>>339492
cont.
Suppose it's the future, and M&Ms come in all sorts of colors, now. They're got yellow and blue and red ones, but also purple ones, black ones, white ones, turquoise ones, etc. In fact, if you grab a bag of M&Ms, you will find there are 20 different colors of them, randomly assorted. Let's look at classic green ones; 1 out of 20, or 5% of any bag will be green ones.
So, one day M&Ms decides to throw a special promotion. A very small number of M&Ms will now be randomly filled with white chocolate instead of milk chocolate. In ever bag there will be a few of them, and they can come in any color shell. You and some of your friends immediately go out and buy a whole bunch of these special bags and take them home. When you empty the bags out and begin to sort them, you notice (as will naturally happen) that some of the M&Ms are broken. This means you can see which of them have white chocolate inside, so you set those aside. Once you've sorted them all out, you have a HUGE pile of unbroken M&Ms, a small pile of broken ones, and a very small pile of broken ones with white chocolate.
Well, that's when you notice something peculiar. Of the broken M&Ms with white chocolate, HALF of them have green shells. Of course, you immediately think that the green M&Ms must be more likely to have white chocolate in them. That's only logical. You're eagar to test out this theory, so your friends grab some hammers and start breaking open ALL the other M&Ms. You eat the ones that prove to not have white chocolate, so it's an enjoyable experience. Once you're finished, you sort all the white chocolate M&Ms into separate color piles, but there's a problem. First of all, there's a lot fewer white chocolate M&Ms than you expected, but worse: the number is consistant across all different colors of shell. There are just as many blue ones, or red ones, or black ones, etc., with white chocolate as there are green ones. How can that be? There should be more green ones here, too. It's a paradox. Maybe your statistical analysis was wrong?
You're curious about this, so you call up the factory that makes the M&Ms, and that's when they let you in on a little secret. It turns out that white chocolate makes the M&Ms more brittle than usual, and the ones with the green shells are much more fragile than all the other colors; it's some kind of interaction between the green ink and the white chocolate. So, that's the M&M Paradox solved!
Going back here, let's extend the analogy back to the original topic. The color of the M&M shells represents sexual orientation: only the green ones are homosexual, or 5% of the total. White chocolate represents a predeliction toward criminal pedophilia, that is: a person very likely to molest a child. Broken M&Ms are the criminals in prison, who were convicted of a crime. Therefore, the broken white-chocolate M&Ms represent the pedophiles convicted of child molestation. In this case, it becomes obvious why it's not correct to apply statistics from one group (broken M&Ms) to the whole (unbroken ones) when the fact was that the white chocolate ones were more likely to be broken in the first place. In fact, supposing ALL the white-chocolate ones were broken, then the statistics applied to those would be COMPLETELY INAPPLICABLE to the unbroken ones; 0% of the unbroken ones would have white chocolate in them, because all the white chocolate ones already broke.
Now it becomes clear how this paradox can be solved. It's not that convicted pedophiles are more likely to be homosexual, it's that homosexual pedophiles are more likely to be convicted. There are a lot of reasons why this may be the case, and they could be similar to ones concerning race (African Americans who commit crimes may be more likely to receive longer prison sentences because of racism, and homosexual child molesters may be more likely to recieve longer prison sentences because of homophobia); or it could be a biological componant, some microcosm of combining homosexuality and pedophilia in the same brain makes a person more likely to commit a crime. Considering pretty much every child molester gets caught and imprisoned, you can now understand why any statistics about them shouldn't be applied to those people who DON'T commit a crime. Any average homosexual plucked off the street is very unlikely to be a pedophile, because almost all the homosexual pedophiles are already in prison.