Donald Trump has run an unconventional campaign. And part of that departure from convention has involved the use of controversial rhetoric that has been widely thought to energize unregistered minority and/or Democratic voters, thus making it more difficult for him to carry critical swing states like Florida and get elected.

Or has he ? To best evaluate this assumption, JMC Analytics and Polling had performed a voter registration analysis on June 3 and determined (using available voter registration data as of May 31) that there was not an appreciable spike in Democratic voter registration since January 1, and that Republicans were signing up more new registrants in Florida, North Carolina, and Pennsylvania than the Democrats were (many states do NOT disclose the race of the voter, so party is the closest available demographic variable that can be used).

Now that several months have passed, and JMC has updated (where available) partisan voter registration data up to the end of August (and in some instances, into September), has anything changed ? This is a valid question to ask, because with Election Day looming closer (in fact, in North Carolina, nearly 1,000 absentee ballots have already been tallied, and in person early voting starts in several states next Friday), voter registration drives conducted by either party are garnering more interest than they would have back in May.

From our analysis of partisan voter registration data, what we see is generally almost a carbon copy of where we were at the end of May, and right now, Florida, Iowa, North Carolina, and Pennsylvania have seen more Republicans than Democrats register since January 1. Below are more specific summary statistics:

(1) There are 10 swing states worth 112 Electoral votes which (with the exception of North Carolina) were all carried by Barack Obama in 2012. In May, 610K more Democrats, 617K more Republicans, and 121K LESS Independents were registered to vote compared to January 1. Three months later, Democrats gained a slight advantage: they out registered Republicans 850 to 833K (with 302K more Independents). In other words, the shift towards the Democrats was truly miniscule –  22K voters out of a 44 million voter base/denominator, or 0.05%;

(2a) There are six non-swing states worth 108 Electoral votes which (with the exception of Alaska and Louisiana) were all carried by Barack Obama in 2012. In May, 683K more Democrats, 208K more Republicans, and 102K LESS Independents were registered to vote compared to January 1. Three months later, Democrats outregistered Republicans 831 to 238K (52K LESS Independents), or a shift towards the Democrats of 117K voters out of 38 million voters;

(2b) Looks can be deceiving with regards to the numbers for the “non swing states”, however: all of that gain both in May and in September came from California alone, which, in fact, has been a Democratic success story this year in several ways. If California is excluded from the “non swing states”, Democrats only outregistered Republicans by 5K in May, and by August, that 5K advantage (out of 20 million voters) was unchanged.

Below are the current (and May) summary numbers:

Partisan voter registration changes: January to May 2016

Partisan voter registration changes between January and September 2016

In conclusion, any assumption that Donald Trump has “spiked” Democratic voter registration is not supported by current voter registration data, and little has changed since this analysis was done three months ago.