Networks within and across Media

1. Networks within and across Media Hyo Kim CIT Ajou U.Gwang Jae Kim Sogang U., Seoul, Korea Han Woo Park YeungNam U., GyeongSan, Korea Ronald E. Rice U. California, Santa Barbara, CA

2. General Problem Statement To what extent can new media maintain, expand, or decrease existing relationships? How might those relationships differ from more traditional FtF relationships?

3. mediated communication can foster “connected presence” (Licoppe & Smoreda, 2005) “communicative readiness” (Nardi, 2005) managing multiple encounters at the same time and across time ongoing awareness of and relationships with specific others cues about one’s availability and attention to the other person a social context (ongoing “common ground” or “field of connection”) how other media choices and uses are interpreted (Nardi, 2005, Wei & Lo, 2006)

4. Nardi, Whittaker, and Bradner (2000): IM used by members of an ongoing work group as a channel to seek and exchange content used as a coordinating tool for managing accessibility and flow of interactions used to maintain a small network of fellow IM users rather than to connect to new others Nardi, Whittaker, and Schwarz (2002): workplace IM buddy lists contained six friends/family and 16 coworkers but the users frequently interacted only with four or five of them

5. Schiano et al. (2002): teenagers communicated regularly with fewer than 5 IM buddies Grinter and Eldrige (2003): fewer than 3 regular communication partners for UK IM Ito (2005): fewer than 3 regular communication partners for Japanese mobile phone users Miyata (2006): mobile phone emails did not affect one’s network diversity but did foster more supportive network ties PC emails increased network diversity, especially through more weak ties

6. Research Questions Individual Level RQ1: How do the number, diversity, type and closeness of communication relationships vary across media, and by respondent’s employment category? Network Level RQ2: How do configurations of relationships vary across media and by employment category? In particular, how do configurations of relationships in new media compare to those in FtF, overall and across employment categories?

7. Method: Korean Context population of nearly 50 million 2003: 98.1% literacy; 7th highest level of information use (consumption flows of ICTs/time period) 61% Internet; 70.1% mobile phone; 53.8% landline 2004: 11th highest Internet penetration (63%) 2005: greatest penetration of mobile phones with color displays in 7 Asia and Pacific countries 83.6% of middle- and 89.3% high-school students used wireless Internet. Youths (aged 12 to 19) who have their own mobile phones used the Internet 6.7 times per week

8. Method: National Web-based Survey and Network Generator Web survey was designed by 10 Korean communication researchers Web sampling designed to represent the Korean adult population, esp. new media users commercial company maintains a large master survey pool registrants provide demographic and media ownership and use information company uses demographic information to create a proportional stratified sampling frame by gender and age categories, appropriate for the specific survey topic entire master pool is solicited by email, survey stays open until each category is filled 1507 people responded

9. Method: Demographics 11 employment categories grouped into 6: salaried, homemaker, middle/high student, college student, IT professional, and others “other” category, with 468 respondents, not included in the analyses 1039 respondents Sex: 44% male and 56% female Age: 29.8% 13-19, 27.9% 20-29, 27.7% 30-39, and 14.5% 40-49 Employment categories, 27.3% salaried, 23.5% home worker, 14.6% middle/high student, 22.9% college student, and 11.6% technical/professional

10. Method: Media, Partners On separate web pages: For each medium -- FtF, EM, IM, MP, SMS: respondent identified up to 5 communication partners with whom they communicate most frequently listed each partner by number (i.e., “person 3”), identified that partner distinctively across the media (so used same number if same person in different media)

11. Method: Social Role, Closeness checked partner’s social role (spouse, children, parent, sibling, other relative, elementary/middle/high school friend, college friend, girl/boy friend/lover, other types of friend, work colleague, work boss/managers, work subordinate, other work related, teacher/professor, or online only) closeness of each communication relationship

12. Analyses: Social Role Network Matrices Example raw initial data: respondent’s number of social roles for Face-to-Face (1039x15)

13. Social Role Co-Occurrences Each 1039x15 media matrix Gk converted into a matrix 15x15 Ak Matrix Ak = Gk*Gk' Ak is a 15x15 social role by social role matrix, aggregated across the 1039 respondents, for medium k.

14. Example for Face-to-FaceDiagonals: total number of mentionsOff-diagonals: total number of co-mentions

15. Transform Ak frequency matrices into correlation matrices Correlation between any two social roles in a medium matrix indicates extent to which those two social roles have similar patterns of frequencies with all other social roles in the Ak matrix Example correlation matrix for Face-to-Face social roles co-occurrence:

17. Results Individual Level: Mean Relationships by Medium and Employment Category FtF: no significant differences in the number of communication partners across employment categories EM: significantly higher number of relationships for worker (salaried and tech professionals) IM: home workers had significantly fewer relationships MP and SMS: middle/high school students have fewer relationships

18. Individual Level: Number, Unique and Duplicate Relationships by Medium and Employment Category

19. Individual Level: Unique Relationships number of unique relationships varied by employment category significant mean differences for: middle/high student (2.7), college student (3.52), homeworker and tech professionals (4.16 and 4.25) tech professionals and salaried (4.25 and 4.67) more work-related partners both FtF and in email more in FtF than email

20. Individual Level: Duplicate Relationships significant difference in number of duplicated relationships in FtF and email among the employment categories mainly between mid/high student and working categories college student and home worker not significantly different from the working categories significant difference in number of duplicated work-related communication relationships in FtF and email 2 work categories differed from the 3 non-work categories

21. Individual Level: Comparison of Social Roles and Employment Category Overall and by Medium Largest number of relationships: overall: with friends, then family, then work, and then online students: friends, through FtF, IM, SMS and mobile phone homeworkers: family relationships, primarily through mobile phone and FtF salaried and tech professionals: work relationships, primarily through FtF and email

22. Individual Level: Relational Closeness by Medium and Employment Category significant medium effect: IM for weak relations MP for close ones significant employment category effect: homeworkers: lower closeness with others compared to all other employment categories primarily due to low closeness through IM significant interaction effect: media by employment category students use email to communicate with less close others in comparison to others

23. Individual Level: Relational Closeness by Medium and Employment Category

24. Network Level: Descriptive Analysis of Configurations of Relationships by Medium correlation (extent of similarity) between pairs of the matrices statistical significance of those correlations Quadratic Assignment Procedure (QAP) in Ucinet 6.0 compute correlation between each media pair; all highly intercorrelated: MP and IM = .57 MP and FtF = .90 enter correlations into a 5x5 (medium by medium) matrix (ignoring the diagonals) provided to Ucinet’s multidimensional and hierarchical procedures

25. Network Level: Descriptive Analysis of Configurations of Relationships by Medium configurations of relationships among the social roles in FtF and mobile phone quite similar (clustered) Email and SMS also tightly but separately clustered IM is less tightly clustered with email and SMS IM least similar to FtF and especially MP

26. Multidimensional Scaling of the Five Media Configurations of Relationships, with Hierarchical Clustering

27. Network Level: Statistical Analysis of Configurations of Relationships by Medium multiple regression quadratic assignment procedure (MRQAP) with semi-partialling (Krackhardt, 1988) compute the overall R2 of FtF matrix by each new media matrix MP is strongest association with FtF configuration swamps all other media due to strong intercorrelations of matrices

28. Network Level: Descriptive Analysis of Configurations of Relations by Medium and Employment Category MDS of relationships among the social roles for FtF, email, and IM three most different media from each other line thickness proportional to the similarity (correlation) between each pair of social roles

29. Figure 2. Configuration of FtF Relationships

30. Configuration of FtF Relationships configuration of relationships in FtF is quite differentiated, with school relations on one side, family relations in the middle (with spouse and children central), and work relations on the other side relations involving spouse and children are most central, and mediate between other relatives and work relations within work relations, co-worker and boss relations are most similar relations among school and friends and one’s professor are quite distinct from relation/ family/ work relations

31. Figure 3. Configuration of Email Relationships

32. Configuration of Email Relationships denser interrelations among workers also among relatives presumably because of ability to overcome distance and time constraints asynchroneity, unlike mobile phone online and teacher/professor relations completely distinct social arena, presumably representing the educational communication domain

33. Figure 4. Configuration of Instant Messenger Relationships

34. Configuration of Instant Messenger Relationships a family medium primarily children, spouse, sibling and parents more centrally but less frequently, relatives and other friends also shared relations involving workers, but somewhat less separated from the family network than in FtF and especially email communication with spouse and work boss somewhat similar in FtF and in IM, but not in Email

35. Configuration of Mobile Phone Relationships relations among the social roles much denser than in the other media except clearly distinct dyad involving college friends and girl/boy friends and lovers still closely related through the spouse, work subnetwork and the family subnetwork distinct stronger similarities/correlations within each subnetworks than in FtF network

36. Configuration of SMS Relationships siblings, children and parents at center of the SMS network used to coordinate the activities and whereabouts of family members strong relations between spouse and work subordinates: people trying to manage both work and family life

37. Network Level: Statistical Analysis of Configurations of Relations by Medium and Employment Category separated original 5 data matrices into 25 - one for each medium (5) by employment category (5) for each, tested influences of the new media configurations on the FtF configuration

38. Network Level: Statistical Analysis of Configurations of Relations by Medium and Employment Category similarity between MP and FtF relations strong for all employment categories students: report most similar relations with social roles in FtF and MP homemakers: relations through email and instant messenger independent of their FtF relations may use CMC media to keep in touch with social roles cannot frequently meet in person -- i.e., old friends, relatives, etc.

39. Network Level: Statistical Analysis of Configurations of Relations by Medium and Employment Category salaried and technical/professional users: greatest similarity between email and FtF mobile phone also similar to their FtF organizational workers: email used to maintain same kinds of relationships managed through FtF old enough to feel comfortable using email email crucial medium supporting work tasks with fellow employees

40. Summary FtF the common medium for the three major employment categories each category has its own distinctive set of complementary media: IM, SMS and mobile for students MP for homeworkers email for organizational workers MP a multiplexing device for maintaining everyday relationships

41. Summary clear divide between: Email by older people MP (and IM and SMS) by younger people organization supportive context for using email to communicate with others homemakers used CMC media (email and IM) to communicate with people who may not readily appear in their FtF communication

42. Summary friendship clearly an important element of Koreans' social lives social context and network approach provides insight into social context for use of traditional and new media aspects of the medium, purposes, and the participants influence how one chooses, uses, and perceives different media (Licoppe & Smoreda, 2005)

43. Thank you!