Did any subsets of whites swing towards Obama?

[Franknburger]

There was a swing towards Obama in several counties on the Texas side of the TX-NM border, which happen to be new targets for natural gas exploration. Gas exploration workers, as well as the land owners, should be happy about anti-coal politics.

The counties in question, however, are pretty thinly populated, so in absolute numbers we are talking about 30-50 people per county having swung to Obama - which may in fact be Latinos that recently cáme in for gas drilling jobs.

P.S: Swings to Obama also occured in several rural oil-producing counties in Oklahoma. However, these counties have significant shares of Native American population, so the swing may also be due to higher native American turnout. 

[Franknburger]

O.k., let's take a more systematic approach, by comparing 2008 and 2012 exit polls. Overall the white vote was swinging 4% against Obama (43 % in 2008, 39% in 2012):

Region & Gender:

Whites overall:  Males 41 / 35 / -6,  Females 46 / 42/ -4

CT:     Whites  51/51/ --;  Males 53/ 51/ -2, Females 66/63/ -3
MA:    Whites  59/57/ -2;  Males 56/ 55/ -1, Females 68/65/ -3

There is a pretty good chance that white men in Connecticut, probably also in other parts of New England, have swung towards Obama.

IA:      Whites  51/51/ --;  Males 50/ 55/ -6, Females 55/59/ +4
WI:     Whites  54/48/ -6;  Males 53/ 47/ -6, Females 60/57/ - 3
OH:    Whites  46/41/ -5;  Males 51/ 45/ -6, Females 53/55/  +2

White women in Iowa should have swung to Obama. For other parts of the Mid-West, the data is less clear, but if there have been local swings to Obama, they should primarily have occured among females.

VA:  Whites  39/37/ -2;  Males 51/ 47/ -4, Females 53/54/ +1
NY:  Whites  52/49/ -3;  Males 59/ 58/ -1, Females 67/68/ +1
NJ:  Whites  49/43/ -6;  Males 56/ 55/ -1, Females 58/62/ +4
AZ:  Whites  40/36/ -4;  Males 45/ 43/ -2, Females 45/50/ +5

While in these states white women overall probly swung lightly against Obama, certain sub-groups of them may have swung towards him.

All other states covered in the exit polls (including  CA, FL, PA,  CO) either had a white swing against Obama of five points and more, or did not show markable gender difference, or both.

Whites by age:

18-29       54 / 44 / -10
30-44       41 / 38 / -3
45-64       42 / 38 / -4
65+          40 / 39/ -1

Since there was little swing among seniors, and higher white male than white female swing against Obama, I interfere that white senior women have probably swung slightly towards Obama. In the following states, Seniors (all races) trended towards Obama: NY (+4), NJ (+1), CT (+10 against 2004, no 2008 data), NH (-1), IA (+1), NV (+2), VA (), OH (). 

In the 30-44 age-group, whites only swung lightly against Obama. This age group (all races) swung towards Obama in the following states: NY (), CA (+1), NJ (), IA (+4), PA (+4), VA (+3), OH (), FL (+3), NC (+3), IN (+1). Considering that the heaviest swings against Obama occured in small cities and suburbs, I interfere that urban whites aged 30-44, and especially females, in the North-East and in the Mid-West are likely to have swung towards Obama. 

Religion

White Catholics:                                     47 / 40 / -7
White non-evangelical Protestants:          44 / 44 / --
White born-again / evangelical                26 / 20 / -6
Jews (all races)                                      78 / 68 / -9
Unaffiliated (all races)                            75 / 70 / -5

The figures speak for themselves: Obama did not lose among white non-evangelical protestants, so white non-evangelical protestant women have most likely swung towards him. This should also be a key factor for Obama's relative strength among whites in the North-East and the Mid-West.

A last observation: Parents (all races) only swung slighly away from Obama:

Mothers               57 / 56 / -1
Fathers                47 / 45 / -2
Women, no kids    57 / 54 / -3
Men, no kids         52 / 47 / -5

Here, I need somebody more familiar with American demographics to clarify whether this pattern is primarily due to ethnic differences (e.g. higher parental rate among Latinos), or may be used to interfere on white voting nehaviour.