Bolts magazine has an article by one of the researchers published in American Political Science Review who found that police interactions directly correlate with lower voter turnout, no matter who or why:
Our study compared the voter turnout of Hillsborough motorists who were stopped by police shortly before and after each election. Drawing on information about each person’s turnout in past cycles, we found that these stops reduced the likelihood that a stopped individual turned out to vote by 1.8 percentage points on average. The effect held when accounting for characteristics like race, gender, party affiliation, past turnout, and prior traffic stops to improve our comparisons. The discouraging effect of stops was slightly higher in 2014 and 2018.
Why would traffic stops make people less likely to show up to the polls? Past research has already established that the most disruptive forms of criminal legal contact, like arrest and incarceration, discourage people from voting. Our study shows that low-level police contact matters in the same way. If a traffic stop makes a motorist fear that the government will harm them, it can prompt a withdrawal from civic life that political scientists call “strategic retreat.” Motorists might worry that a routine traffic stop could escalate into police violence, a more common outcome for Black people in particular. Beyond justified fears of violent victimization, voters might also bristle at the perception of being targeted to raise revenue through excessive ticketing. Accordingly, if incarceration ‘teaches’ would-be voters that their government is an alienating and harmful force in their lives, traffic stops could catalyze a similar form of ‘learning.’
You can read the study here, in American Political Science Review.