I thought it would be criminal not to have written anything from this book.
Do note before I begin, I have not done background research on the writer, David Randall. The book though it seems has gone through 4 editions from Pluto Press (London) – the latest in 2011.
A great majority of us go through statistics. It is all the more pertinent that policymakers and decision makers use these figures and their interpretations wisely. Of course the average citizen is far from being excluded in this exercise of judgment.
The initial segment corresponds with what most the History, Social Studies, Geography and Economics syllabus are doing: source and data/statistical analysis.
Randall recommends that we ask who created the statistic; the reason behind it; and the timing of its revelation. One example (provided later on) was how an employer could claim they suffered 7 million pounds of losses from strikes each month. That huge sum could possibly garner sympathy or encourage government action against the workers. The amount could have included anything from heightened insurance fees (plausibly for breaches of contract in production) to additional money clients pay to get a substitute product!
One aspect raised was the issue of context. A company was reported in 1997 that 29% of its ex-employees were killed by cancer. Yet in the greater scheme of things, 35% of those between age 44-65 die from cancer which actually is 6 percent higher!
The terms or definition used also ‘create’ statistics. Sinusitis as it was included under the definition of ‘chronic diseases’ increased those with ‘chronic diseases’ by 32 million (in the United States)!
Percentages are often attention grabbing but can be rather deceptive. People may declare that they raised profits by 18%. What they failed to tell you was that this was in comparison to 20 years ago (and this is exclusive of inflation mind you). Withal, the denominator for arriving at these percentages are conveniently left out at times.
The average or mean is another concept that needs to be interpreted. Say the total weight of eight persons in the elevator is 560kg. The average is 70kg. But in reality some may contribute 60kg while another 80kg. Thence, the range of values is not clear when using average.
When comparing, raw results are insufficient. There should be a ‘rate per unit’ cross examination. It may be accurate to state that more expired in air travel during 1998 as opposed to 1952. But after one takes into account the total number of miles travelled as well as the travel volumes, one gets a different (and more assuring picture).
On a related note, there needs to be apple to apple comparison. (Reminds me of my days doing benchmarking). Comparing death rates of fit young soldiers in peace time; with elderly, impoverished residents in a small village will hardly generate useful indications.
Next we look at survey or search samples. Sample sizes have to be significant. If you ask one student out of 40 how well the teacher taught, your picture is gravely limited. An interesting example is seen when a survey gives a sample size of 2000 businesses but having only 160 respondents. 80% felt the interest rates were too heavy. But the headline was packaged using those 160 companies and not the 1840 others who did not answer…
Even when the size is adequate, one should note the margin of error. Since the studies are only ‘guides’, the margin allows for a group ‘hit’. For instance, the result may show that 44% prefer the ‘ABC Party’. Yet, with a 5 percent margin of error, in the worst case, the reality is 39%. Of course the converse of 49% may be true. In environmental studies, certain geographical areas may likewise be favoured over other areas. Those that challenge the ‘desired’ results may be excluded altogether. The same principle applies, the data collected must be representative and comprehensive.
Further, we need to check who is consulted. It is less useful asking political theorists whether climate change was a reality. ‘Leading questions’ or poorly constructed ones can also skew the outcomes. From other sources, if you asked “How important is it to getting a good first job?”, in some ways we are preloading the idea that it is important. And if you queried: ‘Do you pick your fart/pass gas frequently?’, how often do you think you would receive the truth? Similar issues affect statistics for injuries since some may exaggerate in order to win insurance claims. (Death statistics as Randall mentioned are usually more reliable).
Moving on, causation and association are concepts that get conflated and confused. Back in my Cross Faculty Module (CFM), my Statistics professor introduced the idea of spurious correlation. The illustration he gave was more highways ‘led to’ more hurricanes. It was probably more roads that allowed one to see the hurricanes and probably also enhanced technology that enabled the efficient discovery of existing hurricanes! Surrounding factors and background have to be considered indeed.
The author ends off by reiterating the importance of statistics. And I agree with him. I would rather have the truth. Amazingly, the truth is stranger than fiction. In countries experiencing harsh winters, there are in fact more accidents during sunny days than ones with heavy snowing. Why? Because more people drive! Thus a heading could be as Randall writes: ‘Snowfall saves lives’.
He recounts how Mao Zedong, during the Great Leap Forward, ordered the elimination of sparrows which apparently reduced grain production. However, by removing sparrows, which had insects as its preferred food and grain as the minor second choice, locusts multiplied. Famine occurred with its correspondent fatalities ranging from 16-30 million. The accomplices in this tragedy were bureaucrats of ‘ignorance’ and the ‘supine media’ who did not check the facts nor raise alarms bells at the impending folly.
No doubt the last lines of the chapter are fairly self seeking; establishing the necessity of the Fourth Estate. Nonetheless, the strength of the argument is not easily erased. The book along with its thought provoking quotations in various chapters is definitely worth a read.