I continue to believe that the poor overall quality of research produced by honest actors is a much, much bigger and insidious problem than outright academic fraud. And that includes poor quality research done to advance narratives about fraud.
It's the "comically overconfident" line that really gets me in this context.
Nate Silver mansplaining what is political science to the sitting President of the Society for Political Methodology is *chef's kiss*.
"Until we solve this problem of intervention validity in independent research about adaptive algorithms...scientists wonât be able to provide reliable answers about the effects of a system that can be used to guide future decisions. And...companies are not incentivized to produce that knowledge"
Does science show as Meta claims that FB algorithms didnât cause much harm in the 2020 election? Or should we believe scientists, who disagree? It became complicated this week with news that Meta changed the algorithm underneath scientists. I unpack the story: natematias.medium.com/how-meta-spi...
PR teams spin scientific results as evidence of safety while also undermining the findings. Where do we go from here?
Some numbers: donors cover 20% more of UW-Madison's budget than the state. Tuition is a bit > gifts/donors, but coordinating action by 50k students is a lot harder than a handful of big donors. 2 biggest gifts were $70M & 75M. Those 2 = 1/6 of 50k tuition payments. budget.wisc.edu/budget-in-br...
When Jeff Flake was in the Senate he was one of the 2 or 3 most conservative members. This is quite something.
Hereâs what I firmly believe about climate catastrophes: the answer is never to place blame or to shrug and say certain people need to âjust move.â There is no place that isnât going to need help in some kind of a catastrophe.
Reminder of this article: "I argue that the project is not a model for future industryâacademy collaborations. The collaboration resulted in independent research, but it was independence by permission from Meta." www.science.org/doi/10.1126/...