Are You Pregnant?
Delhi Charter School in Louisiana enforced a new policy that forces girls to take pregnancy tests, and kicks out any student who tested positive or refuses the test. Not surprisingly, this incited outrage within the community. The ACLU claims that this policy violates Title IX, which prohibit sex discrimination, and the Equal Protection Clause of the Fourteenth Amendment. The ACLU is trying to get the school to retract its policy.
Now you may want to know what this has to do with Networks. The answer is that Bayes Theorem is the solution to this unique problem that Louisiana is facing. We know that pregnancy tests are 99% accurate, which leaves a 1% chance of a false positive. Therefore the Delhi believes pregnancy tests are accurate in telling them if a student is pregnant. This is not the case and can be proven by Bayes Theorem. There is about .5-1% of the population pregnant at any moment in time. So using the theorem, the probability of a girl being pregnant given that she tests positive is 33-50%.
P(pregnant|+)=[P(+|pregnant)P(+)]/P(pregnant)
=[P(+|pregnant)P(+)]/[P(+|pregnant)P(+)+P(+|not pregnant)P(not pregnant)]
=[.99*.005]/[.99*.005+.01-.995]
=33.2%
So there is only a probability of about 33%-50% of a girl being pregnant even if she tested positive. Keep in mind that the numbers I am using are guesses, the accuracy of pregnancy test is never 99% in real life due to user error (some research say it is closer to 90-95%) and the population that is pregnant may to more or less. In conclusion Delhi should get rid of its policy if not for the moral reason, then for the empirical facts and its own reputation. Statistics is saying that Delhi is more likely to kick out students that aren’t pregnant than are. Kicking out non-pregnant students on false bases only undermines what it is trying to do with its policy.
–mz232
Source: http://www.huffingtonpost.com/2012/08/07/delhi-charter-school-in-l_n_1752756.html