12岁洪晃的美国留学生活 纽约空降红小兵(图)
A事件 女性高薪(>十万)
B事件:女性博士
C事件: 嫁高薪男
D事件:男高薪 (10%)
A / B: 女性在已经有博士学位的条件下高薪
C / B: 女性在已经有博士学位的条件下嫁高薪男
P(C) = P(D)
P(A/B) = P(A and B)/P(B)
因为有P(B), P(A/B)跟P(A)无直接关系的。P(A)再小,P(A/B)也可能很大。所以P(A/B)是完全可能大于P(C/B) 的.
B事件:女性博士
C事件: 嫁高薪男
D事件:男高薪 (10%)
A / B: 女性在已经有博士学位的条件下高薪
C / B: 女性在已经有博士学位的条件下嫁高薪男
P(C) = P(D)
P(A/B) = P(A and B)/P(B)
因为有P(B), P(A/B)跟P(A)无直接关系的。P(A)再小,P(A/B)也可能很大。所以P(A/B)是完全可能大于P(C/B) 的.
Last edited by Knowing on 2008-05-06 20:25, edited 5 times in total.
有事找我请发站内消息
嗯,我再来 trytry
P(干的好/女博士) = P(年薪〉=10万/女博士) = a
P(嫁的好/女博士) = P(嫁的好) = P(遇到男人年薪〉=10万) = P(男人年薪〉=10万)= b
Knowing 同学的算法是假设 a =50%, b = 10%,之所以推导出这个架设,是因为 Knowing 同学假设 P(年薪〉=10万/女博士) = P(年薪〉=10万/博士) 并且 P(男人年薪〉=10万) = P(人年薪〉=10万)。
我唯一和 Knowing 同学不同的步骤就是我基于男女不平等的现实做出如下假设:
P(年薪〉=10万/女博士) is slightly smaller that P(年薪〉=10万/博士)
P(男人年薪〉=10万) is slightly bigger than P(人年薪〉=10万)
然后我的 a/b 的数值就比 Knowing 的小。
P(干的好/女博士) = P(年薪〉=10万/女博士) = a
P(嫁的好/女博士) = P(嫁的好) = P(遇到男人年薪〉=10万) = P(男人年薪〉=10万)= b
Knowing 同学的算法是假设 a =50%, b = 10%,之所以推导出这个架设,是因为 Knowing 同学假设 P(年薪〉=10万/女博士) = P(年薪〉=10万/博士) 并且 P(男人年薪〉=10万) = P(人年薪〉=10万)。
我唯一和 Knowing 同学不同的步骤就是我基于男女不平等的现实做出如下假设:
P(年薪〉=10万/女博士) is slightly smaller that P(年薪〉=10万/博士)
P(男人年薪〉=10万) is slightly bigger than P(人年薪〉=10万)
然后我的 a/b 的数值就比 Knowing 的小。
虽然细节可笑了点,但是我觉得从数学上论证了男女不平等直接导致了干的好不如嫁的好以及教育对提高女性社会地位的作用还是很有深度和洞察力的。你们去玩儿吧,我就在一旁笑。
小k说的基本上是对的。
first, let's clarify on the notation. Pr{A|B} is probability even A happens, on the condition that even B has happened for a fact.
Pr{A,C} is probability both A and C happen.
simple model, Pr{income>100k | PhD} = 0.5
Then, Pr{income over100k | female} = Pr{income over100k | female, PhD}* Pr{female, PhD| female} = 0.5 * Pr{PHD|female}
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male} = 0.5 * Pr{PHD|male}
火星狗, if, as you stated Pr{PHD|female}<Pr{PHD|male} ,
then Pr{income>100k | female} < Pr{income>100k | male} , this does not contradict with Pr{income over100k | female, PhD}= Pr{income over100k | male, PhD}
Suppose Pr{PhD| human}= 0.1 (censor data from knowing) , Pr{PHD|female}= ?, Pr{PHD|male}=?
well, we can easily observe Pr{female |PHD}, and Pr{male |PHD}. Suppose, as火星狗presumed, PHD的男性是女性的两倍. the former is 1/3, and the latter is 2/3.
So, Pr{PhD| human}= 10% , Pr{PHD, female}= Pr{female |PHD} * Pr{PhD}= 3.3% , Pr{PHD, male}= Pr{male |PHD} * Pr{PhD}= 6.7%. re-stating 拿PHD的男性是女性的两倍
Pr{PHD|female}= Pr{PHD, female}/ Pr{female} = Pr{female |PHD} * Pr{PhD}/ 0.5= 6.7% , Pr{PHD|male}= 13.4% just re-stating拿PHD的男性是女性的两倍, in a world of equal # of man and woman.
Pr{income>100k | female}= 0.5* 6.7%= 3.3% < Pr{income>100k | male}= 6.7% 拿十万的男性是女性的两倍, as火星狗presumed.
But once a girl already got a PhD,
Pr{income>100k | female, PhD}= 50%
Pr{marry one>100k | female, PhD}= Pr{ income>100k | male }* Pr{marry a man not a woman}* odds {rich man marry a female PHD over marry a girl without }
let Pr{marry a man not a woman} approximate to 1;
well, according to male PHD, the odds { rich man marry a female PHD over marry a girl without} is below one, female PHD has disadvantage in meat market. Lets say 0.8
then Pr{marry one>100k | female, PhD}= 6.7% *0.8 = 5.36% , much lower than Pr{income>100k | female, PhD}= 50%.
-----------------------------------------------------------------------------------
"PHD" can be switched to any other criteria.
As long as Pr{ income over100k | female, some standard } is over 6.7%, a girl is better off invest time on herself.
Of course, if you has a quality that odds {rich man marry a girl with over marry a girl without }= infinity, you are a rich man’s target, then congradulations, your chance of marrying rich = # of rich dudes on earth / # girls with this quality on earth.
Hence, when the quality is beauty and Youth, individual pretty girl's chance is still ugly.
Satisfies, 火星狗?
first, let's clarify on the notation. Pr{A|B} is probability even A happens, on the condition that even B has happened for a fact.
Pr{A,C} is probability both A and C happen.
simple model, Pr{income>100k | PhD} = 0.5
Then, Pr{income over100k | female} = Pr{income over100k | female, PhD}* Pr{female, PhD| female} = 0.5 * Pr{PHD|female}
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male} = 0.5 * Pr{PHD|male}
火星狗, if, as you stated Pr{PHD|female}<Pr{PHD|male} ,
then Pr{income>100k | female} < Pr{income>100k | male} , this does not contradict with Pr{income over100k | female, PhD}= Pr{income over100k | male, PhD}
Suppose Pr{PhD| human}= 0.1 (censor data from knowing) , Pr{PHD|female}= ?, Pr{PHD|male}=?
well, we can easily observe Pr{female |PHD}, and Pr{male |PHD}. Suppose, as火星狗presumed, PHD的男性是女性的两倍. the former is 1/3, and the latter is 2/3.
So, Pr{PhD| human}= 10% , Pr{PHD, female}= Pr{female |PHD} * Pr{PhD}= 3.3% , Pr{PHD, male}= Pr{male |PHD} * Pr{PhD}= 6.7%. re-stating 拿PHD的男性是女性的两倍
Pr{PHD|female}= Pr{PHD, female}/ Pr{female} = Pr{female |PHD} * Pr{PhD}/ 0.5= 6.7% , Pr{PHD|male}= 13.4% just re-stating拿PHD的男性是女性的两倍, in a world of equal # of man and woman.
Pr{income>100k | female}= 0.5* 6.7%= 3.3% < Pr{income>100k | male}= 6.7% 拿十万的男性是女性的两倍, as火星狗presumed.
But once a girl already got a PhD,
Pr{income>100k | female, PhD}= 50%
Pr{marry one>100k | female, PhD}= Pr{ income>100k | male }* Pr{marry a man not a woman}* odds {rich man marry a female PHD over marry a girl without }
let Pr{marry a man not a woman} approximate to 1;
well, according to male PHD, the odds { rich man marry a female PHD over marry a girl without} is below one, female PHD has disadvantage in meat market. Lets say 0.8
then Pr{marry one>100k | female, PhD}= 6.7% *0.8 = 5.36% , much lower than Pr{income>100k | female, PhD}= 50%.
-----------------------------------------------------------------------------------
"PHD" can be switched to any other criteria.
As long as Pr{ income over100k | female, some standard } is over 6.7%, a girl is better off invest time on herself.
Of course, if you has a quality that odds {rich man marry a girl with over marry a girl without }= infinity, you are a rich man’s target, then congradulations, your chance of marrying rich = # of rich dudes on earth / # girls with this quality on earth.
Hence, when the quality is beauty and Youth, individual pretty girl's chance is still ugly.
Satisfies, 火星狗?
MMT
是的。 我干嘛要跟自己过不去。 不过仔细一想,10个男PHD 有n个娶了女PHD,is something that can be observed. Let's say 3 (too optimistic?) then the odds is 3/7 = 0.43.森林的火焰 wrote:女博士在人肉市场上与普通女性比较的竞争力算0.8,相当客气了
then Pr{marry one>100k | female, PhD}= 6.7% *3/7 = 2.87% , much lower than Pr{income>100k | female, PhD}= 50%.
My calculation is correct火星狗 wrote:这个地方有点小问题
Pr{income>100k | female}= 0.5* 6.7%= 3.3% <Pr>100k | male}= 6.7%
Pr{income over100k | female}= 0.5* Pr{PHD|female}= 0.5*6.7%= 3.3%
Pr{income over100k | male}= 0.5 * Pr{PHD|male}= 0.5*13.4= 6.7%
Remember overall Pr{PHD}=10%, Pr{income100k | PHD}= 0.5
MMT
-
- Posts: 265
- Joined: 2007-12-14 16:32
Once constructed correctly, a model can always be modified.火星狗 wrote:另外 Knowing 和猫咪头的推论都严格假设在相同的教育程度下,男女报酬相同。
Suppose female PHD got bias treatment in job market.
lets say Pr{income>100k | female, PhD}= 0.3 , 只有3成, 够低的吧。
Pr{income>100k | dude, PhD}=Pr{income>100k | PhD} = 0.5, man don't suffer bias, so still 0.5.
(yes, they don't add up to 1. They have no reason to.)
Pr{income>100k | female}= 0.3* 6.7%= 2.01%
Pr{income>100k | male}= 0.5*13.4= 6.7%
拿十万的男性increased是女性的3倍还多.
Still, Pr{marry one>100k | female, PhD}= 6.7% *3/7= 2.9%
much lower than biased Pr{making 100k | female, PhD}= 30%
Even if 女博士在人肉市场上与普通女性比较的竞争力算0.8, that is still 5.4% : 30% marrying vs. making high income.
MMT
茂密头你还假设了
Pr{income>100k | PhD} = 0.5
推出
Pr{income over100k | female, PhD}= 0.5
我觉得实际情况相差比较大.
另一个误差来自你假设PHD是挣大于十万必要条件:
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male}
其实应该是:
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male} + Pr{income over100k | male, not a phD}* Pr{male, not a PhD| male}
第二项里也许Pr{income over100k | male, not a phD}比较低但是Pr{male, not a PhD| male}比较高.
Pr{income>100k | PhD} = 0.5
推出
Pr{income over100k | female, PhD}= 0.5
我觉得实际情况相差比较大.
另一个误差来自你假设PHD是挣大于十万必要条件:
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male}
其实应该是:
Pr{income>100k | male} = Pr{income over100k | male, PhD}* Pr{male, PhD| male} + Pr{income over100k | male, not a phD}* Pr{male, not a PhD| male}
第二项里也许Pr{income over100k | male, not a phD}比较低但是Pr{male, not a PhD| male}比较高.
-
- Posts: 265
- Joined: 2007-12-14 16:32
豪情,那么同样也要考虑 Pr{making100k| not PHD, female}
我有意描述一个封闭的系统。 此处 PHD非PHD,泛指一切有助于过得好quality 。
我和小K想说的就是这么回事儿。
男人中有 M%达到一定能力 (包括自身努力,包括爹妈给的),有C %把握衣食无忧 (C 包括就业及一切不可控因素);
有人说 女性只有0。5* M% 才能达到相似能力,而且在男性社会中吃亏,C也要打折扣, 比如3/5=60%的折扣.
那么女性要靠自己过的好,只有男人可能性的30%。 世界上有钱的男人是有钱女人的3倍还多。女大学生应该以结婚为要。
我说且慢, 对于已经达到一定能力的女性, 帐不是这么个算法。
Pr{making money| girls already have quality i}
= 60% * C%
Pr{marry to money| girls already have quality i}
= Pr{rich guy}* odds {rich guys favor this quality vs. not}
= C% * M%* meat-market-odds
Hence, such girls have better chance of achieving better life through marriage if, and only if
M%* meat-market-odds > 60%
If the world is full of guys who could be rich, and they all favor a girl like you in marriage, Ka-ching!
When guys with matching quality are <60% , and they don't favor this type of girls much in meat-market, spending time/energy trying to marry rich would be investing in the wrong portfolio.
小K, 我总结得还明白吧。
当然,不一样想法的人呢,想法不一样。 同样的数据保不定有人总结成, 提高 meat-market-odds 才是王道。
Once 60%*C% reachs a good chance, girls would want more than just money in their guys.
We want to pick rather than be picked.
---------------------------------
火星狗, best case scenario 自然是天才美貌又聪明,事业家庭双丰收。比如象豪情这样的, 还有比如那谁,谁谁,和谁谁谁。
我有意描述一个封闭的系统。 此处 PHD非PHD,泛指一切有助于过得好quality 。
我和小K想说的就是这么回事儿。
男人中有 M%达到一定能力 (包括自身努力,包括爹妈给的),有C %把握衣食无忧 (C 包括就业及一切不可控因素);
有人说 女性只有0。5* M% 才能达到相似能力,而且在男性社会中吃亏,C也要打折扣, 比如3/5=60%的折扣.
那么女性要靠自己过的好,只有男人可能性的30%。 世界上有钱的男人是有钱女人的3倍还多。女大学生应该以结婚为要。
我说且慢, 对于已经达到一定能力的女性, 帐不是这么个算法。
Pr{making money| girls already have quality i}
= 60% * C%
Pr{marry to money| girls already have quality i}
= Pr{rich guy}* odds {rich guys favor this quality vs. not}
= C% * M%* meat-market-odds
Hence, such girls have better chance of achieving better life through marriage if, and only if
M%* meat-market-odds > 60%
If the world is full of guys who could be rich, and they all favor a girl like you in marriage, Ka-ching!
When guys with matching quality are <60% , and they don't favor this type of girls much in meat-market, spending time/energy trying to marry rich would be investing in the wrong portfolio.
小K, 我总结得还明白吧。
当然,不一样想法的人呢,想法不一样。 同样的数据保不定有人总结成, 提高 meat-market-odds 才是王道。
Once 60%*C% reachs a good chance, girls would want more than just money in their guys.
We want to pick rather than be picked.
---------------------------------
火星狗, best case scenario 自然是天才美貌又聪明,事业家庭双丰收。比如象豪情这样的, 还有比如那谁,谁谁,和谁谁谁。
MMT
这个贴子真可爱。 两边都还没有考虑 dropby 说的再就业的因素。火星同学,男女不平等,随着年龄的增长,女性再嫁的好(measured by money)的几率越低,除非因离婚或其他家庭原因自己本身已经有钱。另外高收入的男性并非完全不跟高收入的女性结婚。
dropby wrote:这个概率不能这么算. 自己挣十万, 十万都是自己的, 想怎么花就怎么花. 嫁人挣十万, 最多一半的钱你有支配权, 所以挣十万得至少和嫁二十万比. 另外还有工作的稳定性等问题. 考虑到美国的离婚率大概是百分之五十, 还得算出每次都能嫁二十万的概率。如果是挣十万的话,被lay off的概率虽然也很高,但是重新找到十万工作的概率高得就太多了。
这样算下来的话,即使单从经济角度考虑,估计还是干得好比较容易一些。更不要说其他方面。。。
云浆未饮结成冰
没事干继续搞学术
P(income>100k | female, PhD) smaller than 0.5
P(income>100k | male, PhD) = 0.5
and P(income>100k | dude, PhD) = 0.5,
合理的推论是肯定存在一个比例很高的 {transgender, PhD}种群,而且
P{income>100k | transgender, PhD}>0.5
如果Pr{income>100k | dude, PhD}=Pr{income>100k | PhD} = 0.5, man don't suffer bias, so still 0.5.
(yes, they don't add up to 1. They have no reason to.)
P(income>100k | female, PhD) smaller than 0.5
P(income>100k | male, PhD) = 0.5
and P(income>100k | dude, PhD) = 0.5,
合理的推论是肯定存在一个比例很高的 {transgender, PhD}种群,而且
P{income>100k | transgender, PhD}>0.5
既然M% = 10%, 以上等式为 true 的条件是 meat-market-odds > 600%。所以,各位已经登上了科学高峰的同学,义无反顾的往上爬吧,另一条路已经变成了超小概率事件。M%* meat-market-odds > 60%
http://en.wikipedia.org/wiki/Odds
我就是那大学概率学的两板刷,没有在社会学堂进修过 猫咪头真是 knowledgableIn probability theory and statistics the odds in favour of an event or a proposition are the quantity p / (1 − p) , where p is the probability of the event or proposition.
Generally, 'odds' are not quoted to the general public in this format because of the natural confusion with the chance of an event occurring being expressed fractionally as a probability.