This Blog will be available for two now three weeks. Once again you can earn two blog credits by substantially adding to the conversation at two different times. We will discuss this blog on 11/30.
OK, we probably shouldn't do it but we all constantly make predictions. We, as a society are addicted to predictions and it's easy to make them because incorrect predictions are a dime a dozen and correct predictions, especially unlikely ones, can get you noticed. There is perhaps no place where this comes into play as regularly as politics where predictions are used so often that we start to assume they are reliable. They aren't. But political predictions based on good social science can carry a bit more weight. This past weekend Nate Silver (who now writes a blog for the NY Times and used to run an independent prediction site knows ans fivethrityeight that used extensive and relaible models) wrote a very interesting piece predicting the 2012 race based on three factors (or variables to use good social science speak): Obama's favorability a year before the election, the economic growth measured by GDP, and how extreme the poltiical views of his opponent are. Read (or thoroughly skim) the article here, and play with the interactive feature predicting the popular vote winner here. You may also want to look at the leading market based prediction service known as intrade, where you can wager on just about anything and use the market to predict what will happen. There are a ton of political markets on the site but the Obama reelection market can be found here. be sure to take a look at how their market works and how best to use it.
OK, we probably shouldn't do it but we all constantly make predictions. We, as a society are addicted to predictions and it's easy to make them because incorrect predictions are a dime a dozen and correct predictions, especially unlikely ones, can get you noticed. There is perhaps no place where this comes into play as regularly as politics where predictions are used so often that we start to assume they are reliable. They aren't. But political predictions based on good social science can carry a bit more weight. This past weekend Nate Silver (who now writes a blog for the NY Times and used to run an independent prediction site knows ans fivethrityeight that used extensive and relaible models) wrote a very interesting piece predicting the 2012 race based on three factors (or variables to use good social science speak): Obama's favorability a year before the election, the economic growth measured by GDP, and how extreme the poltiical views of his opponent are. Read (or thoroughly skim) the article here, and play with the interactive feature predicting the popular vote winner here. You may also want to look at the leading market based prediction service known as intrade, where you can wager on just about anything and use the market to predict what will happen. There are a ton of political markets on the site but the Obama reelection market can be found here. be sure to take a look at how their market works and how best to use it.
- After reading this interesting article do you think that this prediction model should matter at all? What is useful? What is not?
- If you looked at the intrade site, do you think that is a better prediction model? Why? Why not?
- What factor(s) (or variables) not included in Nate Silver's model do you think will play a major role in determining who wins in Nov. 2012?
A sports fans favorite part of the year is when the season is right around the corner and sports analysts are making their predictions for the upcoming season. The excitement upon opening up SI’s NFL Season Preview Edition and checking Peter King’s Super bowl predictions is unparalleled. Often predictions are on target but sometimes they are way off. For example in 2008 King predicted the Miami Dolphins would go to the Super bowl, but in reality they went a miserable 1-15. Just like in sports analysts political analyst make predictions on the upcoming election.
ReplyDeleteI think that Silver’s model is very interesting. He touches upon all the important issues that will effect the upcoming election. However I do not think this model is useful. I believe that is wayyyy to early to predict who is going to win the upcoming election. At this point, we do not even know who the republican candidate will be. Bush Sr. had a 79% approval in his third year, who would have ever thought he would not get a second term? Maybe we can start should start predicting who is going to win the election after the republican nomination is clear and we know whom the republican candidate will be. Would it not be ridiculous to predict an NFL season before we know the starting quarterbacks for each team? So it is ridiculous to start making predictions when we do not even know all the candidates!!
This comment has been removed by the author.
ReplyDeleteIn his article, Silver, discusses approval ratings- if it's a proper gauge of the Presidents ability to win in the next election. He illustrated the difference between the times in which we start to predict the presidents future, and which times are a more useful indicator. Where I got lost and where I found Silver to be ambiguous and vague was- what is the distinction between the president being victorious in the next election or his party being victorious in the next election. This brought me to wonder what really is the presidents ultimate goal and at what point does he have to make the decision between himself and his party- or is it even his decision? Do the parties control the president, and he just represents them, or vice versa, is he their boss? Do parties have their own motives separate from the president? Do parties sometimes want a president to not rerun for the next election, for strategic reasons? Maybe the democrats have a better candidate to continue their democratic rule. Do presidents have their own motifs separate from their parties? let's be honest even if they don't claim to, each president always has it in the back of his mind- what should I be doing, for public and professional approval, to win the next election. Or in the case where he is in his second term already- is he thinking about his public and professional approval for his legacy or for his party, who ultimately want to stay incumbent. What really are each parties ultimate goal ( parties being the president and his party)? Do they always share the same one?
ReplyDeleteAlthough I generally do enjoy Silver's analyses, and did enjoy reading the article, I am a bit skeptical of the interactive feature, which attempts to predict the chances of the Republican candidates defeating Obama. I think that although GDP growth is important, there are many, many other factors that will play a significant role in electing a President. This article: http://www.realclearpolitics.com/articles/2011/11/11/the_fuzzy_math_and_logic_of__election_prediction_models_112042-2.html, linked on FiveThirtyEight does a great job of explaining the problems with election predictor models (and keeping with Elie's football theme, mentions- in jest- how all Presidential elections since 1936- besides for 2004- have been predicted by whether the Washington Redskins win their last home game). I don't quite get Elie's analogy when he compares pitting the various candidates against Obama to predicting an NFL season before knowing the starting quarterback. In this case, we are comparing each candidate against Obama individually- kind of like trying to predict an NFL season with all possibilities for quarterbacks accounted for.
ReplyDeleteI think I pointed this out in class, but I still find it interesting how polls consistently are showing Obama significantly leading in the polls when placed against every candidate besides for Romney or a Generic Republican. Perhaps that foretells a serious problem for the Republicans if they nominate a too conservative candidate.
The model in the article was interesting especially because one of its factors was the main factor that has been identified as the most crucial for voters, namely the economy. Depending on what the economy does, Obama's re-election chances improve or decline. However, I think another factor needs to be included.
ReplyDeleteIn predictions for 2012, the models must include a way to incorporate foreign events. For example, whether Iran has a nuclear bomb or not and whether people view it as a threat will influence peoples decisions as they will elect someone who they believe can handle the situation.
Posted for David:
ReplyDeleteAlthough Silver's article was interesting I think that at this point that's all it is. In his article he claims that Americans' opinion of Obama can be measured through polls and approval ratings. I am in no way suggesting that such tools can be used for such a purpose, what I'm really saying is that (aside from even Silver admitting that such methods are "imperfect", and "useful" though not "infallible") It's still too soon to rely on those tools.
Silver himself mentions that such tools are best used toward the end of the third year in office, and in fact we have only recently entered the end of Obama's third year as president. Though Obama has (or hasn't) accomplished much, he still has some time left, time enough to either make things drastically better or worse. Such significant actions as brokering a peace treaty or enacting an effective or detrimental piece of legislation require only a minimal amount of time. Silver even alludes to this last-minute change when he twice mentions "October surprises" such as Romney having a scandalous affair, Sarah Palin poaching enough votes to lower Republican chances, or even Obama sending inappropriate pictures.
I don't know if any kind of "october surprise" will actually occur, but the mere possibility suggests that using current approval ratings in a prediction is shortsighted. Current data on Silver's other two prediction factors (GDP, ideology) are more accurate for such a prediction, since they are less susceptible (though not immune, especially GDP) to such sudden changes.
Posted for David:
ReplyDeleteAside from a couple of problems I had with Silver's three variables I think his predictions, specifically his case studies, are interesting and maybe even accurate. He does a good job of simplifying the issue and minimizing the calculations required. His case studies and the interactive feature suggest what most of us were probably thinking: that Obama has a good chance of winning against every candidate except for Romney, and even then Obama would probably win unless the economy gets even worse. The bottom line in my opinion is that Silver's variables, while straightforward and having led to convincing case studies, are not thorough enough. Silver briefly discusses foreign policy and the unpopularity of the Republican party within his case studies he fails to include them as variables in his model. In addition, he barely mentions other factors such as illegal immigration and Middle East policy. While those variables may not be as prominent as they once were they might still impact the election and should be considered in any analysis, especially since they may be the factors that can aid Obama or his opponent in a close campaign.
The one thing I find conspicuously missing is the fact that "voters" are not a constant. The voters who booted out Carter have aged considerably, and they have been replaced by another generation of people with an entirely new upbringing in a post-USSR age, an entirely different set of priorities and even an entirely different set of cultural backgrounds, since the population of minorities has exploded recently.
ReplyDeleteThen, as the author mentions, there are the inevitable exceptions, such as George H. W. Bush, whose high approval ratings meant little with regards to confidence ratings. And a vote, ultimately, is not a vote of approval, but of confidence. While Obama's performance over the summer was sub-par, his effectiveness as a commander in chief destroys the argument that he is incapable. Voters' memories are short, without doubt, but the death of Osama bin Laden and the imminent end to the wars abroad is hard to forget, even for his opponents.
Most pundits make predictions about the presidential election, but it is inovative to see a formula or a model like the one that Nate Silver is basing his predictions on. However the general basis of his argument is that it pretty much comes down to the economy, which is what pretty much everyone else is saying. If growth stays at around 2%, (and unemployment at 9%) I owuld say that most people agree that Obama is playing underdog in this election. But for Silver to say that Huntsman stands a 76% chance of winning at 2% is a little ridiculous considering he is polling at about 3%, without even counting Gingrich whos sudden surge has led to him the head of the pack. I do like the intrade market as it tracks on more of a daily basis and can account for many different unforseen variables.
ReplyDeleteI agree with Jordan and the comment that it all comes down to the economy. This article sounds like it makes this election a one issue showdown. The article said that even presidents who had bad approval ratings in the year leasing up to the election were able to win because the economy was at least moving in the right direction. I seems as though from this article that if the economy does not pick up, the election will be closer and more issues will come into play. On the flip side it also makes it seem as though should the economy rebound and be on the way up in the next few months leading up to and through the election Obama is likely to win and possibly have a landslide like Reagan did. It seems as though everything is dependent on the economy and all the other factors are secondary.
ReplyDeleteI think its important to point out that this model predicts the general presidential election (its not predicting who will get nominated, its predicting if a particular Republican candidate would get nominated, how he would do in the general election against Obama). I'm not sure how a similar model would work to predict who gets the nomination, or if its even possible to make an accurate prediction using any sort of formula for the nomination.
ReplyDeleteThis article: http://campaignstops.blogs.nytimes.com/2011/11/29/why-are-primaries-hard-to-predict/ got me thinking about this, and pretty much says that the nominations are so volatile (and therefore hard to predict) since the candidates are very similar to each other and people are very likely to change which candidate they support.
As Elie and others said earlier it is rather difficult to predict the outcome of the election before we know who is competing. I don't quite get the whole football thing, as I really don't know anything about the sport but I think it might be something like saying the karate master can beat anyone up before we know who he is fighting. His opponent may be an angry unarmed fool or it might be some weapons expert with a rifle.
ReplyDeleteEven so we do know who most of the competitors are and we can attempt to gauge he probability of Obamas success against each of them if Election Day was today. However to anticipate the changing conditions and accurately predict the political playing field so far in advance seems difficult at best