Difference between Statistics and Biostatistics according to Prof Bernard Rosner

当时上Prof Rosner的课，很好奇统计和生统到底有什么区别，于是很二地发邮件去问 然后Professor居然很nice地回了 最近看到地里很多童鞋在纠结，于是贴出来给大家个参考好了~当时问得很笼统，professor回得也很概括，求内行现身分享细节啊

It is difficult to answer your very broad question.

Biostatistics is concerned with applications of statistics to medical and biological data,.

Many biostatisticians in non-academic settings are mainly concerned with data analysis of biomedical studies (e.g. clinical trials for a drug company).

Biostatisticians in academia often both do theoretical work developing new methods and also participate in analyses of biomedical data. The ideal situation is when the theoretical problems arise from some previous data analysis issue that doesn't have a definite answer from previous literature.
It is also good for biostatisticians to develop some subject matter expertise in the area which the data analysis arises from so they can make a more meaningful contribution.

People in statistics departments are also concerned with developing new methods, but the methods are often not that focused on some particular application (e.g. what is the asymptotic power of some particular test).

Re your last point, regression analysis cannot tell us whether an association between y and x is truly causal. In general, clinical trials are needed to establish causality. Although some research workers (e.g. Don Rubin, Harvard Dept of Statistics) have made some progress in developing causal inference methods for observational studies.

Both statistics and biostatistics are large areas. To make progress you have to focus on some specific sub area and become expert in the literature in this area.

I hope this is helpful.

Thank you,

Bernard Rosner

看星的馨馨 发表于 2013-4-1 14:38
这么一看还是更爱biostat！

哈哈～
Prof Rosner 自己是 BA in Math at Columbia, MS in stat at Stanford, PhD in stat at Harvard
最后还是喜欢也推荐biostat，虽然作为学者总体观点上比较impartial，但是每次office hour 都小黑一下 stat

typical_libra 发表于 2013-4-1 15:07
哈哈～
Prof Rosner 自己是 BA in Math at Columbia, MS in stat at Stanford, PhD in stat at Harvard
...

他黑stat啥呀？说说～

zaqlovegre 发表于 2013-4-1 14:35
thx~~~~~信息很有用啊~~我就是纠结的一员T T

有用就好～加油加油～Good luck～

thx~~~~~信息很有用啊~~我就是纠结的一员T T

这么一看还是更爱biostat！

看星的馨馨 发表于 2013-4-1 15:38
他黑stat啥呀？说说～

他说统计学家很多时候是做出一个模型，然后满世界去找符合这个模型的数据，得到一个好结果，而无视那些算出来结果并不好的数据，有点自欺欺人的感觉；而生统学家一般以数据为中心，针对某一类数据来发明统计方法~

typical_libra 发表于 2013-4-1 18:17
他说统计学家很多时候是做出一个模型，然后满世界去找符合这个模型的数据，得到一个好结果，而无视那些算 ...

黑得漂亮。。

看星的馨馨 发表于 2013-4-1 18:43
黑得漂亮。。

不过在google实习的牛人说学术界的数据确实是结果比较漂亮，他们用工业界的数据跑出来就完全不是那么回事

typical_libra 发表于 2013-4-1 18:53
不过在google实习的牛人说学术界的数据确实是结果比较漂亮，他们用工业界的数据跑出来就完全不是那么回事 ...

这肯定啊