October Weekly Focus: Gender Difference in Four Crucial States

Richard C. Eichenberg

Elizabeth Robinson

(polls through October 28^th)

Note: all percentages based on 2 party vote, dk’s excluded

(Update, 10/31:  Graphic corrected to show VA, not FL, in title)

Our previous posts have shown a very stable division of the gender vote for most of October both nationally and in the so-called swing states (based on polls through October 24^th).  In this post, we update polls through October 28 for four crucial states: CO, OH, NH, and VA. We chose these states because of their obvious importance to the electoral college, but also because in at least three of them (CO, NH, and VA), there have been some signs in late October polling that Obama’s position may have improved slightly, and it is interesting to examine whether this improvement is due to the gender division of the vote.

There have been 69 polls in these four states during October, totalling over 58,000 respondents.  The average is 5,000 to 15,000 respondents per week. Still, even at these large sample sizes, we should treat any change in percentages of 2 percent or less with caution.  Sampling error and other statistical noise does not disappear, even at large sample sizes.

The graphic below shows the weekly average gender division of the vote in the four states combined.  The pattern is very familiar; after leading in September, Obama’s share of both the male and female vote declined in October, reaching a stable level of approximately 55 percent of women and 45 percent of men (one extra word of caution: this includes only 3 polls after Oct 28^th).

Given sampling error and other noise, the best characterization is that these combined percentages have been largely unchanged since October 7^th.

(click on image to enlarge)

State-by-State Summary

The number of polls for each state are relatively small on a weekly basis (monthly totals for each state are presented in other posts), so we do not show them graphically.

Nonetheless, a summary of the weekly progression of the polls in October is as follows:

CO:  Obama’s share of women’s votes in CO have been extremely stable at 54-55 percent since October 7^th , with male vote stable at about 45-46%

NH  polls have been somewhat erratic. With that caveat, it is the only one of these four states that shows some signs of an increase in Obama’s share of the female vote: from 51-53 percent in the first half of October (which was a major drop from September) to 57-59 percent during the last two weeks of October (which approaches the 60 percent among women that Obama enjoyed in late September and the first week of October).

OH:  the gender race in Ohio has been rock solid stable at 56 percent Obama among women and 47 percent among men, with the overall Obama vote in Ohio between 51-52 percent since October 7.

VA:   There is a mixed pattern in VA, with the Obama female share fluctuating between 54 percent (Oct 27)  to as high as 58 percent (Oct 20), but the overall Obama vote in VA has been stable at 51 percent.

Summary Statistics for this Post

Number of polls and respondents by week (four state totals)

week               #polls    #respondents

Sept 30-Oct 6     6          6397

Oct 7-13          17        14980

Oct 14-20        13        11186

Oct 21-27        22        16512

Oct 28- Nov3     3          2192

Total    61        51267

Number of polls in October, by State

state     #polls   #respondents

CO       11         9876

NH       10          6228

OH       23        20106

VA       17        15057

Total    61        51267

Update in response to questions: polling organizations represented in this post

[number of polls and total respondents]

ANGUSREID         1    550
ARG                          6    3600
CNN/ORC                1    722
FOX                           2    2257
GRAVIS                   3    4111
GROVE                     2    1000
GarinHart                1    807
Gravis                        1    645
LAKE                          3    1400
NBC/WSJ/Marist   10    11407
NEWSMAXZOBGY    2    1674
PPP                               11    8624
PURP                             3    1800
QUINN/CBS                1    1548
QUINN/CBS/NYT    2    2542
SUFFOLK                     2    1100
SVYUSA                       3    2030
TIME                               1    783
UNH                                2    1318
WashPost                      1    1228
YOUGOV    3    2121

Total    61    51267

Elizabeth Robinson

Tufts University

(polls through October 24^th)

In an earlier post,  we showed that the approximately two point drop in President Obama’s national poll standing from September to October was not due to a disproportionate decline among women.  In fact, he dropped by exactly two percentage points among both genders.

The question in this post is whether the same pattern characterizes his standing in eleven key states –what some have called “swing states”.

As we shall see, the answer is mixed:  Obama’s standing has on average declined slightly more among women in the key states, but the decline is actually concentrated in just a few states.  In other states, his support has held steady among women or even slightly increased.  And perhaps most importantly, even after these declines, Obama leads among women and overall in almost every swing state.

The Situation in Five Toss-up States

The five states below are currently listed as “toss ups” on most of the professional polling sites.  As the graphic shows, one reason for this is that Obama lost slight leads in all of these states as a result of a decline in his polling numbers in October.  On average, Obama lost -1.2 percent in these states and -2.9 percent among women.

(click on image to enlarge)

As the graphic also shows, however, these averages mask considerable variety.  In North Carolina, Florida and New Hampshire, Obama declined noticeably, while the decline in Colorado and Virginia was less (and may even be indistinguishable from sampling error and other statistical noise).

In summary, the picture provided by these toss-up states is one of diversity.  In three important states, Obama seems to have declined more among women than among men.  In other states, this was not the case.  Why each state takes the pattern that it does will have to be the subject of future research.

The Situation in Six States That Lean Obama

The situation is similarly diverse among states that most professional poll sites consider safe or leaning to Obama as of October 29.  In the graphic below, we show the level of support among men and women in these states for September and October.  Several patterns are evident.  First, in two states –Michigan and Wisconsin– Obama actually increased his lead among women from September to October (or held steady), one reason that he seems to be comfortably ahead in those states.

(click on image to enlarge)

In most of the rest of these states, Obama experienced small declines  of 2-3 percent among women, but in some cases these declines were offset by an increase in support among men (PA, NV). Third (and not shown in the graph), Obama is running ahead of his margin of 2008 victory among women in five important states: FL, IA, OH, MI, and VA. Finally, given the rash of speculation in the press about whether Obama has “lost” his advantage among women in the “swing states”, it is useful to point out that this graphic and the one shown above demonstrates that Obama is ahead among women in all of the 11 eleven states that most observers consider crucial to the outcome in the electoral college –sometimes by very large amounts in important states.

 Caveats

There are several caveats to the points made above.  First, many of the changes from September to October are small, averaging -3.0% in the first graph and -2.5% in the second (and in both graphs there are outliers that have a large affect on the average). Even with a large number of respondents for a pooled sample of surveys for each state and month, some margin of error remains (on the order of 1-2%; see sample sizes below). For this reason, it is probably useful to focus only on states where the change has been larger than 2% or so: FL, NH, IA, NV, and PA).  Second, it is worth noting that in many states,  the decline in Obama’s share of the vote occurs from a large cushion in September, and in some cases (PA) it may simply mean that vote shares are returning to their “normal” mean for that state.  Finally,  the polls reported in this post end on October 24th,  two days after the final presidential debate.  But many professional polls sites have shown a continuing improvement in Obama’s position after this date, and in particular an improvement in several states covered in this post (CO, VA, and possibly FL and NH in particular).  For this reason, our next post will report updated numbers for these four and perhaps two other states (OH and NC).

Number of polls and respondents for October (through Oct 24th)

state #polls #respondents

CO    9         8676
FL    16     13849
IA    7         5750
MI    1            895
NC    6          6395
NH    8         4854
NV    13    10378
OH    20    18058
PA    8          5914
VA    13    11777
WI    8         7099

Total    109    93645

Richard C. Eichenberg

Elizabeth Robinson

Tufts University

October 27, 2012

(polls reported in this post through Oct 24, 2012)

To say that gender is an important factor in the election is to state the obvious.  Beginning as early as last winter, policy debates and legislation affecting women have been a central theme of contention in the campaign, and of course the question of which candidate has an advantage among women voters has become a central preoccupation of poll watchers.

This is nothing new (see the scholarly references in the separate post at right).  There has been a noticeable, if variable, gender difference in presidential elections since at least 1992.  In 2012, however, the voting intentions of women have taken on added significance because of the salience of women’s issues and because the race in the Electoral College is so close.  As a result, each new poll showing a larger or smaller gender difference evokes a cycle of confusing discussion and speculation as to whether the historical “gender gap” has come to an end (for example: here , here, and here).

In this and other posts, we hope to clarify matters.

Technical Details About Our Data Collection

Please see our separate posting on the Technical Details that guided our data collection.  In addition, at the end of this post, you will find the total number of national polls included in our calculations for this post.

The Overall Standing of the Candidates in National Polls

It is useful to begin with the overall state of the national race, as shown in the graphic immediately below.  The figures will be familiar to poll watchers, revealing four phases in the presidential campaign.  First, during the primary season, characterized by fraternal attacks among Republicans, Obama enjoyed a comfortable lead of four to seven percentage points in polls testing his standing against Romney. After April, the race settled into a stable pattern in which Obama led Romney by a margin of 51/49 until September.  In September, Obama moved into the lead, fueled (we think) by the enthusiasm generated by the Democrats’ convention. In October, the situation reversed, a result (we think) of Obama’s dismal performance in the first Presidential debate on October 3^rd.  As of October 24 in our database, the race is essentially tied.

For our purposes, the central question that arises from these trends is whether the shifts and turns during these phases of the horse race are the result of a gender difference.  More specifically, as many have speculated in the press and blog worlds, we might ask: did Obama lose his lead because of a decline in support among women?

Gender Difference in National Polling

The answer is no.

As the graphic below shows, Obama has led among women nationally in every month during 2012, averaging 56.7 percent among women for the entire year (with Romney obviously averaging 43.3 percent).

This gender difference among women for 2012 (Obama leading Romney by +13.1) is exactly the margin that accompanied Obama’s victory in 2008.

Further, to the extent that Obama has a gender problem, it occurs among male voters. The two horizontal lines in the graphic display the percentage among women (top line) and men (bottom line) that Obama won in 2008.  As the lines make clear, he has fared worse among men this year than he did in 2008.  Specifically, Obama is averaging -3.8 percentage points among men for the year compared to 2008. Among women, it is a fairly trivial -.66 percent (55.6 percent in 2012 versus 57 percent in 2008).

Taken as whole, 2012 has been a year in which Obama has lost ground among men –not among women.

But that’s where the good news ends for Obama (to the extent that women provide the margin of victory for Democrats). In October, he lost ground among men and women by the same amount (2 percentage points compared to September). His standing of 46 percent among men and 54 percent among women is the lowest for any time during the year.  More worrisome for Obama supporters, the margin among women of +8 percent nationally  in October has dropped below the level that has provided victory for Democratic candidates in the past (Obama won with +13 points among women in 2008, compared to +11 for Gore in 2000 and +3 for Kerry in 2004). [update 10/30 in answer to reader question. Obama's lead among RV's only in October is 10.1%]

In summary, contrary to the speculation that accompanies one or the other poll released almost daily, Obama has a clear lead nationally among women as of October 24.  This lead is smaller than victorious Democratic candidates of the past, however.  And of course the crucial question is whether the same dynamic characterizes current polling in states that are crucial to the Electoral College outcome.

How is Obama faring in the crucial swing states?  What about Ohio? Or, as we shall see: Colorado and Virginia?

We address these questions in subsequent posts in this series.

Summary Statistics for National Polls Reported in this Post

Number of national polls per month

Jan       7

Feb      8

Mar      8

Apr      11

May     6

June     8

July      16

Aug      15

Sept     21

Oct       25

Total    125

Number of polls from following survey organizations (does not include daily tracking polls)

ABC/WP          3

ANGUSREID  10

ARG                4

CBS                 2

CBS/NYT        1

CNN/ORC       6

DK/PPP            11

DK/SEIU/PPP  3

FOX                 11

GRAVIS          1

IBBD/INV       2

MONM                        4

Marist/Mcc       2

PEW                10

POST/ABC      1

PPP                  6

QUINN                        4

YOUGOV        44

Total    125

Photo credit:  Madeleine Brown Eichenberg (Tufts 2021)

Elizabeth Robinson

Tufts University

We have been following and recording gender difference in state and national polling during 2012 and provide a descriptive overview of the data here.

A few words on the contents of the database:  First, we track only polls for which the gender breakdown is available at no cost (this eliminates Rasmussen polls, to choose one example). Second, we track only polls for which the available cross tabulations are reported in a comprehensible fashion that facilitates rapid updating.  Finally, we have excluded national daily tracking polls because they would overwhelm the virtue of variety that comes from including a large number of different polling organizations (the Gallup and Rasmussen daily trackers are therefore excluded)  After the election has passed, we may add these tracking surveys back to the database for research purposes, perhaps choosing one reading per week. These exclusions have no effect on the state-level polls, as there are no state-level trackers –yet.

Does this introduce a “House Bias” into the data collection?

The excluded polling organizations mentioned above have produced results in 2012 that lean more pro-Romney than other organizations, although we have no way of knowing or assuming that this bias is more or less pronounced for gender differences in the polling results.  Nonetheless, their exclusion here might suggest that our collection is slightly more Obama-friendly than it would be were Rasmussen and Gallup (among others) included.

Partisans of both sides should be aware of that.

On the other hand, based on the sophisticated calculations performed by others, house bias is a smaller problem than one might expect when many polls with different drifts are averaged together, as we do here.  Based on the calculated “house bias” reported by Jackman and by Linzer,  our collection includes a number of polling organizations that lean pro-Obama by about 1 percentage point or slightly more, but it also includes organizations that lean slightly pro-Romney by about the same amount . Finally, the collection includes a large number of polls from organizations whose house bias is close to zero.

A fair guess –and it is only a guess—is that the average house bias in our collection leans only very slightly toward Obama, perhaps half a percentage point.  In fact, it may be less.  As we note below, we estimate the Obama/Romney percentage of the national electorate as of October 24^th at exactly 50/50.  On this date, this is precisely the percentage estimated by the Huffpost’s pollster.com.  Also on this date, the NYT’s Nate Silver estimated a snapshot of the race at 50/49 (Obama/Romney).  In subsequent posts, we will also report the race in important states such as Colorado and Virginia at almost exactly 50/50 as of Oct 24th, which is also the precise number estimated by the professional poll aggregating sites.

Given how close our averages come to the averages yielded in professional model estimates, we have high confidence that our estimates are near the mark.  We will find out –and report– on election day.

Voter Base

The results reported here include likely voters and registered voters in survey samples.  We include registered voters because it allows us to extend the comparison back to January, when few polls had yet included likely voter screens.  As the year progresses, however, far and away the largest number of polls are based on likely voters only.  Specifically, for the entire year of 2012, 55 percent of our polls are Likely Voters.  After July, however, this increases to 84 percent Likely Voters, and it increases to 87 percent Likely Voters during October.

Polling Organizations Represented in the Database and Number of Polls for Each

Chaney, Carole Kennedy, R. Michael Alvarez, and Jonathan Nagler. 1998. “Explaining the Gender Gap in US Presidential Elections, 1980-1992.” Political Research Quarterly, Vol 51, (June), 311-340.

Crowder-Meyer, Melody, “Gender Differences in Policy Preferences and Priorities,” Paper Presented to the Convention of the Midwest Political Science Association, April 2007.

Eichenberg, Richard, “Gender Differences in Attitudes Toward the Use of Force by the United States, 1990-2003″ International Security, 28/1 (Summer 2003), 110-141.

Iversen, Torben and Frances Rosenbluth ,”The Political Economy of Gender: Explaining Cross-National Variation in the Gender Division of Labor and the Gender Voting Gap,” American Journal of Political Science, Vol. 50, No. 1, January 2006, Pp. 1–19.

Kaufman, Karen M., and John R. Petrocik. 1999. “The Changing Politics of American Men:  Understanding the Sources of the Gender Gap.” American Journal of Political Science. Vol. 43.  No. 3 (July). 864-887.

Shapiro, Robert and Harpreet Mahajan. 1986. “Gender Differences in Policy Preferences:  A Summary of Trends from the 1960s to the 1980s.” Public Opinion Quarterly. Vol. 50. No. 1 (Spring). 42-61.

Silver, Nate, “‘Gender Gap’ Near Historic Highs,” New York Times, October 21, 2012.

Wolbrecht, Christina, “Parties and the Gender Gap,”  in: Mischiefs of Faction, October 25, 2012.