Misc. Stats Topics Discussion > How does probability distribution change as time goes forward and left tail probabilities are shown to be zero?
First of all Joe, I love that you are trying to figure this out. That is awesome and a great question... it sounds like something I'd be thinking...LOL.
To answer your question, the probabilities do not change when something hasn't happened, because the event "not happening" was part of the original probability calculation. HOWEVER, if you want to know the new probability of it occuring at a particular time, based on the fact that she is now at 38 weeks and has not had her baby, you wuold need to examine a completely new distribution (as opposed to just re-calculating based on the original distribution). The reason a new distribution would need to be drawn is that you are no longer dealing with the same population. Those original statistics are based on a population of pregnant women (probably newly pregnant) and considers all possible unusual or unexpected occurrences as random error (as your sample size is large enough).
Alternatively, when your wife is at 38 weeks, she is now from a completely different population of women that have certain shared characteristics (e.g. not having their baby extremely premature). In other words, if we view your wife still as from the "normal" distribution of pregnant women, then her probability of having the baby at any given point has not changed, just because she hasn't had it at this point (which is strange to think about, I know).
Jeremy Taylor
Hi Jeremy,
Thanks so much! I really appreciate it! I'm not a statistics person, but I guess I was thinking of that movie 21 where the professor talks about accounting for "variable change" and that one of the doors remains at 33%, whereas the other door goes up to 66% because the third door has been shown to have a goat instead of a car.
Joe
I am trying to figure out the statistical probability of when me and my wife's baby will be born. Due date is June 8. Empirical evidence indicates that there is a bell curve distribution around the 40 week mark (which for us is June 8). From 39 to 40 is 25% probability, from 40 to 41 weeks is 25% probability. From 38 to 39 is 15% probability. From 41 to 42 is 15% probability. The remaining 20% probability is spread out 10% for weeks earlier than 38 weeks and 10% for weeks later than 42 weeks. But at this point, we are at week 38. So we know that in fact there is 0% probability that the birth will happen prior to week 38 because the baby has not been born yet. So how does this fact change the probability distribution? Where does that 10% go? Is it distributed evenly over the remainder of the bell curve?