by Patrick Appel
Last week, John Sides claimed that the youngest millennial voters in 2012 skewed Republican. Daniel Berman casts doubt on this analysis:
When one recognizes that 18 year olds were only 55% white in 2012, that implies either a large shift among younger minorities, or a percentage in excess of 75%-80% among whites.
A more plausible outcome is that the Exit Poll collection method is fundamentally flawed when collecting this data, as its owns creators have claimed, noting their own (ANES) estimate that Obama won 71% of 18-20 year olds. Polls are not collected at all polling stations, but rather at a selection. While efforts are made both to gain representative samples of certain groups, and to weight them accurately, the numbers for groups that were not targeted for representative samples, say 19 year olds, are likely to be off, with margin of errors meaningless.
The Dish 