According to České statistické společnost (Czech Statistical Society), the government’s data keeping track of the spread of COVID-19 is unprofessional, full of holes, and basically make no sense whatsoever.
In an interview with Irozhlas, Ondřej Vencálek, Chairman of the CSS, was careful not to say whether the data was exaggerated or under exaggerated, but did imply that the “numbers” were causing unnecessary fear within the population and that they couldn’t actually be backed up by facts.
“We’ve been monitoring the situation and we don’t like this unprofessional use of statistics. The motive [of ÚZIS] seems to be to make more and more terrifying predictions based on data that we’re not allowed to see. There’s a problem with this; a statistician isn’t a person who plays with numbers, they’re a person who aids in the processing of information.”
Vencálek said that when the Ústav zdravotnických informací a statistiky (Institute of Health Information Statistics, or ÚZIS) published their models, the CSS doubted them.
“The data collected often can’t be used to answer the questions we are asking. The models are often not made public, and if they are public, they haven’t even been professionally reviewed. When ÚZIS published its models, we looked at them and had great doubts. Most other organizations have not published their models either but they still publish these wild predictions. That’s wrong, it has to be transparent.”
Vencálek reiterates that this doesn’t necessarily mean the government’s numbers are overblown, just that it’s impossible to ascertain the accuracy of them. Vladimir Kroc, interviewing Vencálek, notes that cases in January went up just as predicted by ÚZIS, despite Vencálek dismissing those predictions in the prior month.
“Dramatic predictions have consequences; they cause fear in the population. I don’t know if that fear is justified, I’m not saying it didn’t happen. I’m just saying we don’t know at all and these predictions can’t be made like that. It’s not possible to publish something without showing the model. The models used by ÚZIS, and published by them, are a disaster, there are so many bugs.”