5 Comments
User's avatar
Paul Snyder's avatar

Thanks for this, really…

My experience in Stats is from the Health Physics and PH side.

You have a great ability to communicate the logic and legitimacy of “real” data vs biased propaganda.

Sincere gratitude for your efforts.

All the best.

Expand full comment
Haley Dohrmann's avatar

Thank you so much. I really, really appreciate you taking the time to comment. This meant the world to me.

Expand full comment
Paul Snyder's avatar

Even better:

My daughter is in the last year of her MPH for Public Policy and finds your posts a great resource.

We owe you.

Expand full comment
David's avatar

Are people really dumb enough to volunteer information that they get Medicare and just watch TV? Aren't these people supposed to be villainous liars and deabeats?

Expand full comment
Usurper's avatar

Notably, this entire article isn't refuting the study or its results. It's merely acknowledging that, for social science "research", desired outcomes can be engineered by crafting the right inputs and some data manipulation.

Unfortunately, this isn't just true for this particular study, which gives results that Democrats don't like. It's also true for the countless research papers out there in the social sciences which suggest that some completely counter-intuitive outcome is "statistically proven". E.g., the claim that "less policing leads to less crime", that sort of thing, often with R values below .70 (or even .50!). Given that academics are overwhelmingly progressive liberals (and often the research is biased in that direction), casting a critical eye on the rigorousness of these kinds of studies will definitely help conservatives in the long run. These fields are rife with bad research.

Expand full comment