Harvard’s Institute of Politics released their Spring 2016 Youth Survey yesterday, and it has some eye-popping numbers indeed. In the presidential contest, testing Hillary Clinton versus Donald Trump, Clinton has a massive 61-25 lead with likely voters age 18-29. This is even more than Obama’s lead over McCain of 53-32 among likely voters age 18-24 in Harvard’s Spring 2008 Youth Survey.
The chart above shows the demographic breakdown in each of the two years, showing likely voters for 2008 and all voters for 2016. Clearly, we’re not exactly comparing apples to apples here; maybe not quite apples to oranges, but more like apples to crabapples. (There’s more details about the differences below the fold.)
First off, we see Young Republicans in disarray. (Young Democrats, on the other hand, are a little more united than in spring 2008.) The net margin for Clinton among Republicans is only D -44 with Trump as the nominee, far less united than the D -65 in the 2008 numbers or the D -82 for a generic Democrat/generic Republican question included in 2016 (not shown). Ordinarily, I would caution that we should see Republicans rally around their nominee as the general election approaches … but, this is 2016 and Trump is Trump.
One way in which Trump is notoriously Trump is the frequently offensive ways he refers to women and minorities. Perhaps in relation to this, the largest change we see from 2008 to 2016 is among Hispanic youth, with the Democratic margin moving from D +29 to D +62. We also see a large change among young women, moving from D +24 to D +42.
Details on the fine print, and why we should expect the actual fall numbers to be even more favorable to the Democrat than the poll follows.
The fine print
First, the fine print for the chart above. Note that the 2008 numbers are likely voters age 18-24, and the poll included Nader as an option (4 percent). The 2016 numbers, on the other hand, are based on all voters age 18-29.
For the 2016 numbers, switching from all voters to likely voters changes Clinton’s margin from +30 to +36. Among all 2016 voters, Clinton’s margin is 29 for those age 18-24 and 31 for those age 25-29 (who, of course, were 17-21 in 2008).
What does this mean? It means that the 2016 numbers shown in the chart would probably be a little more democratic than shown if they only showed likely voters.
The Spring Survey track record
These numbers are dramatic enough, but when you look at Harvard’s Spring Youth Poll numbers from 2008 and 2012 (shown in dark blue below), you can see in both cases the poll seriously underestimated the final exit poll numbers in the Fall (shown in green).
Should the same thing happen again, we could have a Democratic margin among young voters of D +40 or more come November.
The graph above also includes numbers for 2000 and 2004. However, in these years, unfortunately, only college students were interviewed. The year 2000 is also complicated by a lack of Nader in the polling question.
In 2008 and 2012, college students were broken out separately from other young voters, and in each case they were more Democratic than young voters as a whole. Despite this, the Harvard Youth Survey still showed college voters in the Spring with a lower Democratic margin than the eventual numbers for all youth in the fall for all years except 2000.
