Disclaimer: I’m a Democrat and have been helping both Chris Murphy and Elizabeth Esty’s campaigns. If you know me you probably knew that. This post is not made or authorized at the behest of any campaign, it’s just the result of being stuck home sick during Election Day weekend. Enjoy.

One of the more fun tools I have discovered recently is Google Trends. Google Trends allows you to see relative search volume for search terms on Google in order to ascertain interest in something by the public in general. For example here is the graph showing Linda McMahon and Chris Murphy:

There are several notable spikes and differences here. The first is that overall Linda has a greater online search volume than Chris. I think it is possible that this is due to the difference in online advertising expenditures. According to FEC records the Murphy campaign spent some money on online advertising early in the campaign (around 2011) but in 2012 they have only sent small amounts to outfits like Google  and direct ad buys on some progressive or CT related sites such as CT Capitol Report. Otherwise they spent some money with an outfit called Blue State Digital:

In contrast the McMahon has spent quite a bit on online advertising through an outfit called Harris Media:

The first thing you notice when looking at the trend is that Linda McMahon had some spikes on August 23rd and 29th. Looking at the CTNewsJunkie website this seems to correlate with the Republican National Convention, the exit of Linda McMahon’s spokesman, and the release of conflicting Rasmussen and PPP polls. Then interest in both candidates picks up evenly right after Labor Day. This suggests that when September hits the election finally enters the forefront of voters’ minds. Then we got another spike in interest on September 13th. This is when the DSCC started pouring money into the race and we got a UConn poll showing Murphy ahead of McMahon. After that Linda shows some mini-spikes in interest. September 18th correlates to the controversy over WWE taking down YouTube footage and Romney’s 47% remark. I’m not really sure about the 21st and 23rd, but the spike at the end of September may have to do with the Paul Ryan visit. At this point a Quinnipiac poll shows the race in a dead heat. Then we get a drop-off at the beginning of October until Chris Murphy spikes past McMahon after their first debate. You can again see debate spikes on the 12th and 16th of October, although McMahon surges past Murphy on both of those, Murphy stays slightly ahead in the polls. I’d suggest the debates drove interest and helped Murphy a little, but the effect was not large.

Then McMahon gets a gigantic spike after releasing her ticket-splitters ad. My guess is this is both part of the organic reaction to the ad but also a likely large push of the ad online. We can see the most recent polls after the ticket splitter ad is released shows a plummet in support for Linda McMahon in the Rasmussen, Quinnipiac, and PPP polls.  An October 24th spike for Murphy seems to correlate with his effort to rally women. Then Chris Murphy seems to get a post-Sandy drop-off in search volume while McMahon again spikes.

The analysis for the fifth district is cut short by the fact that there is not enough interest to warrant the creation of a trend graph for search traffic from Connecticut on Andrew Roraback. However we do get one for Elizabeth Esty:

We can see the  spike for Esty based on the primary and then the spike on during the October 14th through 20th is probably based on their first debate. For both Roraback and Esty I did not see any expenses in their finance reports that directly correlated with online advertising and a quick Google search does not show any. My guess is if they are using it, it is geo-targeted to the 5th district, where I am not located, and it is being funded in small amounts through general ad consulting agencies.

For comparison I also offer up the Google Trends graph for Barack Obama and Mitt Romney in Connecticut:

Here you can see Romney garners more search interest and Obama in Connecticut in spite of the fact that Connecticut is considered a solid blue state. I do not believe this is a sign of a potential upset. The visible spikes are for events like the Republican National Convention, release of the 47% video, and the presidential debates. The search volume differential could be due to the fact people already know about Obama and are trying to find out about Romney, or a differential in the use of Google by Romney and unaffiliated voters.

From this I think it is safe to conclude a few things. First that online search traffic is driven by debates, online advertising, and will also get a bump when a significant ad buy hits the television airwaves. Secondarily search interest does not correlate with support of a candidate. While Linda McMahon has driven most of the search interest on Google, she has failed to convert that interest into support. In fact it seems that her Obama ad backfired. Also I think Andrew Roraback should be worried. Although he is not well-known across the state, it seems people are not trying to find out who he is.