Our "Social Network" surveys measured 20 respondents' weak tie access and strong tie discussion networks. Overall, I found interesting correlations between education and network size, spousal connectedness between married husbands and wives, age and new media utilization, and similar patterns on homophily and trends in relationship roles for neighbors, co-workers and members of groups in McPherson's writings. The social networks portrayed in my survey results support Wellman's rise of personal network communities.
Nine out of ten of my respondents aged 18-22 were currently full-time students, which offered no contrast for various education levels and no married respondents. I was able to survey three respondents in the 33+ cohort that had high school level educations or less, and their results contrasted with higher educated respondents provoked questions of class and SES effects on social networks that our survey did not directly ask.
For network size, my average respondent had 4.4 discussion partners, varying slightly by gender (males: 4.; females: 5.3) and age (18-22: 4.8; 33+: 4). My smallest reported discussion network consisted of 2 people, which was listed 4 times. These results do not comply with McPherson's median of 2.08 discussion partners or 25% of respondents with zero discussion partners in "Social Isolation in America." Hampton and Marin also contend that important matters discussion partners measure strong ties, and my findings are closer to their average of 4.8 discussion partners. My data also diverges from McPherson's, in that her 2004 GSS analysis had only 15.3% of respondents list four or five discussion partners. In contrast, 44% of my respondents listed four or five discussion partners, which probably is due to the unequal education distribution compared to the population at large for the GSS.
In my 33+ group, average network size for respondents with a high school education or less was 2 and 5.2 for respondents with some college or more. Depending on the role of their relationship, strong ties provide a wide range of resources and support, such as emotional aid, large and small services, financial aid and companionship (Wellman & Wortley). My results show a positive association between educational attainment and strong-tie discussion partners, thus hinting that lower educational attainment for adults limits access to support. The correlation between strong ties and recovery from heart attacks in Dicken et al's study also would support the claim that these respondents with lower education and subsequently less strong ties are at a greater risk for health issues problems.
My results on roles of ties in discussion networks strongly mirror McPherson's data in Table 2. 21% of the 2004 respondents in McPherson's study listed a parent as a discussion tie and 19% of my respondents did as well. McPherson's decline of group members in discussion networks (11.8%) and coworkers (18%) almost exactly equaled my results (11.25% group members and 18% co-workers). However, only 1.25% of discussion partners in my survey were classified as a neighbor, a smaller percentage than McPherson's 7.9%. However, our data may differ due to the tendencies for full time college students to not classify strong, nonkin ties in their neighborhood as neighbors.
There was no correlation of kin inclusion in discussion network and position sum; however, over 80% of our respondents had 1/3 or more of their discussion network comprised of kin, with the maximum being a 56-year-old respondent's 4 out of 5 members kin. Overall, women were more likely to kin-keep, with 47% of their networks comprised of kin compared to 33% of male networks. As opposed to McPherson's findings, not one respondent in my survey answered only their spouse for discussion partners.
Our survey presented a position generator ranked from descending occupational prestige with 15 positions similar to the randomized position generator Lin uses to measure weak ties. In terms of prestige in the position generator, Nan Ling measures social capital by extensity of positions reached, the upper reachability, and range of positions accessed. The average sum, also known as Lin's extensity, (with each occupation equaling 1 point) was 7.9, with slight variation by gender (males: 8; females: 7.8). There was no significant correlation of position extensity, upper reachability, or range to education or gender in my study, as opposed to Lin's findings of gender significance in upper reachability and range. For my lowest scores, a 22 year-old male scored a 1 and a 21-year-old male scored a 3; the 22-year-old who scored 1 on the prestige generator sum had 3 out of his 6 discussion networks as kin, hinting at a privatized network with little diverse access to weak ties and unique information. The 21 year-old with the position sum of 3 only had 2 discussion partners, neither which were kin. There was a slight positive association between age and position sum, as the average position sum for 18-22 was 7.2, as opposed to 8.2 for 33 years or older respondents. Our sample differed significantly (ours is more educated, composed of less married people, and younger) from Lin's Taiwan networks study, and my average extensity (7.9) was higher than Lin's 6.5. The most frequently accessed position was the lowest, laborer, an interesting finding that shows a common ability to downward reach for all respondents (which follows Milgram's assertions of reaching higher level gatekeeping managers in an organization for trickle down prestige reach). Our 15 occupation position generator differed from Lin's relationship-including position generator in Appendix A of "The Position Generator;" I would have liked ours to have randomized and interspered the prestige of occupations like Lin's did because I did observe some unease and self-consciousness of respondents during this section that could have affected the validity of embedded resources. I also would have liked the inclusion of Van Der Gaag's resource generator to distinguish potential access from actual use of weak ties as well as to ascertain the degree to which different support is coming from weak-tie acquiantances, friends, or kin.
In terms of distance, my results showed a difference in strong tie location according to respondent's gender. 11% of males listed strong ties that resided in their home or dorm compared to 21% of females, an indication that males' networks are not as privatized and feminized yet as Wellman hypothesizes they are trending towards. Males also had more far-flung, geographically dispersed networks, with 55.5% of their strong ties residing outside of their state and in the same country as opposed to 26.2% of female strong ties. Only one respondent,a male, listed a strong tie from another country. However these distant male strong ties may not be as global, as not one male of any age group listed any ties outside of their city but in their state, as opposed to 17% of females and the state category being the modal response for 33+ females.
Barry Wellman argues that new media technology complements instead of replaces face-to-face interactions in a non-zero sum milieu, and my findings support his description. The most interesting findings were in the age distributions and new media usage. 60% of 18-22 year olds use IM, as opposed to only 25% of 33+ respondents. Only one respondent in the 18-22 group used postal mail, and only one 18-22 respondent used the landline phone, as opposed to 62.5% (postal mail) and 87.5% (landline phone) in the 33+ cohort. The discrepancies in age and landline phone usage are most likely due to the transient residential status of the younger full-time students. E-mail and cell phone communication were adopted by all 33+ respondents with some college or higher educational attainment. However, communication medium differs in terms of frequency of interactions in age groups. The 18-22 age group had 530 face-to-face interactions, as opposed to 357 for 33+ age group, probably due to the older cohorts working status. The 18-22 age group had more interactions over cell phone (541 cellphone versus 530 face-to-face) with their strong ties in the past thirty days, a stark contrast to the mere 184 cell interactions (compared to 357 face-to-face) among the 33+ grouping. Younger people are integrating cellular phones into their communication with strong ties more than the older age group, of which 25% did not use a cell phone at all to communicate with a strong tie in the past thirty days. Overall, the lower frequency of contacts the older group had with their strong ties supports Kalmijn’s dyadic withdrawal hypothesis that "the older people are...the fewer friendship contacts they report."
Wellman and Wortley claim that "respondents rarely see their most active network members more than twice per week," which differs from my findings of heavy face-to-face contact for strong tie networks. In educational analysis, two out of my three high school-or-less respondents did not communicate with a strong tie using e-mail, instant messaging, or cell phone, and one of these had only face-to-face contact; indicating a possible digital divide for utilization of new media communication technologies according to education. My results agree with Baym et al.'s that college students use face-to-face interaction more than internet communication, but disagree with the authors that "participants reported using the internet as often as the telephone." (299) Cell phone usage was the most frequently used medium of communication among my college students, with a total of 541 interactions, slightly higher than face-to-face (530) and significantly higher than Internet (e-mail and IM 411 combined) My findings also concur with Chen and Quan-Haase's National Geographic study in which "e-mail was used more with friends than relatives." (Social Interactions 303) New media use is most frequently utilized when strong ties reside outside of the respondent's neighborhood. Wellman's evolved network communities must "actively maintain ties instead of just rely on solidary communities to do this for them." However, two of my male respondents actually had no contact with their nonkin, nongroup strong tie in the past thirty days across all mediums. This finding also opposed Monge's notation that frequency of interaction is an important element in establishing strong ties with others. Both of these unaccessed strong ties were designated as "advisors," thus perhaps indicating an institutional role.
In "Shared Friendship Networks," Kalmijn evaluated the "dyadic withdrawal" hypothesis that friendship networks get smaller as couples cohabitate and that tie networks between partners begin to overlap more. Although we did not ask respondents' relationship status, we can analyze the differences for non-full-time students who listed spouses versus those who did not. Respondents listing a cohabitating partner had larger discussion networks (4.6) compared to those who did not list a spouse (3), but these findings are probably affected by the 2 low education, unmarried outliers with small networks.
For the five respondents who listed a spouse in their discussion network, there was a noticeable difference in husbands' connectedness to wife's discussion partners and wives' less connectedness to their husbands' discussion partners. For the 36-year-old wife, her husband was especially close to 3 of her ties and knew the remaining 2 ties. For the 56-year-old wife, her husband was especially close to 3 of her ties and knew the remaining 1 tie. Neither of the two married women's husbands were strangers to anyone in their network. This finding agrees with McPherson's description in "Birds of a Feather" of young females' greater likelihood to delete tie choice to resolve intransitivity than males. (422) The high rate of spousal connectedness in these women's networks may also be related to their high rate of kin (3/6 for the 36 yr old and 4/5 for the 56 yr old compared to lower rates of 1/5, 1/2, and 2/5 for kin for married men). For the three married men, the 61-year-old's wife was a stranger to 2 of his 4 other ties, the 47-year-old's wife was a stranger to his 1 other tie, and the 54-year-old's wife was a stranger to 1 of his 4 other ties. Two interesting correlations arose in this analysis: All of the husbands' individuated ties, save for the 47-year-old's advisor, were co-workers, which harkens back to Elizabeth Bott's interesting finding on occupation and spousal joint networks. Another finding is that my 47-year old married male respondent's discussion network is an example of Granovetter's forbidden triad. He has two strong-tie discussion partners which are strangers to each other, which violates Granovetter's idea of transitivity among strong ties in networks. Overall for married respondents, women had denser networks composed of more kin than their male counterparts. These dense kin-keeping, privatized networks support Wellman's description of feminized personal networks. While they have high redundancy effective in the diffusion of information, they also lack Burt's weak tie structural holes that can bring unique information and access diverse resources to a network.
In "The Network Community," Wellman asserts that networks become more sparsely knit as people age. In my survey results, my results comparing the age groups (0.7 for know each other or extremely close in 18-22 age group and 0.68 for 33+) disagree with Wellman. There was also little difference in the density of male's networks and female's networks (.25 and .26 extremely close respectively) for my results. In his Toronto study, Wellman found a "density of 0.33, [meaning] that one-third of a person's intimates network have close ties with each other," and that "mean network density declined from 0.33 to 0.13 over a decade." (25) My respondents had more dense networks with a total average of .25 extremely close, .45 knew each other and .30 strangers. My findings do agree with Wellman that "extended kin is rarely supportive," as no respondents listed "other family" in their strong ties. Again, education level proves to be an interesting factor in social networks, and did negatively correlate with network density, as all three of my high school or less respondents had fragmented networks in which all ties were strangers.
We also measured status homophily, in terms of education, gender, and age of the strong ties listed in our surveys. I defined age homophily as within one year below, the same age, or one year above. In "Birds of a Feather," Mcpherson finds "22% of people have no cross-sex confidants," and cites a study by Marsden that found controlling for kin showed "considerable gender homophily." (423) In my 33+ group, no male listed a nonkin strong tie of the opposite gender; Wellman and Wortley claim that women provide emotional support, and my rsults reveal a dangerous trend of investing all emotional support into one vulnerable female kin tie. My results agree with McPherson that "gender homophily is lower among the young and the highly educated” (423), but I wonder if an option for including romantic partner as kin, would have eliminated many of the gender heterophilous friends in the 18-22 age group. McPherson states that "homophily on age can be stronger than any other dimension" and cites Fischer's 38% of Detroit male's close ties being within two years of their age. My results show strong age homophily for the 18-22 group (0.62), but this age homophily significantly drops to 0.09 for the 33+ group. Gender homophily stays consistent at an average of 0.68 for both age groups with no significant variance across the sexes. Educational homophily varies for my age groups, as the younger group had 52% of ties of similar educational attainment and the older group had only 32%. Two of my unmarried, high school-or-less respondents had 100% gender and education homophilous strong ties, and all were located within the same city, suggesting little geographic mobility. The 18-22 group had higher age and educational homophily than the 33+ group, showing the effects of institutional structures of traditional school classes in clustering age groups, especially since all but one respondent are full time students.
McPherson also found that increased education leads not only to a greater number of confidants, but also a lower proportion of kin. My results for males 33+ agree (0.5 kin for high-school or lower and 0.3 for college or higher), but my results for females does not correlate. McPherson finds a positive correlation between education and diversity of ties. Thus, the lower educated respondents not only have less access to strong social support, but also to the unique information and resources that diverse ties bring. When extending educational attainment to designate class distinction, my findings support Gladwell's statement that poverty can be measured by limitation of access to diverse connectors like Lois Weinberg.
In terms of survey design issues, I would have liked our name generator to include additional questions discussed in Hampton and Marin's review of generator options to ascertain type of support accessed from ties. I also found limitations in the survey design, encountering confusion with the exclusion of proper designation for ex-spouses, non-married partners, text-messaging, and appropriate characterization of residential location. Many of my respondents, especially those who reside in the suburbs were perplexed at the gap between "same neighborhood" and "same city," feeling that neighborhood meant a couple of streets away in the standard suburban layout and wanted a farther category for "same county." Perhaps there would have been a measurement of distance in miles instead of category, but I do understand the difficulties some systematically different respondents would have in measuring miles in their head. Elizabeth Bott wrote that "highly developed division of labor in an industrial society produces not only complexity but also variability" in social networks, and I would have wanted a question measuring the population density of residential location and another measuring economic function (white collar, service, creative class, stay-at-home mom) in comparing social networks.
Some interesting aspects of my survey results support Wellman's claim that in modern networks, "community has moved out of its traditional neighborhood base as the constraints of space have weakened." (18) Whereas Wellman found that neighbors comprise 22% of Torontians' personal networks, only one respondent in my surveys listed a neighbor as a strong tie. This does not bode well for proponents of "eyes on the street" social surveillance or local civic engagement. Also the very low numbers of people who listed discussion ties that were members of group gives credence to Wellman's rise of personalized networking replacing group-based mechanical solidarity thesis and Putnam's bowling alone thesis. The two respondents who did list members of group were both 18-22 males and had discussion networks that were heavily concentrated by these comembers (3/6 and 6/6). The two respondents, one with fraternity brothers listed and other with Mormon church members listed, also had very dense networks.
Granovetter uses weak tie linkages to show the bridging of diverse social networks, and Burt praises the accessing of unique information diverse weak ties bring. However the decay of place-based networks engendered by the decline of "third places" and technological advances, presents a danger for networks to become more inclusive and homophilous, which is not good for citizenry.