The First Month of Trump As Presented by Foreign Policy’s Experts

In U.S., national media love presidents. U.S. local and regional media love them even more. Media studies affirm this characteristic of American political communications. Since 1953 about twice as many stories were favorable to the president as were unfavorable - this was true not only for the period as a whole but also for most of its post-Watergate years. However, in the very beginning of his term, President Trump is already conflicting with mainstream media. Trump has devoted a 77-minute news conference to criticizing his press coverage, he wrote on Twitter that the nation’s news media "is the enemy of the American people." Journalists replied that Mr.Trump “shows no understanding of the role of the free press”, they said the president’s language “may be more insidious and dangerous than Richard Nixon’s attacks on the press.”

The debates have risen. Are mainstream media presenting President Trump in a false negative way? Is Trump insulting news media? The president’s team has distributed “Mainstream Media Accountability Survey” once again, asking supporters about their opinion on the issue and for “help to defend our movement from the outrageous attacks from the media.”

Some journalists called this survey “shocking” and “deranged”, others tried to explain that it is actually not a survey, but “a way to fundraise”. However, some of the media have invited their audiences to participate in even more biased surveys-like activities. For example, on February 20, Fox 13 News celebrated the President Day by posting not the most appealing photo of Trump at their Facebook page and has suggested users to express ‘reactions’: ‘like’, ‘love’, ‘haha’, ‘wow’, ‘sad’, and ‘angry’. If Fox 13 News had really wanted to know American’s opinion, the company would have used other means or just looked at results of national surveys. The Pew Research Center published one on February 16, concluding that “the public’s initial impressions of the new president are strongly felt, deeply polarized and far more negative than positive.”

President’s initial impressions are also mixed. Trump said that for the first month he would give himself an “A plus” for effort, an “A” for achievement and “C or a C plus” on messaging. Trump is not satisfied with mainstream media presentation of him and his movement in a negative tone. But what about media targeting leaders from government, business, finance, and the academic world? Do experts have other opinion than journalists? Do they agree to Trump’s “A” grade? I have chosen the Foreign Policy website as a data source to explore this question. Foreign Policy describes its half of a century achievements in terms of 2.4 million online monthly readers, award-winning journalism, design, and the presentation from the world’s leading thinkers. FP claims to be at the forefront of media organizations devoted to the coverage of global affairs. Thus, this source is a natural choice for an international relations scholar.

No surprise that publications on Trump can be found at the FP main page. Sometimes all three top stories are devoted to the new U.S. president. This media also set a special section “Trump’s First 100 Days”, which link is right above the main menu. A fascinating, but still uncomplete, sample. So, for now, I have collected all FP publications on Trump with a search function and this way I have also included pieces not selected for a special section.

As the inauguration took place on January 20 and the first working day of Trump as president was Monday, January 23, I counted four weeks from this date. During ‘the first month of Trump’, from January 23, 2017 to February 19, 2017, there were 384 publications on the FP website which can be found using keywords ‘USA’ or ‘Trump’ in search requests. Out of all texts I have extracted only paragraphs that consist word ‘Trump’ and got 2055 items. I have decided to omit a stage of morphological processing, as I think forms of worlds are interesting in this particular case. Data were scraped and parsed using R and analyzed with package “sentiment”, which functions are based on highly-cited papers by Strapparava and Valitutti for emotions identification [2] and Riloff and Wiebe for polarity [1].

As we can see, 26% of all items were recognized as ‘negative’ and that is not a bad result considering discussions on possible unfair representation of Trump in media. Share of paragraphs with ‘positive’ connotation is 38%, which is slightly bigger than 36% share of ‘neutral’ items. It is important to mention here, that I used an automated quantitative text analysis to produce these results and as any other method it has limitations. Here, for example, paragraphs that stated “Trump fulfilled one of his big campaign promises pulling the United States out of the Trans Pacific Partnership” were coded as “positive” by the program. The TPP topic is very controversial, but in general it is not a error, because to fulfill promises is a right thing for a president to do.

The package ‘sentiment’ allows to code six basic emotions using WordNet-Affect classification. A share of uncategorized paragraphs is big, what was expected considering nature of analyzed texts. If we look at the same item on U.S. withdrawal from TTP, we note that it is actually a fact with positive polarity and with absence of emotions. There should be difference in terms of expression between publications written by journalists and by experts. The three leading emotion categories are joy (32%), anger (14%) and fear (5%).

It is useful to present these results on emotions using traditional content analysis visualization ‘word cloud’ because histogram with so many items in ‘joy’ category can lead to misinterpretations. As it clearly seen, Trump main characteristic is ‘surprise’ during his first month. Quite interesting, especially because RIAC experts that do qualitative analysis came to the same conclusion. TTP went into ‘sadness’ category together with ‘allies’, ‘Obama’ and ‘white’, ‘house’. Seems FP experts think something good happened to China and express anger about Russia in texts on Trump. Unfortunately, terms ‘war’ and ‘sanction’ are very frequent, which made us think that use of hard power would possible increase. Even though Trump didn’t use terms ‘economic weapon’, ‘pressure’ or ‘sanctions’ when he explained a switch from TPP to bilateral treaties, his phrase “if someone misbehaves we will send a letter of termination” did implicitly said about these foreign policy means.

_{Vladimirova A.V. (RIAC :: Political Power Blog, 2017) Data Source: http://foreignpolicy.com/}

Finally, let’s also look at trends in texts polarity. Following a common recommendation to include a step before and a step after the period analyzed, I have increased data set and now it consists of 2724 items on Trump taken from 546 publications between January, 16 and February, 26. The histogram shows the same turn as FP main web page was demonstrating during last few weeks - experts moved from positive assessments to negatives ones, however, this trend is not absolute. The fifth week of Trump in office reveals a different pattern – experts write less and mainly in neutral tone. Probably the first month of Trump has surprised experts (and not only them). Thus, Trump decisions could cause negative feelings. Now it is time to calm down and get used to the new U.S. administration. As things falling into a rut and Mr.Trump is learning how to be President Trump, assessments could change. So, with the FP staff, we are looking forward to the ‘Trump’s First 100 Days’.

1. Riloff E., Wiebe J. Learning Extraction Patterns for Subjective Expressions. // Proceedings of the 2003 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 2003. С. 105–112.

2. Strapparava C., Valitutti A. WordNet Affect: An Affective Extension of WordNet. // Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC’04). 2004. Vol. 4. P. 1083–1086.