Search as Research: Comparing Journalistic Trends in Discourses Surrounding Gun Control During US Elections of 2012 & 2016

We are in an age of the Internet characterized by a shift from the periphery to the center of political, public and media attention. The Internet has replaced newspapers as the primary source of news and public discourse. Online news coverage has been playing a significant role in the United States’ (US) presidential elections. Digitised journalism allows presidential candidates to disseminate information directly and instantly, creating online visibility and shaping public opinion around popular debates (Scharl and Weichselbraun, 122). The topic of election news coverage has attracted much scholarly attention, not only because of the election’s vital position in democratic process, but also given that news coverage provides insight into public debates around prevalent social issues. According to a survey conducted by Horrigan, Garrett, & Resnick, 40% of Internet users over the age of 18 accessed political news online during the 2004 presidential election. This figure is expected to double as the Internet becomes more accessible with time, given the development of search engine technology. Research conducted by Jang and Oh reveals that people prefer issue-based news coverage, and tend to avoid news about campaign trails when reading election news. Thus, it is intriguing to examine the issue-based online news coverage in order to figure out the discourses around popular debates during U.S. election period.

Previous research has paid much attention to discourses around election. From a linguistic perspective, shifting discourse of language education policy in California’s 2016 election is analyzed by going through the text in details (Katznelson and Bernstein). In addition, discourse analysis based on different regions is also elaborated in Porto’s paper. Moreover, much research about online discourses around election periods have been conducted to investigate patterns of user participation, public sentiment and other potential political effects (Robertson et al.; Liu and Lei; Scharl and Weichselbraun).

Previous literature has mainly focused on mapping and analyzing the discourse in one specific time period, but they lack a macroscopic perspective to look through the discourse. Our research tends to fill the gap, arguing that the discourses might change overtime. Thus, the research is designed to compare the shifting discourses around a topical issue during two election periods, based on Google’s search engine results. To determine the “topical issue”, in which discourses might have the most notable changes overtime, we start with a sub-question to examine the journalistic trends in order to select the most suitable issue for discourse analysis. According to our pre-analysis, abortion, climate change, gun control, healthcare and immigration are the most popular issues during the election period, and ‘Gun Control’ was found to be the most relevant topic with the most volatile results. While writing this report, there was yet another school shooting that took place in California, in which two teenagers were killed. Upon learning this, we felt more obliged to address the debates around gun control. The process of how this issue in particular was chosen will be discussed in the following sections.

Research Question: How is journalistic coverage on topical issues affected/influenced by presidential elections in the United States? Are topical issues discussed more before elections for the sake of votes?

Sub RQ1: What journalistic trends are observable during the 2012 and 2016 election campaigns?

Sub RQ2: Does the debate [discourse] around the issue change from the time of Obama’s election to that of Trump’s?

Methodology

In order to effectively attend the research question, a methodology involving research tool Lippmannian device, browser Mozilla Firefox and search engine Google was designed.

Before commencing our medium and social research, a few technical preparations had to be made. What will our ‘medium’ be? The search engine ‘Google’ was decided on the basis of its monopolistic market share in the US. According to StatCounter, a web traffic analysis firm based in Ireland, Google is used by an approximate of 88.25%of the total search engine users in the country. This dominance well suits our research in assessing topics pertinent to the US better.

The second step; a depersonalized research browser was configured. Following the guidance provided in Digital Methods Initiative’s videoon YouTube, browser Mozilla Firefox was configured to minimize engine effects and biases on the sources returned. By installing Firefox as a ‘research browser’, the researcher prepares a clean slate, free of cookies and other engine entanglements such as history and preferences (Rogers), which further adds to the credibility and fairness of the outcomes.

In the current research, we intended to examine if topical poll issues like immigration, education, healthcare, etc. are mere poll gimmicks that politicians employ at the time of elections, or if they do hold importance even after being elected. To know this, we analysed journalistic coverage of the selected topics and the frequency of articles written on specific topics in a selected period of time. Subsequently, based on the results obtained, we selected one particular topic for further perusal that had undergone highly evident changes in the time periods analysed. A discourse analysis of the topic was done to learn the reasons behind the changes it had seen and how divergent were the discourses around it when compared between two different time frames.

In order to go about studying the possible effects of U.S. elections on its topical issues, two different time periods were selected for assessment; the 10 months surrounding second and first presidential elections of Barack Obama and Donald Trump respectively. Since these two personalities were elected as the presidents in the month of November [2012 and 2016], the preceding and succeeding 10 months, i.e. January 2012 to October 2012 and December 2012 to September 2013 for Obama and January 2016 to October 2016 and December 2016 to September 2017 for Trump were ideal for a study. The rationale behind choosing the election time period was an assumption that most of the poll issues would be discussed before the elections, i.e., mainly in the campaigns. The 10 month periods after each respective election were chosen to compare how frequently were the selected topics reported/talked about on media pre and post-elections, thus hinting on answers to our research question.

The research tool Lippmannian Device played a pivotal role in the study. This tool developed by the Digital Methods Initiative, one of Europe’s leading Internet study research group helped us in repurposing the search engine ‘Google’ to become a research tool to know the number of news articles (sources) on a specific queried keyword (topical poll issue) in a specific website (news media) within a set period of time. The tool can also be employed for interpreting bias and gaining a rough sense of a source’s partisanship and distribution of concerns (Rogers).

A long list of topical subjects that were most likely to be discussed in tandem with elections were drafted, this list was further run through the Lippmaninnian Device with the subjects acting as keywords. In the URL box, the top 5 most visited online news websites in the US were added, i.e, New York Times, Washington Post, Huffington Post, NBC News and Fox News (in no specific order). The outcomes portrayed that the topics ‘Abortion’, ‘Climate Change’, ‘Gun Control’, ‘Healthcare’ and ‘Immigration’ had the highest number of sources returned (the date period applied is same as the above). The political nature of the keywords was taken care of, as all the queried topics were ‘neutral’ in nature and not ‘programmes’ or ‘anti-programmes’ as discussed by author Richard Rogers in ‘Doing Digital Methods’.

The shortlisted 5 topics were further queried on the search engine Google with the help of Lippmannian device within the time frames mentioned above. For more apt results,‘United States’ was chosen as the ‘region’ in the advanced settings for Google. The results were obtained in forms of tag clouds on the device and were further visualized in grouped bar charts, using the website DataWrapper. The bar-charts represented the number of sources returned per queried keyword during both the time periods, i.e., 10 months pre and post for Obama, likewise for Trump. One of the 5 topics, i.e., ‘Gun Control’ had intriguing outputs in terms of its source frequency when compared between Obama and Trump’s time frames. Hence, the topic was further pursued to assess the discourse around it. To do so, Lippmannian device and Google were again put to work by analyzing the same 5 news websites and time frames as mentioned above and the queried keyword as “Gun Control”. This way, the top 10 news articles for all the 10 months pre and post Obama and Trump’s elections were picked and converted into dot-txt files for a word-cloud analysis. These text files were further run through the text analysis software AntConc, developed by linguist Lawrence Anthony. AntConc scraped through the text files and highlighted the most frequently used ‘keywords’ (unusually frequent words as per the software) in both the time frames. This process also required a bit of human intervention in terms of selecting the top 50 keywords for the word-cloud as words like ‘this’, ‘that’,‘the’ etc., had to be discarded and words that actually echoed issues around the Gun Control debate were selected. Once done choosing the top 50 keywords around the debate for both Obama and Trump’s time periods, word clouds were made using the website cloud.com to further compare how the debate changed from 2012 [Obama] to 2016 [Trump]. Further details will be discussed in the forthcoming sections.

Findings

Part 1

As our research project consists of two different parts, our findings will be discussed in the same manner. First of all, the findings retrieved through using Google’s search engine as a research tool to investigate what changes in journalistic coverage of topical issues can be observed on Google immediately before and after the election periods of both Barack Obama and Donald Trump were examined. These findings were retrieved by the Lippmannian Device, as discussed in our methodology.

Looking at the US’s journalistic trends that have appeared pre and post Obama, it becomes apparent that the total number of sources returned for the queried keywords have largely remained the same. Figure 1demonstrates that the number of sources returned for the keywords ‘abortion’ and ‘climate change’ almost did not fluctuate at all when comparing the time period before Obama was elected and after. Regarding the keyword ‘immigration’, we see a slight increase, while for the keyword ‘healthcare’ we see a slight decrease for the period after Obama was elected. However, both of these fluctuations do not seem significant. What does seem significant, are the results found for the queried keyword ‘gun control’. The sources returned for this keyword have significantly increased in the time period after Obama was elected.

Considering the periods before and after Trump was elected, it once again appears that the total number of sources returned for the queried keywords have largely remained the same. Surrounding the election of Trump, it becomes apparent that the journalistic trends for ‘climate change’ and ‘health care’ increased slightly in the post period. The number of sources returned for the queried keywords ‘abortion’ and ‘immigration’ have almost remained the same when comparing the before and after period of Trump’s election. Furthermore, the results for the keyword ‘gun control’ stands out, as a significant decrease becomes visible in the post Trump period.

Now it has become clear how the journalistic trends on Google have changed - or remained somewhat the same - when comparing the periods before and after both Obama and Trump’s election, it is of importance to also compare the results surrounding the Obama election to the results surrounding the Trump election. By doing so, two main points have become apparent. First of all, results for the keyword ‘climate change’ have almost doubled when comparing the time of Obama’s election to Trump’s election. Furthermore, the keyword ‘gun control’ seems like an interesting topic for further research, as we see the biggest fluctuations in the frequencies of results for this term. Figure 1shows that in the time of Obama’s election there is an increase in the number of sources returned for the keyword in the post period. This while Figure 2shows that in the time of Trump’s there is a significant decrease in the results for the keyword. This is the reason why we further researched the topic of ‘gun control’, in order to see where this difference might come from. 

Figure 1 - US’ Journalistic trends on Google - Pre & Post Obama (10 months)

Figure 2 - US’ Journalistic trends on Google - Pre & Post Trump (10 months)

Part 2

For the second part of our research, we have focused on the ways in which the discourse has changed regarding ‘gun control’ when comparing Obama and Trump’s presidential times. As already discussed in our methodology, we have created two word clouds - one for Obama’s and one for Trump’s presidential time - wherein the words all have the same sizes (Figure 3 & 4). This is because we are focusing on the narrative framing of the topic of ‘gun control’ rather than frequency of words.

Comparing these two word clouds, it becomes clear that there is indeed a difference in narratives surrounding the topic of ‘gun control’ when comparing Obama and Trump’s time. However, we do see a total of 8 words that are part of both the discourses during Trump and Obama’s election, namely ‘violence’, ‘shooting’, ‘national(ism)’, ‘assault’, ‘NRA’, ‘law(s)’, ‘Sandy’, and ‘Hook’. These terms that are part of both discourses can be ascribed to the fact that most of them are general terms related to the topic of ‘gun control’. The only three terms that stand out here are ‘national(ism)’, ‘NRA’ (National Rifle Association) and ‘Sandy Hook’. This is because these terms refer to an ideology, organization and a specific event that apparently over time have still remained part of the discussions and narratives surrounding ‘gun control’.

Apart from these few common words that are visible in both word clouds, all the other words represent differences between the narratives from Obama and Trump’s presidential times. The ways in which these differences in the two word clouds below can be interpreted will now further be explored in the discussion section of our report.

Figure 3 - Discourses regarding ‘gun control’ during Obama’s presidential time

Figure 4 - Discourses regarding ‘gun control’ during Trump’s presidential time

Discussion

Despite the extant issue concerning the abuse of gun laws in the United States, gun control policies have witnessed negligible changes in the last 25 years (Haltiwanger et al, Zurcher). However, what has fluctuated constantly in this time is the rhetoric concerning gun control. Since Reagan’s presidency in the 1980’s, there has been an unspoken tendency of American presidents to maintain silence with regards to mass shootings (Troy et al). However, this changed when Obama entered the White House. He publicly addressed the mass shootings that occurred in his term time, comforted victims’ loved ones, and pushed for tighter measures of gun control (Haltiwanger et al). His efforts were repeatedly shut down by members of Congress – a majority of whom have affiliations with the NRA (Haltiwanger et al). However, he did manage to implement a policy that required a social security check for mental illness when purchasing guns (Siemaszko). This was done in response to the Sandy Hook elementary school shooting.

On the other hand, Trump’s responses to the shootings that have taken place during his presidency have been a stark contrast to those of Obama’s. In addressing these catastrophes, Trump’s rhetoric principally focused on Islamist terrorism and anti-Muslim immigration [if the shooter was a Muslim], and mental illness, violent video games and the internet [if the shooter was Caucasian] (Wolf). In response to the Dayton and El Paso shootings of 2019, he asserted that “Mental illness pulls the trigger, not the gun” (Siemaszko). Paradoxically, however, he annulled Obama’s social security check legislation (Siemaszko).

We chose to unpack the narrative framing of news articles mentioning gun control in the periods immediately after Obama’s re-election as president in 2012, as well as Trump’s election in 2016. Our aim was to examine how the stark juxtaposition between both presidents’ attitudes and responses to gun control are reflected in the articles published online during their presidency. To do this, we drew conclusions on the narrative framing of these articles, based on the most frequently observed words in the articles. From the respective word clouds created [see Figures 3 & 4], there is, in general, a significant difference between the lexis used in articles published during Obama’s [second] term, and in those published during Trump’s term. Words associated with Obama’s term, such as “legislative,” “drafting,” “lobbying,” and “ban” create a lexical field aligned with sentiments calling for stricter gun control measures. This was expected, given that Obama’s presidency witnessed 24 mass shootings (Troy et al). On the other hand, Trump’s word cloud included words such as “rights,” “arms,” and “convention,” creating a lexical field alluding to the NRA’s annual gun convention, and its protection of the Second Amendment [the right to bear arms]. This was also expected, given Trump’s open affiliation with the NRA: according to the Federal Election Commission data, the NRA spent $30 million on Trump’s campaign in 2016 (Johnson).

However, it should also be noted that certain words, including [but not limited to] “violence,” “shooting,” “assault,” and “NRA” appeared in both word clouds. This may be attributed to the fact that these words are descriptive words, rather than words that may be subject to change depending on perspective. However, it can also be deemed an exemplification of the limitations of our methodology: the mere appearance of words may not be representative of the sentiments expressed in the articles in which they were found. At this juncture, we resorted to techniques of online groundedness (Rogers) to verify our claims: such techniques involved going back to [some of] the scraped articles, to verify how words found in both time periods were being used, and what sentiments were expressed in these articles, in order to make claims about them. Another technique used was fact-checking gun control policies in both presidents’ terms, in search of any changes in gun policies that may justify the differences of words in the word clouds. However, it should be noted that the latter technique was not limited to the 5 news sources that our research focused on. Further limitations of the research include limitations of the time periods chosen to represent both Obama’s and Trump’s term time. We chose the first 10 months after each president was re/elected, which may not be generalizable to their entire presidencies. In addition, the choice of the top 5 news sources was based on their popularity in the U.S., and thus their political stances and potential bias were disregarded.

Despite these limitations, our research provides an interesting insight into the changes in journalistic trends on gun control. Gun control continues to be an issue of great significance in the United States, and proved to be the most volatile in discussions before and after Trump and Obama’s respective re/election periods. Changes in rhetoric regarding an issue of such importance have a substantial impact on the public conscience. Despite Trump’s presidency having experienced the deadliest mass shooting in American history (Siemaszko), Trump’s attitude towards gun control is somewhat nonchalant and often paradoxical. Moreover, his hateful polemic is often criticized as inspiration for more and increasingly catastrophic mass shootings.

Conclusion

Our research aimed to reveal how election campaigning in the U.S. may influence journalistic trends on topical issues. After having scraped the 5 most visited online news sources in the U.S. in the time periods correlating to the 2012 and 2016 presidential elections, we found that gun control had the most volatile and thus interesting results between the different periods. Between Obama’s and Trump’s respective presidencies, we observed a significant shift in the way debates surrounding gun control were being addressed. Moreover, given the increasing proliferation of shootings and abuse of gun laws in the U.S, combined with the fact that in the short time it took to write this report, the U.S. witnessed yet another shooting, we decided that our research revealed an issue of formidable social importance.

Given Google’s dominance as a search engine, repurposing this “epistemological machine” (Rogers, 63) provided a hierarchy of the studied sources, and allowed for the examination of what dominant voices in discourses on gun control.

- - - - -

PS: This article is a result of collaboration between me and 3 other peers of mine from the University of Amsterdam.