The AFU and Urban Legend Archive Science six degrees experiment Select a topic - - - - - - - - Home Searches AFU FAQ - - - - - - - - AFU Animals Books Celebrities Classic Collegiate Death Disney Drugs Food GIF Language Legal Medical Misc Movies Politics Products Religion Science Sex Songs TV - - - - - - - - Other sites

From: pringle@NOSPAM.ccsu.edu (Bill Pringle)
Newsgroups: alt.folklore.urban
Subject: Re: Six Degrees of Separation
Date: 30 Apr 1997 19:43:38 GMT

I finally got to chance to look up the Milgram study. As a bonus, I did not get lost in the stacks or trapped on the elevator, but that's another story.

The cite is:

Milgram, Stanley. "The Small World Problem." Psychology Today, May 1967. pp 60 - 67.

The experiment was designed to see if randomly selected starter-people from all walks of life would be able to find a target-person using only a network of friends. Two separate studies were done, in both cases the target-people were in Cambridge, MA. The starting cities were Omaha,NB. and Wichita, KA.

About 150 people were selected from each of the cities and given a document folder that contained the following:

1. The name and address and some personal data on the target-person.
2. A set of rules. The most important of which was: "if you do not know the target-person on a firstname basis, then pass the document folder on to one friend that you feel is most likely to know the target. That friend must be someone you know on a first-basis."

3 A roster. Each person who got the folder had to put

      their name on the list.  This served to show who it 
      came from and also kept the folder from making any loops.

4.Tracer cards. Each person who transmitted the folder had to fill out and mail the tracer card.

The results: It took a median of 5 intermediate friends

              to go from the starter to the target person.
              The range was from 2 to 10.

Milgram reports some other interesting things learned from the study but I will leave that to the interested reader.

Hope this completes the thread.

Bill "its a small world after all" Pringle email replies: remove the nospam form the address.

From: mig@satlink.comNOSPAM (mig)
Newsgroups: alt.folklore.urban
Subject: Re: Six Degrees of Separation
Date: Wed, 30 Apr 1997 06:26:35 GMT

On 29 Apr 1997 08:16:21 -0400, mpage@panix.com (Madeleine Page) wrote:

>I second the request for details on the experiment, since the two posted
>quotes also raise some questions about the usual interpretation of "six
>degrees of separation". The canonical form of this is "I met the
>President, so anyone who knows me is one degree of separation from
>Clinton", which has always seemed to me a very weak form of the notion.
>
>For instance, I've met the Queen and Mick Jagger (neither one was toting a
>Mars Bar at the time), but I don't think that means that anyone who has
>ever met me is one degree of separation from said worthies. First, I'm
>sure that neither one remembers the occasion. Second, I'm quite certain
>that neither would agree to forward a document for me.
>

Mathematician John Allen Paulos goes over the experiment in his excellent _A Mathematician Reads the Newspaper_. (Paulos is better known for _Innumeracy_, an earlier book, also very good.)

He states it as follows (pp. 55-56)

"In a well-known experiment, the psychologist Stanley Milgram randomly selected pairs of American adults, designating one of them the source and the other the target and directed the source to send a letter to the person in the country who, he or she judged, was most likely to know the target. He also directed that a cover letter be included directing the recipient of the source's letter to do the same thing. This was to continue until the target was reached. (If the target letter lived in Seattle, for example, the source might have sent the letter to an acquaintance in Portland, who might have forwarded it to someone in Tacoma who might [...]"
"Milgram found that the number of intermediate links between source and target ranged from 2 to 10, with 5 being the most common number. This empirical study is even more telling than the a priori probability argument about the mere existence of two intermediates and partially explains why privy information, gossip, and jokes [and ULs...] cascade so rapidly through a population. The number of links would be even smaller if further information about the target were probided to the source."

Paulos covered the "a priori probability argument" in _Innumeracy_ and I believe he even expounded it, along with several other "math tricks" on the Carson show. (No mention was made of the probablity that either the source or the target would be eating a Mars Bar.)

saludos, mig