While I understand that my second point might sound like a cop-out, just consider how the survey findings may have been different, if the respondents had been asked about issues more relevant to social justice narratives, e.g. the prevalence of deadly police shootings of unarmed people of color.

Actual annual figure in recent years: roughly 10 depending on dataset and year.

Median estimates among progressive respondents in several surveys: hundreds or even thousands.

Another example:

Surveys show large fractions of progressive populations believe global poverty has worsened, when the long-term trend has been a substantial decline