This interpretation of my research and the political consequences of MLK’s assassination makes two key mistakes. First, in a kind of soft bigotry of inevitable violence, it assumes Black violent protests are the only plausible response to White supremacist barbarism.
Second, the history of nonviolent direction in the 1960s shows that brutal state and vigilante repression against nonviolent protesters, such as with Bloody Sunday in Selma, injured & traumatized activists but was associated with huge spikes in concern for civil rights. 
Many communities in April 1968 did *not* have violent protests. I use rainfall in April 1968 to approximate random assignment of violent protest (↑ rain = ↓ protest). With that model, counties with post-MLK violent protests saw Democratic vote share decrease ~8% in November.
For more, see:
*Typo in “Second, the history of nonviolent direction” should be “nonviolent direct action.”
