Artillery Row

Does it all add up to you?

The British state can neither gather, interpret, nor explain its own statistics

As headline figures go, the UK and its government have had a tough few days. The worst total Covid-related death rate in Europe and the largest economic crash from April to June of any country in Europe or North America is what used to be known as a double-whammy. No amount of spin can either explain away or divert attention from headlines that are readily understood by experts and laypeople alike. A scan of the overseas media demonstrates that Britain’s dire performance since Coronavirus arrived on its shores is not lost on foreign audiences as well. It is not a great boost for our national prestige.

The UK’s 20.4 percent second quarter fall in GDP also represents Britain’s worst economic nose-dive since it began compiling readily comparable data in 1955. This is a preliminary statistic, which may be revised (in either direction). But no subsequent tinkering is going to alter the reality that over the same period German output fell by half the amount of that of Britain and that the United States dived by less than half the UK’s trailblazing belly-flop – despite the United States also being on the world leader-board for Covid fatalities.

For a government that has presided over these measurable catastrophes, the Conservatives are still remarkably popular. Their lead over Labour has shrunk, but they retain a decent buffer, despite the main Opposition party having found a leader more readily marketable to middle-of-the-road opinion.

An obvious reason for this may be that many voters think that the approach (both on public health and economic assistance) chosen to tackle the pandemic would not have been much different if it was being led by Keir Starmer rather than Boris Johnson. Given Labour’s support for the general strategy (whilst criticising the delivery), it is not an unreasonable assumption.

Another reason is that Rishi Sunak’s ameliorative measures, and in particular the furlough scheme, has created a disconnect between economic collapse and job loss. After 31 October (when the furlough scheme ends) that latter may more closely correlate to the fall in output that has taken place, assuming recovery is not by then happening so rapidly as to ride to the rescue – an optimistic hope.

In the early 1980s, at the peak of the economic downturn, the performance of Margaret Thatcher’s government was effectively weighed in the balance each night by the evening news showing a map of the country with a running meter of where jobs were being lost and created. For almost all of that period, the meter showed a net loss. Every night, therefore, the British public were bombarded with statistics that gave a depressing picture. Given that more than three million were ultimately unemployed, the broad brush captured a grim reality. But by only documenting the big businesses that were opening, closing or laying-off, the nightly figures were misleading because they failed to capture the smaller businesses who were hiring and firing (or the decision of many to become self-employed). Individually these were small changes, but collectively they were also important indicators of economic health that the nightly news missed.

Getting accurate data underpins decision-making and popular perception. Remarkable is the scale of the Johnson government’s failings with statistics – gathering them, interpreting them, explaining them. Given that the Prime Minister’s special adviser, Dominic Cummings, has built his reputation in part upon his advocacy for data science capabilities, this is an especially shocking flaw at the heart of government.

There has been no shortage of coverage about Cummings’s trip to Barnard Castle to test his eyesight, which many (but not the Prime Minister) felt was a resignation issue. Beyond acknowledging that he has attended SAGE meetings, less clear-sighted has been the ability to explain what Cummings’s role has really been in the government’s data-analysis efforts to understand the virus and its wider effects – not only across departmental delivery but on popular behaviours and economic performance.

Whatever the strengths of this government’s communications team, how to present statistics is not one of them. The daily announcement of cases and deaths provides raw facts but no context. When, at the peak of the crisis in early April, daily death figures were hovering between 800 and 950, how did that compare with other major causes of death at that time, or to an average day during a bad winter flu spell? Who amongst us knew?

A sense of how non-contextualised statistics can cause major misapprehension is glimpsed by research undertaken into what proportion of the population the public think have died from Covid-19. Day after day, the recorded numbers were broadcast, printed, and shared. Any concerned citizen had ready access to the raw numbers. But unless they troubled to get out their own calculator and crunch the numbers, they were provided with no means of gauging the proportionate scale of the unfolding tragedy. 46,700 people in the UK have so far died of causes attributed to Covid-19. That is 0.07 percent of the population. But when polled, the British public think 22 percent of their fellow citizens have had the virus (four times larger than the modelled data suggests) and, even more alarmingly, that 7 percent have died from it. In other words, we think the virus has been 100 times more deadly than it has been.

If it is any comfort (and it should not be), the British public are not uniquely incapable of putting statistics in context. Swedish and French respondents were equally far out for their respective national rates. Americans though led the way, imagining that 9 percent of their population had died – a figure equivalent to nearly 30 million fatalities. The actual figure is less than 155,000.

we think the virus has been 100 times more deadly than it has been.

Who is to blame for the public’s massively inaccurate perception of the scale of Covid cases and deaths? The media could have provided context, but – with rare exceptions – has not done so. Perhaps too many journalists either don’t want to diminish the power of a frightening headline or simply have also lacked the gumption to unpack the raw figures given to them. The government comms teams though could have done the basic donkey-work for them. From lobby briefings, to press releases to Downing Street press conferences, no such context has ever been spoon-fed to reporters, readers or viewers.

More extraordinary still is the revelation that Public Health England (PHE) attributed a Covid death to anyone who had tested positive for the virus and subsequently died, even if they did so months afterwards from causes unrelated to Covid. Unless corrected, it would in theory have meant that anyone who had ever tested positive would ultimately be recorded as a Covid victim.

Poorly thought-through decisions are easily made in crisis situations, but it was not a mistake that the devolved authorities in Scotland, Wales and Northern Ireland made. It now transpires that much of Scotland’s supposed superiority over England in Covid fatalities is a statistical error.

Politically though, the damage is done – the SNP administration has built its extraordinary recent popular support by reinforcing a misapprehension of a significant difference between Scotland and England that is now deeply embedded in Scottish self-perception. The difference, per head of population, is actually quite small and may be attributable to influences that have little to do with marginal differences in public health policy between the two nations. But good luck interrupting Nicola Sturgeon with that quibble.

Belatedly, PHE has changed its definition to conform with that already used by Scotland, Wales, and Northern Ireland (anyone who dies within 28 dies of testing positive for Covid). So, instead of 42,072 deaths in England, the figure has been down-scaled to 36,695. Incidentally, the change now means that the UK has a similar death rate per head of population to Italy and Spain and is not the great outlier that the raw data (and popular perception) believes to be Britain’s outstandingly poor performance.

Less susceptible to modification will be the UK’s GDP collapse. But that is, in part, influenced by a lack of consumer and business confidence. And how can that confidence return if Britons think the virus has already killed 7 percent of the country, the equivalent of 4.7 million people?

Enjoying The Critic online? It's even better in print

Try five issues of Britain’s newest magazine for £10

Critic magazine cover