Very soon, @realDonaldTrump is going to announce a pick for #SCOTUS. Journalists will jump al over it and start using social science data to make claims about how conservative the justice is compared to others. Here are some predictions. 1/8
First, people will rely on measures of the nominee’s conservatism that are about the president who last appointed the nominee to the Court of Appeals (I’m assuming it will be a current CoA judge), rather than measures of things the judge has done or said herself. 2/8
Second, as we all know, lots of data suggest that the parties are polarizing. That means past presidents look more moderate than do current politicians. That may or may not be true in the real world, but it’s almost guaranteed to be true in the data they use. 3/8
Third, because of these two factors—-the use of measures of old presidents as proxies for current judges and polarization over time—-it will *look* like the judge is more moderate than Trump or other GOP Senators right now. 4/8
Fourth, journalists (and some academics who are not thinking carefully) will make a big deal of that fallacy to make grand claims about Trump begin strategic, trying to get the median Senator’s vote, etc. 5/8
They might event claim the new nominee isn’t going to change the Court much. But, in fact, Trump’s nominee will be ideologically extreme. The current Senators can’t be lost over ideological extremism. They have shown their willingness to support Trump through more than this. 6/8
In 2 years, after we have lots of data on the new judge’s voting, the same journalists and academics will be shocked and make a whole lot out of the new judge “drifting” to the right and likely even attribute it to the fact that the Dems will (probably) be in power. 7/8
These great claims about our democracy will be driven by poor data literacy, a lack of careful reasoning, and a desire to convince ourselves that this is all more complex that it really is. Trump is a far-right president and has the votes to do it. 8/8
For those you you playing along at home, the whole thing is particularly perverse, because the process I just laid out shows the core logical problem with measuring new judges’ preferences from the president who last appointed them. 9/8
The fact that we keep seeing judges "change" over time, as we move from measures of their appointing presidents to measures of how they vote tells you that it never made sense to use the appointing president as a measure in the first place! 10/8
FWIW, here's how you know that the voting on this nominee will not reflect her ideological orientation. 
• • •
Missing some Tweet in this thread? You can try to
force a refresh
