Social Network Blog - g16

Our Small University Experiment set out to apply Stanley Milgram's small-world study to the social network of the University of Pennsylvania. Stevenson theorizes that the clear boundaries of a university would allow for greater completion rates than world-at-large studied. However, our Antonio Polley group had a traditional 25% success rate along the lines of Milgram's population study, but Susan Yoon's group had a phenomenal 80% completion rate. There are many possible structural reasons for this discrepancy, and other derivations and fulfillments from Stevenson's hypotheses and Milgram's findings on gender homophily, number of chain links, strength of ties, and occupational prestige. While my rationale for choosing my target Neil in Part 1 were all justified in bridging the gap into the Wistar Institute, it was not enough to reach the elusive Mr. Polley.

Stevenson explains that the "relatively clear boundaries" in organizations result in smaller number of intermediaries between the originator and the target and higher completion rates than small world studies of the population at large. While Antonio's completion rate of 25% is close to Stevenson's of 27%, it is not higher than a world at large study like Milgram's small world study of the United States with a 23% success rate. Susan's success rate of 80% does however support Stevenson's comments on the greater success of organizational small world experiments. Stevenson's results of 1.25 mean chain links for completed folders is smaller than our results (4.5 for successful Polley and 3.25 for successful Yoon), but our results are as Stevenson predicted, less than the 5.5 average for the population at large in Milgram and Korte's small world. For both of our groups, the mean intermediary links was lower for incomplete chains, which agrees with Milgram and Korte's results that over 30% of incomplete chains stopped by the third remove. However, my personal intermediary chain links of 4 is more close to the successful Antonio mean (4.5) than the Antonio failures (2.33). My incomplete chain must be an example of Milgram's importance of social distance versus physical distance in his circular failures which reached the Nebraska neighborhood but not the individual target.

Stevenson hypothesizes that "the longer the time at the university, the more likely a student is to initiate a successful chain of communication to a target" and cites many of the structural conditions that prevent freshman from creating successful higher prestige ties than seniors. Our Small University Experiment's methodology differed in that our folders were not distributed equally to freshman, sophomores, juniors, and seniors, but we do see a pattern in years spent at Penn. Eleven of our originating alters have spent four years at Penn, 5 have spent three years at Penn, and 2 lucky people have spent five years at Penn. In our Antonio Polley group, the only two five-year students were the only two successful originating alters. It is also interesting that these two successful fifth year students bucked Stevenson's, Milgram's, and even McPherson's homophilous trends and predictions for success by passing their folders directly to individuals of a lower class who have only spent two years at Penn. As for the Susan Yoon group, 75% of the 4 year students were successful, and 100% of the 3 year students were successful. However these rates are hard to truly generalize because of our very small distribution (Is the difference in exposure between a junior and a senior that great? We did not have any freshman and only two juniors in this group). Also we cannot really draw conclusions from Milgram's experiments, which involved the population at large, instead of a duration-constricted captive population in the university organization. I am also skeptical because of the failure of my own folder. According to my returned postcards, years at Penn does not exactly correlate with success or occupational prestige. My senior status folder began with another four year student and ended with a staff member who has been at Penn for 21 years.

Stevenson's Hypothesis 2 states that "Small world folders are more likely to be passed within a class than between classes and occupation groups in a university." For both of our groups, the majority of student to student transfers were not class-homophilous and do not agree with Stevenson's hypothesis and results of no students passing a folder to a student of a lower class. 60% of Student transfers in Susan's group were to students in a lower class, as compared to 40% in the same class. Almost 57% of student to student transfers in Antonio's group were to students in a lower or higher class, as compared to 42% who remained in their class. As for affiliation, our results somewhat follow Stevenson's hypothesis as 65% of Antonio transfers were affiliation homophilous, but Susan's were an even 50% homophilous, 50% nonhomophilous split. Our results were not homophilous however in terms of transfers to people of the same department/major, which had about a 25% for both groups. This makes sense since both targets were outside of the Communications-dominating originating student field and our goal was to reach a heterophilous target.

Milgram also contended that completed chains are more likely to involve participants with higher occupational prestige and more weak infrequent relations than unsuccessful chains. For Milgram's infrequency of contacts in successful ties, our results showed an interesting quirk in the median time between transfers that may support Milgram's point. While incomplete chains in both Yoon and Polley groups and Polley completed chains average around 4 days, Yoon completed chains took an average of ten days between transfers.

Length at Penn usually correlates with increase in prestige, and thus more in-degrees and greater reachability than an individual who has not been at the university for as long or who occupies a less prestigious position. However, my personal chain somewhat refutes his findings, as my folder died with a staff intermediary who had been at the Wistar Institute for over 21 years. In Part 1 of this assignment: http://www.mysocialnetwork.net/blog/481/g16/2006/09/why_does_the_wistar_institute.html, I had hypothesized that my folder would "pass through staff primarily." This was prescient, as my chain did reach and end in the staff vicinity, from student-student-student-staff-staff. However, I also claimed in question 4 that my aim was to reach someone of high occupation prestige and network surveillance like a "manager" or otherwise "a gatekeeper secretary or security guard." While I did reach staff in the Wistar Institute in my aim to follow Milgram and Korte's "efficient occupational route," (102) I now think that my accessed staff Cheryl must be of lower occupational prestige than the "managers, officials, and sales-clerical personnel" (105) gatekeepers with "maximum surveillance" (107) of the institution in Milgram and Korte's study.

While Milgram's study did not use the same "very weak to very strong" tie continuum we used, he did break down ties as "friend, relative, or acquaintance." (65) In Milgram's Kansas study, 123 recipients passed the folder to friends or acquaintances, compared to only 22 sent to relatives. Our study differs in that an organization like a university has a constructed population where relatives will not be very likely to be included. Also, our study allowed for us to differentiate between friends and acquaintances and classify their exact degree of strength.
Although our two completed folder originators were absent on the day we collected data, their blogs for the Small University Experiment part 1 shows that R45 gave her folder to a strong tie, explaining that "as sisters in a small sorority, we have become very close to each other." G10's Small University Experiment Part 1 blog portrays her first intermediary as a weak tie. Our breakdown for both categories was more evenly distributed than Milgram's findings, and actually favored strong ties instead of Milgram's weakly categorized friends and acquaintances. Out of 18 originating alters (and counting my extrapolated blog tie information for the 2 missing Antonio completed folders), 5 were sent to either very weak or weak ties, 5 were sent to moderate ties, and 8 were sent to either strong or very strong ties. Strong and weak ties had an equal chance of success in Polley's group (50% both), and 50% of the successful ties in Susan's group were to strong ties with 25% each to weak and moderate ties.

My personal choice of a moderate tie (mixing both aspects of weak tie and strong tie) seemed to straddle the best of both worlds. Neil's characteristics of a weak tie allowed me to have relative access to Granovetter's unique non-homophilous ties and for him to act as Burt's structural hole to get my folder from a SAS undergraduate in the Sociology department to Neil's next tie and roster #3, a higher prestige graduate student in the target School of Medicine school. I also think that Neil's strong tie aspects allowed for me to trust him to continue on in the process, instead of ending the folder transfer, as happened with G33, the only originating alter in our group who picked a weak tie. However, these results do not generalize to Susan Yoon's results, with 2 completed weak or very weak ties and 1 incomplete very weak tie.

Stevenson's third hypothesis, which only 44% of his final links followed, states that small world folders will converge on faculty and staff before reaching the target. This is akin to Milgram and Korte's typical status descent for successful chains as the "target typically occupied a lower status than the person who occupied the final link." (105) The funneling effect found in sociometric stars like the Jones, Jacobs, and Browns of Milgram's study, and trickle down occupational prestige of final gatekeepers were not seen in our Small University Experiment. Only one final link, June C., out of 10 completed folders was used more than once to pass along the folder to the target of Susan Yoon. Five of the completed folders for Susan Yoon (3 faculty and 2 staff) had ultimate links of either staff or faculty members, compared to only 3 student final links. However, both of the 2 final intermediaries for the completed Antonio Polley folders were students. This contradicts Milgram's managerial findings, Stevenson's third hypothesis, and the results for Susan Yoon folders, and most likely mean that Antonio Polley has some direct interaction with students in both Wistar and the School of Medicine. Since he is a lab technician, he very well may work with current graduate students on research.

Milgram and Korte's study focused on crossing the racial divide, a challenge which in our experiment can translate to crossing the department divide. Their findings showed that these racial gatekeepers were predominently males of professional status. Although I had trouble exactly evaluating the administration status for Susan Yoon folders, two of her folders crossed the SAS to GSE departmental gap through student to student transfers, and two of hers crossed the gap by student to faculty, in direct opposition to Milgram's trickle down gatekeepers. There was no correlation to the male gender, probably due to overwhelmingly female chains.

Stevenson's final hypothesis claims that small world folders are more likely to be passed to members of the same sex. Stevenson's results showed both sexes were more likely to keep folders within their gender when crossing affiliation boundaries. Our experiment had similar results, with 7 out of 9 affiliation-heterophilous transfers being from female to female in Susan's group. For Antonio's group, three out of four males chose other males when crossing affiliation boundaries, but females split even 1-1 to other females and to males. For overall gender-homophily in a group 7/ 8 female originators trying to reach a male target, Antonio's results were similar for both completed and noncompleted chains: both categories' transfers were slightly over 50% gender homophilous, with an average total of 56.5% of transfers going from male to male or female to female. Of the 7 out of 8 folders that had gender heterophilous transfers (overall chain involved both males and females), only two of the folders had multiple gender switches, returning from female to male back to female or vice versa. Once the remaining five folders crossed the gender gap, they stayed gender homophilous until completion or disintegration. For Susan Yoon, there was more of a discrepancy in the gender homophily of transfers between completed (84.6%) and incomplete chains (100%), but overall still agreed with Stevenson hypothesis and Milgram's assertion that “Participants were three times as likely to send the folder to someone of the same sex as someone of the opposite sex.” Both of the two failed folders remained completely gender homophilous, but only 2 of the completed chains crossed the gender divide. This could be due to target's gender matching the gender of every single originating alter in the Susan Yoon group. There was no implicit need to cross the gender gap for any of the folders, as was necessary for the 94% of originating alters to reach the male Antonio. Our experiment's results mirror those of Milgram's first study, which found that females passed the folder to females 56 times, and to males 18 times.

The Stevenson study's target's prestige level (undergraduate dean of management) differed from my target's occupational prestige (lab technician staff), probably making the dean easier to reach. Stevenson's article goes onto to further describe the selection process, claiming the target was picked because he was "located in the building where most of the classes are held and would be easy to physically access." (3) This was not the case for my target, Antonio Polley, located in the physically isolated Wistar Institute. The ecology of the institutional space on campus definitely harkens back to Wellman's compelling privatization comments, as Wistar has its own separate dining cafeteria, preventing potential bridging interaction among other non-Wistar affiliated Penn individuals at the heterophilous Houston Hall or Wawa.

27% of Stevenson's folders reached their targets, which is similar to Antonio Polley's 25% success rate, Milgram and Korte's 22% (103) and Milgram's 23% (44 out of 160 in Small World Problem page 65). The true outlier here is the very high 80% completion rate for Susan Yoon folders, insinuating that undergraduates at SAS Penn have more structural access to faculty in the Graduate School of Education. Another main reason may be related to the gender breakdown of GSE and the SM, the overwhelming female majority of our class, and homophily of gender in folder transfers. Antonio is a male in the School of Medicine/Wistar, which is 52% male to 48 female: http://www.med.upenn.edu/admiss/2006_class.html
Susan Yoon is a female, like 17 out of 18 of the originating alters, and 78% of the students in the GSE: http://www.gse.upenn.edu/admissions_financial/classprofile.php
This gender discrepancy, along with the physical isolation of Wistar's cafeteria, the integrated nature of the GSE in other undergraduate student life (sharing faculty that also teach undergraduate sociology courses), and Susan's higher ranked role as faculty with direct contact with students as opposed to staff Polley (with less prestigious in-degrees) are all possible reasons why there were greater completion rates for the target of Susan.

Despite my folder's failure in reaching the target, I am satisfied with my transfer choice to the moderate tie Neil. My assessment of Neil's personal characteristics (male, Senior, Bio major) proved to be correct in that he had direct access to a heterophilous prestigous tie and successfully crossed the school gap from SAS into the School of Medicine.