Ref's Electoral Projection for 2016

ACCORDING TO EXIT POLLS: Race very close in Florida, Michigan, Nevada and Colorado. If you live in those states, go vote! The evening vote will make the difference.

Previous Projections: 11/4/16 | 11/5/16 | 11/6/16 |Today's projection resulted from analysis of 142 polls in eleven battleground states from the last two weeks.

I assume Trump will win Maine-2 and all five Nebraska electoral votes

*Why I see a strong chance for a Trump win in Pennsylvania or Michigan

Changes since yesterday:

1. New Hampshire to Clinton: Clinton picked up 1.9 points in the polls only average, which put her ahead of Trump.

Changes since the second projection on 11/5/16:

1. Nevada to Clinton: Clinton actually went down in the polls, but due to the turnout of Democrats in the early vote, the sample bias actually now goes in Clinton's direction by two points. For more on this state, read What is going on in Nevada?

Changes since the first projection on 11/4/16

1. Colorado to Clinton: Clinton picked up 0.5 points in the polls only average, which put her just ahead of Trump.

2. Michigan to Clinton: Clinton also picked up Michigan. She picked up 0.8 points in the poll only average which made the difference.

3. Pennsylvania to Trump: Trump picked up 0.4 points in the poll average, which put him just ahead of Hillary. If public transport workers remain on strike in Philly, Trump will get a bigger bump out of the sample bias. That is not yet incorporated.

Polls Only - Clinton 278, Trump 245

Sample Bias - Clinton 278, Trump 260

Sample Bias + Shy Trump Voter Bias - Clinton 278, Trump 260

REF'S PROJECTION: Sample Bias + Shy Trump Voter Bias + Late breaking voters (Incumbency rule) - Clinton 258, Trump 280

Incumbency rule - normally if incumbents are not above 50%, they are in trouble. I am using 47% as a guide. Where she is below that, she is in danger. This is very conservative assumption that advantages Hillary. She probably should be above 47 to be safe, but I am trying to be conservative.

EVIDENCE for Sample Bias and Shy Trump Voter Bias in Polling

My Twitter account is just getting started. Follow me here!

Florida	Clinton 45.6, Trump 45.7			Trump wins by 1.4% Prediction: Trump 49.2, Clinton 47.8, Other 3.0
Google Consumer Surveys	11/1-11/7	1350LV	Clinton 36.5, Trump 41.3	Polls Only - Trump +0.1 Sample bias - Trump +0.4 Due to high Latino turnout in Florida, but Trump narrowed the Democrat firewall to only 39,000, 130,000 less than 2012. Shy Voter Bias - Trump +0.9 I think it it safe to assume a shy voter bias of 0.5 points. This is a conservative assumption. Late breaking - Trump +1.4 The Comey announcement reduces the late breaking advantage, but it still exists because Hillary is almost two points below 47, where a winning incumbent should be in this race. Trump should yield more than half the undecided. Assuming Johnson and Stein get 3%, that leaves 5.7% undecided. Trump gets 3.1, Clinton gets 2.6.
Trafalgar Group	11/6	1100LV	Clinton 49, Trump 50
Opinion Savvy	11/5-11/6	840LV	Clinton 48, Trump 48
*Quinnipiac	11/3-11/6	880LV	Clinton 46, Trump 45
*Alliance/ESA	11/2-11/6	530LV	Clinton 47, Trump 45
Reuters/Ipsos	10/31-11/6	890LV	Clinton 48, Trump 47
*SurveyMonkey	10/30-11/6	3570LV	Clinton 47, Trump 45
CCES/YouGov as featured on fivethirtyeight.com	10/4-10/6	5370LV	Clinton 44, Trump 43
UPI/CVoter	10/30-11/6	438LV	Clinton 47.3, Trump 47.4
CBS/YouGov	11/2-11/4	1188RV	Clinton 45, Trump 45
*CNN/OpRes	10/27-11/1	770LV	Clinton 49, Trump 47
Gravis	10/31	1200RV	Clinton 49, Trump 46
Remington Research	10/30	990LV	Clinton 44, Trump 48
*Emerson College	10/26-10/27	500LV	Clinton 46, Trump 45
*NYT/Sienna College	10/25-10/27	820LV	Clinton 42, Trump 46
*NBC/Marist College	10/25-10/26	780LV	Clinton 45, Trump 44
*PPP	10/25-10/26	740LV	Clinton 48, Trump 44
Dixie Strategies	10/25-10/26	700LV	Clinton 42, Trump 46
*Bloomberg	10/21-10/24	810LV	Clinton 43, Trump 45

Ohio	Clinton 43.4, Trump 45.9			Trump wins by 3.5% Prediction: Trump 49.7, Clinton 46.2, Other 4.1
*Alliance/ESA	11/2-11/6	530LV	Clinton 44, Trump 43	Polls Only - Trump +2.5 Sample bias - Trump +2.8 Shy Voter Bias - Trump +3.0 Late breaking - Trump +3.5
Reuters/Ipsos	10/31-11/6	550LV	Clinton 47, Trump 46
*SurveyMonkey	10/30-11/6	2530LV	Clinton 42, Trump 45
UPI/CVoter	10/30-11/6	340LV	Clinton 47.1, Trump 48.3
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	3460LV	Clinton 43, Trump 42
Emerson	11/4-11/5	900LV	Clinton 39, Trump 46
*Columbus Dispatch	10/27-11/5	1150LV	Clinton 48, Trump 47
CBS/YouGov	11/2-11/4	1190LV	Clinton 45, Trump 46
*Quinnipiac	10/27-11/1	590LV	Clinton 41, Trump 46
Google Consumer Surveys	10/25-10/31	980LV	Clinton 40, Trump 45
Remington Research	10/30	1187LV	Clinton 43, Trump 48
Trafalger Group	10/24-10/26	1150LV	Clinton 44, Trump 49

North Carolina	Clinton 45.2, Trump 45.2			Trump wins by 1.6% Prediction: Trump 48.4, Clinton 46.8, Other 4.8
Google Consumer Surveys	11/1-11/7	790LV	Clinton 35, Trump 41	Polls Only - TIED Sample bias - Trump +1.2 *The early vote demonstrates a weakness in black and millennial turnout for Clinton. There should be a big sample bias. Shy Voter Bias - Trump +1.5 Late breaking - Trump +1.6
*NYT/Sienna	11/4-11/6	800LV	Clinton 44, Trump 44
UPI/CVoter	10/30-11/6	360LV	Clinton 47, Trump 48
*Quinnipiac	10/27-11/1	600LV	Clinton 47, Trump 44
Reuters/Ipsos	10/31-11/6	370LV	Clinton 46, Trump 47
SurveyMonkey	10/30-11/6	2870LV	Clinton 48, Trump 41
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	2920LV	Clinton 44, Trump 41
Gravis	11/1-11/4	1250RV	Clinton 46, Trump 45
Remington Research	11/1-11/2	2600LV	Clinton 45, Trump 48
*PPP	10/31-11/1	1170LV	Clinton 49, Trump 47
Trafalgar Group	10/27-11/1	1150LV	Clinton 44, Trump 49
WRAL/Survey USA	10/28-10/31	660LV	Clinton 44, Trump 51
*CBS/YouGov	10/26-10/28	990LV	Clinton 48, Trump 45
*Emerson College	10/26-10/27	650LV	Clinton 48, Trump 45
*Elon Univ	10/23-10/27	710LV	Clinton 42, Trump 41

Wisconsin	Clinton 46.7, Trump 41.6			Clinton wins by 2.1% Prediction: Clinton 49.1, Trump 47.0, Other 3.9
*SurveyMonkey	10/30-11/6	1940LV	Clinton 44, Trump 43	Polls Only - Clinton +5.1 Sample bias - Clinton +3.3 Shy Voter Bias - Clinton +2.5 Late breaking - Clinton +2.1
Reuters/Ipsos	10/24-11/6	830LV	Clinton 46, Trump 40
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	1790:V	Clinton 42, Trump 38
Remington Research	11/1-11/2	2720LV	Clinton 49, Trump 41
Clarity Campaign Labs	11/1-11/2	1130LV	Clinton 47, Trump 43
PPP	10/31-11/1	890LV	Clinton 48, Trump 41
*Marquette University	10/26-10/31	1260LV	Clinton 46, Trump 40
UPI/CVoter	10/23-10/29	370LV	Clinton 50,.Trump 46
*Emerson College	10/27-10/28	400LV	Clinton 48, Trump 42

Iowa	Clinton 41.8, Trump 44.3			Trump wins by 3.3% Prediction: Trump 50.3, Clinton 47.0, Other 2.7
*SurveyMonkey	10/30-11/6	1600LV	Clinton 37, Trump 47	Polls Only - Trump +2.5 Sample bias - Trump +2.8 Shy Voter bias - Trump +3.1 Late breaking - Trump +3.3
Reuters/Ipsos	10/17-11/6	416LV	Clinton 44, Trump 44
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	970LV	Clinton 40, Trump 37
*Des Moines Register	11/1-11/4	800LV	Clinton 39, Trump 46
Emerson College	11/1-11/3	700LV	Clinton 41, Trump 44
*The Gazette	11/1-11/2	1080LV	Clinton 41, Trump 44
UPI/CVoter	10/23-10/29	330LV	Clinton 48, Trump 48
*Quinnipiac	10/20-10/26	790LV	Clinton 44, Trump 44

Nevada	Clinton 44.9, Trump 44.8 *Read my article, What is going on in Nevada?			Clinton wins by 1.0 Prediction: Clinton 49.0, Trump 48.0, Other 3.0
Gravis	11/3-11/6	1160RV	Clinton 45, Trump 43	Polls Only - Clinton +0.1 Sample bias - Clinton +1.1 **High Early Latino turnout has offset this bias to -2.0 Shy Voter Bias - Clinton +1.5 Late breaking - Clinton +1.0
*Alliance/ESA	11/2-11/6	530LV	Clinton 44, Trump 44
UPI/CVoter	10/30-11/6	323LV	Clinton 49, Trump 46
*Survey Monkey	10/30-11/6	1120LV	Clinton 44, Trump 43
Reuters/Ipsos	10/17-11/6	340LV	Clinton 45, Trump 42
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	660LV	Clinton 43, Trump 41
Emerson	11/4-11/5	600LV	Clinton 47, Trump 46
Trafalgar Group	11/1/11/4	1100LV	Clinton 45, Trump 50
PPP	10/31-11/1	690LV	Clinton 48, Trump 45
LasVegasNow.com/JMC .	10/28-11/1	600LV	Clinton 45, Trump 45
CNN/OpRes	10/27-11/1	790LV	Clinton 43, Trump 49
Remington Research	10/30	790LV	Clinton 44, Trump 48
*Emerson College	10/26-10/27	550LV	Clinton 44, Trump 42
*Marist	10/20-10/24	710LV	Clinton 43, Trump 43

Arizona	Clinton 42.4, Trump 46.2			Trump wins by 6.4 Prediction: Trump 52.1, Clinton 45.7, Other 2.2
UPI/CVoter	10/30-11/6	350LV	Clinton 43, Trump 50	Polls Only - Trump +3.8 Sample bias - Trump +5.3 Shy Voter bias - Trump +5.7 Late breaking - Trump +6.4
*SurveyMonkey	10/30-11/6	2320LV	Clinton 45, Trump 43
Reuters/Ipsos	10/24-11/6	590LV	Clinton 42, Trump 47
CCES/YouGov as featured on fivethirtyeight.com	11/4-11/6	1570LV	Clinton 38, Trump 46
Data Orbital	11/1-11/2	550LV	Clinton 39, Trump 47
NBC/Marist	10/30-11/1	720LV	Clinton 40, Trump 45
CNN/OpRes	10/27-11/1	770LV	Clinton 44, Trump 49
*Saguaro Strategoes	10/29-10/31	2230LV	Clinton 45, Trump 44
Emerson College	10/28-10/31	700LV	Clinton 43, Trump 47
*CBS/YouGov	10/26-10/28	990LV	Clinton 42, Trump 44
*Monmouth	10/21-10/24	400LV	Clinton 45, Trump 46

Colorado	Clinton 43.6, Trump 40.8			Clinton wins by 1.5% Prediction: Clinton 48.2, Trump 46.7, Other 5.1
*SurveyMonkey	10/30-11/6	2410LV	Clinton 43, Trump 39	Polls Only - Clinton +2.8 Sample bias - Clinton +2.2 Shy Voter bias - Clinton +1.8 Late breaking - Clinton +1.5
Reuters/Ipsos	10/24-11/6	540LV	Clinton 48, Trump 42
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	1560LV	Clinton 42, Trump 37
*PPP	11/3-11/4	700LV	Clinton 48, Trump 43
Keating Research	11-2-11/3	600LV	Clinton 43, Trump 38
Trafalgar Group	10/31-11/3	1150LV	Clinton 45, Trump 44
Gravis	11/1-11/2	1130RV	Clinton 40, Trump 40
*Magellan Strategies	11/1-11/2	500LV	Clinton 44, Trump 43
*Univ of Denver	10/29-10/31	500LV	Clinton 39, Trump 39
*Emerson College	10/28-10/31	750LV	Clinton 44, Trump 41
Remington Research	10/30	950LV	Clinton 45, Trump 44
*CBS/YouGov	10/26-10/28	1000LV	Clinton 42, Trump 39

Pennsylvania	Clinton 45.9, Trump 43.6 *I could see Pennsylvania going to Clinton, just like I could see Michigan going to Trump. The CCES poll and Google polls were significant in my decisions.			Trump wins by 0.2% Prediction: Trump 48.1, Clinton 47.9, Johnson 2.8, Stein 1.2
Google Consumer Surveys	11/1-11/7	1177LV	Clinton 37.7, Trump 37.3	Polls Only - Clinton +2.3 Sample bias - Clinton +1.3 Shy Voter bias - Clinton +0.3 Late breaking - Trump +0.2 *Clinton should be above 45.7 here with all the effort and money they have put into the state. That plus diminished African- American and millennial turnout and few Latinos to offset it, I think Trump wins it.
Alliance/ESA	11/2-11/6	530LV	Cllinton 44, Trump 46
Reuters/Ipsos	10/31-11/6	500.LV	Clinton 48, Trump 45
UPI/CVoter	10/30-11/6	390LV	Clinton 47.8, Trump 47.1
*SurveyMonkey	10/30-11/6	2690LV	Clinton 47, Trump 42
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	3700LV	Clinton 43, Trump 41
*CBS/YouGov	10/4-11/6	930LV	Clinton 45, Trump 43
Trafalgar Group	11/3-11/5	1300LV	Clinton 47, Trump 48
Clarity Campaign Labs	11/1-11/4	1030LV	Clinton 47, Trump 43
*Muhlenberg College	10/30-11/4	410LV	Clinton 44, Trump 40
*Harper Polling	11/2-11/3	500LV	Clinton 46, Trump 46
Gravis	11/1-11/2	1020RV	Clinton 47, Trump 45
*PPP	10/31-11/1	1050LV	Clinton 48, Trump 44
*Susquehanna Polling	10/31-11/1	680LV	Clinton 45, Trump 43
*Monmouth	10/29-11/1	400LV	Clinton 48, Trump 44
*CNN/OpRes	10/27-11/1	800LV	Clinton 48, Trump 44
*Quinnipiac	10/27-11/1	610LV	Clinton 48, Trump 43
Remington Research	10/30	1250LV	Clinton 45, Trump 43

New Hampshire	Clinton 44.4, Trump 41.5 Google only polled 109 likely voters, which is too few to produce a reliable poll.			Clinton wins by 0.8% Prediction: Clinton 48.6, Trump 47.4, Other 4.0
*Univ. of New Hampshire	11/3-11/6	690LV	Clinton 49, Trump 37	Polls Only - Clinton +2.9 Sample bias - Clinton +1.9 Shy Voter bias - Clinton +0.9 Late breaking - Clinton +1.1 I actually think in New Hampshire Clinton will have more late deciding voters breaking her way.
*Alliance/ESA	11/2-11/6	530LV	Clinton 46, Trump 40
UPI/CVoter	10/30-11/6	310LV	Clinton 49, Trump 43
SurveyMonkey	10/30-11/6	672LV	Clinton 49, Trump 38
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	430LV	Clinton 43, Trump 34
Emerson	11/4-11/5	1000LV	Clinton 45, Trump 44
Gravis Marketing	11/1-11/2	1000RV	Clinton 41, Trump 43
*Boston Globe/Suffolk	10/31-11/2	500LV	Clinton 42, Trump 42
ARG	10/31-11/2	600LV	Clinton 43, Trump 48
*UMASS Lowell	10/28-11/2	700LV	Clinton 44, Trump 44
*WBUR/MassInc	10/29-11/1`	500LV	Clinton 39, Trump 40
Inside Sources	10/26-10/28	410LV	Clinton 43, Trump 45

Michigan	Clinton 43.7, Trump 41.2 *I could see Michigan going to Trump, just like I could see Pennsylvania going to Clinton. The CCES poll and Google polls were significant in my decisions.			Clinton wins by 0.4% Prediction: Clinton 48.2, Trump 47.8, Other 4.0
Google Consumer Surveys	11/1-11/7	821LV	Clinton 35, Trump 34	Polls Only - Clinton +2.5 Sample bias - Clinton +1.8 Shy Voter bias - Clinton +1.1 Late breaking - Clinton +0.4
Trafalgar Group	11/6	1200LV	Clinton 47, Trump 49
Reuters/Ipsos	10/31-11/6	340LV	Clinton 46, Trump 45
UPI/CVoter	10/30-11/6	330LV	Clinton 49, Trump 45
CCES/YouGov as featured on fivethirtyeight.com	10/4-11/6	3010LV	Clinton 42, Trump 36
*PPP	11/3-11/4	960LV	Clinton 46, Trump 41
*SurveyMonkey	10/29-11/4	2750LV	Clinton 44, Trump 42
*EPIC-MRA	11/1-11/3	600LV	Clinton 42, Trump 38
*Mitchell Research	11/2	1150LV	Clinton 47, Trump 44
Google Consumer Surveys	10/25-10/31	680LV	Clinton 38, Trump 35
*Strategic National	10/25-10/31	500LV	Clinton 45, Trump 44

*The point of poll averages should be to determine the closest approximation to the actual race going on. Other poll average websites include polls from many weeks, or even months before. That can be okay, but only if the race hasn't shifted. Once a race shifts, only the most recent polls should be averaged.