Monday 28 September 2015

Dodgy Digits - of wobbly numbers as props to "policy"

A ‘scientific’ study – backed by commercial interests - suggested that a drug would be helpful and was administered on a wide scale. The initial indicators were then challenged and the US FDA said not to use it  - but an article says it was still being sold in millions. It says that consumption of the drug led to teenage depression and suicides.  The article does not say or show how this latter link was established. Here is the link to the scenario:

The Independent newspaper reported a couple of days ago a study from which  a projection was made that a precise number of deaths arises from (a form of) air pollution.  How this precise projection was reached was not explained (with the difficulties of partialling out all manner of other associated influences on health).

A long story in a Scottish newspaper combines exact estimates of deaths due to X, with more cautious phrasing in places, and deep in the article even includes a proviso:
David Newbie, professor of cardiology at Edinburgh University, pointed out that the combustion of vehicle fuel produced both NO2 and particulate pollution so it was difficult to separate out their health effects. NO2 could suffer from “guilt by association”, he argued.
The overall gist of the article is to side with shock-horror rather than with a realistic (?) degree of caution over the available data and their interpretation:

The Independent papers reported in late summer of 2015 that :

research predicts the world has entered what could be one of the strongest El Niño events in the past century .

the full link to this ‘story’ is here:

There seems to be a slight oddity in the use of verb tenses in constructing a report – “predicts” seems to my eye to talk about the future while “has entered”  speaks of the present. I am reminded of the situation in very many television documentaries (sadly, many on the BBC) in which “presenters” (how may the meaning from this word have seeped into affecting behaviour?) speak of the past as though it remains part of an ongoing present:
“Julius Caesar then decides to cross the Channel …” 
Such verbal tricks may be worth ignoring as mere elements of style designed to help viewers understand things better (by inviting them to be involved in the same time frame?) – but perhaps the reporting of past and present and future should remain sharply distinct.

In February 2003 protests took place in many countries against the possibility of a western alliance invading Iraq.  According to Wikipedia estimates of numbers varied:
The Wiki review says that the crowd in Rome constituted a world record – recognized by the Guinness Book of Records. 30 trains  brought people in (estimate? 30 x 1500 = 45,000?) and 3000 buses (estimate? 50 x 3000 – 150,000) thus the “brought in” numbers may have reached 200,000 people. These could add to the total population of Rome which wiki gives as 2.9 million, to fulfil the claim at the head of the article that “the protest in Rome involved around three million people”.  The article later says that police estimates were that 650,000 “took place in a rally”. It is perhaps not clear from these different approaches in the Wiki review “exactly” how many expressed their feelings in the city, and/or took part in the rally. There is a tendency for organizers of rallies to report that more people came, than is stgated by police estimates: for example in the same Wiki article it is said : “Police estimated attendance as well in excess of 750,000 people[32] and the BBC estimated that around a million attended.[33] At the finish rally in Hyde Park, the organisers announced 3 million attended” …..

The strongly anti war newspaper the Guardian provides what seems a conservative or low estimate of the numbers of military and civilian deaths attributable to the 2003 invasion.
The BBC reported that the Lancet (medical journal) calculated that well over 600,000 deaths were attributable to the war.

Overall, it seems that in the chaos of war – or even the disruptions of civil protests – it has been difficult to fulfil a notion that “the number” of people involved, or casualties, can be accurately numbered.

From time to time journalists in the press or charities decide to offer a number indicating the suffering from AIDS – especially in Africa. A website from the World Health Organisation is notable for the statement of numbers from countries mired in conflict and where  even in peaceful times access to the hinterland and exact diagnosis would be problematic – thus figures are provided from the Central African Republic, Malawi, Rwanda and other countries where it may be difficult to accurately pinpoint such things while in the same table well-ordered countries like Singapore, and the USA have entered ‘no data’.
A trusted friend informed me:
When we did the natural remedy tests in Zambia in the 1980s we had to involve the President’s Office to select some ten people for treatment.  It was our tests which established whether or not they had HIV/AIDS, that information was not already listed within the system, and it was a huge palaver to get the selected people weighed and checked each week.  In the end I wasn’t at all sure they had put on weight because the herbal treatment had cured them or simply because, for the duration of the test, they were being fed and watered on a regular daily basis….  Zambia did make an effort with its HIV/AIDS treatments and monitoring because Kaunda’s son died of the disease and I think other members of the family also died; not that it made an exemplar for other African countries because the work it did was too spasmodic and inconsistent.

The Independent big and little newspaper decided to headline an accelerating proportion of the population who/which would be demented.  The “story” included the detail:
The latest findings – which uses mathematical formulae to combine life expectancy estimates with projections for the frequency of dementia at different ages – underscore the scale of the potential future crisis caused by an aging population.
I didn’t really understand how the arithmetic had led to the conclusion dramatized in the headline story which can be reached at:

Nevertheless, back to the reporting of science – there are other scenarios in which firm figures are offered as describing results of some process or behaviour, thus:

     Reducing speed limits in itself will reduce (or has reduced) accident numbers;
     Application of laboratory-achieved genetically modified products will cause
          such and such a number of consumer casualties;
     Polar bears (on whom an accurate census has apparently been conducted) will die
           out in some large numbers someone has calculated (how?) ….
     A given number of tigers in the huge nation of India will be reduced to another
           said number, unless particular measures advocated by campaigners, are put in place;

…. This is a beginning to what would likely be a log list of examples where “the figures” may or may not substantiate some (political) case being made. Since skill with figures is not all that widespread, much nonsense develops in “the public discourse”.

No comments:

Post a Comment