Opinions of Obama Follow 2008 Election Results

Editor's note: this is a guest post by Alan Abramowitz, Alben W. Barkley Professor of Political Science at Emory University, and a member of the TDS Advisory Board.

Political observers follow presidential approval ratings obsessively and often interpret them in terms of the daily drama of this or that issue or trend. But amidst much speculation about the stability of Barack Obama's base of support, it's useful to compare his approval ratings in various demographic groups with their support for him last November.

An examination of recent Gallup Poll data shows that Americans’ opinions about the job Barack Obama is doing as president closely mirror the results of the 2008 election. The President's 58% approval rating in the July 6-12 Gallup Poll is slightly higher than the 53% share of the vote that he received last November, but his approval rating among various demographic groups correlates almost perfectly with his vote share among the same groups.

The following figure shows the relationship between Obama’s 2008 vote share in 24 demographic groups and his current approval rating in the same groups: the correlation between the two is a near perfect .99.

The implication of these results is that when it comes to opinions about the President, little has changed in the past eight months. Despite the continued weakness of the economy and the steady drumbeat of attacks on the President’s policies from the right, the coalition of groups that put him in office last November remains solidly behind him today.

4 Comments

| Leave a comment

James Vega | July 14, 2009 8:26 PM | Reply

There is an additional point that should be made about the recent polling that shows Obama trending down

In February Obama was in his honeymoon, the Republicans were in complete disarray and the big issues were the stimulus package and the budget - on neither of which the Republicans could get any real traction.

Since then Obama has had to deal with the Auto bailout, Cheney-Gitmo, the size and ambition of his health care and climate change initiatives as well as sniping about his "weak" approach to the Middle East, Iran, etc -- a whole series of issues that would inevitably peel away some of the "soft" honeymoon support he had in February.

Given this, the slippage we're seeing is actually remarkably small, not large. He's now put out most of the most controversial aspects of his progressive program and -- not only has he held his coaltion together, as Dr. Abramowitz' data shows -- but his support is still in the mid-50's overall and above 50% among independents.

Paleoprof | July 14, 2009 10:11 PM | Reply

Ok maybe I'm missing something here but: I'm not sure what relevance the r squared value has. If Obama's support fell wouldn't that just shift the line down but maintain the same r squared value assuming his support fell off equally among all the demographic groups? Or is that the point. If his support was falling it wouldn't be falling in all the demographic groups so the r square should go down?

napster | July 15, 2009 6:01 PM | Reply

@James Vega

The r-squared is a measure of the variation from a straight line. 100% or 1.00 means a perfect linear fit, and 0% or 0.0 means not linear at all, or completely scattered.

The rhetorical questions you raise are correct. Namely,(and I quote from your statement) "If Obama's support fell wouldn't that just shift the line down but maintain the same r squared value assuming his support fell off equally among all the demographic groups? Or is that the point. If his support was falling it wouldn't be falling in all the demographic groups so the r square should go down?"

If the r squared was less than it was before that would mean that some of the prior data points have changed significantly, so as to cause more deviation from a straight-line than existed in the original data-set measured back in February.

However, since the r squared value (or coefficient of determination) is unchanged, that means one of two things: 1) small deviations from the prior data set have been offset by other small deviations, or 2) the random nature of polls are within the same statistical normative range that existed in February 2009.

So nothing is really changing, according to the r squared value.

napster | July 15, 2009 6:05 PM | Reply

I'd also like to add further about the first possibility, namely that "1) small deviations from the prior data set have been offset by other small deviations".

If you look at the graph below, the Republican dot is below the line, and the Conservative dot is above. This means that the rise of the Conservative dot is offsetting the fall of the Republican dot.

To me this is a clear indication that true conservatives are getting sick of the insane politicization of real issues by the Republican party.

The Democratic Strategist's comments section welcomes intelligent discussion and debate from individuals representing every sector of the Democratic community.

Because of the spam problem, the first time you leave a comment you will have to sign up for a username by filling out a brief form. This just takes about two minutes and after that you will always be able to join the discussion just using your username and password.

Also, please note that all comments must be expressed in a mature and civil tone of voice. Individuals posting rude or otherwise inappropriate material will lose their access to the discussion.

Thank You, TDS staff