LaVar Hosting Jeopardy July 26-30 Stan Account 👁

This appears to be another case of misleading reporting of data by the Census Bureau, not evidence of vote suppressionhttps://twitter.com/kdrum/status/921338975599644672 …

I've got a longer write-up of the issues here, which even Justice Roberts ran afoul of in his Shelby County decision http://www.electproject.org/home/voter-turnout/cps-methodology …

The shorter version: The Census Bureau counts all persons with missing data as "did not vote" in their survey. This is a terrible practice. No other pollster treats their missing data this way

In 2012, 7 of 78 (9%) WI African-American respondents had missing CPS vote data In 2016, 22 of 61 (36%) had missing data When the Census Bureau counts missing data as "did not vote" WI African-American appears to alarmingly declined when, in fact, fewer answered the survey

When I correct for non-response and over-report bias, I estimate Wisconsin African-American turnout declined 7.5 points, while nationally the decline was the same 7.5 points

Now, I wouldn't recommend using the CPS to estimate the turnout effect of voter id laws in Wisconsin, anyway (unless the effect is really big). The margin of error of the sub-group will be large with 78 respondents in 2012 and 61 in 2016

I and others have discussed these issues with the folks who compile the CPS voting and registration report. They are aware of them, but haven't taken action yet (obviously)