Underpinning the government’s plans for more open data are two big assumptions. Both are about the public. One is that people will take a positive view of data sharing and see the advantages in departments and agencies mixing and matching the information they hold.
The other is that people will make sense of what is released. That they will take in the numbers and percentages and understand what they are being told. Though Cabinet Office minister Francis Maude did not use the phrase when he published the white paper on open data, he clearly has in mind the ‘armchair auditors’ painted by his colleague Eric Pickles – citizens able to hold government more closely to account thanks to the voluminous information about the state that is becoming available.
But our colleague, the stats postdoctoral research fellow at the University of Cambridge Robin Evans, sounds a warning. (he blogs at itsastatlife.blogspot.com ) Say doctors are now going to release information about their patients’ cancer treatments. From it you could easily construct a rough and ready performance table showing the proportion of patients who recover. A fair appraisal would, among other things, ask how many cancer cases a doctor sees, and grade them according to severity - but that complicates the picture. The trouble is, that without the caveats, the open data result may be misleading.
Who’s going to insert the caution? Evans says there’s a challenge here for the statistically literate. ‘We have to ensure that naive and misleading interpretations of data are not allowed to predominate.’
But the white paper does not have much to say about who might step up and educate the public, shooting down erroneous interpretations and inserting qualifications into the picture.