Romney Should Respond Like Obama’s 2008 “Race Speech”

The post is co-written with Anthony Daniels (Wayne State University).

I am working on a paper with a peer (Anthony Daniels) which suggests that Mitt Romney should combat this negative attention so that his previous remarks are specifically irreverent with like party members and independents.

Our model suggests Romney should do something, the problem is Romney may have a loosing hand.  This may be the type of story that goes beyond simple agency.  Even so, from a game theory perspective, he has to do something to mitigate damage among moderates and to stop the negative media coverage from snowballing.  The only way for him to do this would be a Sister Souljah moment of standing up to the right, or perhaps an honest apology.  Can Romney lead if GOP voters are not interested in following him back toward the middle?

Our data suggests that Obama brought the race discussion back toward the middle (after losing polling points after the Wright controversy threatened his image as a uniter). Obama managed to bring back his polling numbers without losing his base, while also satisfying independents and moderates. He had some tough words for both sides, and he chided Wright.  On the other hand, Romney seems to be able to only double-down, you know “no apologies!”

Can Romney initiate an Obama type race speech, but now on taxes and entitlements? Maybe some type of debate surprise is in order (almost always more risk than reward.)  Does he give a speech on entitlements where he tries to heal the class divisions of Americans?  Gives more specifics?  If he could only come out and say that the Grover no-tax pledge and entitlements must both give!

Considering the 2010 election: How do you neutralize this among moderates, independents, and weak Republicans without offending the Tea Party voters?

Our data shows that Obama’s race speech was indeed successful in turning the negative media frame into a positive media frame, and that only Republicans thought the Wright controversy was important when going into the election. On the other hand, we find that John Kerry’s relative inaction regarding the Swift Boat campaign hurt his electoral chances of winning because some independents and same party voters did think that the Swift Boat campaign was important in their vote choice. Clearly, the examples we use in our paper are external factors that a candidate may choose to combat via agency, while Romney must combat his own comments–but our answer is the same. Mitt Romney better do something to make this irrelevant for Republicans and Independents, otherwise this controversy will impact the vote choice.

Here is our abstract for the Midwest Political Science Association conference:

This study examines the media coverage surrounding the Jeremiah Wright controversy during the 2008 primary race between Hillary Clinton and Barrack Obama.  We adopt a game theory framework to assess the role of political actors in responding to and shaping negative media frames.  Previous research on media effects finds two competing descriptions: minimal and selective effects.  We find support for both, media effects are subtle but can impact campaigns.  We adopt the subtle effects assumption that media coverage can have a meaningful influence on political campaigns primarily though shaping the narrative of public discourse.  Consistent with theory, we find the impact to be limited, subtle, and the likely product of cumulative exposure.  Daily events, the public’s response, and the actions of political actors all combine to shape a media narrative.  Employing a multi-methodological approach, we find that political actors can shape the media narrative and that frames impact the public.  However, these effects are mitigated by partisanship and other heuristics.   Our analysis suggests that political actors should combat negative media frames in order to make the controversy unimportant in the eyes of same party members and independents; otherwise, the candidate’s electoral chances of winning will diminish.

We caution that context and fundamentals matter. Gaffes may not have any significant effect on actual vote choice. But if Romney allows this gaffe to become a negative media frame (Romney is out of touch with reality and lower-income people) in the mindset of same-party members and Independents (i.e., those he seeks in the video), then our model predicts that Romney’s chances of winning election will diminish.

The smart play would be to admit that he was wrong about the millions of non income tax paying citizens who will vote for him, and, initiating a real conversation about the Grover no-tax pledge and entitlements for the median voter to accept or reject.

In evidence, see this graph of public opinion in light of Obama’s race speech:

Our rational choice framework, in light of Romney’s opportunities:

Pos. is positive. Neg. is negative.


33 thoughts on “Romney Should Respond Like Obama’s 2008 “Race Speech”

  1. What opportunity will he have to respond other than Stump Speeches at this point? The debates he will be asked these questions, but there isn’t enough time to really develop his argument at that point.

    • Hi Craig, good question. Romney could use the media the way Obama did with his race speech. Everyone was playing the Obama race speech and talking about it (free media coverage) because it was a great sensational story that was affecting Obama’s chances of winning [the primary] election. Romney needs to give a great speech on economic diversity in America and entitlements–that the deserving deserve entitlements. The long speech–thanks to billion dollar elections–could be a really big event. Obama’s race speech was a long conversation about race and political change. Romney will have to say that he was wrong about the numbers–that millions of people that pay no income tax will actually be voting for him… but he should also go after the Grover pledge if he wants to win Republicans who were Reagan Republicans or subscribe to the Republican label somewhere from the middle to the far right–since the Pledge is extreme right and is a sign of corruption to the public realm (my hypothesis). And the Grover pledge forces Republicans to join the Tea Party–to become the Tea Party. So it’s a double-whammy, really, for moderate Republicans and Independents.

  2. It may be more important to note the differences between types of elections. Obama’s gaffe/problem occurred in a primary election where voters are less decided. Romney’s current gaffe occurred during a presidential election where most people have already made up their minds.

    You drew on the example of Kerry, but most (if not all!) forecast had him losing the election and his non-response to the swift-boat ads had nothing to do with that. It was all fundamentals.

    Finally, how do you square this with the general idea that gaffes do not matter?

    Also, would love to see the full paper or just more details on the model.

    • Hello Sighs, great point. There are meaningful differences between the primary and general campaigns, and I’m glad that you said that most voters have made up their minds–and not all. So, there are voters to be swayed, to one of the two candidates. And it’s not in Romney’s interest to push them to Obama.

      The suggestion that Romney should give a speech is also due to some regression analysis we compiled (in the paper) which asked voters whether or not the Wright [and Swift Boat] controversy mattered in their vote choice. We found that the same-party members and independents did not find the Wright controversy meaningful [for Obama] in their vote choice and that there was some significance in the vote choice for Kerry regarding Swift Boat among like party members and independents. That should help you understand more about the model and why we drew that conclusion in our blog post–based on data–that Romney should give a speech like Obama’s. The regression analysis therefore suggests that political actors can do something to influence voters. I caution, though, that we don’t have enough evidence to show causality. However, our paper does contribute to the field and shows that political candidates’ choices can produce subtle effects that impact the voters’ decision in the booth.

      So leaving out the regression, I assume, caused your great question. I’m glad you asked. You clearly have a good handle on political science.

      Many of times, gaffe’s don’t matter in the actual vote decision because a voter generally considers other aspects in his or her vote decision, like the economy, foreign wars (retrospective and prospective voting). But this gaffe may seriously matter, we believe, if it reinforces a negative image with voters (i.e. frame) that Romney is out of touch with reality–since his comment about 47% of the population dedicated to voting for Obama is so out of touch with reality.

      Hence our conclusion–what Romney could do, given his choices (go negative, do nothing, go positive). He should change the conversation by having a serious conversation (go positive in a manner similar to Obama’s race speech)–not let the negative frame continue.

      The full paper is in the final stages of prep., but not ready to post.

  3. Many Sighs, our model does let us keep track of new observable data–to expand the conclusions and implications. Consider (sorry for the long reply–I assume you see the negative frame in the media altogether now):

    Title: Are Romney’s ’47 percent’ comments beginning to move the polls? (+video)

    Main Point: At first glance, the latest polls don’t look good for GOP presidential nominee Mitt Romney. A Gallup survey shows his ’47 percent’ remarks cost him some support among independent voters. But it’s weeks until the election, and that effect may not persist.


    Title: Obama solidifies lead over Romney, ahead by five points: Reuters/Ipsos poll

    Main Point: (Reuters) – President Barack Obama maintains a lead of 5 percentage points over Republican Mitt Romney as he solidifies his advantage in the U.S. presidential race, according to a Reuters/Ipsos poll released on Thursday.


    Title: The First Reaction To Mitt Romney’s ’47 Percent’ Comments Look Pretty Brutal

    Main Point: Reuters/Ipsos and Gallup have both polled the reaction to Mitt Romney’s controversial comments on supporters of President Obama, and, unsurprisingly, both firms found some backlash from voters.
    Both polls found a fair amount of negative shift in opinion toward the Republican nominee, particularly among Independents.
    Forty-three percent of voters in the Reuters poll said that they viewed Romney less favorably after hearing the comments.
    About 60 percent said he “unfairly dismissed” the 47 percent he was talking about. And in the Gallup poll, 36 percent said it’d make them less likely to vote for Romney.
    Among Independents is where Romney gets hit particularly hard. In the Reuters poll, 36 percent said the leaked video made them view Romney less favorably.
    In the Gallup poll, 29 percent said it would make them less likely to vote for Romney (compared with 15 percent that said they were more likely to vote for him after the remarks).
    The Reuters poll, meanwhile, shoots down what has become a popular conservative notion over the past couple days: That Romney’s remarks won’t hurt him because they won’t realize he’s talking about them. Overall, 67 percent of voters said they more identified with the “47 percent” he was talking about than Romney’s audience at the fundraiser.


  4. I wonder how much of this an be contributed something like Zaller’s theory of public opinion. So something like this: 1) Something negative happens 2) People do not really know/care what is happening and just hear the negative noise 3) This shows up in opinion polls for a few days and then goes away with no real consequence.

    It is possible that talking about it might even make it worse! because it keeps it in the news for a few days.

    What Obama did was a bit different. He transitioned (and he did it well, because he is good at oratory and is likable where Romney is not really either of those) from a negative thing (his pastor is/was a nut) to a positive (narrative of being black and having a hopeful view of future race relations). I am not sure there is a possible similar transition for Romney.

    I am still very skeptical of your model. It is very difficult by comparing the results from two regression analysis. I realize that the paper is still in the prep stages, but it sounds like the analysis is done! Perhaps you could post the output, or perhaps post a do-file so interested parties can see what it is that you are up!

  5. Ben Franklin had to put a lot of wire together to see if the light bulb would go on when lighting struck the rod. Our model is the recognition that we need some wire–what can political actors do to positively impact their polling numbers? Certainly, sometimes it feels like the easiest way for a political candidate to rise is to drop the bottom out from an opponent(s). Our model gives a rational choice approach and allows for future research to run regression (which is an accomplishment in itself)–to see exactly which mode(s) of agency significantly helps a candidate win election (it’s not all Party ID–like the Truman election). For example, I can use this model to further understand what I study: the three main American political languages (liberalism, republicanism, and Biblical thought) in America. So this model (perhaps The Agency Response Model) won’t likely be ready to pass on to someone who builds election forecast machines; because it’s not wide enough yet. Meaning, I need a way to measure thousands of past cases that are specifically appropriate for our model–instead of our narrow exemplar; Barack Obama’s Wright frame. Our multi-methodological approach does help to advance evidence–to be, perhaps, worth explaining 2-3% of any election forecast model for scientific prediction. Hope that helps.

  6. Oh sorry, that was Thomas Edison… But you know, many Americans still think Obama’s a Muslim, that Iraq had Weapons of Mass Destruction, and that Edison got the light bulb from his apprentice.

    Isn’t it true that some political scientists see the vote choice as based on the economy and fundamentals? I think so, but I also see people being impacted by the American political languages. Many have strong liberal tendencies (e.g., the Louis Hartz Liberalism poem I’ll post tomorrow), and others feel strongly about republicanism…I think. Measuring all this and seeing it this way is a lot different that what today’s professional election forecasters machine. This is because they use states’ total votes for the electoral college, and many states already can be almost assured to vote for either Romney or Obama. Our model is mostly for the candidate running for election–helps provide a game plan based on evidence. Fortunately, it looks like campaigns will get more positive over time–deliberative democracy’s on it’s way (in a while).

    Of course the knee-jerk reaction to my last line might be… “no way”…
    But let’s remember the election of 1800: “With Jefferson as President,” so warned one newspaper, “Murder, robbery, rape, adultery, and incest will be openly taught and practiced, the air will be rent with the cries of the distressed, the soil will be soaked with blood, and the nation black with crimes.” Others attacked Jefferson’s deist beliefs as the views of an infidel who “writes aghast the truths of God’s words; who makes not even a profession of Christianity; who is without Sabbaths; without the sanctuary, and without so much as a decent external respect for the faith and worship of Christians.”

  7. I’m not sure what you mean. You seem to equate Zaller’s theory as unmeaningful noise–so the media doesn’t matter–as long as it’s not against you in the last days, perhaps.

    We find that some people indeed took Kerry’s Swift Boat ads into the vote decision (not just Republicans–which we would expect), and that only Republicans took the Wright speech into the vote decision. So on it’s face, there is some type meaningful media coverage–ads, campaign responses, strategies; which means that I should move past the way you portray Zaller and figure out a model to measure which stuff matters.

    But I already said all this, so I’ll update this post next when I have the paper ready–or, hopefully, published. May be a while. If you left a real email address–that would make things easier.

  8. I am not saying the media does not matter. I am saying polling shortly after a gaffe/mistake does not really matter. When polled people will recall the last thing they heard about a candidate, and respond with that. Since most people probably view the 47% comments negatively, that will show up in the polls for a few days until something else happens or people just stop caring about it. Then the polls will stabilize again. If Romney were to try to give a speech it would probably keep the 47% comments in the news for a few more days and possible create a longer lasting frame. This is because there is not really any positive way for Romney to transition the comments into a positive thing, like Obama did.

    I still would like to see your regression output. Could you either post the output itself or a dofile? That would be great. We are talking about an effect and I am kind of disadvantaged here because I cannot see all the details.

    • 1. Look at the type of speech we suggest–a speech that looks at much deeper American issues–which would change the media-poll narrative altogether. We do provide a content analysis of the news–not just the amount of news out there (part of the model). Obama’s act of employing agency did change the conversation altogether.

      2. The argument is not that don’t Gaffes matter–see the original post, it’s the narrative that voter’s use as a heuristic when going to vote.
      *Other polls have Obama up 10 points (and the final results will be the final proof)–which means a lot of tracking for political scientists who use our model.

      4. All that said, our model is only a starting point for understanding how agency works in campaigns.

      5. We agree that the fundamentals matter, but so to do campaigns.

      6. The Obama case does show an actual response in the Gallup tracking poll–albeit a primary. Independents did take Swift Boat into their vote decision–which matters if you are in a campaign to win an election.

      7. Finally, again, our model presumes a narrative and not a singular gaffe. The Obama case was unique in that it fit into a small window where not much was going on, and he employed positive agency to change the narrative–it wasn’t so much the tones of his voice, as it was the content of his speech.

      I’m triple checking our regression with our department’s stats specialist–but professors are busy, and I need to be sure. I hope you understand.

    • Try using the RCP average excluding Ras and Gallup. You will see Romney is not tied with Obama. Also the swing state numbers also show movement. Of course, with Libya, the conventions, it is not so easy to parse out. That said, I would venture to say that a look at the polling average will show movement toward Obama over the past two weeks. A good example is Silver’s model (which is over hyped but does a good job on direction of momentum).

  9. The chart does not show they are tied. It shows Romney around 45%. I should have charted Obama too for a better comparison, but as I said I just quickly threw this together! I used the pollster average for (national)polls given on the day. Perhaps something more is happening in swing states, but I doubt it can be attributed to Romney’s comments. And I still think it would probably be bad for him to respond (especially that it is almost past now) because it would just get the media’s attention to the comment again.

    • Many Sighs, Craig and Ads, thanks for the comments.

      The main point from our model is that a candidate’s response may be measured as meaningful or insignificant… we can now build a database to figure out what matters. We know that the Swift Boat inaction (on Kerry’s part) mattered in the vote decision. Now, we suggest that the 47% comment (original post) will be like the Swift Boat attack IF it adds to the negative frame in the electorate: Romney is out of touch with reality. We know he’s not the greatest speaker, but we suggest that Romney should give a speech (there’s still time) to combat that negative frame, decisively, in order to improve his chances of winning the election, through positive agency. The speech, “like” the Obama race speech exemplar, would show that he is very much in touch with reality–that’s what he needs to combat. He could do so through a serious discussion about entitlements while saying that elected officials are responsible only to the people–not an interest “no-new tax pledge signing” group. The content of the speech could speak to the middle–to the median voter’s reality. So, the speech doesn’t bring it back up, but transforms the negative frame [I’m so out of touch with reality] into a positive frame and leaves the negative frame in the past–indeed, makes the past 47% totally irrelevant–because of the more serious contextual speech. This is exactly what Obama did–combated the negative frame that he was a radical and not a uniter by being a moderate and telling the radicals that they were wrong, but uniting and saying that everyone has a chance to be heard… can’t disown my grandmother because she said racist remarks…but we have to come together as Americans.

      I need to write another poem now… feel free to have a discussion about election models here, which I won’t get into…but am happy to follow.

  10. And could you please just post the output?! It is hard to get a good understanding of everything without seeing the numbers from your model and the other variables it includes.


    • I’ll talk it over with my co-author and check the progress with the professor looking our work over (stats guru). Perhaps I could write a new post altogether considering the many misunderstandings in order to clarify too. I’m glad that you are interested.

  11. Mr. Romney, so 47% is everywhere now–absolutely everywhere. All I can say is that positive agency, sometime before the first debate, would be really, really smart of you. Otherwise, someone will ask voters, “Did the 47% comment affect your vote decision for Mr. Romney?” …and a significant amount of independents will say, “Yes.” Change the narrative through positive agency. Get specific in order to show, “I’m in touch with reality.”

  12. Looks like something else happened (the debate) and the polls have swung in favor of Romney! Do not worry, something else will happen and the polls will shift again. Then things will normalize and people will vote with the fundamentals.

    Any word on that posting details to your model yet? Still waiting.

  13. Here is what you ought to do. FYI, these commands require stata and perhaps a download of a particular command package. Run a logit model with all your variables (and I am assuming your dependent variable is a dummy for vote choice, you still have not said anything about it). run a prchange, save with all your variables set to their mean or modal value and have the variable that captures whether the respondent was influenced by Obama/Kerry’s remarks set to zero (if it is not a dummy, make it a dummy) and then run prchange, diff with the same but the variable that captures whether the respondent was influenced by Obama/Kerry’s remarks set to one. See the difference in predicted probability, I am sure it is very small.

    HELL! I would be willing to run this for you if you mail me your data. Will respond quickly with do-file and log, if you are willing.

    Let me know.

    • thanks for your interest. we are not willing to send data out to invalid email addresses…just sayin’

      We will likely redo the post with the regression in a week or so… where I just had my second oral exam and am composing an NSF proposal (as my dissertation proposal), Anthony is still preparing for his second oral exam –and completing his dissertation proposal. Hang in there, many sighs.

  14. Will provide real email, if you are willing to send the data.

    Please be quick with these things. Transparency is important.

  15. I thought it was all fundamentals… 😉 So now you agree that people–agency–can do something to change frames and public opinion and you want to calculate how much? You do realize that our point is about agency, right? And that our stats come at it backwards–that Kerry’s non-action affected vote choice, and that Obama’s positive agency caused no effect on vote choice (leaving it to the fundamentals, perhaps).

    Thanks for the laugh about transparency (but I get your point too). We’ll post a new entry when we are ready.

    I do appreciate your comments.

  16. I just wanted to say that it is kind of irresponsible to even make this post in the first place if the analysis is not ready to shown. It literately would take less than five minutes to update this post with information on your model (I am not even sure at this point what kind of model you are using! You just say “regression” there are so many different types of regression). the dependent variable, the independent variables and then the output from the model.

    • Hi Many Sighs, let me disagree. First, regression is not an entire answer, but may only be part of a scientific answer in the field of political science (some political scientists even think that regression is based on bias results–especially without integrating culture or institutions).

      Second, our post does not depend on regression analysis, but agency in combating a negative frame. I still don’t think that you recognize that that is our main point. Tellingly, many political science election models only integrate certain factors and leave out many others, like political actor agency, voter suppression, etc, but they are still worth posting about.

      Third, Romney was confronted with a negative frame (out of touch with reality) and though he finally came around to saying the comment was wrong (after the first debate), I sure would like to see a compiled video of him initially backing up the comment, going negative, and then finally using positive agency–after the 1st debate. In fact, I would hypothesize that this was a good part of his recent polling bump (though conventional wisdom will leave it to “winning the debate”). So, considering that our model is narrow about using agency to combat negative frames in order to change the media message–and thereby the media effect, AND by using agency in order to talk to voters directly (and perchance changing their internal narrative regarding the candidate)–the two pictures on the original post tell you a great deal.

      Forth, our post adds to the debate from the narrow aspect of agency in combating negative frames. This post is an attempt at deliberative democracy–deliberating what Romney could have done to combat the negative frame. That is worth posting about.

      Fifth, What is telling here, in this deliberative space, it the “Youtube Effect.” I mean, I don’t know who you are and you take no responsibility for your comments–so you say anything you want–even going to the extent of calling a post irresponsible. I hope in the future, as you clearly have an understanding for political science, that you will begin to post as yourself–to have a real record of your commitment to serious conversations about politics. That would be responsible of you, I think. And if you think about it, the post has made you think in a deliberative manner, for yourself, and so I hope that you gained some insight that you would not otherwise have reflected on (i.e. positive agency in combating negative frames). We would all benefit from a real network of deliberation–instead of many masks in the Political Science Cloud (internet).

      Finally, just to let you know, we will wait for a response from MPSA (poli-sci conference) before updating the post with our regression–which isn’t a slam dunk, but does add to the evidence that political candidates should utilize positive agency in combating a negative frame.

  17. But without details to the analysis, we do not even know if a problem exists that needs agency. Sure the 47% comments gave Romney a rough few days in the media, but now he is polling at his highest point of the election. I think that is what you are missing in my criticisms. That is why I want to see the analysis you two have done– which it appears that you both have little faith in due to your unwillingness to post the details– because I do not believe their is a problem that needs agency.

    Most political science models do not include these variables because they do not have a significant effect on election outcomes. So there is no need to include them. If they were important, they would be in the models.

    But anyways, hope to see a post from you two soon! :-p

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s