We undoubtedly live in the information era, where we have 24 hr rolling news and can Google any topic in a matter of milliseconds, but how accurate is this information and indeed can we use it in a significant way?
Neuroscientist Daniel Levitin tries to address these themes in “A field Guide to lies & Statistics” and his book enables me to put some flesh onto the bones of my initial guide to good writing.
The bottom line is whether the book does what it sets out to do, and if so, are there any useful techniques or conversely examples of practices that the novice writer should avoid.
Levitin is described as a Neuroscientist, though I found out he has flirted with a career as a stand-up comedian, so I was expecting as a minimum, high levels of communication skill if not Nobel Prize winning prose.
A second factor is the subject matter; Statistics. One of the truths laid bare in the book is that in the West, despite our worship of computers and algorithmic methods, most of us are curiously number phobic (Google’s CEO doesn’t even know his own salary … apparently). My experience of filling in my Mother’s tax return brought this inconvenient truth home to me with frustrating poignancy.
Rather than launch a long & eliptical diatribe I’m going to apply my benchmarks outlined in the first part of this blog, to measure how well Levitin’s tome and indeed my metrics stand up.

Alex Sink Webpage

That Sinking feeling; Identifying Expertise

    1. Research

The book seems well researched on the brief of how we deal with/ interpret statistics; but hey I’m no statistician so my endorsement of the depth of his knowledge, is not copper-bottomed, but he is heavily (academically) credentialised on his dustcover. The one area I found unconvincing was the Internet, even though he gave the quagmire of Social Media a well advised body-swerve. Levitin is definitely mining the right seam as in an ideal world our decision making would be improved if we based it on appropriately accurate information (WMD’s comes to mind).

    1. Voice

Again I can’t fault the voicing which is conversational and not overly jargonised. As in songwriting the opening chords are often critical, and his stand-up skills are evident here and but he eschews the warmup anecdote in the first paragraph, with a standard instructional tone aka ; tell em what yo gon tell em. The logic behind this always fails me as surely the reader would have read the Amazon  synopsis blurb,   if not the dust jacket or review ?  There were a few passages which left me scratching my head but these were often due to my poor grasp of the Mathematics being exposed, rather than an inherent faultline in authorship. Again the dryness of the subject matter is difficult to camouflage, and this is definitely not a page turner.

    1. Structure

This is the one area where I would raise an objection. Levitin’s stated objective is enable the reader to make sense of a complex world by improving critical thinking. An ambitious objective which in my own mind was never achieved. The book is split into three main sections (I) Evaluating Numbers (II) Evaluating Words (III) Evaluating the World ; there’s a heck of a lot of evaluation gwan. There’s even a short conclusion section, which is … inconclusive.The structure seems very arbitrary and in fact the book can be read pretty much any order, as there is no narrative thread. For a technical reference this structure would be appropriate, but within the sections the content is a numbing interlace of some heavy maths, and candy floss related anecdotes.
For a book written by an academic it is very anecdotal, if not contradictory and hence unmemorable. I was particularly looking forward to an explanation of the much heralded Bayesian theory, so read keenly the chapter entitled Probabilities. Although three types of probabilities are discussed, Classical, Frequency, and Bayesian, I was none the wiser on how you could actually apply them to everyday problems; the Chapter is rambling and confusing, with a lot of bewildering probability math. Educationally glaring are the lack of examples on which to reinforce/practise your understanding of probabilities exposed by Levitin.
Another problem is the clunky gear change half way through(Identifying Expertise) when the Internet as a information resource is dealt with. As TP Barnum famously said, there’s a sucker born every minute, and the exploitation of gullibility is right up there with the world’s oldest profession, but Fake news or Levitin’s Counterknowledge has become a hot topic (de nos jours). From personal experience trying to work out the veracity of the Internet, is a bit like tilting at windmills; a bridge too far for a statistics backgrounder. Statistics and Probabilities are a difficult enough subject without throwing in the shapeshifting, chaotic curve ball of the Internet. His assertion that you can determine the veracity of a web page simply from its Url is simply doesn’t stand up to inspection!.

    1. Clarity and Conciseness.

As alluded to in 2 (Voice) Levitin’s prose is clear and easy on the eye. So he passes the first hurdle, however he’s very prone to meandering with often contradictory anecdotes, which muddy the message he is trying to convey. There are numerous examples of Statistical gotchas, but rarely does he elucidate a cohesive strategy to deal with this bad data. I concurred with his overall mantra that we must be more robust when presented with statistical information, but am still uncertain as to how to go about this proactively. As I stated previously opportunites to practice the learning are minimal, as are “tell them what you told ’em” summary sections, which a badly needed with such an alien subject matter. So clear, but rambling. Here’s a typical example of the opacity of a classic Levitin ramble from the Identifying expertise section:

“Individuals with similar training and levels of expertise will not necessarily agree with another, and even if they do, these experts are not always right.”

If there is a third Act in this book I blinked and missed it.
For a book written by a ex. Stand-up comic I found the book did not reach the Kermodian 6 laughter level.

    1. Mechanicals

I can’t fault the mechanicals as the book has few typos, a good font size, and use of whitespace and is largely comprehensible, with many graphics. It has been well edited. There is an Appendix on Bayes’s rule, a Glossary, and extensive Notes (cf.1): a lot of effort has gone into the book’s mechanics. Whilst the prose is workmanlike it is often uninspiring, and narratively very granular; short little stories that are often literal cul-de-sacs, which leave the reader non-plussed, rather than building to its conclusion; Discovering your Own which basically I interpret as having taken 5 hrs of your life … you sort it out!.

    1. Polish

Lies and Statistics is nothing if not polished, and fully up to the standard you would expect from a Penguin Random House/Viking production. On a technical level, the book is complex with a plethora of figures, tables, diagrams and photos, so great credit has to be given to the production team and his long-suffering editor Stephen Morrow. A professional job. Notwithstanding the efficiency of the prose there are few memorable literally flourishes, or indeed phrases.



Ok time to fess up; Lies & Statistics should be a manual for the rest of us (the innumerate) to navigate a world where we drown under the barrage of figures, aimed at getting us to buy into certain stories. This can be as simple as buying a product, following your favourite footballer, or as nuanced as voting for a politician or party, or indeed approve a life changing decision to go to war.
The truth is that in a lifetime, whether it’s a lawyer, a plumber, a priest or politician, we are repeatedly lied too, and Levitin outlines various strategies we can can use to test these assertions. The Internet has become the tool of choice of the modern medicine men, but even a bus can be used to propagate dodgy statistics. The most notorious case of a prospectus of falsehoods was the dodgy dossier, used to hoodwink the British Parliament to vote on a disasterous war in Iraq.
An interesting aside has been the recent fallibility of the bookies & pollsters in trying to predict the probability of election results & referenda, but this is more eloquently explored by Nate Silver’s blog.
Levitin is a skilled and experienced writer, who in his anecdotes displays some narrative (story-telling) savvy, but he’s clearly not targeting my demographic as I feel you need some basic grounding in Statistics to understand his book. This strategy seems all the more bizarre in that the demographic of counters is a minority to start off with.
The book is well written, with plenty of visual presentations & interesting anecdotes. However paradoxically some parts are very complex with others (Internet) bordering on the banal.
A key writing skill is identifying your audience. The problem with a Field Guide to Lies and Statistics is not particularly that the writing is bad, but it is inappropriate for a lay audience, and I could only recommend this for an academic audience; a classic case of preaching to the converted and hence a massive missed opportunity.
Levitin only superficially deals with why misinformation is so prevalent. Is it simply a human trait, or a result of the ease with which it can be accomplished, and the low chances of being caught out?
One aspect of this is Levitin’s reluctance to “follow the money” in Wire parlance. Frequently there is a financial component to misinformation. I can also think of situations where the damage of this misinformation is bought off by a financial settlement, or ends in financial disaster (credit crunch). For me this is a logical extension of Levitin’s work and may have proved more illuminating in explaining the motivations and consequences of misinformation. In itself the financial dimension requires an additional layer of vigilance.

Leave a Reply