Kevin Hassett, a White House economic adviser, has come up with a “model” that shows that there will be no more coronavirus deaths in the country by May 15, which is ten days from now.
On Tuesday, Hassett appeared on CNN to answer a few questions regarding his prediction. The interview did not go well for Hassett who appeared to struggle to answer the questions.
“Why did you do it?” CNN’s Poppy Harlow asked. “Did anyone ask you to do this and why are you creating these?”
Hassett explained that he began creating his chart daily because of the ventilator shortage. According to the economic adviser, his chart is made to track the accuracy of other scientific models.
“If you have really wiggly data and you wonder how a curve that someone else has drawn fits the data then you can smooth through the data so your eyeball can compare it,” Hassett said. “This is a very standard statisticians toolkit kind of thing to visualize how models are performing relative to the data.”
“The Washington Post is reporting Jared Kushner really grasped on to this,” Harlow observed. “You said it didn’t go to the president or the task force but it was used to push forward a more rapid opening of the economy.”
“I never did that,” Hassett argued. “I was in the Oval with Dr. Fauci and Birx and always agreed with their forecasts.”
“It is true that if you don’t understand what cubic fits are and you don’t understand the purpose of data visualization,” he added, “if you’re not a statistical guy that maybe the people who wrote about this, maybe you can’t misunderstand what’s going on.”
Take a look at the interview in the video clip below:
Hassett’s so-called “cubic model” was panned by experts.
I would bet $538 that the White House's "cubic model" is literally just an MS-EXCEL trendline with a third-degree (cubic) polynomial. https://t.co/TvrHm25dB6
— Nate Silver (@NateSilver538) May 5, 2020
Like this took me 45 seconds and it seems to match the description of the "cubic model" quite well. pic.twitter.com/s91O6CTkw1
— Nate Silver (@NateSilver538) May 5, 2020
Hassett on @cnn right now trying to explain that “cubic fits” are part of standard methods for forecast evaluation, and people who don’t know statistics might be confused.
As someone who has studied forecast evaluation for ~yrs – No.— Roger Pielke Jr. (@RogerPielkeJr) May 5, 2020
For the nontechnical:
Curve fitting (what Hassett did) is *not* building a model. All he's doing is getting a curve to go through some data points. That doesn't predict the future in any dynamic process. It doesn't work in the stock market, and definitely not for a pandemic. https://t.co/crg8uTxucp
— Sam Wang (@SamWangPhD) May 5, 2020
