Scan barcode
bootman's review against another edition
5.0
I cannot express how important I think it is for everyone to read books about being skeptical of data and research. This book from Gary Smith was recommended while I was reading another book on skepticism and critical thinking, and Standard Deviations didn't disappoint. The chapters in this book are short and sweet to give you examples and overviews about how data can be misinterpreted. I really enjoyed the book, but as someone who isn't as into numbers as some people, there were a few parts that went over my head. Regardless if you're a numbers person or not, this is a great book to open your mind and realize the flaws in our thinking and how data can be twisted to push agendas or make us make dumb decisions.
anna_hepworth's review
DNF; was likely to be 1 or 2 stars. I did start skimming in the hope that at least I'd get the gist, but I was too frustrated to keep going.
There are better books that discuss these topics. No need to read this one.
Many of the examples given were shallow, and missed things I would consider relevant when discussing why the conclusions were likely to be skewed. The writing style--which I've found reasonable when reading about fields I'm not familiar with--grated, although how much that had to do with the frequency with which I said WTAF about some statement, I'm not sure (I am very familiar with data and statistics). Examples are Very USian, from a particular cultural milieu, and I often find books written from that particularly alien-to-me culture very off-putting. I suspect that Black and/or working class USians would find this full of micro-aggressions.
Reading one of the early examples, my thoughts were 'wow, this has not aged well', and was flabbergasted to discover that this was published in 2014.
I did find some of the section on reproducibility of science to be good, but as I'd hit the 'I am not going to read all of this in detail, I'm going to skim' by then, I can't be sure it didn't have similar issues to the previous sections.
There are better books that discuss these topics. No need to read this one.
Many of the examples given were shallow, and missed things I would consider relevant when discussing why the conclusions were likely to be skewed. The writing style--which I've found reasonable when reading about fields I'm not familiar with--grated, although how much that had to do with the frequency with which I said WTAF about some statement, I'm not sure (I am very familiar with data and statistics). Examples are Very USian, from a particular cultural milieu, and I often find books written from that particularly alien-to-me culture very off-putting. I suspect that Black and/or working class USians would find this full of micro-aggressions.
Reading one of the early examples, my thoughts were 'wow, this has not aged well', and was flabbergasted to discover that this was published in 2014.
I did find some of the section on reproducibility of science to be good, but as I'd hit the 'I am not going to read all of this in detail, I'm going to skim' by then, I can't be sure it didn't have similar issues to the previous sections.
branniboob's review
3.0
This was an informative book. It's a good introduction to typical mistakes one makes when drawing conclusions from data, (self selection bias, survivor bias etc) and has got me thinking about some conclusions I make when in discussion with other people.
I feel though that Smith rather ran out of steam towards the latter half of the book. The material began to feel unoriginal and I felt like there wasn't much to take away unlike in the first half.
I'd recommend a quick flick through the earlier chapters for a new lens through which to view observations made about the world.
I feel though that Smith rather ran out of steam towards the latter half of the book. The material began to feel unoriginal and I felt like there wasn't much to take away unlike in the first half.
I'd recommend a quick flick through the earlier chapters for a new lens through which to view observations made about the world.
belovedsnail's review against another edition
informative
medium-paced
3.5
A good run down of data errors. It may just be me, but I felt it could have been a much shorter book and not every error required so many examples.
therightprofile's review against another edition
3.0
An insightful and necessary book on how to be suspicious and wary of studies which are biased and where data may have been ransacked and molded in order to fit a certain outcome. There are many useful tips to keep in mind to navigate through the uninterrupted flow of information and useless studies. It does get a little repetitive and loses its engagement and yet there's a wealth of humour and the information is presented in a very clear, engaging and concise manner.