programs Excel SAS, SPSS, R, STATA UCINET + NetDraw Analytic procedures Specific to social network packages Export SN data to be used in descriptive and inferential statistical models Li et al. (2013)

26 Errors to consider Omission: missing data; false disconnects Commission: erroneous inclusion of nodes or ties Attribution: behavior or trait incorrectly assigned to node Data collection: memory failures in retrospective recall Ego bias: variability in how well people report relationships Active/central/powerful egos more accurate

27 Application: Examples Published social network literature

28 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

29 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

30 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

31 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

32 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

33 Published social network studies* Study Topic Use of SN data Barr, Ensminger, & Johnson (2009) Casciaro (1998) Chow & Chan (2008) Cultural differences in trustworthiness Personality and network recognition accuracy Social networks and knowledgesharing IV: social capital DV: network accuracy IV: social capital Christakis & Fowler (2007) Obesity IV: ties (links) to obese actors at Time 1 Johnson, Booster, & Palinkas (2003) Lomas et al. (1991) Rosenquist, Fowler, & Christakis (2011) Tanjasiri & Tran (2008) Zohar & Tenne-Gazit (2008) * Not a mutually-exhaustive list Group dynamics in secluded worksites Opinion leader effectiveness (VBAC) Spread of depression Improving health information dissemination (cancer) Transformational leadership and safety climate IV: network structure IV: opinion leader promotion of VBAC IV: network ties DV: linkage rate IVs: talking frequency, network density, centralization

34 Application: Examples In-process NIOSH research projects

35 Scientific impact and co-authorship ties Whole network example Purpose: To understand how the composition of the universe of Journal of Occupational Health Psychology (JOHP) authors affects researchers impact on the field of occupational health psychology Relational capital Prolific author connections International connections Structural capital Degree centrality Research Impact Article count Research Impact Citation count Cognitive capital Tenure Training program Source: Li et al. (2013) [modified]

36 Scientific impact and co-authorship ties Whole network example Evolution of JOHP network:

37 Scientific impact and co-authorship ties Whole network example Evolution of JOHP network:

38 Scientific impact and co-authorship ties Whole network example Evolution of JOHP network:

39 Scientific impact and co-authorship ties Whole network example Evolution of JOHP network:

40 Scientific impact and co-authorship ties Whole network example Evolution of JOHP network:

41 Scientific impact and co-authorship ties Personal (ego) network example Purpose: To explore the association between college students employment characteristics and composition of social support systems Sample: Midwest college students (n = 212) Measures: Employment: # jobs; main job sector; second job sector; weekly paid hours Network: relationships; years known; concordance/discordance; communication frequency; closeness; familiarity; advice exchange Procedure: Secondary analysis of survey data collected to examine factors influencing human papillomavirus (HPV) vaccine uptake

42 Scientific impact and co-authorship ties Personal (ego) network example

43 Scientific impact and co-authorship ties Personal (ego) network example Trends investigated Rate of employment Significant predictors of paid working hours Homogeneity of social networks on personal attributes Sex, age, race, living situation, grade, major # jobs, occupation, employment sector Interconnectedness of network How many connections know one another in an ego network? Activation of social network What are the relationship categories of most salient supporters (alters)? Factors influencing feelings of closeness to network ties

44 Closing thoughts

45 Takeaways Social network data provides useful information about connectedness Examining relationships can complement studies of individual behaviors and attitudes Connectedness (e.g., social circles, social capital) is important factor to consider when studying a variety of work-related topics: Organizational processes and practices Group dynamics Information exchange Behavior change Health conditions and exposures Psychosocial factors (e.g., social support) Many different future directions for research using social network approaches

46 References Barr, A., Ensminger, J., & Johnson, J. C. (2009). Social networks and trust in cross-cultural economic experiments. In K. Cook, R. Hardin, & M. Levi (Eds.), Whom can we trust? How groups, networks, and institutions make trust possible. New York: SAGE. Borgatti, S. P., Everett, M. G., & Johnson, J. C. (2013). Analyzing social networks. Thousand Oaks, CA: SAGE Casciaro, T. (1998). Seeing things clearly: Social structure personality and accuracy in social network perception. Social networks, 20, Chow, W. S., & Chan, L. S. (2008). Social network, social trust and shared goals in organizational knowledge sharing. Information & Management, 45, Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32 years. New England Journal of Medicine, 357(4), Johnson, J. C., Booster, J. S., & Palinkas, L. (2003). Social roles and the evolution of networks in isolated and extreme environments. Journal of Mathematical Sociology, 27(2-3), Li, E., Liao, C. H., & Yen, H. R. (2013). Co-authorship networks and research impact: A social capital perspective. Research Policy, 42, Lomas, J., Enkin, M., Anderson, G. M., Hanna, W. J., Vayda, E., & Singer, J. (1991). Opinion leaders vs. audit feedback to implement practice guidelines: Delivery after previous cesarean section. Journal of American Medical Association, 265,

47 References Rosenquist, J. N., Fowler, J. H., & Christakis, N. A. (2011). Social network determinants of depression. Molecular Psychiatry, 16(3), Streit, J.M.K., & Rosen, B. L. (2018, May). Exploring social support network structures for multi-role workers. Poster presented at the Second International Symposium to Advance Total Worker Health, Bethesda, MD. Streit, J.M.K, & Ward, R.M. (2017, June). How do coauthorship relationships affect research impact? A Social Network and Citation Analysis of the Journal of Occupational Health Psychology, Paper presented at Work, Stress, and Health 2017, Minneapolis, MN. Tanjasiri, S. P., & Tran, J. (2008). Community capacity for cancer control collaboration: The Weaving an Islander Network for Cancer Awareness, Research, and Training. Cancer Detection and Prevention, 32, S37-S40. Valente, T. W. (2010). Social networks and health. New York: Oxford University Press. Zohar, D., & Tenne-Gazit, O. (2008). Transformational leadership and group interaction as climate antecedents: A social network analysis. Journal of Applied Psychology, 93(4),

48 Acknowledgements Dr. Brittany Rosen, CHES, FASHA University of Cincinnati Cincinnati, OH Dr. Rose Marie Ward Miami University Oxford, OH

49 Thank you! Presenter contact: For more information, contact CDC CDC-INFO ( ) TTY: The findings and conclusions in this presentation are those of the author and do not necessarily represent the official position of the National Institute for Occupational Safety and Health, Centers for Disease Control and Prevention.