I learned a lot this year about the impact of missing data or information about data from governments in a fraught environment.
Sometimes governments worry about making data or information public because they have concerns about the impact of doing so. Publishing data and information has a cost to it; technical costs may be low but person costs frequently aren’t, if you want to do it well. Governments might worry about the quality of data or the level of uncertainty around some figures. They might feel that the public or journalists will not be able to understand or communicate the reasons behind the numbers. Or the data might tell a story that they are themselves concerned about, and they know there will be follow up questions that they’ll have to handle.
But when governments make those decisions, they frequently fail to consider the impact of not making that data or information available, or doing so after a delay or in response to a fuss. They tend to think that doing nothing is cost free. It isn’t.
How not sharing data lands
People take the lack of data and information about data as a signal. The interpretation of that signal varies depending on the person and context. Here are some examples of the assumptions people will make.
The incompetence assumption
Some people will assume governments aren’t sharing data because they aren’t collecting it, because they are incompetent. Data is essential to keep a handle on an evolving environment, so people interpret not sharing data as a sign governments don’t have basic facilities in place to handle the situation. This can apply to whole datasets (eg case data, vaccination data) or when there is a lack of granularity within data (eg to low level geographies, or split by significant characteristics such as gender, age or ethnicity).
Concern that government don’t have their hand on the tiller leads people to distrust their judgement and route around your direction and guidance. It also may entail firefighting comms from government to reassure people they do have the situation under control.
Missing data is only one indicator of incompetence, of course, so publishing data is unlikely to entirely fix this problem. But minimising the number of incompetence indicators you’re displaying as a government is a good idea.
The prejudice assumption
When the data involved is most relevant to a particular community, particularly one that is historically neglected or maltreated, some people will assume governments aren’t sharing data because the needs and concerns of that community are not of interest to them, because they don’t care about that community due to prejudice.
People might assume governments aren’t sharing data because they’re not collecting it, which implies they don’t want to see what’s happening to that community. Or they might assume it’s because governments don’t think it’s worth sharing data with that community, because they do not think they are competent to be able to understand or use it. Both imply prejudice.
During the Covid-19 crisis in the UK we have seen this in relation to data about coronavirus in care homes (lack of care about old people), data about the impact on people of colour (racism), and data at local levels (London-centrism).
Concern that governments are prejudiced leads to people from affected communities distrusting their intentions in other areas. This can widen existing inequalities through differential public health and economic impacts when, for example, fewer people from those communities use an official app or they choose not to take up official economic or health support such as screening tests or vaccinations. Again, governments need to shoulder the cost of firefighting comms to counter these concerns when they hit the press.
And again, missing data is only one indicator of prejudice. That doesn’t mean governments shouldn’t try to address it.
The obfuscation assumption
Some people will assume that the reason data or information about data is missing is that there is something nefarious going on that governments are trying to hide, or that the data reflects poorly on them. We saw this in the UK when there were moves to stop publishing data daily early in the pandemic, and in reactions to the NHS’s relationship with Palantir and other big tech firms.
The concern that governments are hiding something leads people to assume the worst. This might be that the figures are really awful, or that they are sharing personal data with firms with poor reputations. Lack of data and information about data creates a Petri dish for misinformation and conspiracy theories, again increasing general distrust, reducing take up of official support, and leading to bad health outcomes and firefighting comms.
Maybe governments are actually doing something they feel ashamed of. It’s unlikely to be worse than the stories that people imagine in a fact vacuum, and governments would be able to manage the consequences a lot better with proactive, controlled comms.
How to manage not being able to share data
As I’ve described, people take the lack of sharing of data or information about data as a signal. So what should governments if they can’t share data because they actually don’t have it, or because it’s too difficult to publish, or because they have to prioritise other tasks?
If it’s not due to incompetence, prejudice, or a cover up, then governments need to communicate the real reason, rapidly and proactively. Ideally this should include a description of what they are doing about it. For example, how are they working to collect missing data? When do they think they will be able to share the data or information? If they are currently deprioritising this work, what are they prioritising instead, and when will they review that decision?
Missing data limits how well we can understand the world, and how informed our decisions can be. But the fact that it’s missing is also problematic because people interpret it as a signal. Governments, and other organisations, should take this into account when assessing when and whether to publish data and information about data, and invest in good proactive comms.