Thread by @JohnFPfaff

As someone whose work relies almost entirely on government-gathered data, there are times I just feel like throwing in the towel.

We cover up bad policy by screwing with the data (this, CDC and police killings, etc etc), then use that bad data to justify the policies.

https://twitter.com/JamesSurowiecki/status/1480304142858223616

Then too many journalists cover that data as (roughly-enough) legitimate--though seeing some encouraging pushback recently w crime data.

And, honestly? Too many academics throw the shoddy data into a fancy model and make bold claims with it (see, e.g., everything using the UCR).

I have a paper I'm completely frozen on bc I can't figure out if it's even okay to really use UCR data to say anything anymore.

And we're going to spend YEARS trying to figure out exactly what happened w Covid--meaning, what the numbers really ARE.

Sometimes, it's just... really f***ing depressing. And infuriating. And exhausting.

So much computational power. So many methodological advancements.

And so many people screwing up the inputs. Sometimes out of honest mistake. But too often for far more craven reasons.

My dad wrote this paper 45 years ago running punch-cards through a mainframe (jstor.org/stable/2326815).

I have more computational power than he could have imagined then sitting on my lap right now.

But I sometimes really wonder if I can actually say more than he did?