Talk:'Pregnancy pact' grabs international attention for small Massachusetts town

Statistics

One thing the news articles haven't mentioned is the statistics behind this to answer the question How unlikely is it that 17 students would be pregnant by chance?" The news says a quarter as many were pregnant last year. Assuming a wikipedia:Biniomial distribution with n=1162 and assuming last year had 5 pregnancies making p=0.0043, then the chance of 17 pregnancies is 0.00002—two in one hundred thousand. That is a tiny number. Even taking into account that there are on the order of ten thousand high schools in the US (I couldn't find a hard number [1][2]. But if you apply the Binomial distribution at a school level, assuming 10,000 schools, you would expect to find one school in the country with a pregnancy rate this abnormal every 5.5 years (n=10,000, p=0.00002).

This is a sad story. But, with the caveat that it is probably indicative of problems in Gloucester and that pregnancies in one school in one year are probably not statistically unrelated, this story seems less newsworthy if you consider that you would expect to see a similar story twice a decade. I can't remember a similar story previously.

Comments? BenFrantzDale (talk) 10:45, 24 June 2008 (UTC)