- 2012 precinct returns were downloaded from here: http://statewidedatabase.org/index.html
- 2016 precinct returns were gathered form here: https://github.com/datadesk/california-2016-election-precinct-maps
- Latino CVAP at the precinct level was downloaded from here: http://statewidedatabase.org/index.html
We use precinct level vote returns to estimate the Latino vote share for each candidate. These are the best data currently available to interrogate the exit polls and how Latinos actually voted in 2016. We gathered 2016 returns from 9,793 precincts across 16 counties in California. To this data we added 2010-2014 Census Bureau estimates of the number of adult Latino citizens in each voting precinct. Next, we merged the 2012 presidential vote returns so we can make comparisons to the 2016 election results. The scatterplots show weighted loess regression lines where the weights are the proportion of voters in the precincts over the total number of voters in the county. We then used ecological inference to determine how Latinos votes. This method is further explained here: http://gking.harvard.edu/eiR
These are the counties in the analysis: Alameda, Butte, El Dorado, Los Angeles, Marin, Mono, Nevada, San Bernardino, San Diego, San Francisco, San Luis Obispo, San Mateo, Santa Barbara, Santa Clara, Santa Cruz, and Sierra.
If you have any questions, please contact bwa@ucla.edu