President Obama's rising approval rating nationally may bode well for Hillary Clinton in the general election.
The relationship between approval for Obama and support for Clinton is very strong, according to an analysis by the NBC News Data Analytics Lab.
Republicans argue that a Clinton presidency would amount to a third term for Obama, but in states where his job approval is high, that's good news for Clinton's campaign.
It can be difficult to forecast what a national approval number might mean for general election outcomes for several reasons, especially because the election is fought on a state-by-state basis.
To characterize the breadth of support for Obama - and perhaps also identify states that are favorable for Clinton in 2016—we used modern statistical methods called multilevel regression post-stratification (MRP)* to analyze 55,739 survey responses from the NBC News|SurveyMonkey Weekly Election Tracking Poll and produce state-level estimates of support for Obama.
The state-by-state Obama approval map below presents the results of this analysis for the past three weeks, from August 8 through August 28, and shows clear differences by state.
Whereas Obama's approval is estimated to be higher than 60 percent in states such as Vermont, Massachusetts, New York and California and elsewhere, it is below 40 percent in North Dakota, Oklahoma, Wyoming and West Virginia.
Clinton, who served as secretary of state under Obama, has embraced the policies of the current administration, which could help explain why the relationship between approval for the president and support for Clinton is so strong.
The chart above plots the relationship between Obama's approval ratings and support for Clinton and reveals that there is a close correlation. The data strongly suggest that Obama is not harming, but instead is actually helping Clinton. The differences are informative—states plotted above the 45-degree line indicate that Clinton support is slightly higher in those states than Obama's approval level. Almost every state is above the 45-degree line, which is a very promising sign for her electorate prospects.
*Multilevel regression post-stratification -- sometimes called MRP -- involves two steps. First we model presidential approval as a function of demographics and geographic regions to estimate the average approval for various demographic groupings and geographies. Second, we then determine how many people belong to each of those demographic combinations in every state using Census data. To produce a state-level average, the weights from the second step are then applied to the estimated group averages. Because of how it uses information, MRP has been shown to provide more precise estimates than aggregating respondents separately by state under reasonable conditions.