In their article, the search string of choice is ‘will I get deported’, as only people who are at risk of being deported are likely to use this phrasing on Google’s search engine. Using Google Trends to survey populationsĬhykina and Crabtree use a search string that only people who may be affected in the present or in the future are likely to use. This finding is important given today’s polarized political climate and matches well with other results in the literature on polarization. Second, and more substantively, political cleavages are more likely to determine people’s attitudes toward certain social phenomena than factual evidence. First, confirming Chykina and Crabtree’s main intuition, GT can be used effectively to survey populations, provided the search terms used are representative. With these findings, this article provides two takeaways for current and future research. Conversely, the correlation between people’s worries about the virus and the actual number of cases and deaths is weaker than the relationship between worries about the virus and partisanship across US states. States where Clinton’s vote share was high in 2016 tend to show greater concern about the coronavirus, while states where she lost are significantly less concerned. Interestingly, we find that there are strong differences along party lines in how people become concerned about the virus, confirming that the pandemic has indeed become a partisan issue. This is a strong indication that GT data can be used to survey the general population regarding their level of concern for the virus. We find evidence that GT searches for ‘will I die from coronavirus’ are highly correlated with both political preferences as well as coronavirus cases and deaths. Second, if they believe the virus to be a general health threat, they will be more concerned. First, if people see increases of cases and deaths in their state, they will become more concerned about the virus. We expect people to become worried about the virus on two grounds: epidemiology and politics. Taking advantage of abundant data around the COVID-19 pandemic, we analyze searches on Google in different American cities and states and ask the following questions: can GT tell us the extent to which people are worried about the coronavirus? And, more importantly, are these worries created by high levels of cases and deaths in these locations or are they politically motivated? Since the publication of Chykina and Crabtree’s piece, few works have expanded on their findings or probed whether they apply to different populations or issues. This paper focuses on the first avenue of research, namely, the potential for GT to serve as an alternative to survey populations. Timoneda and Wibbels argue that incorporating variance in GT search interest can help us forecast protests. The third application for GT is forecasting, which has been subject of long and inconclusive academic research. Chadwick and Şengül show evidence that searches for ‘unemployment’ in Turkey closely mirror the unemployment rate (see also for similar applications). A second role for GT in research is generating alternative proxies for useful variables. Their example is the search term “Will I be deported?”, which only people susceptible to deportation use. Chykina and Crabtree aptly show how specific search terms reflect the preoccupations of certain groups of people. There are three promising avenues of research for political scientists using Google Trends (GT) –a service by Google that aggregates search data by input term, geographical location, and time.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |