Table 7.

Four groups of counties with varying interactions between mobility and infection.

Effect	Group of counties	Explanation
$\mathit{inf}\stackrel{+ve}{\leftarrow }mob$ (low, low) (“pink” in Fig. 9)	Proactively restrictive (cautious, took actions before surge in cases)	These counties had a lower infection as well as lower mobility indicating a proactive restriction on mobility as a precaution. They belonged to mostly East and West coast regions (e.g., northern portion of California, Oregon, Washington, Vermont, Maine, New Mexico) and are mostly metropolitan counties (54% of them).
$\mathit{inf}\stackrel{+ve}{\leftarrow }mob$ (high, high) (“green” in Fig. 9)	Loose enforcement (less reactive to rising cases)	These counties experienced a higher number of infections and at the same time, a higher degree of mobility, suggesting that they might be behind activating mobility restrictions. These were mostly rural counties spanning mostly in the Southeast region. Notably, counties from Georgia, Alabama, and Mississippi were included here and these states did not continue ‘stay at home’ order after April 30.
$\mathit{mob}\stackrel{-ve}{\leftarrow }inf$ (low, high) (“blue” in Fig. 9)	Comfortably relaxed (no surge, no action)	These counties had a lower infection but with higher mobility. That means, no mobility restriction was in place in those counties (or people perhaps did not comply with them as there were fewer reported cases). These were mostly rural counties located in the U.S. Midwest, West and Southwest regions.
$\mathit{mob}\stackrel{-ve}{\leftarrow }inf$ (high, low) (“orange” in Fig. 9)	Reactively restrictive (respond after surge in cases)	These coastal areas metro counties experienced higher infection and lower mobility. They got severe infection rate by the pandemic and mobility restrictions were realized.

Note: Values for degree of mobility were estimated by the model