“But does the fact that most of the let­ters in the orig­i­nal exper­i­ment were never deliv­ered sug­gest that while when links exist that there may be only 6 degrees, some peo­ple are just never con­nected with oth­ers i.e. an infi­nite degree?”

Interesting question. I suspect the failure rate says a lot about people’s willingness to participate in unsolicited psychology experiments, but if you check out the Stats NZ report above you’ll see that almost all NZ employees are connected, so it would be very surprising if there were unconnected social components out there. On the other hand, who would know … :-)

There is another property of social networks that come into play in the Milgram experiment: your social network must be searchable in some way. That is, you need to have some idea of who in your social network might be closer to the target than you are. Obviously in this case, geography plays some role. I may post on this in more detail at a later date.

“Evi­dence from LEED (the Linked Employer-Employee Data­base) would sug­gest that the con­nec­tiv­ity between peo­ple in the work­force is 3.6 degrees.
(Page 10 of the PDF at: http://www.stats.govt.nz/publications/workknowledgeandskills/leed-reports/degrees-of-separation-in-nz-workforce.aspx )”

OK, so our friends at Stats NZ have sorted that one out for us. If the workforce is connected by 3.6 degrees, then the entire country will be separated by something very close to that. I guess we should consider this an upper bound, since I am led to believe that some people have acquaintances outside of work …