What do you think?
Rate this book
匕賵丕賯丞 丕賱卮丕賷: 賰賷賮 兀丨丿孬 丕賱廿丨氐丕亍 孬賵乇丞 毓賱賲賷丞 賮賷 丕賱賯乇賳 丕賱毓卮乇賷賳
"毓賱賷賳丕 丿乇丕爻丞 丕賱廿丨氐丕亍 賱賳賮賴賲 爻賳賳 丕賱賰賵賳, 賮賯丿 鬲亘賷賳 賱賳丕 賴匕賴 丕賱爻賳賳"
兀賮丕丿鬲 爻賷丿丞 賮賷 丨賮賱丞 卮丕賷 亘賰丕賲亘乇賷丿噩. 廿賳噩賱鬲乇丕 兀賳 賲匕丕賯 丕賱卮丕賷 賷禺鬲賱賮 廿匕丕 氐亘亘賳丕賴 毓賱賶 丕賱丨賱賷亘 賲賳賴 廿匕丕 氐亘亘賳丕 丕賱丨賱賷亘 毓賱賷賴, 賮兀孬丕乇 賰賱丕賲賴丕 爻禺乇賷丞 毓賯賵賱 丕賱賲噩賲賵毓丞 丕賱毓賱賲賷丞. 丕賯鬲乇丨 兀丨丿 丕賱囟賷賵賮 賵賷丿毓賶 乇賵賳丕賱丿 丌賷賱賲乇 賮賷卮乇 兀賳 賷禺鬲亘乇 賮乇囟賷鬲賴丕 毓賱賲賷丕. 賮賱賷爻 賴賳丕賰 賲賳 賴賵 兀賰孬乇 賰賮丕亍丞 賲賳賴 賱賱賯賷丕賲 亘賲孬賱 賴匕丕 丕賱丕禺鬲亘丕乇, 賱兀賳賴 賵噩賴 賲賷丿丕賳 丕賱廿丨氐丕亍 毓賱賶 賳丕丨賷丞 囟亘胤 賰賷賮賷丞 丕賱丨氐賵賱 毓賱賶 丕賱亘賷丕賳丕鬲 賵兀賴賲賷丞 賮賴賲賴丕. 廿賳 胤乇賷賯丞 鬲噩賲賷毓 丕賱亘賷丕賳丕鬲 賵丕爻鬲禺丿丕賲賴丕 毓賳丿賴 鬲賵丕夭賷 兀賴賲賷丞 丕賱亘賷丕賳丕鬲 匕丕鬲賴丕.
兀賮丕丿鬲 爻賷丿丞 賮賷 丨賮賱丞 卮丕賷 亘賰丕賲亘乇賷丿噩. 廿賳噩賱鬲乇丕 兀賳 賲匕丕賯 丕賱卮丕賷 賷禺鬲賱賮 廿匕丕 氐亘亘賳丕賴 毓賱賶 丕賱丨賱賷亘 賲賳賴 廿匕丕 氐亘亘賳丕 丕賱丨賱賷亘 毓賱賷賴, 賮兀孬丕乇 賰賱丕賲賴丕 爻禺乇賷丞 毓賯賵賱 丕賱賲噩賲賵毓丞 丕賱毓賱賲賷丞. 丕賯鬲乇丨 兀丨丿 丕賱囟賷賵賮 賵賷丿毓賶 乇賵賳丕賱丿 丌賷賱賲乇 賮賷卮乇 兀賳 賷禺鬲亘乇 賮乇囟賷鬲賴丕 毓賱賲賷丕. 賮賱賷爻 賴賳丕賰 賲賳 賴賵 兀賰孬乇 賰賮丕亍丞 賲賳賴 賱賱賯賷丕賲 亘賲孬賱 賴匕丕 丕賱丕禺鬲亘丕乇, 賱兀賳賴 賵噩賴 賲賷丿丕賳 丕賱廿丨氐丕亍 毓賱賶 賳丕丨賷丞 囟亘胤 賰賷賮賷丞 丕賱丨氐賵賱 毓賱賶 丕賱亘賷丕賳丕鬲 賵兀賴賲賷丞 賮賴賲賴丕. 廿賳 胤乇賷賯丞 鬲噩賲賷毓 丕賱亘賷丕賳丕鬲 賵丕爻鬲禺丿丕賲賴丕 毓賳丿賴 鬲賵丕夭賷 兀賴賲賷丞 丕賱亘賷丕賳丕鬲 匕丕鬲賴丕.
490 pages, Hardcover
First published April 1, 2001
420 people are currently reading
5998 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 245 reviews
April 6, 2021
賷兀鬲賷 毓賳賵丕賳 丕賱賰鬲丕亘 賲毓鬲賲丿丕 毓賱賶 賯氐丞 丨賯賷賯賷丞 賮賷 睾丕賷丞 丕賱胤乇丕賮丞
賮賷 賷賵賲 氐賷賮賷 丨丕乇 賮賷 廿賳噩賱鬲乇丕 噩賱爻 賲噩賲賵毓丞 賲賳 兀爻丕鬲匕丞 噩丕賲毓丞 賰丕賲亘乇丿噩 鈥徺堌操堌ж囐� 賱丕丨鬲爻丕亍 丕賱卮丕賷 鈥�
賵禺賱丕賱 噩賱爻鬲賴賲 丕賱賱胤賷賮丞 鬲賱賰 匕賰乇鬲 廿丨丿賶 丕賱爻賷丿丕鬲 兀賳 賲匕丕賯 丕賱卮丕賷 鈥徺娯勝� 廿匕丕 氐亘亘賳丕賴 賮賵賯 丕賱丨賱賷亘 毓賳 賲匕丕賯賴 廿賳 氐亘亘賳丕 丕賱丨賱賷亘 賮賵賯賴
亘乇睾賲 兀賳 丕賱鬲乇賰賷亘丞 丕賱賰賷賲賷丕卅賷丞 賴賷 匕丕鬲賴丕 賮賷 丕賱丨丕賱鬲賷賳
賵賲賳 丕賱丨丕囟乇賷賳 賰丕賳 丕賱毓丕賱賲 乇賵賳丕賱丿 賮賷卮乇 賵丕賱匕賷 丕賯鬲乇丨 丕禺鬲亘丕乇 賮乇囟賷丞 丕賱爻賷丿丞 鈥徺傌ㄙ� 丕賱鬲爻乇毓 亘丕鬲禺丕匕 賯乇丕乇 賲丕 賵爻胤 爻禺乇賷丞 丕賱丨丕囟乇賷賳鈥�
鈥� 賮賯丿賲 賱賴丕 兀賰賵丕亘賸丕 亘毓囟賴丕 氐亘 賮賷賴 丕賱卮丕賷 毓賱賶 丕賱丨賱賷亘.. 賵丕賱兀禺乇賶 氐亘 賮賷賴丕 鈥徹з勜勝娯� 毓賱賶 丕賱卮丕賷 乇丕噩賷丕 賲賳賴丕 兀賳 鬲丨丿丿 兀賷賸丕 賲賳賴丕 賷賳鬲賲賷 賱兀賷 賲噩賲賵毓丞 鈥徺堌蒂嗁� 丕賱兀賰賵丕亘 鬲亘毓丕 賱匕賱賰鈥�
爻賲賷鬲 賴匕賴 丕賱鬲噩乇亘丞 賵丕賱鬲賷 賳賵賯卮 賮賷賴丕 賮乇囟賷丕鬲 禺丕氐丞 亘毓賱賲 丕賱廿丨氐丕亍 亘鬲噩乇亘丞
匕賵丕賯丞 丕賱卮丕賷
鈥� 鈥�
賱賯丿 丕毓鬲賲丿鬲 鬲賱賰 丕賱鬲噩乇亘丞 毓賱賶 賲丕 賷爻賲賶 亘 鈥忊€� process-product 鈥巖elationship鈥� 鈥�
兀賷 丕賱毓賱丕賯丞 亘賷賳 丕賱毓賲賱賷丕鬲 賵丕賱賳鬲丕卅噩 鈥�
賵賲賳 賵丨賷 丕賱賲賳丕賯卮丕鬲 丕賱鬲賷 丿丕乇鬲 丨賷賳賴丕 鬲賲 鬲兀賱賷賮 賰鬲丕亘 毓賳 鬲氐賲賷賲 丕賱鬲噩丕乇亘
賰丕賳 賱賴 丿賵乇丕 賴丕賲丕 賮賷 廿丨丿丕孬 孬賵乇丞 毓賱賲賷丞 賮賷 丕賱賯乇賳 丕賱毓卮乇賷賳 賮賷 鈥徹官勝� 丕賱廿丨氐丕亍 丕賱丨丿賷孬鈥�
***
賷丨丿孬賳丕 丕賱賰鬲丕亘 毓賲賵賲丕 毓賳 賲丕赖賷丞 丕賱賳賲丕匕噩 丕賱丕丨氐丕卅賷丞
賵毓賳 丿賵乇賴丕 賮賷 丕賱丨賷丕丞 鈥徹з勝堌з傌官娯�
賵兀丨賯丕 賴賷 賵氐賮 丨賯賷賯賷 賱賱賵丕賯毓責
賷乇賵賷 丕賱賰鬲丕亘 賯氐氐丕 賲禺鬲賱賮丞 毓賳 賰賷賮賷丞 丕賰鬲卮丕賮賴丕
鬲丐賰丿 毓賱賶 丨賯賷賯丞 賲賴賲丞 賱丕 賳毓賷乇賴丕 丕賴鬲賲丕賲丕 賮賷 賲噩鬲賲丕毓鬲賳丕 兀亘丿丕
廿賳 胤乇賷賯丞 鬲噩賲賷毓 丕賱亘賷丕賳丕鬲 賵丕爻鬲禺丿丕賲賴丕 鬲賵丕夭賷 兀賴賲賷丞 丕賱亘賷丕賳丕鬲 賳賮爻賴丕
***
賷丐乇禺 丕賱賰鬲丕亘 賱賰孬賷乇 賲賳 丕賱毓賱賲丕亍 賰丕賳鬲 賳馗乇鬲賴賲 廿賱賶 丕賱毓丕賱賲 賲禺鬲賱賮丞鈥�
賵兀丨丿孬鬲 兀毓賲丕賱賴賲 孬賵乇丞 賮賷 丕賱丨賷丕丞 亘卮賰賱 兀賵 亘丌禺乇
廿賳賴 賷氐賮 丕賱賵賯丕卅毓 丕賱鬲賷 兀丿鬲 廿賱賶 鬲卮賰賱 賳馗乇賷丕鬲 丕賱毓賱賲丕亍 鈥�
賵丕賱兀賴賲 丕賱兀爻丕賱賷亘 丕賱廿丨氐丕卅賷丞 丕賱賲爻鬲禺丿賲丞 禺賱丕賱 匕賱賰 鈥�
***
賲賳 亘毓囟 丕賱兀賲孬賱丞 丕賱賲匕賰賵乇丞
(1)賯賵丕賳賷賳 賳賷賵鬲賳 賱賱丨乇賰丞 賵亘賵賷賱 賱賱睾丕夭丕鬲
鬲賲 丕爻鬲禺丿丕賲賴丕 賮賷 丕賱鬲毓亘賷乇 毓賳 丕賱賵丕賯毓 鈥徺堌嗀ㄘ� 丨賵丕丿孬 丕賱賲爻鬲賯亘賱
賲丕 賰丕賳鬲 鬲丨鬲丕噩賴 賴匕賴 丕賱鬲賳亘丐丕鬲 賴賷 賲噩賲賵毓丞 賲賳 鈥徹з勝呚关ж勜ж�
賵亘乇睾賲 匕賱賰 丕爻鬲睾乇賯鬲 丕賱丨囟丕乇丞 丕賱爻丕卅丿丞 兀乇亘毓賷賳 毓丕賲丕 賱鬲賵丕賰亘 鬲賱賰 丕賱賳馗乇丞 鈥徹з勜官勝呝娯�
鈥徺佡娰呚� 亘毓丿 氐丕乇鬲 賳馗乇賷丕鬲 賳賷賵鬲賳 丕賱乇賷丕囟賷丞 鬲爻鬲禺丿賲 賱賱亘丨孬 毓賳 賰賵丕賰亘 兀禺乇賶
亘賱 廿賳賴 鬲賲 鈥徹з冐簇з� 賳亘鬲賵賳 賮賷 丕賱賲賵賯毓 丕賱匕賷 丕賮鬲乇囟鬲賴
---------------------------------
(2)丕賱毓丕賱賲 丕賱賮賱賰賷 亘賷賷乇 賱丕亘賱丕爻
賯丕賲 亘賳卮乇 賰鬲丕亘賴 丕賱囟禺賲 毓賳 丨爻丕亘 丕賱賲賵丕賯毓 鈥徹з勝呚池傌ㄙ勝娯� 賱賱賰賵丕賰亘 賵丕賱賳噩賵賲
毓賳 胤乇賷賯 賲乇丕賯亘鬲賴丕 賲賳 爻胤丨 丕賱兀乇囟 鈥�
賵兀禺亘乇賴 賳丕亘賱賷賵賳 匕丕鬲 賲乇丞 亘兀賳賴 賱賲 賷噩丿 匕賰乇丕 賱賱廿賱賴 賮賷 丕賱賰鬲丕亘 鈥�
賮兀噩丕亘賴 丕賱毓丕賱賲 亘兀賳賴
" 賱賲 賷噩丿 丨丕噩丞 賱賲孬賱 賴匕賴 丕賱賮乇囟賷丕鬲"
賵賱賰賳賴 丕丨鬲丕噩 賲丕 兀爻賲丕賴 " 毓丕賲賱 丕賱禺胤兀 "
賮賲乇丕賯亘丞 丕賱賰賵丕賰亘 賵丕賱賳噩賵賲 賲賳 毓賱賶 爻胤丨 鈥徹з勜Ｘ必�
賱賲 鬲鬲賮賯 鬲賲丕賲丕 賲毓 賲賵丕賯毓賴丕 丕賱賲賮鬲乇囟丞 鈥� 賮賷 賰鬲丕亘賴
***
丕賱賰鬲丕亘 賰丕賳 丕爻鬲毓丕乇丞 賲賳 丕賱賲賰鬲亘丞 丕賱毓丕賲丞
噩匕亘賳賷 睾賱丕賮丞 丕賱噩賲賷賱 丕賱賮丕禺乇 賰賲丕 噩匕亘賳賷 毓賳賵丕賳賴
賵賴賵 賱賷爻 爻賴賱丕 毓賱賶 賲亘鬲丿賶亍
匕賵 賱睾丞 氐毓亘丞 亘毓囟 丕賱卮賷亍
賵賱賰賳 賲丕 廿賳 鬲賳丿賲噩 賮賷賴 丨鬲賶 鬲丨亘賴
賮賷 賷賵賲 氐賷賮賷 丨丕乇 賮賷 廿賳噩賱鬲乇丕 噩賱爻 賲噩賲賵毓丞 賲賳 兀爻丕鬲匕丞 噩丕賲毓丞 賰丕賲亘乇丿噩 鈥徺堌操堌ж囐� 賱丕丨鬲爻丕亍 丕賱卮丕賷 鈥�
賵禺賱丕賱 噩賱爻鬲賴賲 丕賱賱胤賷賮丞 鬲賱賰 匕賰乇鬲 廿丨丿賶 丕賱爻賷丿丕鬲 兀賳 賲匕丕賯 丕賱卮丕賷 鈥徺娯勝� 廿匕丕 氐亘亘賳丕賴 賮賵賯 丕賱丨賱賷亘 毓賳 賲匕丕賯賴 廿賳 氐亘亘賳丕 丕賱丨賱賷亘 賮賵賯賴
亘乇睾賲 兀賳 丕賱鬲乇賰賷亘丞 丕賱賰賷賲賷丕卅賷丞 賴賷 匕丕鬲賴丕 賮賷 丕賱丨丕賱鬲賷賳
賵賲賳 丕賱丨丕囟乇賷賳 賰丕賳 丕賱毓丕賱賲 乇賵賳丕賱丿 賮賷卮乇 賵丕賱匕賷 丕賯鬲乇丨 丕禺鬲亘丕乇 賮乇囟賷丞 丕賱爻賷丿丞 鈥徺傌ㄙ� 丕賱鬲爻乇毓 亘丕鬲禺丕匕 賯乇丕乇 賲丕 賵爻胤 爻禺乇賷丞 丕賱丨丕囟乇賷賳鈥�
鈥� 賮賯丿賲 賱賴丕 兀賰賵丕亘賸丕 亘毓囟賴丕 氐亘 賮賷賴 丕賱卮丕賷 毓賱賶 丕賱丨賱賷亘.. 賵丕賱兀禺乇賶 氐亘 賮賷賴丕 鈥徹з勜勝娯� 毓賱賶 丕賱卮丕賷 乇丕噩賷丕 賲賳賴丕 兀賳 鬲丨丿丿 兀賷賸丕 賲賳賴丕 賷賳鬲賲賷 賱兀賷 賲噩賲賵毓丞 鈥徺堌蒂嗁� 丕賱兀賰賵丕亘 鬲亘毓丕 賱匕賱賰鈥�
爻賲賷鬲 賴匕賴 丕賱鬲噩乇亘丞 賵丕賱鬲賷 賳賵賯卮 賮賷賴丕 賮乇囟賷丕鬲 禺丕氐丞 亘毓賱賲 丕賱廿丨氐丕亍 亘鬲噩乇亘丞
匕賵丕賯丞 丕賱卮丕賷
鈥� 鈥�
賱賯丿 丕毓鬲賲丿鬲 鬲賱賰 丕賱鬲噩乇亘丞 毓賱賶 賲丕 賷爻賲賶 亘 鈥忊€� process-product 鈥巖elationship鈥� 鈥�
兀賷 丕賱毓賱丕賯丞 亘賷賳 丕賱毓賲賱賷丕鬲 賵丕賱賳鬲丕卅噩 鈥�
賵賲賳 賵丨賷 丕賱賲賳丕賯卮丕鬲 丕賱鬲賷 丿丕乇鬲 丨賷賳賴丕 鬲賲 鬲兀賱賷賮 賰鬲丕亘 毓賳 鬲氐賲賷賲 丕賱鬲噩丕乇亘
賰丕賳 賱賴 丿賵乇丕 賴丕賲丕 賮賷 廿丨丿丕孬 孬賵乇丞 毓賱賲賷丞 賮賷 丕賱賯乇賳 丕賱毓卮乇賷賳 賮賷 鈥徹官勝� 丕賱廿丨氐丕亍 丕賱丨丿賷孬鈥�
***
賷丨丿孬賳丕 丕賱賰鬲丕亘 毓賲賵賲丕 毓賳 賲丕赖賷丞 丕賱賳賲丕匕噩 丕賱丕丨氐丕卅賷丞
賵毓賳 丿賵乇賴丕 賮賷 丕賱丨賷丕丞 鈥徹з勝堌з傌官娯�
賵兀丨賯丕 賴賷 賵氐賮 丨賯賷賯賷 賱賱賵丕賯毓責
賷乇賵賷 丕賱賰鬲丕亘 賯氐氐丕 賲禺鬲賱賮丞 毓賳 賰賷賮賷丞 丕賰鬲卮丕賮賴丕
鬲丐賰丿 毓賱賶 丨賯賷賯丞 賲賴賲丞 賱丕 賳毓賷乇賴丕 丕賴鬲賲丕賲丕 賮賷 賲噩鬲賲丕毓鬲賳丕 兀亘丿丕
廿賳 胤乇賷賯丞 鬲噩賲賷毓 丕賱亘賷丕賳丕鬲 賵丕爻鬲禺丿丕賲賴丕 鬲賵丕夭賷 兀賴賲賷丞 丕賱亘賷丕賳丕鬲 賳賮爻賴丕
***
賷丐乇禺 丕賱賰鬲丕亘 賱賰孬賷乇 賲賳 丕賱毓賱賲丕亍 賰丕賳鬲 賳馗乇鬲賴賲 廿賱賶 丕賱毓丕賱賲 賲禺鬲賱賮丞鈥�
賵兀丨丿孬鬲 兀毓賲丕賱賴賲 孬賵乇丞 賮賷 丕賱丨賷丕丞 亘卮賰賱 兀賵 亘丌禺乇
廿賳賴 賷氐賮 丕賱賵賯丕卅毓 丕賱鬲賷 兀丿鬲 廿賱賶 鬲卮賰賱 賳馗乇賷丕鬲 丕賱毓賱賲丕亍 鈥�
賵丕賱兀賴賲 丕賱兀爻丕賱賷亘 丕賱廿丨氐丕卅賷丞 丕賱賲爻鬲禺丿賲丞 禺賱丕賱 匕賱賰 鈥�
***
賲賳 亘毓囟 丕賱兀賲孬賱丞 丕賱賲匕賰賵乇丞
(1)賯賵丕賳賷賳 賳賷賵鬲賳 賱賱丨乇賰丞 賵亘賵賷賱 賱賱睾丕夭丕鬲
鬲賲 丕爻鬲禺丿丕賲賴丕 賮賷 丕賱鬲毓亘賷乇 毓賳 丕賱賵丕賯毓 鈥徺堌嗀ㄘ� 丨賵丕丿孬 丕賱賲爻鬲賯亘賱
賲丕 賰丕賳鬲 鬲丨鬲丕噩賴 賴匕賴 丕賱鬲賳亘丐丕鬲 賴賷 賲噩賲賵毓丞 賲賳 鈥徹з勝呚关ж勜ж�
賵亘乇睾賲 匕賱賰 丕爻鬲睾乇賯鬲 丕賱丨囟丕乇丞 丕賱爻丕卅丿丞 兀乇亘毓賷賳 毓丕賲丕 賱鬲賵丕賰亘 鬲賱賰 丕賱賳馗乇丞 鈥徹з勜官勝呝娯�
鈥徺佡娰呚� 亘毓丿 氐丕乇鬲 賳馗乇賷丕鬲 賳賷賵鬲賳 丕賱乇賷丕囟賷丞 鬲爻鬲禺丿賲 賱賱亘丨孬 毓賳 賰賵丕賰亘 兀禺乇賶
亘賱 廿賳賴 鬲賲 鈥徹з冐簇з� 賳亘鬲賵賳 賮賷 丕賱賲賵賯毓 丕賱匕賷 丕賮鬲乇囟鬲賴
---------------------------------
(2)丕賱毓丕賱賲 丕賱賮賱賰賷 亘賷賷乇 賱丕亘賱丕爻
賯丕賲 亘賳卮乇 賰鬲丕亘賴 丕賱囟禺賲 毓賳 丨爻丕亘 丕賱賲賵丕賯毓 鈥徹з勝呚池傌ㄙ勝娯� 賱賱賰賵丕賰亘 賵丕賱賳噩賵賲
毓賳 胤乇賷賯 賲乇丕賯亘鬲賴丕 賲賳 爻胤丨 丕賱兀乇囟 鈥�
賵兀禺亘乇賴 賳丕亘賱賷賵賳 匕丕鬲 賲乇丞 亘兀賳賴 賱賲 賷噩丿 匕賰乇丕 賱賱廿賱賴 賮賷 丕賱賰鬲丕亘 鈥�
賮兀噩丕亘賴 丕賱毓丕賱賲 亘兀賳賴
" 賱賲 賷噩丿 丨丕噩丞 賱賲孬賱 賴匕賴 丕賱賮乇囟賷丕鬲"
賵賱賰賳賴 丕丨鬲丕噩 賲丕 兀爻賲丕賴 " 毓丕賲賱 丕賱禺胤兀 "
賮賲乇丕賯亘丞 丕賱賰賵丕賰亘 賵丕賱賳噩賵賲 賲賳 毓賱賶 爻胤丨 鈥徹з勜Ｘ必�
賱賲 鬲鬲賮賯 鬲賲丕賲丕 賲毓 賲賵丕賯毓賴丕 丕賱賲賮鬲乇囟丞 鈥� 賮賷 賰鬲丕亘賴
***
丕賱賰鬲丕亘 賰丕賳 丕爻鬲毓丕乇丞 賲賳 丕賱賲賰鬲亘丞 丕賱毓丕賲丞
噩匕亘賳賷 睾賱丕賮丞 丕賱噩賲賷賱 丕賱賮丕禺乇 賰賲丕 噩匕亘賳賷 毓賳賵丕賳賴
賵賴賵 賱賷爻 爻賴賱丕 毓賱賶 賲亘鬲丿賶亍
匕賵 賱睾丞 氐毓亘丞 亘毓囟 丕賱卮賷亍
賵賱賰賳 賲丕 廿賳 鬲賳丿賲噩 賮賷賴 丨鬲賶 鬲丨亘賴
October 22, 2020
First, my old review from reading it back in 2007 before my MS in Statistics. Below that, my newer review from re-reading it in 2013 before my PhD.
~~~2007~~~
I loved the overviews of fascinating philosophical problems surrounding the use of statistics. (For example, hypothesis testing is used pretty much everywhere but even the mathematicians who came up with it had doubts about its validity and usefulness in most situations... What does a 95% confidence interval REALLY mean, in terms of real life, when you come right down to it?)
The author also paints a great picture of how rich and varied the field of statistics is, and what interesting people have contributed to it. As someone who wishes I were a real mathematician, it's a little depressing for me to read about a genius like Kolmogorov who came up with interesting results in every field he touched ever since childhood... But it's also heartening to hear about the many non-"genius" people who, unable to find an existing technique to evaluate their data, ended up creating original, useful statistical tools that became widely used and spawned whole new fields of mathematical research.
I wish the author had used at least a few equations or graphs - it's hard to explain complicated math in natural English - but it's still a great and informative read.
~~~2013~~~
I'm getting a lot more out of it on the 2nd reading, now that I know a lot of the theory he describes in handwavy terms for lay readers.
This book is definitely a history of statistics through anecdotes: the author's stories about the time he heard Savage lecture or was a discussant for Neyman, his colleagues' memories of Fisher and Pearson, etc.
It's especially fascinating if you already recognize these names from their papers and are curious about their personalities, what inspired their work and how it fit into their historical context, etc. But you won't learn the technical details here.
If you want something from a more formal historian of statistics, Stigler seems to be the guy: and .
One book I wish someone would write is the history of statistics education and stats textbooks. If Bayes hid his famous theorem, Fisher discouraged unthinking use of alpha=0.05, Neyman disavowed the hypothesis tests he created, etc. ... then how did these things come to be standard methods applied mechanically by scientists everywhere?
Specific points:
* p.4-5: Before Fisher's , "experiments were idiosyncratic to each scientist." Although Fisher obviously didn't invent experimentation, he put it on a consistent and rigorous footing. And researchers only published their conclusions with a small supportive subset of the data, not the whole experiment. (Gregor Mendel dropped inconvenient data from his famous pea experiments!) Even today we often see researchers hesitant to share not only their data but their code, despite the reproducible research movement.
* p.7: "any useful experiment has to be one that allows for estimation of those outcomes," i.e. Fisher realized that your parameters will always be estimated imperfectly, but if you design your experiment and choose your estimator carefully, the estimates can be precise enough to be useful.
* p.10: Pearson's is apparently still worth reading today. (I also recall that it inspired young Neyman.)
* A brief academic family tree of early statisticians working in the UK: Francis Galton came up with correlation and regression to the mean. He then (taught? managed? influenced?) Karl Pearson, who promoted the idea that distributions are the thing of interest to science, and developed his family of skew distributions. Karl Pearson influenced many in the next generation, including Fisher, Gosset, Neyman, and Pearson's own son Egon. From there on, the field of statistics exploded.
* p.17: Pearson proposed that what's "real" (or of interest to science) is not each measurement we take, but the abstract distribution they come from. But it took Fisher to make clear the distinction between a parameter's true value, and your estimator for that parameter (the formula you feed data into), and your estimate of the parameter (the actual value you got using a particular dataset). He also clarified that even if you collect a ton of data, your estimates of that distribution will still be just estimates---and that Pearson's estimators do not have good properties (the resulting estimates won't tend to be as close to the truth as they could be), but better estimators can be derived.
* p.19: Galton and K. Pearson's Biometrika was originally founded with the goal of measuring distributions of biological measurements about species around the world, showing that those distributions change over time, and thus providing proof of Darwin's theories. It was very expensive to print, containing complicated typesetting for math formulas and being the first journal to include full color photos. Its correspondents sound like Indiana Jones, traversing jungles and deserts and rain forests to measure native tribes and little-known animal species. It sounds like the primarily-mathematical articles (which it is now known for) were used mostly as filler at first.
* p.28-29: I love the story of Gosset, or "Student," working at Guinness and keeping his research a trade secret. Hotelling tried to meet Gosset in the 1930s and "arrangements were made to meet him secretly, with all the aspects of a spy mystery." We also often forget that Gosset did not derive the mathematical t-distribution (that was Fisher) but rather he ran an early and very laborious Monte Carlo experiment, shuffling stacks of cards with numbers on them and recording the averages etc. Gosset was driven by the need to work with small samples. As he told K. Pearson: "If I am the only person that you've come across that works with too small samples, you are very singular."
(Paul Velleman gave debunking some of the myths around the Gosset story; I hope to find his slides somewhere.)
* p.39: Cramer's book seems to be a key link in the history of stats textbooks: Fisher wrote his as a practical manual, without proofs. Then Cramer wrote to fill in gaps and write entire proofs as needed. "Cramer's book was used to teach a generation of new mathematicians and statisticians, and his redaction of Fisher became the standard paradigm."
This is not unlike the quote from economist Paul Samuelson: "Let those who will write the nation's laws if I can write its textbooks."
* p.49: Maybe I find the "degrees of freedom" concept confusing because it "was Fisher's discovery and was directly related to his geometric insights and his ability to cast the mathematical problems in terms of multidimensional geometry"---sadly not one of my strongest areas.
* p.51: Fisher gave a 1947 series of talks about science on the BBC. I would love to find recordings but googling does not help, although some transcripts might be in The Listener magazine if I can find a library with access to this in its database.
* p.59: Gumbel's "is a magnificently lucid presentation of a difficult subject, filled with references to the development of the subject. The first chapter ... alone is an excellent introduction to the mathematicals of statistical distribution theory. ... Although I first read the book after I had received my Ph.D. in mathematical statistics, I learned a great deal from that first chapter."
* p.66: "Since the statistic is random, it makes no sense to talk about how accurate a single value of it is. ... What is needed is a criterion that depends upon the probability distribution of the statistic" and Fisher was the one who first proposed a few such criteria (Consistency, Unbiasedness, Efficiency).
* p.70-71: "In the late 1960s, I had a programmable desk calculator. ... One afternoon, I programmed the machine, checked the first few steps to make sure I had not made an error in my program, turned off the light in my office, and left for home. Meanwhile, the programmable calculator was adding and subtracting, multiplying and dividing, silently, mumbling away in its electronic innards. Every once in a while it was programmed to print out a result. The printer on the machine was a noisy impact device that made a loud sound like "BRRRAAAK."
The nighttime cleaning crew came into the building and one of the men took his broom and wastepaper collector into my office. There in the darkness, he could hear a humming. He could see the blue light of the calculator's one eye waxing and waning as it added and subtracted over and over again. Suddenly, the machine awoke. "BRRAAK," it said, and then, "BRRAAK, BRRAAK, BRRAAK, BRRRRAAAAK!" He told me later that it was a terrifying experience and asked that I leave some sort of sign up the next time, warning that the computer was at work."
This delightful story reminds me of , former Director of the US Census Bureau.
* p.75: "The reader may recall those terrible moments in high school algebra when the book shifted into word problems. Mr. A and Mr. B were set rowing in still water or against a steady current, or maybe they were mixing water with oil, or bouncing a ball back and forth. Whatever it was, the word problem would propose some numbers and then ask a question, and the poor student had to put those words into a formula and solve for x. The reader may recall going back through the pages of the textbook, desperately seeking a similar problem that was worked out as an example and trying to stuff the new numbers into the formulas that were used in that example.
In high school algebra, someone had already worked out the formulas. The teacher knew them or could find them in the teacher's manual for the textbook. Imagine a word problem where nobody knows how to turn it into a formula, where some of the information is redundant and should not be used, where crucial information is often missing, and where there is no similar example worked out earlier in the textbook. This is what happens when one tries to apply statistical models to real-life problems."
* p.84: "The central limit theorem states that this distribution can be approximated by the normal probability distribution regardless of where the initial data came from." Well, not quite! There are very important constraints on the original data that must be met before you can apply a CLT. For example, the mean of iid Cauchy random variables is another Cauchy, not approximately Normal. See some other CLT counterexamples in .
* p.95-96: Nice example of how statistics differs from another mathematical approach, chaos theory, which can also be used to describe the world and make predictions(?), but (unlike statistics) has no measure of how well the model fits reality.
* p.98: The word "significant" used to mean "that the computation signified or showed something"---not necessarily something very important! Sadly, a shift in the English language changed the general usage of this word, making it a confusing term for students and users of statistics.
* p.99: Fisher's succinct explanation of significance, from 1929: "An observation is judged significant, if it would rarely have been produced, in the absence of a real cause of the kind we are seeking."
And from the same paper: "The test of significance only tells him what to ignore, namely all experiments in which significant results are not obtained."
In other words, a nonsignificant result doesn't mean there is no effect, just that the effect wasn't measured well enough in this experiment. And a single significant result doesn't mean you've proven the effect exists---you must be able to "design an experiment so that it will rarely fail to give a significant result."
* p.100: Salsburg's summary of Fisher's guidelines: "If the p-value is very small (usually less than .01), he declares than an effect has been shown. If the p-value is large (usually greater than .20), he declares that, if there is an effect, it is so small that no experiment of this size will be able to detect it. If the p-value lies in between, he discusses how the next experiment should be designed to get a better idea of the effect."
I love this advice: if it's not significant, then you design a better experiment, not assume that the effect doesn't exist! Sadly this is not the way p-values are used in much of science nowadays.
* p.102: I'd love to read the letters between Neyman and Egon Pearson, but they don't seem to be collected and published as far as I can tell.
* p.108: Again from Fisher: "tests of significance, when used accurately, are capable of rejecting or invalidating hypotheses, in so far as they are contradicted by the data: but ... they are never capable of establishing them as true"
* p.112: I know of Keynes as an economist, but didn't realize he also studied probability and wrote which "demolishes [the frequentist definition of probability] as a useful or even meaningful interpretation, showing that it has fundamental inconsistencies that make it impossible to apply the frequentist definition in most cases where probability is invoked."
* p.113-115: Unfortunately, Neyman found frequentism the easiest way to build a mathematically tractable & consistent theory of hypothesis testing, and that's the version that became entrenched in textbooks everywhere, even though "as early as 1935 ... he raised serious doubts" and "Neyman seldom made use of hypothesis tests directly." It seems that hypothesis tests became popularized through Wald's work on decision theory and through Lehmann's textbook .
* p.115-116: I greatly admire Neyman and am proud to share his first name, but all this about how nice he was is a bit of a hagiography. He was a pretty nice guy but could be a jerk too and had some serious troubles at home, estranged from his wife and distant from his son. His biography is very good.
* p.118: Nice explanation of how interval estimates help us decide whether the estimate is precise enough (i.e. the resulting policy decisions be the same whether the truth is near the lower or higher bound) or whether we need more data and better precision (i.e. the right decision would differ based on whether the lower or upper bound is true).
* p.123: "Fisher never got far with his fiducial distributions" and I thought this was an abandoned dead-end after Fisher died, but it turns out people still study fiducial inference, including CMU's own .
* p.142: "Godel once said that the gist of human genius is the longevity of one's youth."
* p.143: I used to wonder why we bother studying measure theory and foundations of probability---it's just proving things that seem obvious, right?---but before Kolmogorov put this all in order, all these "obvious" results were very much ad hoc, instead of being rigorously tied together. Likewise with Lebesgue's work on the foundations of calculus. Although the links seem obvious to us know, it was very different before Lebesgue and Kolmogorov, and I can't imagine what a change it must have been to read their work for the first time (without having already been exposed to the fruits of their labor like we have today).
* p.146-147: Kolmogorov tried tackling the real-life interpretation of probability, in a very different way from his work on measure theoretical foundations, but apparently did not complete this project before his death and sadly nobody has been able to figure out where he was going with it.
* p.148-150: Statistics can be highly political! It seems laughable today to think that Soviet planners would dismiss statistical work because "random variable" translates as "accidental magnitude" and the central planners felt insulted that their work could be considered accidental... But this lack of proper experimentation and evidence-based decisions led to massive starvation and economic weakness. Even apart from such extremes, governments have always tightly controlled the release of national statistics. Soviet statisticians were being threatened during the Cold War, and even today there are reports of for publishing damning inflation estimates.
~~~2007~~~
I loved the overviews of fascinating philosophical problems surrounding the use of statistics. (For example, hypothesis testing is used pretty much everywhere but even the mathematicians who came up with it had doubts about its validity and usefulness in most situations... What does a 95% confidence interval REALLY mean, in terms of real life, when you come right down to it?)
The author also paints a great picture of how rich and varied the field of statistics is, and what interesting people have contributed to it. As someone who wishes I were a real mathematician, it's a little depressing for me to read about a genius like Kolmogorov who came up with interesting results in every field he touched ever since childhood... But it's also heartening to hear about the many non-"genius" people who, unable to find an existing technique to evaluate their data, ended up creating original, useful statistical tools that became widely used and spawned whole new fields of mathematical research.
I wish the author had used at least a few equations or graphs - it's hard to explain complicated math in natural English - but it's still a great and informative read.
~~~2013~~~
I'm getting a lot more out of it on the 2nd reading, now that I know a lot of the theory he describes in handwavy terms for lay readers.
This book is definitely a history of statistics through anecdotes: the author's stories about the time he heard Savage lecture or was a discussant for Neyman, his colleagues' memories of Fisher and Pearson, etc.
It's especially fascinating if you already recognize these names from their papers and are curious about their personalities, what inspired their work and how it fit into their historical context, etc. But you won't learn the technical details here.
If you want something from a more formal historian of statistics, Stigler seems to be the guy: and .
One book I wish someone would write is the history of statistics education and stats textbooks. If Bayes hid his famous theorem, Fisher discouraged unthinking use of alpha=0.05, Neyman disavowed the hypothesis tests he created, etc. ... then how did these things come to be standard methods applied mechanically by scientists everywhere?
Specific points:
* p.4-5: Before Fisher's , "experiments were idiosyncratic to each scientist." Although Fisher obviously didn't invent experimentation, he put it on a consistent and rigorous footing. And researchers only published their conclusions with a small supportive subset of the data, not the whole experiment. (Gregor Mendel dropped inconvenient data from his famous pea experiments!) Even today we often see researchers hesitant to share not only their data but their code, despite the reproducible research movement.
* p.7: "any useful experiment has to be one that allows for estimation of those outcomes," i.e. Fisher realized that your parameters will always be estimated imperfectly, but if you design your experiment and choose your estimator carefully, the estimates can be precise enough to be useful.
* p.10: Pearson's is apparently still worth reading today. (I also recall that it inspired young Neyman.)
* A brief academic family tree of early statisticians working in the UK: Francis Galton came up with correlation and regression to the mean. He then (taught? managed? influenced?) Karl Pearson, who promoted the idea that distributions are the thing of interest to science, and developed his family of skew distributions. Karl Pearson influenced many in the next generation, including Fisher, Gosset, Neyman, and Pearson's own son Egon. From there on, the field of statistics exploded.
* p.17: Pearson proposed that what's "real" (or of interest to science) is not each measurement we take, but the abstract distribution they come from. But it took Fisher to make clear the distinction between a parameter's true value, and your estimator for that parameter (the formula you feed data into), and your estimate of the parameter (the actual value you got using a particular dataset). He also clarified that even if you collect a ton of data, your estimates of that distribution will still be just estimates---and that Pearson's estimators do not have good properties (the resulting estimates won't tend to be as close to the truth as they could be), but better estimators can be derived.
* p.19: Galton and K. Pearson's Biometrika was originally founded with the goal of measuring distributions of biological measurements about species around the world, showing that those distributions change over time, and thus providing proof of Darwin's theories. It was very expensive to print, containing complicated typesetting for math formulas and being the first journal to include full color photos. Its correspondents sound like Indiana Jones, traversing jungles and deserts and rain forests to measure native tribes and little-known animal species. It sounds like the primarily-mathematical articles (which it is now known for) were used mostly as filler at first.
* p.28-29: I love the story of Gosset, or "Student," working at Guinness and keeping his research a trade secret. Hotelling tried to meet Gosset in the 1930s and "arrangements were made to meet him secretly, with all the aspects of a spy mystery." We also often forget that Gosset did not derive the mathematical t-distribution (that was Fisher) but rather he ran an early and very laborious Monte Carlo experiment, shuffling stacks of cards with numbers on them and recording the averages etc. Gosset was driven by the need to work with small samples. As he told K. Pearson: "If I am the only person that you've come across that works with too small samples, you are very singular."
(Paul Velleman gave debunking some of the myths around the Gosset story; I hope to find his slides somewhere.)
* p.39: Cramer's book seems to be a key link in the history of stats textbooks: Fisher wrote his as a practical manual, without proofs. Then Cramer wrote to fill in gaps and write entire proofs as needed. "Cramer's book was used to teach a generation of new mathematicians and statisticians, and his redaction of Fisher became the standard paradigm."
This is not unlike the quote from economist Paul Samuelson: "Let those who will write the nation's laws if I can write its textbooks."
* p.49: Maybe I find the "degrees of freedom" concept confusing because it "was Fisher's discovery and was directly related to his geometric insights and his ability to cast the mathematical problems in terms of multidimensional geometry"---sadly not one of my strongest areas.
* p.51: Fisher gave a 1947 series of talks about science on the BBC. I would love to find recordings but googling does not help, although some transcripts might be in The Listener magazine if I can find a library with access to this in its database.
* p.59: Gumbel's "is a magnificently lucid presentation of a difficult subject, filled with references to the development of the subject. The first chapter ... alone is an excellent introduction to the mathematicals of statistical distribution theory. ... Although I first read the book after I had received my Ph.D. in mathematical statistics, I learned a great deal from that first chapter."
* p.66: "Since the statistic is random, it makes no sense to talk about how accurate a single value of it is. ... What is needed is a criterion that depends upon the probability distribution of the statistic" and Fisher was the one who first proposed a few such criteria (Consistency, Unbiasedness, Efficiency).
* p.70-71: "In the late 1960s, I had a programmable desk calculator. ... One afternoon, I programmed the machine, checked the first few steps to make sure I had not made an error in my program, turned off the light in my office, and left for home. Meanwhile, the programmable calculator was adding and subtracting, multiplying and dividing, silently, mumbling away in its electronic innards. Every once in a while it was programmed to print out a result. The printer on the machine was a noisy impact device that made a loud sound like "BRRRAAAK."
The nighttime cleaning crew came into the building and one of the men took his broom and wastepaper collector into my office. There in the darkness, he could hear a humming. He could see the blue light of the calculator's one eye waxing and waning as it added and subtracted over and over again. Suddenly, the machine awoke. "BRRAAK," it said, and then, "BRRAAK, BRRAAK, BRRAAK, BRRRRAAAAK!" He told me later that it was a terrifying experience and asked that I leave some sort of sign up the next time, warning that the computer was at work."
This delightful story reminds me of , former Director of the US Census Bureau.
* p.75: "The reader may recall those terrible moments in high school algebra when the book shifted into word problems. Mr. A and Mr. B were set rowing in still water or against a steady current, or maybe they were mixing water with oil, or bouncing a ball back and forth. Whatever it was, the word problem would propose some numbers and then ask a question, and the poor student had to put those words into a formula and solve for x. The reader may recall going back through the pages of the textbook, desperately seeking a similar problem that was worked out as an example and trying to stuff the new numbers into the formulas that were used in that example.
In high school algebra, someone had already worked out the formulas. The teacher knew them or could find them in the teacher's manual for the textbook. Imagine a word problem where nobody knows how to turn it into a formula, where some of the information is redundant and should not be used, where crucial information is often missing, and where there is no similar example worked out earlier in the textbook. This is what happens when one tries to apply statistical models to real-life problems."
* p.84: "The central limit theorem states that this distribution can be approximated by the normal probability distribution regardless of where the initial data came from." Well, not quite! There are very important constraints on the original data that must be met before you can apply a CLT. For example, the mean of iid Cauchy random variables is another Cauchy, not approximately Normal. See some other CLT counterexamples in .
* p.95-96: Nice example of how statistics differs from another mathematical approach, chaos theory, which can also be used to describe the world and make predictions(?), but (unlike statistics) has no measure of how well the model fits reality.
* p.98: The word "significant" used to mean "that the computation signified or showed something"---not necessarily something very important! Sadly, a shift in the English language changed the general usage of this word, making it a confusing term for students and users of statistics.
* p.99: Fisher's succinct explanation of significance, from 1929: "An observation is judged significant, if it would rarely have been produced, in the absence of a real cause of the kind we are seeking."
And from the same paper: "The test of significance only tells him what to ignore, namely all experiments in which significant results are not obtained."
In other words, a nonsignificant result doesn't mean there is no effect, just that the effect wasn't measured well enough in this experiment. And a single significant result doesn't mean you've proven the effect exists---you must be able to "design an experiment so that it will rarely fail to give a significant result."
* p.100: Salsburg's summary of Fisher's guidelines: "If the p-value is very small (usually less than .01), he declares than an effect has been shown. If the p-value is large (usually greater than .20), he declares that, if there is an effect, it is so small that no experiment of this size will be able to detect it. If the p-value lies in between, he discusses how the next experiment should be designed to get a better idea of the effect."
I love this advice: if it's not significant, then you design a better experiment, not assume that the effect doesn't exist! Sadly this is not the way p-values are used in much of science nowadays.
* p.102: I'd love to read the letters between Neyman and Egon Pearson, but they don't seem to be collected and published as far as I can tell.
* p.108: Again from Fisher: "tests of significance, when used accurately, are capable of rejecting or invalidating hypotheses, in so far as they are contradicted by the data: but ... they are never capable of establishing them as true"
* p.112: I know of Keynes as an economist, but didn't realize he also studied probability and wrote which "demolishes [the frequentist definition of probability] as a useful or even meaningful interpretation, showing that it has fundamental inconsistencies that make it impossible to apply the frequentist definition in most cases where probability is invoked."
* p.113-115: Unfortunately, Neyman found frequentism the easiest way to build a mathematically tractable & consistent theory of hypothesis testing, and that's the version that became entrenched in textbooks everywhere, even though "as early as 1935 ... he raised serious doubts" and "Neyman seldom made use of hypothesis tests directly." It seems that hypothesis tests became popularized through Wald's work on decision theory and through Lehmann's textbook .
* p.115-116: I greatly admire Neyman and am proud to share his first name, but all this about how nice he was is a bit of a hagiography. He was a pretty nice guy but could be a jerk too and had some serious troubles at home, estranged from his wife and distant from his son. His biography is very good.
* p.118: Nice explanation of how interval estimates help us decide whether the estimate is precise enough (i.e. the resulting policy decisions be the same whether the truth is near the lower or higher bound) or whether we need more data and better precision (i.e. the right decision would differ based on whether the lower or upper bound is true).
* p.123: "Fisher never got far with his fiducial distributions" and I thought this was an abandoned dead-end after Fisher died, but it turns out people still study fiducial inference, including CMU's own .
* p.142: "Godel once said that the gist of human genius is the longevity of one's youth."
* p.143: I used to wonder why we bother studying measure theory and foundations of probability---it's just proving things that seem obvious, right?---but before Kolmogorov put this all in order, all these "obvious" results were very much ad hoc, instead of being rigorously tied together. Likewise with Lebesgue's work on the foundations of calculus. Although the links seem obvious to us know, it was very different before Lebesgue and Kolmogorov, and I can't imagine what a change it must have been to read their work for the first time (without having already been exposed to the fruits of their labor like we have today).
* p.146-147: Kolmogorov tried tackling the real-life interpretation of probability, in a very different way from his work on measure theoretical foundations, but apparently did not complete this project before his death and sadly nobody has been able to figure out where he was going with it.
* p.148-150: Statistics can be highly political! It seems laughable today to think that Soviet planners would dismiss statistical work because "random variable" translates as "accidental magnitude" and the central planners felt insulted that their work could be considered accidental... But this lack of proper experimentation and evidence-based decisions led to massive starvation and economic weakness. Even apart from such extremes, governments have always tightly controlled the release of national statistics. Soviet statisticians were being threatened during the Cold War, and even today there are reports of for publishing damning inflation estimates.
April 21, 2012
I think this should be required reading for every young statistician. All the other majors seem to have some sort of History of [insert program name here], but I don't remember one from when I was working on my major (in statistics). I felt this book was exactly what it claimed to be--a description of how statistics revolutionized science in the 20th century. Some people seem to think that this book is supposed to describe statistical methods like an introductory textbook. If you want that, you should go read an introductory statistics textbook. It's not like there aren't plenty of those out there. This is a historical/philosophical look at how statistics has influenced science and vice versa.
The book is organized according to topics in statistics including biographical sketches of the people important to the development and application of each of the topics. This can make it a little difficult to keep everything in perspective for how it fits in the timeline. But I don't think there would have been a better way to organize it anyway.
I loved reading this book and found it entertaining, witty, and enlightening.
The book is organized according to topics in statistics including biographical sketches of the people important to the development and application of each of the topics. This can make it a little difficult to keep everything in perspective for how it fits in the timeline. But I don't think there would have been a better way to organize it anyway.
I loved reading this book and found it entertaining, witty, and enlightening.
July 2, 2020
丕賱丕丨氐丕亍 鈦︹櫏锔忊仼 賵丕丨丿 賲賳 丕賱賮乇賵毓 丕賱賲賮囟賱丞 毓賳丿賷 賮賷 丕賱乇賷丕囟丞 丕賱賱賶 賴賵 丕賱毓賱賲 丕賱賲賮囟賱 亘丕賱賳爻亘丕賱賷 ..
*丕賱乇賷丕囟丞 毓賱賲 賮賱爻賮賷 亘丨鬲
亘賲丕 廿賳 賰賱 丨丕噩丞 賮賶 丕賱賰賵賳 賳爻亘賷丞 賵 賱賷賴丕 丕賰鬲乇 賲賳 賵噩賴丞 賳馗乇 賮丕賱廿丨氐丕亍 賴賵 丕賱毓賱賲 丕賱兀賴賲 賱兀賳 賲賮賷卮 丨丕噩丞 賳賯丿乇 賳丨胤賴丕 賮賷 賲賰丕賳賴丕 賲賳 睾賷乇 賲丕 賳賯丿賾乇 丕賱賳爻亘丞 丕賱氐丨賷丨丞貙 賳爻亘丞 丕賱賲丕丿丞 丕賱賮毓丕賱丞 賮賶 丕賱丿賵丕亍貙 賳爻亘 丕賱賲乇囟賶 賮賷 丕賱丨乇賵亘 賮賶 丕賱兀賵亘卅丞貙 賳爻亘 丕賱賲賵鬲賶貙 丕賱賯賷丕爻丕鬲 丕賱賴賳丿爻賷丞 賵 丕賱賲亘丕賳賷 丨鬲賶 賮賶 丕賱丿賷賰賵乇 賵 賮賶 兀亘爻胤 丕賱兀卮賷丕亍...
丕賱賰鬲丕亘 毓賳賵丕賳賴 毓賳 賯氐丞 丨賯賷賯賷丞 毓賳 爻鬲 亘鬲賵囟丨 丕賱賮乇賯 亘賷賳 丕賱卮丕賷 亘丕賱賱亘賳 賵 丕賱賱亘賳 亘丕賱卮丕賷 " 廿賳賴丕 丕賱鬲賮丕氐賷賱 " 馃槄
丕賱胤乇賷賮 廿賳 丕賱賯氐丞 丿賷 賰丕賳鬲 爻亘亘 鬲胤賵乇 毓賱賲 丕賱廿丨氐丕亍 賮賶 丕賱毓氐乇 丕賱丨丿賷孬 ..
丕賱賰鬲丕亘 亘賷丿賵乇 丨賵賱 鬲胤賵乇 丕賱毓賱賲 賵 丕賱賳馗乇賷丕鬲 賵 馗乇賵賮 馗賴賵乇賴丕 ..
馗乇賵賮 丕賱毓賱賲丕亍 賵 丕賱丕囟胤賴丕丿 丕賱爻賷丕爻賷 兀孬賳丕亍 丕賱丨乇賵亘 丕賱毓丕賱賲賷丞 賵 卮賰賵賰 丕賱爻賵冥賷鬲 賮賶 丕賱兀賲乇賷賰丕賳 賵 丕賱毓賰爻 賵 鬲兀孬賷乇 丿丕 賮賶 鬲毓丕賲賱賴賲 賲毓 丕賱毓賱賲丕亍 賵 鬲兀孬賷乇 丿丕 毓賱賶 馗乇賵賮 賲毓賷卮鬲賴賲 賵 丕賱氐乇丕毓 亘賷賳 丕賱賳馗乇賷丕鬲 賵 丕賱毓賱賲丕亍 賳賮爻賴賲 ...
亘丿賵賳 兀賮賰丕乇 賲毓賯丿丞 兀賵 賳馗乇賷丕鬲 兀賵 兀乇賯丕賲
*丕賱乇賷丕囟丞 毓賱賲 賮賱爻賮賷 亘丨鬲
亘賲丕 廿賳 賰賱 丨丕噩丞 賮賶 丕賱賰賵賳 賳爻亘賷丞 賵 賱賷賴丕 丕賰鬲乇 賲賳 賵噩賴丞 賳馗乇 賮丕賱廿丨氐丕亍 賴賵 丕賱毓賱賲 丕賱兀賴賲 賱兀賳 賲賮賷卮 丨丕噩丞 賳賯丿乇 賳丨胤賴丕 賮賷 賲賰丕賳賴丕 賲賳 睾賷乇 賲丕 賳賯丿賾乇 丕賱賳爻亘丞 丕賱氐丨賷丨丞貙 賳爻亘丞 丕賱賲丕丿丞 丕賱賮毓丕賱丞 賮賶 丕賱丿賵丕亍貙 賳爻亘 丕賱賲乇囟賶 賮賷 丕賱丨乇賵亘 賮賶 丕賱兀賵亘卅丞貙 賳爻亘 丕賱賲賵鬲賶貙 丕賱賯賷丕爻丕鬲 丕賱賴賳丿爻賷丞 賵 丕賱賲亘丕賳賷 丨鬲賶 賮賶 丕賱丿賷賰賵乇 賵 賮賶 兀亘爻胤 丕賱兀卮賷丕亍...
丕賱賰鬲丕亘 毓賳賵丕賳賴 毓賳 賯氐丞 丨賯賷賯賷丞 毓賳 爻鬲 亘鬲賵囟丨 丕賱賮乇賯 亘賷賳 丕賱卮丕賷 亘丕賱賱亘賳 賵 丕賱賱亘賳 亘丕賱卮丕賷 " 廿賳賴丕 丕賱鬲賮丕氐賷賱 " 馃槄
丕賱胤乇賷賮 廿賳 丕賱賯氐丞 丿賷 賰丕賳鬲 爻亘亘 鬲胤賵乇 毓賱賲 丕賱廿丨氐丕亍 賮賶 丕賱毓氐乇 丕賱丨丿賷孬 ..
丕賱賰鬲丕亘 亘賷丿賵乇 丨賵賱 鬲胤賵乇 丕賱毓賱賲 賵 丕賱賳馗乇賷丕鬲 賵 馗乇賵賮 馗賴賵乇賴丕 ..
馗乇賵賮 丕賱毓賱賲丕亍 賵 丕賱丕囟胤賴丕丿 丕賱爻賷丕爻賷 兀孬賳丕亍 丕賱丨乇賵亘 丕賱毓丕賱賲賷丞 賵 卮賰賵賰 丕賱爻賵冥賷鬲 賮賶 丕賱兀賲乇賷賰丕賳 賵 丕賱毓賰爻 賵 鬲兀孬賷乇 丿丕 賮賶 鬲毓丕賲賱賴賲 賲毓 丕賱毓賱賲丕亍 賵 鬲兀孬賷乇 丿丕 毓賱賶 馗乇賵賮 賲毓賷卮鬲賴賲 賵 丕賱氐乇丕毓 亘賷賳 丕賱賳馗乇賷丕鬲 賵 丕賱毓賱賲丕亍 賳賮爻賴賲 ...
亘丿賵賳 兀賮賰丕乇 賲毓賯丿丞 兀賵 賳馗乇賷丕鬲 兀賵 兀乇賯丕賲
December 27, 2012
I really wanted to like this book. I love science history books, and while I am not a technical person, I appreciate the "Physics for Poets" level description that are a feature of many science history books. My problem with this book, and ultimately why I gave up is precisely due to the author's inability to handle the technical details. He says that he wife reminded him not to be too detailed, and ultimately he wasn't detailed enough. He described major changes in statistics, but it was hard to tell what those changes were. And frankly, many of the people described were not that interesting. They were statisticians after all.
February 9, 2019
賷爻乇丿 賵賯丕卅毓 賵 丕丨丿丕孬 賱毓賱賲丕亍 丕賯鬲氐丕丿 賵 丕丨氐丕亍 賵賲丕 氐丕丿賮賴賲 兀孬賳丕亍 賵囟毓 賲毓丕丿賱丕鬲賴賲 賵 賳馗乇賷丕鬲賴賲
賰鬲丕亘 賲賴賲 賱賲賳 賷丿乇爻 丕賱丕賯鬲氐丕丿 賵 丕賱丕丨氐丕亍
賰鬲丕亘 賲賴賲 賱賲賳 賷丿乇爻 丕賱丕賯鬲氐丕丿 賵 丕賱丕丨氐丕亍
December 29, 2024
Un recorrido a trav茅s de la revoluci贸n estad铆stica que tuvo lugar en el siglo XX. Salsburg escoge a cerca de una treintena de estad铆sticos (con alguna estad铆stica) y nos cuenta algo de su vida y sus aportaciones al campo. Su intenci贸n es que sea un libro que pueda leer cualquiera, al margen de sus conocimientos en matem谩ticas, pero para disfrutarlo de verdad creo que es necesario conocer algo del tema, aunque sea al nivel de una estad铆stica de 2潞 de bachillerato de Ciencias Sociales (mejor si es un poco m谩s). Y, a煤n superando ese nivel (tampoco por mucho), hay cosas que no terminado de pillar.
Pero eso no ha evitado que sea una lectura muy amena, que me ha dado un contexto que agradezco sobre cosas que, como profesor, he explicado alguna vez en clase (aunque eso se limite a los primeros cap铆tulos). Tambi茅n ha sido una fuente de an茅cdotas curiosas, como al raz贸n por la que William S. Gosset firmaba sus art铆culos con el pseud贸nimo de Student y su relaci贸n con la cerveza Guinness. Una lectura, en fin, que recomiendo si ten茅is alg煤n inter茅s en conocer algo m谩s sobre la historia de la Estad铆stica.
Pero eso no ha evitado que sea una lectura muy amena, que me ha dado un contexto que agradezco sobre cosas que, como profesor, he explicado alguna vez en clase (aunque eso se limite a los primeros cap铆tulos). Tambi茅n ha sido una fuente de an茅cdotas curiosas, como al raz贸n por la que William S. Gosset firmaba sus art铆culos con el pseud贸nimo de Student y su relaci贸n con la cerveza Guinness. Una lectura, en fin, que recomiendo si ten茅is alg煤n inter茅s en conocer algo m谩s sobre la historia de la Estad铆stica.
November 4, 2014
A leitura 茅 t茫o interessante quanto a do livro "o andar do b锚bado", contudo, uma senhora que toma ch谩 茅 um pouco mais acad锚mico. Saber como e em qual situa莽茫o surgiram as distribui莽玫es t de student, o f de fisher, as distribui莽玫es n茫o param茅tricas, al茅m de conhecer a import芒ncia de diversos estat铆sticos matem谩ticos (homens e mulheres) para o desenvolvimento da ci锚ncia ao longo dos anos com uma leitura bastante leve e did谩tica.
October 14, 2014
賯乇兀鬲 丕賱鬲乇噩賲丞 丕賱毓乇亘賷丞 賱乇賳丕 丕賱賳賵乇賷.
賷丨丕賵賱 丕賱賲丐賱賮 鬲亘爻賷胤 毓賱賵賲 丕賱廿丨氐丕亍 賱賱毓丕賲丞貙 賱賰賳 丕賱賰鬲丕亘 氐毓亘 丕賱賮賴賲 賲丕 賱賲 賷賰賳 賱丿賶 賱賯丕乇卅 賲毓丕乇賮 賲賯亘賵賱丞 賮賷 賲噩丕賱 丕賱乇賷丕囟賷丕鬲 賵丕賱毓賱賵賲 丕賱丨丿賷孬丞 毓丕賲丞.
賷丨丕賵賱 丕賱賲丐賱賮 鬲亘爻賷胤 毓賱賵賲 丕賱廿丨氐丕亍 賱賱毓丕賲丞貙 賱賰賳 丕賱賰鬲丕亘 氐毓亘 丕賱賮賴賲 賲丕 賱賲 賷賰賳 賱丿賶 賱賯丕乇卅 賲毓丕乇賮 賲賯亘賵賱丞 賮賷 賲噩丕賱 丕賱乇賷丕囟賷丕鬲 賵丕賱毓賱賵賲 丕賱丨丿賷孬丞 毓丕賲丞.
February 3, 2024
As a statistician, I found much of this book fascinating. Many of the anecdotes shared about some of the field's most famous characters gave me an even greater appreciation for their achievements. I also especially appreciated the forays into the philosophical underpinnings of statistics and the focus on the big picture as Salsburg traced and connected many of the breakthroughs in statistics throughout the 20th century.
My critiques are all about the writing, which came across as very stilted. Transitions, both within and across chapters, were frequently abrupt or nonexistent. Roughly the first third of the book seemed to have very little organizational structure, and while I enjoyed some of these individual chapters, it was very hard to follow. Later in the book, the structure became clearer: each chapter focused broadly on one or a few related statisticians and their contributions to the field. This was fine but had the disadvantage of not being chronological, so that later chapters frequently had to remind the reader that the statistician being discussed was unaware of the breakthroughs covered in previous chapters. There were also too many personal anecdotes about Salsburg meeting this or that statistician at a conference.
Finally, the book seemed ill suited for the target audience. Salsburg intended the book for a non-mathematical audience, deliberately using plain English as opposed to mathematical notation to describe the statistical methods discussed. Often, however, the descriptions of these methods were vague and I have no confidence that someone not already familiar with statistics would be able to understand these descriptions. I do, however, think it is a great read for statisticians and perhaps other scientists who use and at least generally understand statistics.
My critiques are all about the writing, which came across as very stilted. Transitions, both within and across chapters, were frequently abrupt or nonexistent. Roughly the first third of the book seemed to have very little organizational structure, and while I enjoyed some of these individual chapters, it was very hard to follow. Later in the book, the structure became clearer: each chapter focused broadly on one or a few related statisticians and their contributions to the field. This was fine but had the disadvantage of not being chronological, so that later chapters frequently had to remind the reader that the statistician being discussed was unaware of the breakthroughs covered in previous chapters. There were also too many personal anecdotes about Salsburg meeting this or that statistician at a conference.
Finally, the book seemed ill suited for the target audience. Salsburg intended the book for a non-mathematical audience, deliberately using plain English as opposed to mathematical notation to describe the statistical methods discussed. Often, however, the descriptions of these methods were vague and I have no confidence that someone not already familiar with statistics would be able to understand these descriptions. I do, however, think it is a great read for statisticians and perhaps other scientists who use and at least generally understand statistics.
April 2, 2019
This was so interesting! It was so cool to hear about the actual people behind all of the names my stats training taught me - Pearson, Fisher, Tukey, Box, Cox, and more. It also served to show how young this field of statistics is in some ways, but how classic it is in others.
This book does suffer from the law of misonomy that Salisbury mentions often - 鈥渢he lady tasting tea鈥� is in about three lines. I鈥檓 not sure what the reasoning was there, and it threw me for a bit early on.
The author claims this is for non-technical, non-mathematical people, and he doesn鈥檛 include any formulas or proofs for that reason.
That being said, I think statisticians, particularly of my own generation, are the best audience here, as I found the development of my chosen field super interesting.
This book does suffer from the law of misonomy that Salisbury mentions often - 鈥渢he lady tasting tea鈥� is in about three lines. I鈥檓 not sure what the reasoning was there, and it threw me for a bit early on.
The author claims this is for non-technical, non-mathematical people, and he doesn鈥檛 include any formulas or proofs for that reason.
That being said, I think statisticians, particularly of my own generation, are the best audience here, as I found the development of my chosen field super interesting.
September 16, 2011
I love what David Salsburg attempts to do here: explain the basic concepts of statistics by guiding the reader through the history of its development as a discipline. Too often we learn concepts and methods that are popular today without understanding why we use them or how they developed. But however much I appreciate Salsburg's approach, I cannot recommend his book. It is inconsistently paced, lacking in any real explanations of the statistics, and peppered with "when I met [so-and-so famous person]" and "when I invented this statistical term with [so-and-so famous person]" name-dropping.
The first chapters offer a mangled, difficult-to-follow history of the genesis of statistics. Salsburg introduces some basics of statistics, such as regression to the mean and skew distributions, but he wedges them into the narrative as afterthoughts. He literally spends a mere one to two sentences to explain a concept. I understand that this is not a statistics textbook, but a breakthrough new idea has no meaning to me unless I halfway understand what the idea is. Needless to say, there was a dissonance between Salsburg's excitement and my dull incomprehension of what was so exciting.
In later chapters, the book becomes more of a biography per chapter, which was easier for me to take in but not how I would have chosen to organize a book on the development of statistics. My overall impression is that Salsburg made an outline of thoughts he jotted down, rearranged a few of the points, then fleshed out his half-baked outline into a book. The result is a book that isn't explanatory enough for a beginner and isn't detailed enough for an expert.
The first chapters offer a mangled, difficult-to-follow history of the genesis of statistics. Salsburg introduces some basics of statistics, such as regression to the mean and skew distributions, but he wedges them into the narrative as afterthoughts. He literally spends a mere one to two sentences to explain a concept. I understand that this is not a statistics textbook, but a breakthrough new idea has no meaning to me unless I halfway understand what the idea is. Needless to say, there was a dissonance between Salsburg's excitement and my dull incomprehension of what was so exciting.
In later chapters, the book becomes more of a biography per chapter, which was easier for me to take in but not how I would have chosen to organize a book on the development of statistics. My overall impression is that Salsburg made an outline of thoughts he jotted down, rearranged a few of the points, then fleshed out his half-baked outline into a book. The result is a book that isn't explanatory enough for a beginner and isn't detailed enough for an expert.
June 30, 2017
It's a book about statistics... but it doesn't actually talk about how to do stats. Instead, it's about the evolution of the practice of statistics told by someone who was in the front lines of its evolultion. Each chapter is dedicated to a person or development so that we see the field evolve over time. It's really a fantastic meditation on what we can do -- and should do -- with stats and what we can't.
My favorite part relates to the lowly p-vale. Where on earth did this thing come from? Salsburg gives us a hint (p.99) -- "The closest [Fisher] came to defining a specific p-value that would be significant in all circumstances occurred in an article printed in the Proceedings of the Society for Psychical Research in 1929." He states in this article: "It is a common practice to judge a result significant, if it is of such a magnitude that it would been produced by chance not more frequently than once in twenty trials. This is an arbitrary, but convenient, level of significance for the practical investigator, but it does not mean that he allows himself to be deceived once in every twenty experiments. The test of significance only tells him what to ignore, namely all experiments in which significant results are not obtained. He should only claim that a phenomenon is experimentally demonstrable when he knows how to design an experiment so that it will rarely fail to give a significant result."
My favorite part relates to the lowly p-vale. Where on earth did this thing come from? Salsburg gives us a hint (p.99) -- "The closest [Fisher] came to defining a specific p-value that would be significant in all circumstances occurred in an article printed in the Proceedings of the Society for Psychical Research in 1929." He states in this article: "It is a common practice to judge a result significant, if it is of such a magnitude that it would been produced by chance not more frequently than once in twenty trials. This is an arbitrary, but convenient, level of significance for the practical investigator, but it does not mean that he allows himself to be deceived once in every twenty experiments. The test of significance only tells him what to ignore, namely all experiments in which significant results are not obtained. He should only claim that a phenomenon is experimentally demonstrable when he knows how to design an experiment so that it will rarely fail to give a significant result."
November 27, 2019
丕賱廿丨氐丕亍 賵賲丕 兀丿乇丕賰 賲丕 丕賱廿丨氐丕亍 賰賱 卮賷亍 廿丨氐丕亍 鬲賱賰 丕賱廿毓賱丕賳丕鬲 丕賱鬲賷 鬲馗賴乇 賱賰 毓賱賶 賲賵丕賯毓 丕賱鬲賵丕氐賱 丕賱廿噩鬲賲丕毓賷听 賵賰賲賷丞 丕賱賲賵丕丿 丕賱丿丕禺賱丞 賮賷 鬲乇賰賷亘 丕賱兀卮賷丕亍 丨賵賱賳丕 鬲賱賰 丕賱兀亘丨丕孬 丕賱鬲賷 鬲賯乇兀賴丕 丕賱賲毓賱賵賲丕鬲 丕賱賲鬲賳丕孬乇丞 賮賷 賰賱 丨丿亘 賵氐賵亘 賴賳丕 賰賱賴丕 廿丨氐丕亍 賵兀賰孬乇 賵兀賰孬乇 賵兀賰孬乇....... 廿賳 賱賲 鬲賰賳 賲丿乇賰 兀賴賲賷鬲賴丕 賮毓賱賷賰 賯乇丕亍丞 賴匕丕 丕賱賰鬲丕亘
賴賱 鬲毓賱賲 兀賳 賴賳丕賰 賮乇賯 賮賷 丕賱胤毓賲 廿匕丕 賵囟毓鬲 丕賱卮丕賷 兀賵賱賸丕 孬賲 丕賱賱亘賳 毓賳 丕賱毓賰爻責....
賷毓鬲亘乇 賴匕丕 丕賱賰鬲丕亘 爻乇丿 鬲丕乇賷禺賷 賱毓賱賲 丕賱廿丨氐丕亍 賵賱賷爻 賲毓賱賵賲丕鬲 乇賷丕囟賷丞 賲賰鬲卮賮 丕賱賯丕賳賵賳 賵鬲丕乇賷禺賴 賵丕賱賳賯丿 丕賱匕賷 賯丿賲 賱賷賴丕 孬賲 丕賱鬲毓丿賷賱丕鬲 丕賱鬲賷 鬲賵丕賱鬲 毓賱賷賴 賵賴賰匕丕.
賴賱 鬲毓賱賲 兀賳 賴賳丕賰 賮乇賯 賮賷 丕賱胤毓賲 廿匕丕 賵囟毓鬲 丕賱卮丕賷 兀賵賱賸丕 孬賲 丕賱賱亘賳 毓賳 丕賱毓賰爻責....
賷毓鬲亘乇 賴匕丕 丕賱賰鬲丕亘 爻乇丿 鬲丕乇賷禺賷 賱毓賱賲 丕賱廿丨氐丕亍 賵賱賷爻 賲毓賱賵賲丕鬲 乇賷丕囟賷丞 賲賰鬲卮賮 丕賱賯丕賳賵賳 賵鬲丕乇賷禺賴 賵丕賱賳賯丿 丕賱匕賷 賯丿賲 賱賷賴丕 孬賲 丕賱鬲毓丿賷賱丕鬲 丕賱鬲賷 鬲賵丕賱鬲 毓賱賷賴 賵賴賰匕丕.
May 5, 2009
The first three chapters were the best. He started out with some really good stuff that was both biographically interesting and statistically informative. But is seemed like he lost steam. That said, there were still some good chapters and interesting anecdotes, and I generally enjoyed the book. I had to read about two-thirds for a class, and I finished the rest of it after the class was over, so that says something.
January 26, 2018
賷丨賰賷 丕賱賰鬲丕亘 賯氐丞 賳卮賵亍 丕賱廿丨氐丕亍 賵鬲胤賵乇賴貙 賰賱 賮氐賱 鬲賯乇賷亘賸丕 亘丕爻鬲禺丿丕賲 兀丨丿 丕賱乇賵丕丿. 賵賱賰賳貙 賱兀賳 丕賱兀賲乇 鬲丨鬲丕噩 賱賱賰孬賷乇 賲賳 丕賱鬲丿丕禺賱丕鬲 賵丕賱賳馗乇賷丕鬲 賵丕賱胤乇賯 丕賱賲鬲賳丕賮爻丞貙 賮賱丕 賷禺賱賵 丕賱賰鬲丕亘 賲賳 丕賱賯賮夭丕鬲 丕賱夭賲賳賷丞 賱賱兀賲丕賲 賵丕賱禺賱賮.
鬲乇噩賲丞 丕賱賰鬲丕亘 鬲禺賱賵 賲賳 丕賱爻賱丕爻丞 賮賷 賲毓馗賲 丕賱賵賯鬲貙 賵丕賱兀賮賰丕乇 丕賱廿丨氐丕卅賷丞 氐毓亘丞 亘毓囟 丕賱卮賷亍 賱睾賷乇 丕賱賲禺鬲氐貙 賱賰賳賴 賰丕賳 乇丨賱丞 賲賲鬲毓丞 賮賷 毓丕賱賲 兀噩賴賱 丕賱賰孬賷乇 毓賳賴貙 賵兀卮丕乇 亘賮囟賵賱 賱丕 賷賳鬲賴賷 鬲噩丕賴賴 賮賷 丕賱賵賯鬲 匕丕鬲賴.
鬲乇噩賲丞 丕賱賰鬲丕亘 鬲禺賱賵 賲賳 丕賱爻賱丕爻丞 賮賷 賲毓馗賲 丕賱賵賯鬲貙 賵丕賱兀賮賰丕乇 丕賱廿丨氐丕卅賷丞 氐毓亘丞 亘毓囟 丕賱卮賷亍 賱睾賷乇 丕賱賲禺鬲氐貙 賱賰賳賴 賰丕賳 乇丨賱丞 賲賲鬲毓丞 賮賷 毓丕賱賲 兀噩賴賱 丕賱賰孬賷乇 毓賳賴貙 賵兀卮丕乇 亘賮囟賵賱 賱丕 賷賳鬲賴賷 鬲噩丕賴賴 賮賷 丕賱賵賯鬲 匕丕鬲賴.
November 24, 2017
Fascinating history of statistics. (Doug says that's an oxymoron.)
August 13, 2022
The Lady Tasting Tea is to statistics what Sophie鈥檚 World is to philosophy.
March 29, 2022
5-6 different partially written books combined into a single manuscript that turns out to be a mostly shallow biographical survey of early statisticians. there are some gems in here about the milieu of early statistics, but doesn't really deliver anything more substantive than an interesting footnote or two
February 8, 2020
It was very interesting to read about the people behind the known statistical methods! Also, the author has a nice writing style, it does not feel dry. Sometimes, he even builds up the expectation for the next chapter so I just wanted to know what happened...
October 19, 2023
It's good, but could be better: more details of the statistics, and more overview of the times (20th century) to give everything more context.
I feel like the subject matter really deserves an excellent book... but this is a good start
I feel like the subject matter really deserves an excellent book... but this is a good start
May 28, 2019
An excursion through history of statistics in the 20th century. It's interesting to learn how the ideas related to each other, as well as a little about the statisticians' personal lives.
Read
January 28, 2024Goed wi
October 18, 2020
賰鬲丕亘 丿爻賲貙 賱丕 賷賲賰賳 賯乇丕亍鬲賴 賲乇丞 賵丕丨丿丞 賵賱丕 賷賲賰賳 賯乇丕亍鬲賴 亘賲賮乇丿賴 丿賵賳 卮乇丨 兀賰孬乇 賱賱賲氐胤賱丨丕鬲 丕賱賵丕乇丿丞 :D
賷乇賰夭 丕賱賰丕鬲亘 毓賱賶 丕賱兀賮賰丕乇 賵鬲胤賵乇 丕賱廿丨氐丕亍 丿賵賳 匕賰乇 兀賷 賲毓丕丿賱丕鬲 乇賷丕囟賷丞 兀賵 鬲毓賯賷丿丕鬲貙 賱賰賳 亘毓囟 丕賱賮氐賵賱 賱丕 賷賲賰賳 賮賴賲賴丕 丿賵賳 賲夭賷丿 賲賳 丕賱鬲賵囟賷丨 賱賱賲氐胤賱丨 兀賵 丕賱賳馗乇賷丞 丕賱廿丨氐丕卅賷丞 丕賱鬲賷 賷丿賵乇 丨賵賱賴丕 丕賱賮氐賱.
噩賴丿 賵丕囟丨 賮賷 丕賱鬲乇噩賲丞.
賷乇賰夭 丕賱賰丕鬲亘 毓賱賶 丕賱兀賮賰丕乇 賵鬲胤賵乇 丕賱廿丨氐丕亍 丿賵賳 匕賰乇 兀賷 賲毓丕丿賱丕鬲 乇賷丕囟賷丞 兀賵 鬲毓賯賷丿丕鬲貙 賱賰賳 亘毓囟 丕賱賮氐賵賱 賱丕 賷賲賰賳 賮賴賲賴丕 丿賵賳 賲夭賷丿 賲賳 丕賱鬲賵囟賷丨 賱賱賲氐胤賱丨 兀賵 丕賱賳馗乇賷丞 丕賱廿丨氐丕卅賷丞 丕賱鬲賷 賷丿賵乇 丨賵賱賴丕 丕賱賮氐賱.
噩賴丿 賵丕囟丨 賮賷 丕賱鬲乇噩賲丞.
March 15, 2017
Amazing book on the history of statistics and experimentation. a must read for all.
May 25, 2018
The most popular image of Statistics we have is from Mark Twain鈥檚 re-tweet of the quote attributed to Benjamin Disraeli, "There are three kinds of lies: lies, damned lies, and statistics.". With the advent of computers and vast amount of storage, ever more data is available for crunching by scientists. Consequently, we have ever more conclusions based on data, not all of them unbiased. Politicians, environmentalists, businesses and scientists have all been guilty of selectively choosing data to push their agendas under the garb of 鈥榮cientific conclusions based on real statistical data鈥�. However, if we reflect carefully, we see that our well-being itself nowadays is understood only in terms of numbers and indices given to us by the science of Statistics. Without numbers like GDP growth rates, Consumer Price Index, Inflation rates, Unemployment rates, currency exchange rates etc, life as we know today would be a stumble in darkness. So, it is important to understand the role of Statistics in the modern world, what it means to us and how we can productively use it to improve our lives. This book by David Salsburg takes us through the important ideas and developments in Statistics during the past hundred years and more. It shows us the towering figures of this discipline, their contributions, their collaborations with one another as well as their profound disagreements and how it fundamentally changed the way science itself looked at understanding Nature.
Statistics was one my subjects in the University. I used to have an understanding of it as a branch of Mathematics/Science where one collects and analyzes numerical data in large quantities. I understood its purpose to be the extraction of values, called parameters, out of this mass of data so that we can make sense of the reality that this data represents. The preface to this book, by the author himself, showed me how primitive this understanding is. He shows how Statistics moved the philosophical vision of Science away from a deterministic model of the Universe to a probabilistic model. In the nineteenth century, Science viewed the Universe as working on clockwork precision. A small number of Mathematical formulas, like Darwin鈥檚, Newton鈥檚 and Boyle鈥檚 laws, could be used to describe reality and predict future events. What was needed were a set of such formulas and measurements with precision. But in practice, measurements lacked precision. The more the instruments were refined, the more scientists became aware of greater variations. The differences between what the models predicted and what was observed and measured grew bigger. The picture of the 鈥榗lockwork universe鈥� lay in shambles. Science started moving towards a new paradigm - the statistical model of reality. Because statistical models of reality are mathematical, we can understand reality through the ideas of randomness, probability and statistics. In the twentieth century, the rise of Quantum Mechanics reinforced it substantially. I found this view of the evolution of Science in the twentieth century fresh and insightful.
The book is mainly a selective history of statistics. Giants like Ronald Fisher, Karl Pearson, William Gosset, Francis Galton, Jerzy Neyman and W.E. Deming are all extensively covered for their seminal work as well as the struggles they had to wage to get their ideas accepted and at times, rejected. We see extensive biographical information and some gossip, at times. The work of many scientists is set in the social context of their times, because their work was carried out in totalitarian and post-colonial societies. For example, in the USSR, during the 1930s, communist orthodoxy was hostile to applied statistics. It affected the work of eminent scientists like Arnold Kolmogorov, who founded the axioms of probability. Indian statistical giants like P.C. Mahalanobis and C.R. Rao found themselves in more exciting times in a newly independent India in the 1950s, collecting and sorting important demographic data on the Indian population for the benefit of planning by the Nehru administration, which believed in using Science for development. W.E. Deming鈥檚 work on Quality control was given short schrift in his native USA, but the Japanese embraced it to emerge as the premium automakers of high quality in the 1980s. There is a special chapter in the book covering the contributions of women scientists like Stella Cunliffe, Judith Goldberg and others.
The book details their advancements in various fields, which include more reliable pharmaceuticals, higher quality beer, econometrics, superior quality control in manufacturing, social policy and medical diagnostic tests. There are interesting discussions on whether there is a direct link between recidivism and the length of sentence of a prisoner. The accepted wisdom is that 鈥榯he longer the sentence, the less the recidivism鈥�. The author discusses Stella Cunliffe鈥檚 analysis of this question which exploded the myth of this association. The chapter 鈥橳he Man who remade Industry鈥� has compelling details on the great contributions of W.E. Deming on quality control and how it revolutionized the Japanese automobile industry. However, I shall just touch upon one fundamental insight the author outlines in the chapter, 鈥楽moking and Cancer鈥�, which captured my imagination.
The chapter on 鈥楽moking and Cancer鈥� is centered on a philosophical and analytical discussion on what is 鈥榗ause and effect鈥�. Author Salsburg says that Prof. Bertrand Russell effectively showed in the early 1930s that there is no such valid scientific concept as 鈥榗ause and effect鈥�! It is a vague, common notion that does not stand up to pure reason. It contains an inconsistent set of contradictory ideas and is of little or no value in scientific discourse! If it is so, what does it mean for us in society? Did Agent Orange cause those health problems in Vietnam and after? Does smoking cause cancer? The statistics giant, Ronald Fisher, a pipe smoker himself, did not believe smoking caused cancer. He pointed out that studies showed that people who did not inhale the smoke had a higher incidence of lung cancer than those who inhaled. This is inconsistent with the conclusion. Additionally, he mused, suppose that there was something genetic that induced some people to smoke than others. Suppose this same genetic disposition involved the occurrence of lung cancer. It was well known that many cancers have a familial component. Suppose this link between smoking and cancer was due to the same genetic disposition. To prove his case, Fisher assembled data on identical twins and showed that there was a strong familial tendency for both twins to be either smokers or non-smokers. He challenged others to show that lung cancer was not similarly genetically influenced. Fisher鈥檚 objections are motivated by science. Studies of smoking use data from what is called opportunity samples, or people who were smoking already. Ideally, one must do a study that asks half the participants to start smoking two packs or more a day and make observations. This is known as a double-blind study to prevent bias. But this is ethically untenable. Though a lot of evidence exists that smoking is bad, each one of them is in some way flawed as well.
I found this analysis fascinating because we routinely accept so many 鈥榗ause and effect鈥� claims by environmentalists and other social scientists without much scrutiny. Is the thinning of arctic sea ice really the cause of polar bears dying of starvation? Did DDT really cause cancer? Did the CFCs from refrigerators really cause the Ozone layer to vanish over the Antarctic?
The final chapter titled 鈥楾he idol with feet of clay鈥� is a philosophical look at the future. Salsburg says that the progress of Science implies that eventually the statistical revolution also will be overthrown in favor of a better one. Science produces a model that fits available data and uses it to predict results of new experiments. But, no model is fully accurate. So, more and more data results in more and more complicated models and their exceptions. At some point, it no longer serves the purpose and new thinkers emerge to create a revolution. One can see the Einsteinian revolution as one such event. In this sense, Salsburg says that the science of statistics also stands on feet of clay and that the revolution which may overthrow it, is perhaps already germinating amongst us.
I found it an enjoyable book to read. I learnt a lot as well.
Statistics was one my subjects in the University. I used to have an understanding of it as a branch of Mathematics/Science where one collects and analyzes numerical data in large quantities. I understood its purpose to be the extraction of values, called parameters, out of this mass of data so that we can make sense of the reality that this data represents. The preface to this book, by the author himself, showed me how primitive this understanding is. He shows how Statistics moved the philosophical vision of Science away from a deterministic model of the Universe to a probabilistic model. In the nineteenth century, Science viewed the Universe as working on clockwork precision. A small number of Mathematical formulas, like Darwin鈥檚, Newton鈥檚 and Boyle鈥檚 laws, could be used to describe reality and predict future events. What was needed were a set of such formulas and measurements with precision. But in practice, measurements lacked precision. The more the instruments were refined, the more scientists became aware of greater variations. The differences between what the models predicted and what was observed and measured grew bigger. The picture of the 鈥榗lockwork universe鈥� lay in shambles. Science started moving towards a new paradigm - the statistical model of reality. Because statistical models of reality are mathematical, we can understand reality through the ideas of randomness, probability and statistics. In the twentieth century, the rise of Quantum Mechanics reinforced it substantially. I found this view of the evolution of Science in the twentieth century fresh and insightful.
The book is mainly a selective history of statistics. Giants like Ronald Fisher, Karl Pearson, William Gosset, Francis Galton, Jerzy Neyman and W.E. Deming are all extensively covered for their seminal work as well as the struggles they had to wage to get their ideas accepted and at times, rejected. We see extensive biographical information and some gossip, at times. The work of many scientists is set in the social context of their times, because their work was carried out in totalitarian and post-colonial societies. For example, in the USSR, during the 1930s, communist orthodoxy was hostile to applied statistics. It affected the work of eminent scientists like Arnold Kolmogorov, who founded the axioms of probability. Indian statistical giants like P.C. Mahalanobis and C.R. Rao found themselves in more exciting times in a newly independent India in the 1950s, collecting and sorting important demographic data on the Indian population for the benefit of planning by the Nehru administration, which believed in using Science for development. W.E. Deming鈥檚 work on Quality control was given short schrift in his native USA, but the Japanese embraced it to emerge as the premium automakers of high quality in the 1980s. There is a special chapter in the book covering the contributions of women scientists like Stella Cunliffe, Judith Goldberg and others.
The book details their advancements in various fields, which include more reliable pharmaceuticals, higher quality beer, econometrics, superior quality control in manufacturing, social policy and medical diagnostic tests. There are interesting discussions on whether there is a direct link between recidivism and the length of sentence of a prisoner. The accepted wisdom is that 鈥榯he longer the sentence, the less the recidivism鈥�. The author discusses Stella Cunliffe鈥檚 analysis of this question which exploded the myth of this association. The chapter 鈥橳he Man who remade Industry鈥� has compelling details on the great contributions of W.E. Deming on quality control and how it revolutionized the Japanese automobile industry. However, I shall just touch upon one fundamental insight the author outlines in the chapter, 鈥楽moking and Cancer鈥�, which captured my imagination.
The chapter on 鈥楽moking and Cancer鈥� is centered on a philosophical and analytical discussion on what is 鈥榗ause and effect鈥�. Author Salsburg says that Prof. Bertrand Russell effectively showed in the early 1930s that there is no such valid scientific concept as 鈥榗ause and effect鈥�! It is a vague, common notion that does not stand up to pure reason. It contains an inconsistent set of contradictory ideas and is of little or no value in scientific discourse! If it is so, what does it mean for us in society? Did Agent Orange cause those health problems in Vietnam and after? Does smoking cause cancer? The statistics giant, Ronald Fisher, a pipe smoker himself, did not believe smoking caused cancer. He pointed out that studies showed that people who did not inhale the smoke had a higher incidence of lung cancer than those who inhaled. This is inconsistent with the conclusion. Additionally, he mused, suppose that there was something genetic that induced some people to smoke than others. Suppose this same genetic disposition involved the occurrence of lung cancer. It was well known that many cancers have a familial component. Suppose this link between smoking and cancer was due to the same genetic disposition. To prove his case, Fisher assembled data on identical twins and showed that there was a strong familial tendency for both twins to be either smokers or non-smokers. He challenged others to show that lung cancer was not similarly genetically influenced. Fisher鈥檚 objections are motivated by science. Studies of smoking use data from what is called opportunity samples, or people who were smoking already. Ideally, one must do a study that asks half the participants to start smoking two packs or more a day and make observations. This is known as a double-blind study to prevent bias. But this is ethically untenable. Though a lot of evidence exists that smoking is bad, each one of them is in some way flawed as well.
I found this analysis fascinating because we routinely accept so many 鈥榗ause and effect鈥� claims by environmentalists and other social scientists without much scrutiny. Is the thinning of arctic sea ice really the cause of polar bears dying of starvation? Did DDT really cause cancer? Did the CFCs from refrigerators really cause the Ozone layer to vanish over the Antarctic?
The final chapter titled 鈥楾he idol with feet of clay鈥� is a philosophical look at the future. Salsburg says that the progress of Science implies that eventually the statistical revolution also will be overthrown in favor of a better one. Science produces a model that fits available data and uses it to predict results of new experiments. But, no model is fully accurate. So, more and more data results in more and more complicated models and their exceptions. At some point, it no longer serves the purpose and new thinkers emerge to create a revolution. One can see the Einsteinian revolution as one such event. In this sense, Salsburg says that the science of statistics also stands on feet of clay and that the revolution which may overthrow it, is perhaps already germinating amongst us.
I found it an enjoyable book to read. I learnt a lot as well.
August 10, 2024
Great insights into the man and woman who built the future of science for our generation
August 7, 2008
The full title here is The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century. This book by David Salsburg is pretty much what the title suggests: part history of the rise of statistical methods in scientific research and part biography about the people responsible for it. This probably isn't a book for anyone not already versed in inferential statistics and related subjects. It won't, for example, teach you much about statistics, so you'll be pretty lost or at best unimpressed by most of the stories and adulations the book contains. I would have appreciated a bit more exposition and explanation, but for those of us with a background in stats, it keeps things at a sufficiently high level so that we're not forced to pull out our old textbooks just to know what's going on.
And it's pretty interesting stuff. While Salsburg lacks (or at least holds in reserve) the panache and wit necessary to make this a really entertaining read, he does give glimpses into both the absurdity and mundanity of scientific process in this area. I was amused to learn, for example, that many august statistical techniques like analysis of variance were created so that someone could figure out how much artifical cow poop to spread over an acre of farm land. The book also tracks some of the more interesting personalities in the field, relating tales about how William Gossett created a now common and relatively simple procedure known as "Student's t-test" while working for a beer brewery (Guiness, no less) whose strict policies about sharing research forced him to publish under the (perhaps unimaginative) psudonym "Student." And then there were the cat fights and irrational, career-long grudges that these men and women slung around at each other. Though not quite on the level of say Bill Bryson's A Short History of Nearly Everything, this book does a decent job of layering those pedestrial and alltogether human eccentricities over the enormity of the scientific accomplishments they created.
So while not exactly light reading and not for the uninitiated, it's a pretty interesting read.
And it's pretty interesting stuff. While Salsburg lacks (or at least holds in reserve) the panache and wit necessary to make this a really entertaining read, he does give glimpses into both the absurdity and mundanity of scientific process in this area. I was amused to learn, for example, that many august statistical techniques like analysis of variance were created so that someone could figure out how much artifical cow poop to spread over an acre of farm land. The book also tracks some of the more interesting personalities in the field, relating tales about how William Gossett created a now common and relatively simple procedure known as "Student's t-test" while working for a beer brewery (Guiness, no less) whose strict policies about sharing research forced him to publish under the (perhaps unimaginative) psudonym "Student." And then there were the cat fights and irrational, career-long grudges that these men and women slung around at each other. Though not quite on the level of say Bill Bryson's A Short History of Nearly Everything, this book does a decent job of layering those pedestrial and alltogether human eccentricities over the enormity of the scientific accomplishments they created.
So while not exactly light reading and not for the uninitiated, it's a pretty interesting read.
October 12, 2019
I saw the book as divided into the early chapters where he covers the formative history of modern statistics, focusing on Karl Pearson and Fischer, the middle chapters, in which he gives a series of biographical sketches of important contributors to statistics and finally the last chapter in which he discusses the philosophical implications and problems of statistics. I enjoyed the first and last part of the book, but I really wonder whether the short biographical sketches would interest anyone not already familiar with the statisticians involve. The chapter dedicated to Deming was of great interest to me, but that's because I already knew something of him--from a reader's perspective he deserved it. But did the Lady in Black deserve a whole chapter? Overall, the book does help the layperson understand that statistics is in fact a controversial field which skirts some philosophical topics as well. The last chapter in particular will hopefully spur readers into finding out more about the deeper problems and interpretations of probability and statistics, as it has for me. I realize the book is intended for laypersons, but I feel the author could have at least tried to make some of the ideas more concrete. For example, he mentions the forbidding topic of measure theory without the slightest effort at giving some verbal explanation of what it involves. What I had in mind is some more explanations such as the one he does in fact give of a probability space, using the probability of rain as an example, breaking down the different interpretations of this seemingly simple statement. I wish he had divided the material more clearly by the impact of statistics on science, business, politics and the military.
Displaying 1 - 30 of 245 reviews
Can't find what you're looking for?
Get help and learn more about the design.