I read the link. I have some major issues with the methodology as far as I understand it. I'll explain why.
His response variables are the number of deaths and the number of cases. He's using the response strategy as an independent binary variable to explain the number of deaths and number of cases. That doesn't make sense to me if I'm interpreting that correctly, and I'm pretty sure I am. Here's the issue:
The number of deaths is one of the primary reasons states chose their Covid-19 response strategy. The seven states that did not adopt a shelter-in-place strategy are obviously not interested in doing so because they have a low death count. So if we consider the number of deaths that occur in a state with shelter-in-place, like Michigan, against a state without shelter-in-place, like Arkansas, we would see more deaths in Michigan. In his regression model, that is an indication that shelter-in-place doesn't work, because there are more deaths in Michigan than Arkansas.
That makes no sense.
Michigan is implementing shelter-in-place because of the number of deaths while Arkansas is lax about this because of their lack of number of deaths. The same goes for cases. That doesn't show whether the strategy is working so much as it indicates prior history of how badly covid-19 has affected that state.
Furthermore, he still gets a high p-value because several states are taking proactive approaches to these strict measures despite not having the high number of cases or deaths. States like Alaska and Hawaii (not verified but implied by the words of the writer).
"As of 6 April, seven US states had not adopted shelter-in-place orders...Those seven states are Arkansas, Iowa, Nebraska, North Dakota, South Dakota, Utah and Wyoming."
It would be nice if he posted the data he's using instead of asking people to request it from him. He should post the regression model that he's generating from running these numbers as well, not just the p-values.
Population showed up as significant in his model, which leads me to believe that I'm interpreting this correctly, that his response variables are the number of deaths and cases. Which, again, doesn't make sense.
That's not how I would have ran this regression model. Looking into his background further, I noticed that he's a political scientist, not a statistician.
My $0.02.