MOMO mumbo jumbo (#COVID2019)

Tension is brewing between those who want to measure a nation’s Covid-19 deaths by aggregating individual records and those who want to model an ‘excess deaths’ figure. This blog looks at the MOMO numbers which are being used by some. The data are worth looking at but are inaccessible (due to being poorly presented) and seem to be giving a misleading impression. As they appear from a largely unexplained black box it’s hard to understand them. I can only hope they will not be given any real role in the debate until we know what the numbers mean. Continue reading MOMO mumbo jumbo (#COVID2019)

Dodgy data and ‘bankrupt’ models

If the data are useless – and they pretty much are (see my other blogs) – there is a temptation to turn to experts, and their models. Sadly they are offering no greater help to a confused person looking for insight. One reader comment on the FT site makes the point (a little unfairly): “Why is the UK the only country to be churning out these intellectually bankrupt models and ideas?” Continue reading Dodgy data and ‘bankrupt’ models

Died “with” or died “of” Covid-19

The accuracy of the estimates of deaths attributed to Covid-19 is a hot topic. I fear that the numbers are overstated. First, there is a (long running) issue related to the accuracy of the information put on the death certificate. Where there is flexibility and uncertainty in the system of registration, I think that balance of probability is that the pressures on those completing the forms will lead to over not under-estimation of covid-19. Second, the figures that exist are not presented fairly as they mix up the “underlying” cause of death with a mention of coronavirus anywhere on the form. ONS must present underlying covid-19 deaths on the same basis as they present other deaths. Continue reading Died “with” or died “of” Covid-19

New and open data can help tackle Covid-19

I’ve already explained that the publicly available data is of little use in tracking the progress of the virus and the consequences of related policy. We’ve had glimpses of the data that the government (see below, COBR) is using. It is now time to discuss openly what new or high frequency data the government has, how it can be shared and what other data might be sourced from inside and outside government to tackle the virus. Continue reading New and open data can help tackle Covid-19

Covid-19 – fighting a virus without data

What a difference a month makes. We’d heard of covid-19 deaths in China and the Diamond Princess cruise ship drama was acting out but little did we imagine …….. This blog summarises my thoughts on the data and statistics of the crisis so far. I suspect in a few weeks it might be instructive to look back at where we were. Even if the data and communication of the underlying facts are now very poor – and they are really very poor – the judge of our institutions and professions will be the response. Read the summary points below as they show that a response is now needed. (Link to full document.) Continue reading Covid-19 – fighting a virus without data

A new way to present death statistics – ONS

It is good that the ONS is improving its publication of weekly deaths but the improvements cannot mask that it is very out of date data in a fast moving environment and that, still, a lot of people who died with not due to the virus will be included, exaggerating the deaths number. Continue reading A new way to present death statistics – ONS

Public Health England should stop publishing misleading data

We should expect important people in positions of power to publish good quality data or failing that ensure that poor data is presented fairly. I wrote today to Public Health England and the UK Statistics Authority to ask them to ensure that policy makers and the public are given the information they need and deserve and that the published data is no longer presented in a misleading way.   Continue reading Public Health England should stop publishing misleading data

Describing datasets – the “Highgate 12”

People are confused about what data about them exists, who owns it and how it is being used. We all know that our health records are not the same as what Facebook has on us, and a store card is not the same as mobile phone tracking, but just how are they different? Ultimately, is the arrival of data into so many corners of our life something to celebrate or, as the media often portrays, something to fear? This framework sets out some key features of datasets including, for example, ownership, transparency, consent and access. Identifying the features allows both the risks and opportunities from data to be better understood. Once understood, there’s less likely to be a knee jerk reaction against data – and that would limit the scope to improve both our public services and our personal decision making.

Continue reading Describing datasets – the “Highgate 12”

Treasury’s options on RPI?

Wednesday is set to be a sad day for anyone who believes that the UK’s statistical system is “independent”. The Treasury, yes, the Treasury, will pronounce on the future of the RPI, and/or the future of UKSA and its various ill-defined and overlapping parts. The Cabinet Office is actually responsible for UKSA, UKSA has a reporting line to Parliament, and you might think that the Office for Statistics Regulation would do what its name suggested but no, the good old Treasury will impose its not-at-all independent self on the data. Continue reading Treasury’s options on RPI?

How poor governance led to the problems with the RPI

Wednesday morning (it has been announced) will see the publication of the Government’s response to the critical House of Lords report on “Measuring inflation“. This blogs identifies a series of failings in governance that led to the extraordinary and stupid situation everyone finds themselves in. Continue reading How poor governance led to the problems with the RPI