10 June 2011

Voter Registration - ALL 50 STATES plus DC

Because of federalism, there is no standard system in the USA for recording statistics pertaining to elections (voter registration, voter turnout, election results) according to any criteria. This means that there is no uniform method for attaining information about elections from the states and compiling it together. Therefore, because not all states have voter registration by party affiliation, it is impossible to come up with complete partisan statistics for the Union. For instance, ND has no voter registration at all! We can, however, come up with a composite of each state based on its voter registration trends and past electoral history.

Some important things to look at, if available:

a.) Voter registration by: party, age, gender or race. I believe that with the major hispanic factor, VR by race will play a role in the SW of the Union; the incorrect prognostication of the the hispanic vote probably accounted for some very off-polling in the SW in the 2010 mid-term elections. So, I suspect that record keeping including race as a sub-factor may come more and more into play in the near future.

b.) Changes in voter registration in a state. PA is an outstanding example of a state that does very  thorough VR/VT statistics work, including VR/VT change.

c.) The INTERVAL of VR reporting (bi-yearly, once yearly, semi-annually, quarterly, monthly, weekly). As a great example, the state of North Carolina reports the most current VR statistics by party (raw numbers) in a balkan at the top of  its main SOS page. This is an example I wish the other 49 states would follow.

d.) The most current election results and an archive of past election results by:
year, office, raw vote votals, percentages (to the 1s, 10th or 100th of a percentage point), margins.
Election results statewide, by CD, by county, or by precinct, sometimes by city/municipality, sometimes by state legislative district. Often the end results are still called „unofficial“ in spite of the fact that they are indeed the end totals.

Some states provide a huge amount of information and it is easy to get to. Other states provide little information and some make it hard to get to the information. IL is a good example of a state where information is hard to come by. My point is that this mish-mash is extremely inconsistent.

I have put together a handy table of links to sites for each state and DC for current and past information on VR (Voter Registration), VT (Voter Turnout) and election results, wherever they may appear. These are the current links, most of which go to the respective SOS websites for the states, but not always. In the „Notes“ column I have made some commentary that you may find very helpful. I also indicate in which form the user can download the desired information.

Closer to election day 2012 I will be updating these links. Many states provide a special link to their presidential election night results, which will then later be transferred out to another link.

Why the need for this information?

Well, massive changes in voter registration in a state can indeed indicate a possible partisan tilt in the coming General Election. And I will remind, it is always better to compare VR to a previous similar cycle, i.e. it is better to compare VR in a presidential cycle with other presidential cycles, i.e. 2012 to 2008 and 2004, not to 2010 and 2006, for presidential cycles, which are not base elections, bring out a larger and more diversified electorate than mid-term elections, which are indeed base elections. In other words, don't mix apples and deer meat.

That being said, voter registration is not automatically predicative of the tilt of a state in a presidential election: in WV, for instance, there are more registered Democrats than Republicans, but this state has voted reliably GOP in national elections since 2000, and with increasing margins.

However, there are some states where the disparity in VR is so great that is it practically a given that that state will go for the party that has the lead in registration. The two best opposing examples of this are RI for the Democrats and WY for the Republicans.

RI currently has an enrollment of:

40.97% Democratic / 10.37% Republican / 48.61% undeclared. That is a +30.60% advantage for the Democratic party in Rhode Island. And the vast majority of „undeclareds“ vote democratic (similar story in MA).

Conversely, in WY, the enrollment is:

67.50% Republican / 22.13% Democratic / 0.38% Libertarian / 9.99% undeclared, +45.37% registration advantage for the Republican Party in Wyoming. These respective disparities are so large that it is practically impossible for the minority party to have a real shot of winning that  state in a national election.

It pretty much takes a win of over 59% in the National Popular Vote nationally for the opposing party to win these types of states and practically run the entire electoral board.FDR won 46 of 48 states in 1936 with 60.80%, Nixon won a 49 of 51 "state" (inc. DC) sweep with 60.67% in 1972 and Reagan won a numerically identical sweep with 58.77% in 1984. The only exception to this is NPV record holder LBJ, who won with 61.05% of the PV but lost 5 states in the south plus Goldwater's AZ.


Sometimes outside factors influence things more than we wish for. In 2007, before it became apparent that Barack Obama would be a strong contender for the Democratic nomination for President, the enrollment in PA looked like this in May 2007:

47.76% Democratic / 40.20% Republican / 12.04% „other“, a +7.56% registration edge for the Democrats.

Then Rush Limbaugh decided to encourage GOP voters to change their affiliation to vote in the DEM primary in April of 2008 (PA has closed primaries) with his so-called „Operation Chaos“, which resulted in a massive shift in registration in the Keystone State one year later, in May 2008:

50.43% Democratic / 38.26% Republican / 4.50% unafilliated / 6.81% „other“, a +12.17% registration edge for the Democrats and a +4.61% registration margin gain over 2007.

Limbaugh thought that those Republicans would switch back once the primary was over. Some did, but many did not and many more new voters were added to the rolls in the fall of 2008. Here was the last VR report from PA at the time of the election in November 2008:

51.62% Democratic / 37.04% Republican / 11.34% „other“, a +14.58% registration edge for the Democrats and a +7.02% registration margin gain over 2007. That translated into 1.2 million more registered Democrats in PA than registered Republicans, an edge the Democratic party had not seen in years. And since PA has traditionally very high turnouts, those 1.2 million more registered Democrats played a decisive role in an Obama landslide in this state. Of this there can be no doubt.

What was the partisan shift in PA in 2008 over 2004?

Obama 2008: +10.31%
Kerry 2004: +2.50%
Partisan shift, PA, 2008 over 2004: DEM +7.81%

To compare, this is how the enrollment looked at election time in 2004 (where Kerry won the state by a lean +2.50%):

47.64% Democratic / 40.70% Republican / 0.41% Libertarian / 0.19% Green/ 11.06% „other“, a +6.94% registration edge for the Democrats, slightly less than it's edge at this time in 2007 (May). And yet, with this slight edge, Kerry won PA in 2004. This is exactly why, having seen the monstrous shift from +500,000 DEM voters in 2004 to +1,200,000 DEM voters in 2008, I knew that Obama would win the state in a landslide, and indeed, he did.

Difference in registration edge at election time, 2008 over 2004:

2008: +14.58%
2004: +6.94%
Registration shift: DEM +7.64%

Anyone see a correspondence here? Of course! The large shift in VR to the Democrats' advantage played out percentually in the GE, almost to a tee.


From the table, we can glean some important facts:

28 states provide VR statistics by party affiliation.
5 states provide VR statistics by gender.
5 states provide VR statistics by age.
4 states provide VR statistics by race. AL, GA, LA, SC.
Only 3 states provide information according to 3 of these four categories: AK, GA, SC
Only one state provides information for all four categories: LA. And it also divides race by partisan identification. LA is the only state in the Union to do this.

14 States provide no information in these four categories: AR, IN, MI, MN, ND, (OH) TN, TX, UT, VT, VA, WA, WI, WY. Ohio is in parenthesis as it provides the partisan data individually per name on a 200 Megabyte TXT file that anyone can download, but presents no partisan grand totals. Ohio is the only state in the Union to provide the names, addresses and partisan affiliation of all of it's voters for download on the internet, a phenomenon that disturbs me greatly and would be a good discussion topic. These 14 (15) states do provide raw numbers (totals) of registered voters, nothing more.

I am still researching six states for more information, for I am sure I saw hard data on RV from these states in the past, but see no such data at their SOS websites right now: HA, ID, IL, MS, MO, MT

Some states have very user friendly websites. Other states have very difficult websites. I personally think that IL and NH have the two worst websites for elections on the web. I also think that the user interface and graphic set up at the websites for AK, LA, WA are outstanding. NC has an upper balkan at its SOS website that always has the latest RV raw numbers, which I think it absolutely right. CO has a very user friendly set up for its RV downloads. Those are just some examples.

Here is the table. The states are in bold and color coded based on their performance in the GE 2008:

State VR/VT link Elections Past Elections VR: party VR: age VR: race Interval Notes
VR/VT current 2000-2008 N N Y year VR format: xls, pdf, clickable download link. Data from 1980-present. Also a special link for female elected officials. Election format: html, previous results as pdf.
VR Current past Y Y, Gender N month Format: htm: historical VR data published yearly. Election results in pdf or htm. Results tabulated to 100th of a percentage point. Votes are called „cards“ in Alaska.
Past (to 2010) Y N N quarter Format: pdf. Registration by party and county, CD or LD. Links in english and spanish.
NVRA current Past (1976-) N N N year Formats: pdf, xls. Only lists new applications according to public help agency.
VR, VP current Past (1990-) Y N N 15, 60 and 154 days before reg close. Format: pdf. Extensive report broken into many parts. By party affiliation, including pie graph. Also party registraion by county.
VR current Past Y Gender N month Formats: all reports together as pdf. Individual reports as EXCEL.
VR Current past Y N N year Format: pdf. VR by party/town and county. Also a NET CHANGE report.
VR current past Y N N Month Format: shmtl. Party affiliation by county and precinct.
VR Current Past Y N N Month Format: pdf. Party by ward and precinct.
Current past Y N N Month Format: asp. Pie graph with running numeric totals. Also past totals since 1972, with vertical bar graph.
VR Current past N Y, Gender Y month Format: pdf. Individual reports by age, race, gender by county, municipality or district. Historical breakdown from 1962-present. But NO VR by party identification.
current past ?? ?? ?? ?? Still researching
Info Page current past N ?? ?? ?? The website explicitly states that ID does not record VR according to party affiliation. Format elections: htm.
Info Page Balkan
?? ?? ?? ??
Worst state elections website of all 50 states!
VR, VT Current past N N N Cycle Format: pdf
VR 2000-2011 1938-1999 Y N N Month Format: pdf. Party affiliation by county, CD or state district.
VR 1980-present 1980-present Y N N Irreg. Format: EXCEL. Party affiliation (raw numbers) by county.
VR 1973-present 1973-present Y Gender N Month Format: TXT. VR by county.
VR Current / graphical / text 1998-2011 statewide / 1998-2011 by parish Y Y, Gender Y N/A Format: html, txt, aspx. Interactive Parrish map (clickable). Election results as .xls data.
VR 1990-2011 1990-2011 Y N N yearly Formats: txt, pdf
VR Current past Y N N monthly Format: pdf for VR, website (html) for election results.
VR-Current /
2000-2012 2000-2012 Y N N Irreg. Format: pdf. Independents listed as „unenrolled“
VR-VT - Primaries
1998-2011 1998-2011 N N N yearly Format: html. Only raw totals, no sorting by party affiliation, race, age or gender.
VR , VT 2001-2011 2001-2011 N N N yearly Format: html,aspx. Only raw totals, no sorting by party affiliation, race, age or gender. Link to previous years at the bottom of the page.
N/A current researching ?? ?? ?? ??
VT Current 2002-2011 ?? ?? ?? ?? Formats: asp (elections), pdf (VT)
VT 1992-2010 1992-2010 ?? ?? ?? ?? Formats: asp, pdf, in some cases, .zip.
VR - past Current? 2000-2008 Y N N cycle Format: pdf.
VR 1998-2011 1998-2011 Y Y N month Format: aspx (VR)
VR current 1992-2008 Y N N Irreg. Format: pdf / website is not well organized.
VR 1924-2011 1924-2011 Y N N Irreg. Format: pdf. VR and election results all on same http.
VR 1990-2010 1990-2010 Y N N Quarter Format: pdf, registration by CD, JD, LRD, LSD, County Precincts available.
VR 1994-2010 1994-2010 Y N N Quarter Format: pdf, registration by county, election district, assembly district or CD available.
Election night link is usually provided 1992-2010 Y N N Weekly Format: asp. On the main page, the most current stats are presented clearly in an upper balkan (raw totals)
Info Page 2000-2010 2000-2010 N N N N/A ND is the only state that does NOT have VR.
Voter Files
VT analysis
An election night link is usually provided later. 1940-present N N N N/A Ohio provides an enormous list download (.txt) of the names/adressess of every registered voter in the state. The party affiliation is listed with the name. No official tallies are presented at the OH SOS website. This is the only state in the Union to do this this way. Lists provided by county, CD, state rep district, state sen district, and state overall. The statewide list is 200 MB in size. There is also a voter file layout sheet (.doc) that describes it all). Election returns in .pdf format.
VR 2010 2004-2008 Y N N Yearly (Jan) Format: pdf
VR, Graph, Map (2000),
1996-2010 Y N N Yearly Format: htm. Close to elections, OR also provides running VR statistics, but after an election cycle, provides a yearly total.
VR, Current Stats, VR Changes,
Most current 2000-2011 Y N N Apr, May, Nov of each year Format: xls. PA provides some of the most comprehensive VR data by party affiliation and keeps track of party registration changes, inc. Party switches, additions and deletions. PA election returns also with a clickable county map.
Election FAQ , VT in GE's Current (2010) 1976-2008 Y N N unknown Format: pdf, by CD or City/Town. Links to .pdf files on the right side of the FAQ page.
VR , Past VR Election night link 1996-2010 N Y, Gender Y unknown Formats: htm, html. Sorted by counties/precincts, CD, state senate or state house seats. No VR by party affiliation in SC.
VR , VT Election night link 1889-2010! Y N N yearly Format:shtm
VR and VT 1995-current 1995-current N N N bi-annual Format: pdf. VR nur by active /inactive, not by party affiliation, age, gender or race.
VR and VT (VAP) Info /
1992-current N N N Election cycle Format: shtml. VR nur by RV also compared to VAP, not by party affiliation, age, gender or race.
VR 1960-2010 1960-2010 N N N unknown Formt: html. VR nur by active /inactive, not by party affiliation, age, gender or race.
VR , past VR current past N N N Bi-yearly (even numbered) Formats: pdf or EXCEL. VR and all election results links are at the same webpage.
VR/VT (1976-) 1995-2011 1995-2011 N N N yearly Formats: html (VR/VT), Java (Website) or CSV for elections.
VR/VT current 1900-2010 N N N yearly Format: html. RV total is listed on the main elections page (right). Special VT report by age and gender.
2008, 1976-2006 2008-2010
Y N N Bi-yearly (even numbered) Format: php (Web-text). The VR pages also have links to clickable county maps to show partisan strength.
RV 2008-2010)
WI politics info
current 2000-2010 N N N 2008-2010 (new WI website) Format: mostly pdf. WI switched over to a new website with the new Republican administration, so the data is now spread over two large websites.
RV 1996-2010 1996-2010 N N N monthly Format: pdf

I would also like to point out some characteristics of the better websites. Here is a screenshot of the NC website, with this balkan at the top with the VR raw totals, 24/7. Very handy, indeed!

Here, at the LA VR website, the running VR totals, plus an interactive map.

Click on a Parish and those totals then appear. Very practical:

Similarly, at the WA website, the election totals are provided with a mouse-over map of the state:

OR takes it to another level and even provides a VR change graphic:

The AK website uses drop down menus for you to pick the election information you want:

Both CA and FL provide graphs to show VR. Here is FL:

Here is CA from its latest VR report:

Conversely, the IL website is lots of todo about nothing. Can you find the election results here?

Right, under Vote Totals, but only two years are offered, with no links to any archives. And there is no VR information from IL to be found anywhere.


I will be updating this table as needed. Should anyone have further information they wish to contribute then feel free to comment or send me an email.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.