Showing posts with label issue evolution. Show all posts
Showing posts with label issue evolution. Show all posts

Tuesday, April 14, 2009

2008 GOP Candidate Emergence, Part 3

This is part three in a series of examinations of the fluctuations in the volume of Republican candidate Google searches during the 2008 presidential election cycle. You can find part one (the invisible primary trends among the top six candidates) here and part two (the invisible primary trends minus the Ron Paul skew) here.

Last week FHQ had a look at the development of GOP presidential candidate searches in Google throughout the 2008 invisible primary period (2005-2007). When the 2008 search data is added to the full time series a much deeper glimpse at the significant jump John McCain made heading into the 2008 contests is gained. Also, Ron Paul's 2007 gains peak once primary season commences and then decay rather quickly as a McCain nomination becomes highly likely following the Super Tuesday contests on February 5.

When the Paul numbers are suppressed (see figure above), we see that the two tracks argument mentioned in the previous post (a Thompson/Huckabee track and a McCain/Romney track) breaks down as the contests get underway. Recall, that once Thompson's candidacy failed to take off, Mike Huckabee essentially filled the void entering 2008. But that more social conservative track peaks and collapses after the Iowa caucuses, leaving a two person battle (in terms of Google searches) among the moderate/fiscal conservatives on the McCain/Romney track. Until...

Super Tuesday. Once we zoom in to look at just the 2008 portion of the time series, it is apparent that (again, in terms of Google searches) Huckabee's inability to back up the Iowa win with anything prior to Super Tuesday hurt the former Arkansas governor's chances at the nomination. Romney, despite the money spent, didn't win Iowa but was able to manage victories in several states (Wyoming, Michigan Nevada and Maine) between that point and Super Tuesday. That seems to have kept him viable in Google searches until Super Tuesday when Romney bested McCain in an Obama-esque run through the caucus states while falling further behind McCain in the delegate count because of the Arizona senator's wins in larger, winner-take-all states. The former Massachusetts governor's searches plummet after that point, coinciding with his withdrawal from the race.

In that intervening Iowa to Super Tuesday period, though, the race was on that McCain/Romney track in regard to Google searches. And though Romney dropped below Huckabee upon his withdrawal, Huckabee was more an afterthought in comparison to McCain at that point anyway. We don't, for instance, see Huckabee's search levels go up following Romney pulling out of the race. And that's what we'd expect given the way these nomination campaigns have gone in the recent past: a nominee quickly emerges and everyone else falls by the wayside.

Just for a bit of perspective, let's zoom in a bit further and include just the January to August data (dropping the general election search data). McCain searches don't reach the point at which Ron Paul was at the beginning of 2008 until after the GOP convention. But what is really striking is how much that Paul presence online deteriorates as McCain is sealing the deal on the nomination (Super Tuesday to March 4). Yes, both candidates are quite similar in their search trajectories over that period, but the key is looking at where each began the year. Once the contests started McCain searches took off and Paul searches dropped precipitously.

The other thing we gain from this is that 2008 on the GOP side provides us with a case not of invisible primary candidate emergence, but of primary season candidate emergence. And that's not something that Americans have been able to witness too often in the post-reform era. It is too bad we don't have comparable data for the Ford-Reagan race in 1976.

Recent Posts:
GOP Going the Caucus Route in Oklahoma in 2012?

No Caucuses? North Dakota in 2012

2008 GOP Candidate Emergence, Part 2

Friday, April 10, 2009

2008 GOP Candidate Emergence, Part 2

This is part two in a series of examinations of the fluctuations in the volume of Republican candidate Google searches during the 2008 presidential election cycle. You can find part one (the invisible primary trends among the top six candidates) here.

As the figures in part one showed, once we get to the 2007 segment of the invisible primary time series, the exponential growth of Ron Paul's search volume detracts from our ability to see the trends among the viable Republican candidates vying for the GOP's 2008 nomination. When the Ron Paul data is suppressed, Fred Thompson and Mike Huckabee appear to be bigger players. Both at various points during the latter half of 2007 tower over their remaining competitors in terms of their Google search traffic. You can see that in the full (2005-2007) figure above, but the 2007 snapshot below is more indicative.

In many regards we can see both candidates spikes as inter-related. Huckabee didn't jump all that much after his win in the September Iowa straw poll, but his win there coupled with the failed roll-out of Fred Thompson's candidacy seemed to boost Huckabee's profile entering the caucuses in Iowa. Essentially though, we have two tracks going here; one representing each side of the Republican Party (in its simplest binary terms). Huckabee and Thompson best typified the social conservative wing of the party; the segment of the Republican base that seemed least represented by and least enthusiastic about this pool of candidates. So much of the Republican invisible primary was about either how McCain was attempting to appeal to those voters or who the alternative would be. Through that lens, Thompson/Huckabee was that alternative.

But there was another track here as well; a more moderate or economic conservative track. This was a three person race between McCain (once the Arizona senator's campaign hit the wall in the summer of 2007), Giuliani and Romney. Giuliani's progression throughout the year though isn't all that variable, and as such, this was more of a two person battle between Romney and McCain.
[This is where Glenn's point on polling the other day is interesting: Giuliani was leading many of the polls during 2007 (see below), yet had basically flatlined in terms of search volume.]

[Click Figure for Link to Actual Polls at]

And what we see in that 2007 Google search snapshot is that Romney rises and passes McCain as the Arizona senator is falling throughout 2007. The former Massachusetts governor, then stays ahead of McCain until the calendar turns and the contests begin. With those two simultaneous trends, the GOP side is a little more interesting from a candidate emergence perspective than the Democratic side. And that the search time series doesn't comport with the polling going on during the 2007 portion of the Republican invisible primary raises some questions for us to pursue here in the future, especially regarding Giuliani. Why is it that America's mayor was polling so well, but got so little interest online?

There seems to be a better match between the two metrics concerning the other candidates. Yes, there is a Huckabee lag from search data to polling and Romney is consistently behind, though, close to McCain during the final quarter of 2007. On the whole though, the rising and falling of each candidate roughly corresponds across both measures (it is a matter of shifting the whole line that differs across both) with the exception of Giuliani, whose numbers are widely divergent between the two.

These data are designed to elicit questions rather than answer them, and I think we've got that here.

Up Next: the GOP candidates in 2008

Recent Posts:
2008 Republican Presidential Candidate Emergence: The View Through Google Trends

What About 2008? Democratic Presidential Candidates Through the Lens of Google Trends

From One Contest to Another: Ohio Redistricting Competition

Thursday, April 9, 2009

2008 Republican Presidential Candidate Emergence: The View Through Google Trends

To avoid chart saturation here, I'll break this examination of 2008 GOP candidates into three parts. The first part will focus on Republican candidate emergence during the invisible primary, the second part will drop Ron Paul in the context of 2007 to better ascertain search volume shifts during the latter half of that year, and finally part three will look at how things changed once primary season began.

As I mentioned in the series of posts investigating the shifts in Democratic candidate search volumes, the early speculation following the 2004 election and entering the 2008 invisible primary centered on a potential John McCain-Hillary Clinton general election. Well, the US got half of that last November, but early on Google searches favored both the New York senator and the Arizona senator overall. On the Republican side, though, there certainly is McCain red hovering over the other colors across much of the 2005-2006 period. And that's somewhat in keeping with the "next one in line" nominations that the GOP has had more often than not throughout the last generation.

But the full invisible primary (2005-2007) time series does not really provide us with the true nature of McCain's search volume relative to his most viable competitors for the nomination (for reasons that seem obvious simply by looking at the charts, but that FHQ will get into momentarily). If we zoom in on each of the three years individually, though, we get a better glimpse of what McCain's real advantage was. Again, McCain's volume is ahead of the other five candidates through 2005 (other than when Fred Thompson shot up in the summer when Sandra Day O'Connor announced her retirement from the Supreme Court and the former Tennessee senator was named by the Bush administration as the head of an informal group to guide her replacement, John Roberts, through the confirmation process.). Now, it is important to note here that there are no doubt endogeneity issues here as the media coverage of events in the political realm certainly has an impact on the volume of search traffic for a particular keyword. In other words, a search for anyone of these candidates is not necessarily a presidential run-related search.

However, that certainly changes somewhat as attention shifts toward the presidential race at the conclusion of the 2006 midterms (see above). All six candidates see at least a modest jump following the elections that brought the Democrats back into control of both houses of Congress. Again though, McCain is ahead across much of that year.

Heading into 2007 that progression continues. Mitt Romney's stock rises and finally surpasses McCain during the summer 2007 low point for the Arizona senator's campaign. More interestingly, though, Fred Thompson's search volume increases upon the formation of his presidential candidacy exploratory committee. The online chatter behind his potential candidacy continued into the summer months. As McCain's prospects waned, Thompson's grew. But Thompson's strength was as a potential candidate. Upon officially entering the race in September 2007, the former Tennessee senator quickly underwhelmed hopeful conservatives, losing ground online.

Of course, much of this Thompson spike is -- which is really quite an interesting case of candidate emergence -- is clouded by the sudden and consistent growth of Ron Paul online. The Texas congressman's close in 2007 dwarfed even Barack Obama and Hillary Clinton's heading into the election year. And those were two of the top three candidates in a Democratic field most Democratic voters were very enthusiastic about. Despite the following online (and this was something the power of which FHQ readers were recently reminded of), Paul just wasn't a serious candidate, and his numbers affect our ability to see the movement among the other five candidates who were all viable options heading into 2008.

Before examining the 2008 context, though, we'll look at the GOP invisible primary sans Ron Paul. Fred Thompson searches will appear much more significant and we'll better see the movement around other candidates (especially Mike Huckabee after his Iowa straw poll win and Fred Thompson's false start.). That's where we'll turn our attention next.

Recent Posts:
What About 2008? Democratic Presidential Candidates Through the Lens of Google Trends

From One Contest to Another: Ohio Redistricting Competition

Newt on 2012: "We'll See"

Wednesday, April 8, 2009

What About 2008? Democratic Presidential Candidates Through the Lens of Google Trends

Yesterday, FHQ looked at the Google Trends search volume for the 2008 Democratic presidential candidates between 2005 and 2007. The goal was to look for the emergence of the candidates during the invisible primary. And what happened, on the Democratic side at least, was more a case of candidate displacement than candidate emergence. Barack Obama basically overtook John Edwards as the Hillary Clinton alternative.

Invisible primary aside, though, what does the search volume look like for each of the top five candidates once primary and caucus contests begin providing tangible results in January 2008?

Well, for starters, tacking on that extra year and the election day spike really dwarfs some of the earlier data. What's important, though, is that red line (Obama) spikes in early 2008 and stays above the yellow line (Clinton) through mid-June when Clinton's volume trails off. Sure this gives us some scale, but that election day jump for Obama is deterring us from seeing some of the changes from the invisible primary.

If we cut off the data at September 2008 -- just after the Democratic convention -- we lose some of the skew from that election day spike. Instead of one multi-colored line running across the bottom of the time series from January 2005 to November of 2006, we can actually see Hillary Clinton hovering above the other potential candidates instead of appearing to be one in the crowd (while still appearing to be at the top).

The picture is clearer still when just the 2008 data is isolated. Obama still is clearly ahead of Clinton throughout the time series, but there are certainly some fluctuations given the events on the ground. The space between Obama and Clinton is widest following Super Tuesday in early February and Obama's streak of wins to close out the month. Wright, bitter-gate and losses to Clinton in Texas and Ohio in March, however, closed that gap. Obama, though, maintains that lead, diminished though it is until the nomination contests end and Clinton withdraws.

Again though, much of this tracks with the events that were happening in the contest at the time, but Glenn does raise an interesting point. What does polling look like during this period? That's obviously another layer to add into this; one that Cohen, et al. (cited in original post) considered. In the original study, (actually Karol et al. 2003), they found that endorsements had three times the impact on polls than polls had on endorsements during the invisible primary period. No, that doesn't answer the question in the context of the primaries and caucuses, but it does indicate that polls (like the data here) are likely next in line on the causal chain behind events in the nomination race.

Recent Posts:
From One Contest to Another: Ohio Redistricting Competition

Newt on 2012: "We'll See"

2008 Democratic Presidential Candidate Emergence: The View Through Google Trends

Tuesday, April 7, 2009

2008 Democratic Presidential Candidate Emergence: The View Through Google Trends

The other day I speculated about value of watching Google search trends as a means of tracking presidential candidate emergence. As I said in that post, it is one thing to look at that in real time, but quite another to look at how this looks over the course of an entire invisible primary period. Fortunately Google Trends has archived search data back to January 2004 and that affords us the opportunity to put this idea to the test in the context of 2008 presidential candidate emergence.

Now, keep in mind that this is an idea still very much in its infancy (and it may stay there given the limitations of the data and other complications). First of all, what you'll see below may not be tracking an organic growth and solidifying of support behind a candidate (or viability behind a candidacy) so much as a media-triggered urge to go find out something about a candidate. If we're looking for a causal chain, then, it may be something like:
endorsement/fundraising total --> news story --> internet search
In the context of a modern campaign built on social networking (via technology especially) the chain of events isn't as clear and regardless, all of the points in that chain have something of a recursive relationship anyway.

The other caveat is much more easily accounted for. Obviously, we'd expect the volume of searches to go up as a presidential election year approaches. That kicks the chain of events (in whatever order) into hyperdrive.

All caveats aside, though, how did this look in the context of the Democratic field of candidates as they emerged, announced and ran for the Democratic nomination between 2004 and 2008?

The expectation going in is that Hillary Clinton would dwarf all the other included candidates (and I just included FHQ's estimation of the top five candidates on the Democratic side) and at some point be passed by Barack Obama. And that is generally what we see in the full chart at the top. But it is easier to see Clinton's lead in search volume across all of 2005 and well into 2006 in the yearly snapshots. Around the time of the midterm elections in 2006 we see Obama's searches shoot up. The Illinois senator's numbers increased most likely because of his appearance on Meet the Press, where he discussed the possibility of a White House run. You'll also note that Edwards numbers also jump at the end of 2006 when he announced he would be seeking the 2008 Democratic nomination. There were similar spikes for Clinton and Obama when they announced during the first couple of months of 2007.

We don't see candidate emergence here so much, but we do see some candidate displacement. Hillary Clinton was very much a factor in the 2008 presidential campaign. In conversations I had here at UGA as early as 2005 the discussion centered on a Clinton-McCain general election in 2008. To some degree then, the story is more about the emergence of an alternative to Clinton. For instance, John McCain emerged as the alternative to George W. Bush in the 2000 Republican primaries. Edwards actually runs ahead of Obama through 2005 and 2006 (minus the Obama MTP blip), but once Obama announced his bid, he consistently ran ahead of Edwards for most of 2007. Obama, then, displaced Edwards as the Clinton alternative and that was solidified by Edwards opting into the federal matching funds system for the primaries in the late summer/early fall of 2007.

Now, we are limited by this data to some degree. These are weekly snapshots of the candidates' positions relative to each other in terms of their individual search volumes. Daily accounts are available and would provide us with a richer story (especially vis a vis the "which came first the news or the search" conundrum), but that's a something for another day.

Up next? The 2008 GOP candidates.

Recent Posts:
And Your 2012 GOP Presidential Nominee Is...


76,914-76,817 77,017-76,934: Murphy Tedisco Leads in NY-20

Sunday, April 5, 2009

Presidential Candidate Emergence: An Alternate Measure

I had this link come into my inbox the other day and it really got me thinking about using this Google search data to track presidential candidate emergence during the invisible primary.

[Image Courtesy of Click to Enlarge]

Now sure, Google itself warns against using their Labs-designated (read: not quite ready for primetime) Trends tool data for heavy duty research, which this isn't, so I couldn't help myself. The good folks at Irregular Times got the ball rolling on this in terms of tracking the 2012 Republican candidates' emergence in real time, but that only tells us a little bit of the story. Google Trends stretches back to January 2004 and that affords us the opportunity to track the fluctuations of the 2008 candidates on both sides as a baseline for comparison.

But here's the thing: I actually prefer the Google data over the Cafe Press search data. Yes, Irregular Times makes the point that Google search data pulls in all the search data regardless of whether you were looking up John McCain in 2006 in the context his 2008 presidential bid or some legislative work he was doing on the Hill. I can buy that. And while the benefits of using the Cafe Press search data (searching for actual candidate-related merchandise) are that we are gaining strength of attachment, the drawback is that we are potentially losing out on data concerning searches that while not as strong, are still related to these candidates in terms of the presidency. In other words, I'd like to take the larger view and try to narrow the scope somehow than narrow things unnecessarily right off the bat and miss something important.

[Fine FHQ, what's the point?]

This actually settles quite nicely into the realm of political science. The very first thing I thought of when I saw this data was issue evolution. The classic model constructed Carmines and Stimson (1981) looked at issue changes (such as on racial issues during the 20th century) on two planes. First, issue stances change over time, but secondly, their evolution takes place at the elite level within the party (in terms of perception and actions in Congress) and works its way down to the mass level affecting perceptions on the issues there.

This obviously has a link to the invisible primary period we are in now ahead of 2012. No, it isn't terribly active right now. Not at the mass level, at least. But there's no doubt there is jockeying going on at the elite level and that ultimately finds its way down to the masses. This approach has already seen some attention within the literature. Cohen, et. al (2003, 2005, 2008) have examined this at the elite level, tracking candidates' efforts to woo donors and high-profile endorsements. It strikes me, though, that this Google Trends data is an interesting means of tracking the level to which this permeates the masses. Now granted, the Cohen argument is that the system is set up in a way to allow for party autonomy over the nomination decision, but this data seems like an alternate means of investigating this as opposed to focusing on polling (which may have some endogeneity issues with internet searches) or waiting for vote outcomes in the primaries.

This week, then, we'll be focused on this relationship (among other things). Ideally I'd be able to roll this out in one big post, but I don't have the time tonight (and I suppose I've been sitting on this for a couple of days already anyway) to put it all together. We all may be better served having it broken down into its component parts. Regardless, this should be fun to look at.

Recent Posts:
The 2012 Presidential Primary Calendar (4/4/09)

Georgia in 2012: Back to March?

Championship Set in NPR's 2012 Bracket