[UNDER DISCUSSION – REVISED]

Abstract

We are researching Swiss residents’ interest in the two candidates of the presidential election 2012, Barack Obama and Mitt Romney. Based on search data, our research will allow to make (relative) quantitative statements as well as provide qualitative analysis about the nature of Swiss residents’ information demand about the US elections.

Interest of Swiss residents in the US presidential candidates

The US presidential elections are an internationally highly mediated political happening, impacting politics on a world-wide level (cf. Metzler and Nischik, 2010, p. 113-114). Almost two-thirds of Swiss residents aged 14 years or older have declared being overall “interested in politics” (Latzer (ed.), 2012). Therefore, we can assume that many Swiss residents are interested in the US presidential election 2012, taking place on November 6.

Modern election campaigns being generally rather personalized (in the sense of person-centered, cf. Holtz-Bacha, 2006), the US presidential election campaigns are no exception: Metzler and Nischik (2010) analysis of the 2008 campaigns has inspired the authors to compare election and candidates to a pop-event with “polit-stars” (p. 126).

Therefore, we would like to know more about how Swiss residents’ general political interest translate into specific interest in the two presidential candidates of the US, Barack Obama and Mitt Romney. Whom are they more curious in? What do they want to know? We are especially interested in what happens during election day and afterwards: How does interest in each candidate change once one of them has been elected?

Interest: searching for (more) information online

“As people increasingly turn to the Internet for news, information, and research purposes, it is tempting to view online activity at any moment in time as a snapshot of the collective consciousness, reflecting the instantaneous interests, concerns, and intentions of the global population” state Goel et al. (2010, p. 17486). People’s web searches often indicate their interests. Using search engines to look for specific or general information can be considered a main internet usage (Quirmback 2013, p. 1). Previous studies have shown, for instance, that people affected with flu are highly over-proportionally interested in information related to flu, symptoms of flu and/or flu medication – which makes the number of such queries, and its variation, indicative of flu pandemics (Choi and Varian, 2012; cf. p. 3 for such and other examples; cf. also Goel et al., 2010, p. 17486). Most relevantly to our research is recently published study validating the use of search data for state politics research by showing a correlation between the search for ballot measure’s names and topics to actual turn-out during the US presidential election 2008 (Reilly et al., 2012). Therefore, seems logic that Swiss residents interested in the US-presidential election 2012 would be searching online for information about the two candidates.

Indeed, at least 40% of Swiss grown-ups search for political information online (Latzer (ed.), 2012). Internet usage in Switzerland is widespread (according to the Swiss Federal Office of Statistics (2012) over 80% of Swiss residents use internet on a regular basis) and broadly distributed: the category of people over 69 years not withstanding, there are only few variations regarding age, income, geographic location etc. (Swiss Federal Office of Statistics, 2012).

With their search queries, Swiss residents express their information demand about the US presidential candidates. The query “Barack Obama”, for example, indicates a demand of information related to Barack Obama. Idem for “Mitt Romney”. We are particularly interested in knowing what other expressions the candidate’s names are searched with: “Barack Obama Twitter”, for instance, shows interest of another nature than “Barack Obama Muslim”.

Retrieving search data

Like in most European countries, Google has a market share of around 90 per cent (D. Lewandowski, 2012, citing Lunapark, 2011). We have therefore focused on Google search data, provided by Google itself.

What is search data?

We define “search data” broadly as any piece of information related to enquiries via a search engine, explicitly including singular pieces of data as well as aggregated data. This information includes what I. Bordino et al. (2012) call query data . However, search data goes beyond query data, since it might not only be related to a specific query (time, term, user etc.), but also to the architecture of the respective search engine. Search data thus includes (but is not limited to): query terms, query log data (time/place/… of search query), device/browser/operating system/… used for accessing the search engine, volume of queries of a specific word, evolution of specific search queries over time, suggested “related searches” by the search engine etc.

Google Trends and Google Adwords

Google Trends consist of the two tools Google Trends (former version) and Google Insights For Search, which have very recently been merged into a single tool (Google, 2012a). It “analyzes a portion of Google web searches to compute how many searches have been done for the terms you enter, relative to the total number of searches done on Google over time” (Google, 2012b). Google Trends provides normalized search data about one to maximum five expressions going back to 2004. The platform allows geographical restrictions.

On Google Trends, we have been retrieving daily search data from Monday the week before election day (29th October 2012) until Sunday the week after election day (18th November 2012) for the expressions “obama”, “romney”, “barack obama”, “mitt romney”, as well as “obama romney” (on Google Trends identical to “romney obama”) with settings set to “Switzerland”. This will allow us to identify daily changes in the number of searches before and after election day, as well as compare volume and evolution between the two candidates.

Google Adwords is Google’s “ad agency” with an integrated “rating system”: the Google Adwords Keyword Tool (Lee, 2011). This Keyword Tool informs (potential) advertisers about possible costs, average figures about global and local monthly searches (up to 12 months back) as well as “related searches”. The later two are particularly interesting for our research: the figures because of the numeric values they provide; the “related searches” because of its keyword “ideas” and “suggestions”.

On Google Adwords, we have been retrieving daily search data from Monday the week before election day (29th October 2012) until Sunday the week after election day (18th November 2012) for the expressions “obama”, “romney”, “barack obama”, “mitt romney”, “obama romney” as well as “romney obama” (on Google Adwords distinct expressions) with settings set to “Switzerland – All languages”, “Switzerland – English”, “Switzerland – German”, “Switzerland – French”, each with two options: “Desktop and Laptop” and “All mobile devices”. This will allow us to describe the nature of information demand.

The nature of information demand can not only be analyzed as shifting over time, but also by comparing the two candidates. (For instance, Google considers the search query “michelle obama fashion” related to “barack obama” – but there is no “ann romney fashion” query associated with “mitt romney”. On the other hand, a related search to “romney” is ” romney bain capital”.)

Issues and Limitations

On an epistemological level, we acknowledge that the model of free and unbiased information access can be considered an “utopia” (E. Goldmann, 2008). Potentially, search engines themselves are the new gatekeepers (J. Grimmelmann, 2009).

With this in mind, it cannot be underlined enough that Google is a for-profit company and does not reveal its business secrets and the way its algorithms work. This proves also to be a challenge for this study: our research depends on the functionalities provided by Google’s tools, and the company’s declarations about them. But then again, the lack of other empirical sources legitimates a cautious exploration of the data, with the goal that our findings will in the near future be confirmed, maybe corrected and certainly expanded. Also, we are – in parallel to the research at hand investigating the potential and limitations of the Keyword Tool for scholars in order to know more about how it works.

Another point worth considering is the fact that interest in the US election 2012 is probably biased by affinities to politics (cf. Rothenbühler, 2012, p. 21 for an overview). And we do not actually know the profile of the people whose searches we analyze – which would be an interesting approach, but unfortunately outside of our scope. Our aim, however, is to provide first insights into aggregated information demand in order to analyse how Swiss residents’ interest in the US presidential candidates translates into different web search queries before and after election day. Further research might compare a Swiss media coverage analysis to our data in order to find out more about the Swiss dynamics between (individual) interest, media and politics.

Research in progress

Since the study is still ongoing, this workshop proposal presents preliminary considerations as well as first observations. At the time of the workshop, research will have progressed notably. The workshop itself would present the most recent findings of our research.

References (also hyperlinked within the article):

Bordino, Ilaria, Stefano Battiston, Guido Caldarelli, Matthieu Cristelli, Antti Ukkonen, and Ingmar Weber. “Web Search Queries Can Predict Stock Market Volumes.” PLoS ONE 7, no. 7 (July 19, 2012): e40014.
Choi, Hyunyoung, and Hal Varian. “Predicting the Present with Google Trends.” Economic Record 88 (2012): 2–9.
Goel, Sharad, Jake M. Hofman, Sébastien Lahaie, David M. Pennock, and Duncan J. Watts. “Predicting Consumer Behavior with Web Search.” Proceedings of the National Academy of Sciences 107, no. 41 (October 12, 2010): 17486–17490.
Goldman, Eric. “Search Engine Bias and the Demise of Search Engine Utopianism.” Yale Journal of Law & Technology (2006 2005).
Google (a). “Insights into What the World Is Searching for — the New Google Trends – Inside Search.” Insights into What the World Is Searching for — the New Google Trends, September 27, 2012. http://insidesearch.blogspot.co.il/2012/09/insights-into-what-world-is-searching.html. (accessed: 9th November 2012)
Google (b). “How Is the Data Derived? – Google Trends Help”, 2012. http://support.google.com/trends/bin/answer.py?hl=en&answer=92768. (accessed: 9th November 2012)
Grimmelmann, James. “The Google Dilemma.” New York Law School Law Review 53, no. 939 (2009). http://works.bepress.com/james_grimmelmann/19.
Holtz-Bacha, Christina. “Personalisiert Und Emotional. Strategien Des Modernen Wahlkampfes.” Aus Politik Und Zeitgeschichte, no. 7 (2006): 11–19.
“Indikatoren”, February 10, 2010. http://www.bfs.admin.ch/bfs/portal/de/index/themen/16/04/key/approche_globale.indicator.30106.301.html?open=1,302&close=302.
“Indikatoren”, February 10, 2010. http://www.bfs.admin.ch/bfs/portal/de/index/themen/16/04/key/approche_globale.indicator.30106.301.html?open=1,302&close=302.
Latzer, Michael (ed.). Internet Und Politik in Der Schweiz. Forschungsbericht – Abteilung Medienwandel & Innovation! Zürich: Universität Zürich, IPMZ, März 2012.
Lee, Micky. “Google Ads and the Blindspot Debate.” Media, Culture & Society 33, no. 3 (April 1, 2011): 433–447.
Lewandowski, Dirk, ed. Web Search Engine Research. Library and Information Science. Bingley: Emerald Group Publishing Ltd., 2012.
Metzler, Gabriele, and Reingard M. Nischik. “Die Präsidentschaftswahl in Den USA Im November 2008.” In Politikwissenschaft Und Politische Bildung, edited by Markus Gloe and Volker Reinhardt, 113–142. VS Verlag für Sozialwissenschaften, 2010.
Quirmbach, Sonja Monika, and Sonja Monika Quirmbach. “Über Die Bedeutung Von Suchmaschinen.” In Suchmaschinen, -. X.media.press. Springer Berlin Heidelberg, 2013.
Reilly, Shauna, Sean Richey, and J. Benjamin Taylor. “Using Google Search Data for State Politics Research An Empirical Validity Test Using Roll-Off Data.” State Politics & Policy Quarterly 12, no. 2 (June 1, 2012): 146–159.
Rothenbühler, Martina, Ehrler, Franziska, and Kissau, Kathrin. CH@YOUPART Politische Partizipation Junger Erwachsener in Der Schweiz. Schweizer Kompetenzzentrum Sozialwissenschaften FORS. Staatssekretariat für Bildung und Forschung SBF, 2012.