idrees2022's review against another edition
3.0
In a key case study in this otherwise useful book, the author himself engages in the kind of statistical misuse that the book is meant to warn against. The speaks about the partisan uses of stats and cites the debates over deaths in Iraq as an example. He rightly rejects the Iraq Body Count numbers as a methdologically flawed undercount. He then mentions the 2006 Lancet study and notes that its method had greater rigor. But then adds that the Lancet's 650,000 mortality figure was contradicted by the subsequent WHO/IMH study which only arrived at a figure of 150,000. Except, the Lancet numbers referred to excess deaths -- that is death by all war related causes, not just violence -- whereas the WHO/IMH numbers refer specifically to violent deaths. The two numbers are not comparable. And this is precisely what the kind of common mistake that the author is warning against elsewhere in the book.
More...