In the month before the March 25th New York registration deadline, 120,311 voters registered in the Democratic party. The chart above shows daily totals, which peaked at over 20,000 on March 26th. That’s one day past he deadline, but since mail-in registrations only had to be postmarked by March 25th, perhaps many or most of those voters are eligible to vote in the April 19th primary.
Data for the chart comes from a voter list from the NY Secretary of State, containing 16 million voter records.
Of the total, 69% of the newly-registered Democrats were under the age of thirty. Females were 53% of the total.
Here’s the surge in a little broader context, from mid-2015:
The last contested presidential primary was in 2008 so it’s interesting to look at the latest data in that context. The registrations in the chart below are grouped by week (7-day periods):
Note the 2008 and 2012 peaks above are for the November general election, not the primary as in 2016. So in 2012 there was little interest in the primary and no early-year surge. In 2008, a hotly-contested primary there was a pre-primary surge, but not as much as 2016.
Given the strong support of Sanders in the under-30 demographic, it’s plausible that the large number of new Democratic registrations is due in large part to Sanders backers.
If Sanders were to net, say, 60,000 votes from the 120,311 new registrations (i.e., Sanders 90156 to 30155 or 75%), and if turnout is the same as the 1.8 million votes in the 2008 primary, that would account for 3.3% extra margin in the total vote. This is perhaps a high estimate as maybe 75% is a bit high even for young voters in this state, and not all the newly registered voters will in fact turn out. One could expect though that people in a rush to register by the deadline are highly motivated to vote.
If the 20,000 registrations marked with a registration date of March 26th are in fact not valid to vote in the primary, that represents a large number of potentially thwarted voters.
UPDATE: On the question of expected turnout for the newly-registered voters I looked at 2008 for comparison. Of the newly-registered Democrats under 30 in that year for the month before the registration deadline, only 24% actually voted in the primary. This compared to 27% of those over 30.
So while the turnout of the newly-registered voter was above average as would be expected, it was still pretty low, and their effect correspondingly less.