When We Practice to Deceive:
Features of Fake News
Susan Spano
Middlebury Institute of International Studies at Monterey
Susan Spano is a Master’s Degree candidate in the TESOL Department at Middlebury
Institute of International Studies.
Correspondence concerning this article should be addressed to Susan Spano, Middlebury
Institute of International Studies, 460 Pierce St., Monterey, CA 93940. Contact:
susan.spano@gmail.com
FAKE NEWS 2
Abstract
Deception is as old as Satan. Lately, though, it has taken an alarming spin: “Fake News,” (FN).
The phenomenon appeared on the public radar screen in 2016, skewing presidential election
results, some claim, motivating violent conspiracy theorists, and substituting factionalizing
fabrications for free, unbiased news reporting. Media studies showed that Americans believed
FN news reports to be accurate 75% of the time during the last presidential election cycle
(Jackson, 2016), and Facebook’s top 20 FN stories were clicked on more often last year than the
top 20 real ones (Pogue, 2017). But while much has been written about FN, there is little
empirical evidence that can help us better understand it. The purpose of this study is to provide
such evidence by using Genre Analysis, as well as the findings of psychological, sociological,
and communications research studies on the topic of deception, to identify features common in
FN. BuzzfeedNews.com’s top 50 FN stories (Silverman, 2016), which will serve as a primary
data source. Surveys provide additional data for analysis and triangulation. The results of this
study will help news consumers separate the wheat from the chaff, and give media literacy
educators a useful tool for helping young people and ESL learners exercise critical thinking skills
when they go online for the news.
Keywords: fake news, genre analysis, deception, lies, media literacy
FAKE NEWS 3
When We Practice to Deceive:
Features of Fake News
President Obama signs executive order banning the Pledge of Allegiance in schools nationwide—abcnews.com.co
ISIS leader calls for American Muslims to support Hillary Clinton— worldnewdailyreport.com
Morgue worker arrested after giving birth to a dead man’s baby—tmzworldstar.com
These are headlines from three of the most frequently-viewed, Facebook-disseminated
“Fake News” (FN) stories published last year (Silverman, 2016). Many people knew at a glance
that they were blatantly untrue. But many others didn’t. A poll conducted by Ipsos Global, a
marketing and the digital media company, and BuzzfeedNews (2016) found that Americans
believed FN stories to be accurate 75% of the time during the last presidential election cycle.
Moreover, Facebook’s top 20 FN stories were clicked on more often last year than the top 20 real
ones (Pogue, 2017).
There might be little danger in the proliferation of FN if it were always as obviously false
and innocuous, as in the story about the morgue worker. But claims persist that FN helped to
swing the 2016 presidential election against candidate Hillary Clinton. Last December, a man
driven by an Internet-propagated conspiracy theory fired a rifle into a Washington, D.C., pizzeria
where, according to FN reports, high-ranking Democrats were operating a child sex ring.
FN spreads, as well, in non-Internet form, for instance when presidential counselor
Kellyanne Conway made repeated public references to a terrorist massacre in Bowling Green,
KY, that never happened. And it appears to have gone global, with the Chinese state media now
using the FN terms to refute western news reports about the use of torture in the PRC
(Hernandez, 2017).
FAKE NEWS 4
The invasiveness of FN has been driven partly by the public’s increasing reliance on
curated Internet news sites—The Huffington Post, LifeHacker, Reddit, etc.—instead of primary
sources like The New York Times and The Economist. Another factor feeding FN is the tendency
among readers to favor news aligned with their own beliefs over opposite reports by reputable
media outlets. As President Obama (2017) said, “Increasingly, we become so secure in our
bubbles that we start accepting only information, whether it’s true or not, that fits our opinions,
instead of basing our opinions on the evidence that is out there.” Balmas (2014) stated that
people who often expose themselves to FN are more likely to believe it than people who rely on
news outlets generally deemed more reliable.
In other words, it’s buyer beware.
But when news readers are young or novice English language learners, with as yet
unsharpened critical thinking skills, coming of age in the brave new world of digital and social
media, the danger of FN is heightened. Misinformation and ignorance has reached such a point
that there is now a cross-disciplinary field to study it: Agnotology, investigating why people
don’t know things like the scientific facts supporting climate change and the health hazards of
cigarettes. Bedford (2010) noted that the effort to misinform can take several forms, including
blurring distinctions between what is known and what is not known and claiming disagreement
or uncertainty about facts where none exists.
Croissant (2014) saw Agnotology more as a study of ignorance than of misinformation,
putting it in three categories: willful ignorance in relation to something known, unconscious
ignorance, and finally not knowing what is truly unknown. However, both intentionally
propagated falsehood (as in FN) and ignorance sometimes share a cognition-based feature, the
conflation of cause and effect. Moreover, both lies and the seeding of ignorance “may be
FAKE NEWS 5
similarly motivated by a multiplicity of factors linked together by considerations of social
power” (p. 7).
This study will reference Agnotology, theories and findings about deceit from
psychology, sociology, information management, and applied linguistics to peer into the
phenomenon of FN. The project’s purpose is to seek through texts for common FN features.
Therefore, this study will take up these two research questions:
1. What are the features of FN?
2. Do educated, adult news consumers know FN when they see it, and if so, what
features do they use to identify it?
Background
To understand FN in its broadest context, it is worth considering a number of cross
disciplinary findings about lying in general, including psychological research on how people
communicate when lying, often avoiding the first person singular, using cognitively simple words and
phrases, and conveying an over-riding sense of negativity (Newman, Berry & Richards, 2003).
Psycholinguists Arciulu, Mallard, and Villar (2010) added that deceptive discourse often reflects
a paucity of detail and the absence of fillers like um. People lie at least once a day, the authors
reported, but the ability to recognize a lie remains basically a matter of chance.
In a review of a wide range of psychological studies on deception, Adenzato and Bucciarelli
(2010) said that children begin telling simple lies by the age of three. The authors went on to
describe their own research on how children from age 4 to 11 recognize mistakes and deceit. The
study’s findings suggest that youngsters get better at spotting lies and mistakes as they get older,
and more surprisingly, that they are best at distinguishing complex, rather than simple lies.
FAKE NEWS 6
Another phenomenon that has long fascinated psychologists and neuroscientists is the
Misinformation Effect which makes people remember things that never happened when they are
exposed to falsehoods. Loftus (2005) said that a seminal study using neuroimaging technology to
track how subjects recalled an event they had viewed after getting false information about it
found that “misinformation was remembered as being part of the original event 47% of the time”
(p. 361).
Deceit research in psychology, sociology, and communications has often attempted to
analyze and categorize falsehoods by apparent intent and by types, such as lies of commission
and lies of omission. According to Merriam-Webster’s Collegiate Dictionary (2003), the verb to
lie means “to make an untrue statement with intent [my italics] to deceive.”
Intentional fabrication has long played a role in American politics (known as
mudslinging). Stories about Ulysses S. Grant’s sexual peccadillos proliferated in the 1872
presidential campaign. But driven by the Internet, FN today has “gone viral,” so to speak. It
appeared on the radar screen during the 2016 election, as stories about Clinton paying people to
protest at Trump rallies and getting support from ISIS reverberated in the echo chamber of the
Internet. Legitimate news sources had to make difficult determinations about whether to report
shocking FN falsehoods propagated on the Internet and in public discourse.
Snopes.com, Factchecker.org, and other groups attempted to keep tabs on the veracity of
news reports, and BuzzFeedNews went a step farther by tracking down websites, servers, domain
codes, and hit patterns to locate the sources of some of the most egregious FN stories. In one
case, BuzzFeedNews found hundreds of false pro-Trump stories coming from a small town in
Montenegro where people with computer skills had figured out how to make money by creating
FN websites (with cleverly-disguised URL’s) linked to advertising engines like Google AdSense
FAKE NEWS 7
(Silverman, 2016). In another case, an L.A. entrepreneur who was tracked down by journalists
said he made $20,000 to $30,000 a month writing anti-Clinton FN stories at the height of the
presidential campaign last year (Smith & Sydell, 2016). Perhaps sadder still, Justin Coler, an FN
writer interviewed by the CBS news show, 60 Minutes, said that seeing the audience for his
made-up stories explode in social media was “like an addiction” (Croxton & Gonsalves, 2017).
Facebook is not the only purveyor of FN. But it is one of the biggest. According to
Silverman (2016), two billion people logon to Facebook every month and 44% of adult
Americans get their news from it. So, when FN began making news itself last fall, commentators
started urging Facebook to do something about it. Founder Mark Zuckerberg was initially
reluctant to get involved, saying that Facebook is a tech not a media company. But both
Facebook and Google Adsense eventually instituted policies using sophisticated computer
algorithms to try to find and block FN postings, though the accuracy of the algorithm approach
has been criticized (Silverman, 2016) and FN keeps flowing, as evidenced by BuzzfeedNew’s
list of last year’s top 50 FN stories (Appendix A).
Definition of FN
The term FN is now so widely used for such a variety of phenomena that even the
producers of 60 Minutes had trouble defining it (Croxton & Gonsalves, 2017). To narrow it
down and put it in the context of this study, it is useful to begin with what it is not. Political
satire of the sort produced by Saturday Night Live and The Daily Show will not be considered
FN. Uses of the term as argumentative rebuttal and derogation in political discourse, as when
President Trump Tweeted that the mainstream media deals in FN (Grynbaum, 2017), will not
meet this paper’s FN criteria.
FAKE NEWS 8
The present study will address itself exclusively to Internet-based FN stories on political
topics, beginning with the BuzzfeedNews top 50 FN stories of 2016. Since the list’s publication,
however, some of the stories have been removed from the Internet and others are non-political,
tabloid-style reports of oddities, as in “Morgue worker arrested after giving birth to a dead man’s
baby” (Silverman, 2016). Consequently, data samples have been added to reach a total of 40
stories from websites identified by multiple sources (The New York Times, CBS News,
Scientific American, etc.) as disseminators of FN (Appendix B).
A definition that suits the study’s purposes can be induced from the headlines of those
stories and from understanding the background of FN, as rehearsed above. In this study, then, FN
can be defined as:
News reports based on provably false information, disseminated on the Internet for
political purposes.
Genre Theory
With a primary aim of identifying recognizable linguistic features in recent samples of
FN, this study fits snugly under the aegis of Genre Theory. Coming of age with the rise of
English for Specific Purposes (ESP) in the 1970’s and 1980’s, the theory looks at written and
oral texts as genres shared by communities of learners with special interests (like science and
technology). Defined early on as a communicative event with a common purpose (Swales, 1990),
genre has since then been reinterpreted through myriad lenses. Bazerman (2004) called genres
“psycho-social recognition phenomena that are parts of processes of socially organized
activities” (p. 317), and Wang (2008) said that genre study “attempts to capture how writers
achieve their social purposes by using various structural forms, constructing different focuses,
and manipulating topics and readers by using various linguistic devices” (p. 363).
FAKE NEWS 9
The theory has further evolved, with the study of genres subdividing into numerous, but
not mutually-exclusive strands. One important subdivision is Discourse Analysis, which
interprets communication “beyond the sentence,” partly inspired Grice’s Cooperative Principle.
Grice proposed four maxims for successful communication: the maxim of quality (speech based
on truth and evidence), the maxim of quantity (suitably informative speech, neither too long or
short), the maxim of relevance (on-topic speech), and the maxim of manner (clarity and
directness of speech). Dilmon (2009) said that deception obstructs all four maxims when an
interlocutor is made to believe that a deceiver is being faithful to the Cooperative Principle in
general (p. 1154).
Discourse Analysis has been termed Text Analysis when focused on the written word in
its cultural setting. As such, analysis of a written genre can move beyond semantical, syntactical,
pragmatic interpretation to contextual feature like what has been omitted, which, as we have
seen, is a factor of interest to psychological deception researchers (Bazerman & Prior, 2004).
Swales (1990) broadened Genre Analysis further by proposing that samples of a
particular genre follow a number of common steps— “moves,” he called them; ergo, Move
Theory. Millar (2011) said the moves involved in the recipe genre, for instance, include “the
name of the dish, a list of ingredients and amounts, the steps for cooking, and the number of
people the dish serves” (p. 4). Move Theory has great relevance to FN, a pseudo-form of the
news story. This well-established journalistic genre, as all novice reporters learn, covers who,
what, when, where, and how, in the first paragraph, and progresses from the most important facts
to less vital ones toward the article’s end.
Move Theory is closely related to the macro-level textual exploration of Rhetorical
Analysis. As defined by Selzer (2004), Rhetorical Analysis is “an effort to understand how
FAKE NEWS 10
people within specific social situations attempt to influence others through language” (p. 281).
Among the textual features assessed in Rhetorical Analysis are semantic connotation, hyperbole,
ethos/pathos, and epideictics (an Aristotelian form of declamation invoking concepts related to
virtue and vice).
Intertextual Analysis takes the fascinating approach of looking at a text in relation to
other texts, especially how outside sources are cited (Bazerman, 2004, p. 87). Using the
intertextual approach, Wang (2008) studied Chinese news commentary about 9/11, finding that
authors attempted to distance themselves from their sources in response to socio-cultural factors
like the role played by the media in Chinese politics (Wang, 2008).
In English language and literature, Williams (2016) postulated rules for prose writing
based on ethics. Echoing the Gricean idea that communication depends on commonly-held
principles, Williams said, “The social contract between thoughtful writers and readers implies a
goodwill exchange fair to everyone and in the long-term best interest of both” (p. 222). Stylistic
features he identified, such as abstraction, overuse of the passive voice and nominalized verbs,
oversimplification, and intended obscurity, tend to break the reader-writer contract and may
signpost deception.
On a methodological level, almost all the fields of Genre Analysis mentioned above have
tapped Corpus Linguistics to assess lying texts. Research in psychology, sociology, and
communications (Van Swol, Braun & Malhotra, 2011; Allen & Blinder, 2013; Baker,
Gabrielatos & McEnery, 2013) has also relied on corpus analysis, often focused on lexical
grammatical commonalities in deception, usually by seeking collocates. However, most studies
of this kind look at dyadic, oral forms of lies and posit that underlying tensions in the mind of the
liar reveal themselves linguistically in liar texts (Newman, Berry, & Richards, 2003). As a
FAKE NEWS 11
written form, composed for publication, FN would seem less likely to show itself in that way,
though some of the features social psychologists have studied in deceptive texts—low cognitive
complexity at the sentence level, minimal use of the first person singular pronoun, and a high
occurrence of words that have to do with negative emotions—can play a role in the textual
analysis that lies at the heart of this study (Newman, Berry & Richards, 2003).
Another textual feature often assayed through corpus analysis is total word count, easy
enough to determine digitally, but much harder to understand as a possible feature of lying texts.
Van Swol, Braun & Malhotra (2011) reported varying speculation about word count. Some
studies have found high word counts for fabrications, especially when the deceiver knows that
the listener can’t confirm or disprove what he or she hears. Other studies report that low word
counts are more common in lying texts, particularly when the liar says as little as possible to
avoid touching upon something the listener might already know.
Research Method
Research Question #1: What are the features of FN?
Bhatia (1993) outlined a genre analysis system that the present study will adopt:
1. Describe the context. 2. Conduct a literature review. 3. Define the text’s contextual framework in further detail. 4. Locate samples. 5. Analyze the samples. 6. Consult a specialist in the genre about the findings.
The author will begin inductively by qualitatively examining FN samples for
characteristics that stand out and recur. This process will be informed by the deception features
touched upon in the literature of lying reviewed above, including rhetorical moves, use of
sources, authorial distancing by avoidance of the first person singular, negative emotion words
and profanity, word count, etc. Other factors arise from FN context, for instance, disguised
FAKE NEWS 12
URL’s, links to more FN, ads appearing on the same page as the FN sample, along with
sometimes hyperbolic punctuation.
Seeking out features of FN is a way of identifying and studying it as genre clearly related
to, but not at all the same as, the news story, a well-established journalism genre. In writing news
stories, reporters follow such tried-and-true rules (or moves) as:
• The first paragraph, or lede, should include all the vital information, the who, what, when, where, and how. • The story should have the form of an “Inverted Pyramid,” with all the most important facts at the top, followed by less important detail. • Facts and figures must be substantiated. • All quotes should be verified and clearly identified. • Writers must be impartial and objective; they should not editorialize. • Advertisement must be kept separate from news.
The absence or presence of factors such as these will also play a role in my initial search for FN
features.
Once a critical mass of samples has been assessed inductively, a list of possible FN
features is expected to arise. The list will be winnowed down, prioritizing features that have
occurred in the highest frequency, and then used in the analysis of the rest of the data.
This precise approach has been used in Discourse Analysis. Barton (2004) described it as
“qualitative and inductive, with basic quantitative verification” (p. 63), a process that involves
identifying features by qualitatively scrutinizing samples, confirming their occurrence over a
corpus of texts, analyzing their occurrence quantitatively, and finally drawing thematic
conclusions.
Research Question #2: Do educated, adult readers know FN when they see it? What
strategies do they use to identify it?
Brief surveys (Appendix C) will be given to the author’s Facebook friends and to
students at the Middlebury Institute of International Studies at Monterey. In order to maximize
FAKE NEWS 13
data, the surveys will be easy-to-take, beginning with a brief FN text, the number one FN news
story of 2016, “Obama signs executive order banning the Pledge of Allegiance in schools
nationwide” (Silverman, 2016). Respondents will be asked to say whether it is true or false.
Next, they will be asked whether they can recognize FN, using a 5-point Likert scale (1 never to
5 always). Finally, they will be invited to list and describe the features they use to identify FN
stories. While participants will constitute a convenience sample, they are expected to be savvy
Internet media users. The responses will undergo quantitative and qualitative analysis, hopefully
providing human confirmation of the study results.
Results
Research Question #1: What are the features of FN?
Identifying features to use in FN analysis was a dynamic process. Some items that at first
seemed important (for instance, dated quotes and non-functioning links) proved less significant
the more samples I assessed. I looked for features intrinsically by studying 10 FN texts and
extrinsically by checking to see if features mentioned in studies and articles were evident in
them. Ultimately, I arrived at 22 features in 4 categories:
Table 1. FN feature categories
Then I looked for these features in a total of 40 FN samples, including 25 articles from
Buzzfeed’s top 50 FN stories and 15 from other FN sources.
1. Word and Sentence—involving lexis and superlative/hyperbole 2. Contextual—features arising from the context 3. Intertextual—pertaining to quotations and the use of sources 4. Strategic—macro-level “moves” or rhetorical approaches
FAKE NEWS 14
Table 2. Frequency of FN features _____________________________________________________________________________________
FN Features Frequency/ Sample/Description Number of Samples in which the Feature Occurred ______________________________________________________________________________
1. Negative words 40 kill, prison, bigot 2. Hyperbole 32 greatest children out there 3. False quotes 32 Fabricated source or words 4. Unattributed sources 29 experts, sources, data 5. News-like detail 28 Facts, figures, names, dates 6. Passive voice 26 were taken, was found 7. Editorializing 26 he admitted, surprisingly 8. Negative reference, insult 26 (9 Islam) Abusive characterization 9. Extended quotes 22 Article is basically one quote 10. Famous names, sources 21 Michael Jordan, Fox News 11. News format 19 Looks like a news story 12. Sex words, phrases 18 masturbation, hot girls 13. Red herring 16 Distracting argument 14. Platitudes 16 In God We Trust 15. Oblique defamation 15 Triangular derision 16. Name-calling 14 Crooked Hillary 17. Poor copyediting 13 Spelling, grammar 18. Idiomatic expressions 11 wanna, gonna 19. Unbiased appearance 10 Looks fair, 2 sides 20. Extreme punctuation 9 Caps, exclamation points 21. Profanity 9 hell, fart 22. False Dilemma 3 Two choices only ______________________________________________________________________________
FAKE NEWS 15
Table 3. Description of top 10 FN features
1. Negative words evoke negative emotions or reactions in readers: • “Individuals who violate this order can face fines of up to $10,000 and up to one year in federal prison.”— Jimmy Rustling, abcnews.com • “People all over the country have been flooding my office with calls, telling my staff of horror stories about being harassed and intimidated by poll workers,” Obama told reporters. 2. Hyperbole/superlative involve clearly exaggerated statements or claims. • “The Pledge of Allegiance helps our children be the greatest children out there, the best there is, greater than all the other children in this world.”—Trump • “I personally went to two Donald Trump rallies and I can say with 100% certainty that none of the protesters were getting paid.”—Jimmy Rustling, abcnews.com 3. False quotes are made-up statements, either obviously false or shown false in a Google search. • “I cannot allow him [Trump] into this chair, with his finger so close to the button. The power would go to his head immediately.”—Obama • “Hillary Clinton looks like someone took a leprechaun, put make-up on it, and told him to accept money from whomever it wants if that get him the presidency.”—Pence 4. Unattributed sources are quotes, facts, and statements that lack specific source citation. • “With such few members, but billions in real estate holdings, the Church of Scientology has been called ‘the most famous small business in the world.’”—abcnews.com • “Christian Times Newspaper has learned that Guccifer, the Romanian hacker currently being held on charges for hacking Hillary Clinton’s personal email server, has been found in his Virginia jail cell, dead of an apparent suicide.”—Dave Hodges 5. New-like detail is an FN feature that shows itself in blatantly false stories, made to look true by inclusion of facts, figures, and the kind of specific detail that is a critical component of journalism. • “The liberal Democrats are openly hostile to Christians,” said Graham in an interview on WEHW radio. • “A new survey released today from the Make Michigan Great Again PAC shows that residents across the state rated Flint tap water higher on the ‘Trustworthiness’ scale than Democratic nominee for president, Hillary Rodham Clinton.”—nationalreport.net 6. Use of the passive voice obscures the doer of the action, i.e. the subject of the verb. • “Many have even said that they were flat out denied entry into the voting booths…”—usanews.com • “She [Beyoncé] was seen storming off stage and leaving the venue in a limousine shortly thereafter.”— thelastlineofdefense.org 7. Editorializing has to do with expressing an opinion, sometimes covertly. • “Barack Obama has sensationally told CNN’s Wolf Blitzer that he will not vacate the Oval office if Donald J. Trump is elected.”—burrandstreetjournal.com • “And vote many times he did.”—thelastlineofdefense.org 8. Negative reference to a minority or special interest group, or insult to an individual. • “…women are still devoid of a soul but have been shown to possess qualities common to the mammal species.”—worldnewsdailyreport.com • “…do we really need to listen to this [Beyoncé] ghetto trash music?”—thelastlineofdefense.org 9. Extended quotes occur in stories that are basically one extended quotation, or reference to a quotation. • “Actor Bill Murray Announces 2016 Presidential Run”—abcnews.com • “Trump Claims America Should Never Have Given Canada Its Independence”—burrardstreetjournal.com 10. Famous names, sources are extraneous references to famous people or news sources. • “[George] Soros-funded left wing groups…”—dangerandplay.com • “…by making short jabs at the likes of Bill O’Reilly…”—theearthchild.co.za
FAKE NEWS 16
Table 4. Feature frequency by type
Word and Sentence 1. Negative words 2. Hyperbole 3. Passive voice 4. Sex words 5. Idioms 6. Profanity
Contextual 1. News-like detail 2. Negative reference 3. Famous names 4. News format 5. Platitudes 6. Name-calling 7. Poor copyediting 8. Extreme punctuation
Intertextual 1. False quote 2. Attribution 3. Extended quote
Strategic 1. Editorializing 2. Red herring 3. Oblique 4. Unbiased 5. False dilemma
Total word counts were tabulated for each sample, but ultimately seemed of uncertain
value as an FN indicator. I compared the average story length of 499 words found in the samples
to the length of two front-page New York Times stories from April 28, 2017, both of which had
word counts of just over 1,000. This suggests that FN stories tend to be shorter than real ones,
but further research would be needed to draw any firm conclusions about the relationship
between length and FN.
Table 5. FN samples word count
It was possible to distinguish the political leaning of 31 out of 40 samples: 20 Anti
Democratic (Obama and Clinton), 2 Pro-Democratic (Obama and Clinton), 6 Pro-Trump, and 3
Anti-Trump. The remaining 9 samples had to do with religious, health, race, and women’s
issues.
Mean words—499
Median words—414
Highest—1,739 words
Lowest—38 words
FAKE NEWS 17
Table 6. FN samples political leaning by top feature in each category
Leaning Negative words News detail False quote Editorializing
Anti-Democratic (Obama, Clinton) 22 samples total
Yes in 20 samples (1-31 instances per sample) 91%
Yes all Yes in 19 samples Yes in 17 samples
Pro-Democratic (Obama, Clinton) 2 samples total
Yes in 1 sample (5 instances) 50%
Yes all Yes in 2 samples Yes in 1 sample
Pro-Trump 6 samples total
Yes in 6 samples (1-28 instances per sample) 100%
Yes all Yes in 7 samples Yes in 5 samples
Anti-Trump 3 samples total
Yes in 3 samples (3-7 instances per sample) 100%
Yes all Yes in 2 samples Yes in 2 samples
Two additional factors were noted but not specifically addressed in the FN sample
assessment, the first URL’s. Given that these are created by FN purveyors, it’s no surprise that
they can be misleading. Many are reputable-sounding, for instance http://empireherald.com,
http://worldnewsdailyreport.net, and http://nationalreport.net. Others are extremely close to
established news sources; compare the FN site http://abcnews.com.co to the real ABC News site
http://abcnews.go.com. Others are clearly satirical or doubtful, as in http://tmzhiphop.com and
http://straightstoned.com.
Finally, the look and content of the website surrounding a piece of FN can be telling.
Almost all FN samples in this study had links for additional FN headlines, some potentially
credible, others of a ridiculously vulgar tabloid sort. In many cases, when you click such a link,
you’re led eventually to an advertisement for pornography or super-vitamins. The products alone
could make a reader suspicious. However, keeping news separate from paid advertising is an
FAKE NEWS 18
inviolable principle of traditional journalism, so the blurring of the line between ads and news in
these samples can be seen as another FN marker.
Figure 1. FN ads and links
Research Question #2: Do educated, adult readers know FN when they see it? What
strategies do they use to identify it?
Survey I, sent to my Facebook friends, reported findings from 27 respondents, almost all
female and over 61. This group was highly savvy about identifying FN.
(1 Very likely true—5 Absolutely untrue) Figure 2. Survey I respondent ability to identify an FN article
FAKE NEWS 19
(1 Never—5 Always) Figure 3. Survey I respondent self-assessment about ability to decide if an article is true or false
Table 7. Survey I respondent FN features
• Look for a byline and I check the Internet address. • Source (is it biased), quotes from knowledgeable people, inflammatory language. • Reputable media outlet and author. Credible spokespersons. Reasonable content and facts. More facts than opinion. However, a lot of phony press is out there and sometimes it's important to check articles on Snopes or FactCheck.com. • Citing to various sources; common sense • I get most of my news from sources I know and trust to have high standards of responsibility and factchecking. If I don't know the source, I evaluate the quality of the writing and the author's ability for logical thought. I compare the claims being made with knowledge I already have. The quotes used in a fake news story are often the biggest giveaway, as in the story below. The Trump quote in particular is an obvious parody. Capturing a specific public figure's syntax is the hardest thing to fake. If I'm ever not sure of an article's validity after reading it carefully, I check with Snopes or a news source I trust. • Look for the media resource • Source. Emotion. Grammar.
Then I aimed the survey at a younger demographic by giving it to students at the Middlebury
Institute of International Studies in Monterey (MIIS), CA. Survey II reflects the opinions of 35
MIIS respondents, 23 females and 12 males. They were almost all age 21-40, except for two age
41-60. Findings in this younger demographic also showed sharp skills in recognizing FN.
FAKE NEWS 20
(1 Very likely true—5 Absolutely untrue) Figure 4. Survey II respondent ability to identify an FN article
(1 Never—5 Always) Figure 5. Survey II respondent self-assessment about ability to decide if an article is true or false
FAKE NEWS 21
Table 8. Survey II respondent FN features
• I look for reputable sources that I know (NY Times, BBC, etc.) • I look for other sources posting the same story. • News source; prior knowledge/context; are there other articles that confirm it (preferably from verified sources)? If there is a study or survey referenced? I either look at the original myself or take the article with a huge grain of salt. • Publications with a strong editorial reputation. Recognizable URL. "About Us" section for sources I'm not familiar with. Red flags include sensationalist headlines, or headlines that demonize a person or group. • I might look for news reported by news agencies from different countries especially regarding certain transnational controversial issues.\ I consider the logic, legality of the claims in the article. • I consider the logic, legality of the claims in the article. • I check the source--is it someone's blog or a reputable source like WaPo, the NYT, CNN, The Economist, The Hill, etc.? I look at the language--if it's too strong, the article is likely either false or heavily biased. If I still have questions, I search for reputable news sources that report on the same story. • I read sources that have a history of telling complete stories and that provide argumentation from all sides. I read a variety of different sources, left- and right-leaning, and sources at different levels, state, national, and international. I look for sponsorship or sources of funding, if possible/applicable. If something "seems" untrue, I try to verify it. • I look for multiple reputable news agencies and verified sources. I also do resources on factors that stand out to me in the news • Double check different sources for the same info. I do not trust any news agency • I look for information I can confirm using other sources, how it interacts with my understanding of the facts. • Source • Grammar and spelling matter. If there are errors, believability goes down. If it's big news I usually check 2 other stories. • The issue being discussed is highly unlikely to happen. • I read news from other sources to compare information and talk to people I trust. • Source, talk to people and get their opinions • Spelling errors, contradiction to previous knowledge • I look for references, the author, cite of publication, and Snopes • The source (if it's from Fox News or another dubious source--religious, etc.) I usually doubt it. If it comes from a name I trust (Reuters, NY Times, etc.) I tend to trust it. • Other articles about the same issues, read about the author and the publisher • You need to look at the source. The source of the source. The author. The methodology used. When it was written. What kind of article it is. Opinion paper, research paper. • Who posted/published it. Why. When was it posted or up-dated? • Check the website it's on; double check for a site description listing articles as satire or not; check sources • References, citations, quotes, publisher • Authoritative source. Opinions from third party. • Not sure; I don't generally come across much fake news, so far as I can tell. • Sources, does it come from an authoritative source? Discussion with friends. Google to search if there are similar content online. • I look for strong personal opinions/biases, or absolute statements (untrue). Inclusion of data, links to reputable sources/being from a reputable source, look of bias (true). • I check the source--is it someone's blog or a reputable source like WaPo, the NYT, CNN, The Economist, The Hill, etc.? I look at the language--if it's too strong, the article is likely either false or heavily biased. If I still have questions, I search for reputable news sources that report on the same story.
FAKE NEWS 22
Interpretation and Conclusions
FN is widely-discussed and news consumers like the ones surveyed in this study have
strategies for recognizing false reporting when they see it. Moreover, Facebook, one of the major
FN sharing websites, recently gave users a mechanism for marking false news posts, along with
a list of generalized tips for spotting them (Price, 2017). This study, however, has attempted to
find empirical evidence about FN’s nature, yielding a list of 22 common characteristics that were
tracked in 40 samples. To recap, the 10 most frequently-occurring FN features identified in this
study were:
1. Negative words 2. Hyperbole/superlatives 3. False quotes 4. Unattributed sources 5. New-like detail 6. Passive voice 7. Editorializing 8. Negative reference, insult 9. Extended quotes 10. Famous names, sources
Numbers 1 (Negative words) and 6 (Passive voice) were predicted by psychological
studies on deception mentioned earlier (Newman, Berry & Richards, 2003). Their occurrence in
study samples was consistent and marked, with negative words especially abundant in Anti
Democratic and Pro- and Anti-Trump samples. However, further research is needed to compare
the occurrence of negatives lexical items to positive ones in FN, and to compare negativity in
false news to negativity in bona fide news. For example, two sample articles from the front page
of the New York Times on April 28, 2017, showed relatively high negative word occurrence,
especially in a story about the possibility of war with North Korea (suggesting that the subject of
a story might have as much to do with negative word frequency as deceptive intent).
FAKE NEWS 23
The use of the passive voice in FN samples tended to obscure the doer of an action,
allowing for unprovable statements like “…they were denied entry,” leaving uncertainty about
who wouldn’t let them in. Without a clear agent for the action, a reader can’t even begin to
assess the validity of a statement. Journalists are taught to avoid passives, not just because they
can be confusing, but because they tend to weaken the force of the prose.
While numbers 2 (Hyperbole/superlatives), 8 (Negative reference, insult), and 10
(Famous names, sources) may seem unsurprising, these features weren’t identified in deception
literature. All three of them go against fundamental journalistic principles, in fact may be
libelous. But anything goes on the Internet and in social media, it seems.
Number 6 (News-like detail) is less a feature of FN than a method used by writers to
disguise FN as trustworthy journalism, similar to the disguised URL’s noted earlier. The same
effort to disguise can be seen in feature numbers 11 (News format) and 19 (Unbiased
appearance). When laid out with a headline, byline, and dateline, like a print newspaper article,
with supporting facts and figures and the semblance several different viewpoints, an FN item can
all too easily seem real.
The presence of all three Intertextual features—3 (False quotes), 4 (Unattributed
sources), and 9 (Extended quotes)—in top frequency positions reflects the way a news article
interacts with other written or spoken texts made by experts, involved parties, or witnesses, i.e.,
quotes. The news story genre has strict conventions about the use of other peoples’ words,
observations, and ideas, for instance, that the writer must quote and attribute anything he or she
didn’t observe at first hand. At the same time, quotes are thought to lend color and credibility,
with one quote per paragraph a common reporter’s rule of thumb. Only on rare occasions do
FAKE NEWS 24
editors allow quotes from generic un-named sources (a spokesperson, expert, scientific report,
etc.)
Based on the loose handling of quotes observed in study samples, it could be claimed that
FN is an altogether different genre of communication than real news. Unattributed or
generically-sourced quotes occurred commonly in study samples, and several FN stories were
really nothing more than extended false quotes, as in “Everyone Says They Want to Go to Africa
Or Mexico, Well Here’s Your Chance! Make America Great Again and Go Back to Your
Country for Free—Trump.” The samples were also replete with information reported at second
hand, unattributed, such as “Obama Passes Law for Grandparents to Get All their Grandchildren
Every Weekend.” Even more often, samples brazenly included quotes that readers didn’t even
have to Google to know untrue, as in Trump calling Obama “an illegitimate Muslim traitor.”
Survey respondents reported the necessity of validating an article by checking others news
websites for corroboration, but were far less focused on the need for skepticism when internal
quotes are mishandled.
Only one FN feature from the Strategic category appeared on the top 10 list:
editorializing (number 9), which means to blur the border between fact and opinion, a
journalistic taboo. However, frequency may not be the best way to judge strategy in writing
because each story generally has only one or two basic “moves,” and it takes concerted effort to
analyze and identify them, which the casual news reader may be uninclined to do.
Two of the Strategic features found in FN samples (Red herring and False dilemma)
commonly appear on lists of logical fallacies, which teachers warn writing students to avoid.
Logical fallacies have been defined by Weber and Brizee (2013) as “common errors in reasoning
that will undermine the logic of your argument, either illegitimate arguments or irrelevant points,
FAKE NEWS 25
identified because they lack evidence that supports their claims.” They typically include “straw
man” arguments, or oversimplifying a viewpoint in order to attack it; “ad hominem,” which
means to assail a person’s character rather than the view he or she espouses; and “either-or”
simplification of a far more complicated issue. In the world of FN, however, strategies for
avoiding composition problems are turned inside-out and used to create the problems they were
intended to avoid.
Taken in sum, the FN features identified and assessed for frequency in this study describe
a sub-genre of Internet-based news stories, though journalists and experienced news consumers
would strenuously resist its admission into the journalism category. Perhaps, FN is more a sub
genre of deception or of writing, in general. In any event, using the study’s 22 FN features to
define and categorize a genre may be less important, on a pedagogical level, than teaching the
features to learners to sharpen their reading, writing, and critical thinking skills.
Surveys conducted for this study show that highly-educated people, both young and old,
are perfectly able to separate true from false news. But it should be recalled that American
Internet news consumers, at large, believed FN stories to be accurate 75% of the time during the
2016 elections (Silverman, 2016). And much more concerning are the findings of a recent
Stanford University study showing that 82% of middle school students couldn’t tell the
difference between an ad labeled “sponsored content” on a website and news, and 40% of the
young people studied decided the area around Fukushima, Japan, was toxic because of a false
headline and a photo of a deformed daisy (Shellenberger, 2016).
The danger, according to K-12 educators, is that teenagers whose only exposure to news
is through their smartphones grow up lacking the critical ability to determine what is true and
what is not (Barron, 2017). Silverman, Buzzfeed media editor, said that “we need our schools to
FAKE NEWS 26
start thinking about how we integrate more media literacy and critical thinking education so that
people can make better judgments for themselves” (Davies, 2016).
Fortunately, many educational institutions and organizations have responded to the
problem of media illiteracy in an age of FN, including Stanford History Education Group, which
offers a free, downloadable social studies curriculum that teaches teenagers how to assess the
trustworthiness of sources; the education technology corporation Newsela; and Stony Brook
University’s Center for Media Literacy. Their modules on media literacy align well with
established standards, such as the State of California’s Common Core objectives for English
language arts and literacy in history/social studies (Appendix D).
The present study suggests a number of teaching methods that can be added to media
literary curriculums, not least the list of FN features identified through sample analysis, which
can serve as a useful tool for helping learners recognize doubtful news and sources. Students
could be asked to follow simplified genre analysis procedures to deduce their own lists of FN
characteristics from samples, in the process sharpening critical thinking and reading skills.
It is possible, as well, to incorporate FN recognition in general writing classes by having
students consider how deceptive news stories confound composition taboos against red herrings,
editorializing, oversimplification, and other logical fallacies. Lessons about correct use of
sources can involve examples of some of the ways FN mishandles them. Finally, learners can
and should be exposed to fundamental ethical issues that frame the relationship between honest
writer and reader. As William (2016) elegantly said, “The social contract between thoughtful
writers and readers implies a goodwill exchange fair to everyone and in the long-term best
interest of both” (p. 222).
FAKE NEWS 27
References
Adenzato, M. & Bucciarelli, M. (2010). Recognition of mistakes and deceits in communicative interactions. Journal of Pragmatics, 40(4), 608-629.
Allen, W. & Blinder, S. (2013). Migration in the news: Portrayals of immigrants, migrants, asylum seekers, and refugees in national British newspapers, 2010 to 2012. Retrieve from: http://www.migrationobservatory.ox.ac.uk/wp-content/uploads/2016/04/ReportMigration_News.pdf
Baker, P., Gabrielatos, C. & McEnery, T. (2012). Sketching Muslims: A corpus driven analysis of representations around the word “Muslim” in the British press 1998-2009. Applied Linguistics, 34(3), 255-278.
Balmas, M. (2012). When fake news becomes real: Combined exposure to multiple news sources and political attitudes of inefficacy, alienation, and cynicism. Communication Research, 41(3), 430-454.
Barron, J. (2017). In an era of fake news, teaching students to parse fact from fiction. Retrieved from: https://www.nytimes.com/2017/03/20/nyregion/fake-news-brooklyn-middle- school.html?_r=0
Bazerman, C. (2004). Speech acts, genres, and activity systems: How texts organize activity and people. C. Bazerman & P. Prior (Eds.)., What writing does and how it does it: An introduction to analyzing texts and textual practices (309-339). Mahwah, NJ: Lawrence Erlbaum.
Bazerman, C. & Prior, P. (Eds.). (2004). What writing does and how it does it: An introduction to analyzing texts and textual practices (p. 2). Mahwah, NJ: Lawrence Erlbaum.
Bedford, D. (2010). Agnotology as a teaching tool: Learning climate science by studying misinformation. Journal of Geography, 109(4), 1-9.
Belcher, D., Johns, A.M. & Paltridge, B. (2011). New directions in English for specific purposes research. Ann Arbor, MI: University of Michigan Press.
Bhatia, V. (1993). Analyzing genre: Language use in professional settings. New York, NY: Longman.
Bhatia, V. (2009). Intertextual patterns in English legal discourse. In D. Belcher (Ed.)., English for specific purposes in theory and practice (pp. 186-204). Ann Arbor, MI: University of Michigan Press.
FAKE NEWS 28
California State Board of Education (2013). Common Core state standards: English language arts & literacy in history/social studies, science, and technical subjects. Retrieved from: http://www.cde.ca.gov/be/st/ss/documents/finalelaccssstandards.pdf
Croissant, J.L. (2014). Agnotology: Ignorance and absence or toward a sociology of things that aren’t there. Social Epistemology, 28(1), 1-19.
Croxton, W., & Gonsalves, R.C. (2017). What’s “fake news?” 60 Minutes producers investigate. Retrieved from: http://www.cbsnews.com/news/whats-fake-news-60-minutes-producersinvestigate/
Davies, D. (2016). Fake new expert on how false stories spread and why people believe them. Retrieved from: http://www.npr.org/2016/12/14/505547295/fake-news-expert-on-howfalse-stories-spread-and-why-people-believe-them
Dilmon, R. (2009). Between thinking and speaking—linguistic tools for detecting a fabrication. Journal of Pragmatics, 41(6), 1152-1170.
Grynbaum, M. (2017, February 17). Trump calls the news media the ‘enemy of the American people.’ The New York Times. Retrieved from https://www.nytimes.com/2017/02/17/business/trump-calls-the-news-media-the-enemy- of-the-people.html?_r=0
Hernandez, J. (2017, February 3). China’s response to reports of torture: Fake news. The New York Times. Retrieved from https://www.nytimes.com/2017/03/03/world/asia/chinafake-news-xie.html?_r=0
Huckin, T. (2004). Content analysis: What texts talk about. What writing does and how it does it: An introduction to analyzing texts and textual practices (p. 13). Mahwah, NJ: Lawrence Erlbaum.
Jackson, C. (2016). Ipsos/BuzzFeed poll – fake news. Retrieved from http://ipsos-na.com/news-polls/pressrelease.aspx?id=7497
Lie. (2003). Retrieved March 7, from https://www.merriam-webster.com/dictionary/lie
Loftus, E.F. (2005). Planting misinformation in the human mind: A 30-year investigation of the malleability of memory. Learning & Memory, 12(4), 361-366.
Millar, D. (2011). Promoting genre awareness in the EFL classroom. English Teacher Forum 49(2), 4.
(Newman, L., Berry, D., & Richards, J. (2003). Lying words: Predicting deception from linguistic styles. PSPB, 29(5), 665-675.
FAKE NEWS 29
Obama, B. (2017). Read the full transcript of President Obama’s farewell speech. The Los Angeles Times. Retrieved from http://www.latimes.com/politics/la-pol-obama-farewell- speech-transcript-20170110-story.html
Paltridge, B. & Starfield, S. (2013). The handbook of English for specific purposes. Malden, MA: John Wiley & Sons.
Pogue, D. (2017) How to stamp out fake news. Scientific American, 316(2), 24.
Price, R. (2017). Facebook will teach you how to spot fake news. Business Insider. Retrieved from: http://www.businessinsider.com/facebook-how-to-spot-fake-news-2017-4
Selzer, J. (2004). Rhetorical analysis: Understanding how texts persuade readers. C. Bazerman & P. Prior (Eds.)., What writing does and how it does it: An introduction to analyzing texts and textual practices (309-339). Mahwah, NJ: Lawrence Erlbaum.
Shellenberger, S. (2016). Most students don’t know when news is fake, Stanford study finds. Retrieved from: https://www.wsj.com/articles/most-students-dont-know-when-news-is- fake-stanford-study-finds-1479752576
Silverman, C. (2016). Here are 50 of the biggest fake news hits on Facebook from 2016. Retrieved from https://www.buzzfeed.com/craigsilverman/top-fake-news-of- 2016?utm_term=.siz2aOvQo#.uiE6kGA9y
Smith, R., & Sydell, L. (2016). Planet Money finding the fake news king. Retrieved from: http://www.npr.org/sections/money/2016/12/02/504155809/episode-739-finding-thefake-news-king
Swales, J. (1990). Genre analysis: English in academic and research settings. Cambridge, UK: Cambridge University Press.
Tarone, E., Dwyer, S., Gillette, S., & Ickes, V. (1981). On the use of the passive in two astrophysics journal papers. English for Specific Purposes, 1(1), 123-140.
Van Swol, L, Braun, M. & Malhotra, D. (2011). Evidence for the Pinocchio Effect: Linguistic differences between lies, deception by omissions, and truths. Discourse Processes, 49(2), 1-37.
Wang, W. (2008). Intertextual aspects of Chinese newspaper commentaries on the events of 9/11. Discourse Studies, 10(3), 361-381).
Weber, R. & Brizee, A. (2013. Logical fallacies. Owl Purdue Online Writing Lab. Retrieved from: https://owl.english.purdue.edu/owl/resource/659/3/
Williams, J. (2016). Style: Lessons in clarity and grace. New York, NY: Pearson.
FAKE NEWS 30
Features of Fake News
Susan Spano
Middlebury Institute of International Studies at Monterey
Susan Spano is a Master’s Degree candidate in the TESOL Department at Middlebury
Institute of International Studies.
Correspondence concerning this article should be addressed to Susan Spano, Middlebury
Institute of International Studies, 460 Pierce St., Monterey, CA 93940. Contact:
susan.spano@gmail.com
FAKE NEWS 2
Abstract
Deception is as old as Satan. Lately, though, it has taken an alarming spin: “Fake News,” (FN).
The phenomenon appeared on the public radar screen in 2016, skewing presidential election
results, some claim, motivating violent conspiracy theorists, and substituting factionalizing
fabrications for free, unbiased news reporting. Media studies showed that Americans believed
FN news reports to be accurate 75% of the time during the last presidential election cycle
(Jackson, 2016), and Facebook’s top 20 FN stories were clicked on more often last year than the
top 20 real ones (Pogue, 2017). But while much has been written about FN, there is little
empirical evidence that can help us better understand it. The purpose of this study is to provide
such evidence by using Genre Analysis, as well as the findings of psychological, sociological,
and communications research studies on the topic of deception, to identify features common in
FN. BuzzfeedNews.com’s top 50 FN stories (Silverman, 2016), which will serve as a primary
data source. Surveys provide additional data for analysis and triangulation. The results of this
study will help news consumers separate the wheat from the chaff, and give media literacy
educators a useful tool for helping young people and ESL learners exercise critical thinking skills
when they go online for the news.
Keywords: fake news, genre analysis, deception, lies, media literacy
FAKE NEWS 3
When We Practice to Deceive:
Features of Fake News
President Obama signs executive order banning the Pledge of Allegiance in schools nationwide—abcnews.com.co
ISIS leader calls for American Muslims to support Hillary Clinton— worldnewdailyreport.com
Morgue worker arrested after giving birth to a dead man’s baby—tmzworldstar.com
These are headlines from three of the most frequently-viewed, Facebook-disseminated
“Fake News” (FN) stories published last year (Silverman, 2016). Many people knew at a glance
that they were blatantly untrue. But many others didn’t. A poll conducted by Ipsos Global, a
marketing and the digital media company, and BuzzfeedNews (2016) found that Americans
believed FN stories to be accurate 75% of the time during the last presidential election cycle.
Moreover, Facebook’s top 20 FN stories were clicked on more often last year than the top 20 real
ones (Pogue, 2017).
There might be little danger in the proliferation of FN if it were always as obviously false
and innocuous, as in the story about the morgue worker. But claims persist that FN helped to
swing the 2016 presidential election against candidate Hillary Clinton. Last December, a man
driven by an Internet-propagated conspiracy theory fired a rifle into a Washington, D.C., pizzeria
where, according to FN reports, high-ranking Democrats were operating a child sex ring.
FN spreads, as well, in non-Internet form, for instance when presidential counselor
Kellyanne Conway made repeated public references to a terrorist massacre in Bowling Green,
KY, that never happened. And it appears to have gone global, with the Chinese state media now
using the FN terms to refute western news reports about the use of torture in the PRC
(Hernandez, 2017).
FAKE NEWS 4
The invasiveness of FN has been driven partly by the public’s increasing reliance on
curated Internet news sites—The Huffington Post, LifeHacker, Reddit, etc.—instead of primary
sources like The New York Times and The Economist. Another factor feeding FN is the tendency
among readers to favor news aligned with their own beliefs over opposite reports by reputable
media outlets. As President Obama (2017) said, “Increasingly, we become so secure in our
bubbles that we start accepting only information, whether it’s true or not, that fits our opinions,
instead of basing our opinions on the evidence that is out there.” Balmas (2014) stated that
people who often expose themselves to FN are more likely to believe it than people who rely on
news outlets generally deemed more reliable.
In other words, it’s buyer beware.
But when news readers are young or novice English language learners, with as yet
unsharpened critical thinking skills, coming of age in the brave new world of digital and social
media, the danger of FN is heightened. Misinformation and ignorance has reached such a point
that there is now a cross-disciplinary field to study it: Agnotology, investigating why people
don’t know things like the scientific facts supporting climate change and the health hazards of
cigarettes. Bedford (2010) noted that the effort to misinform can take several forms, including
blurring distinctions between what is known and what is not known and claiming disagreement
or uncertainty about facts where none exists.
Croissant (2014) saw Agnotology more as a study of ignorance than of misinformation,
putting it in three categories: willful ignorance in relation to something known, unconscious
ignorance, and finally not knowing what is truly unknown. However, both intentionally
propagated falsehood (as in FN) and ignorance sometimes share a cognition-based feature, the
conflation of cause and effect. Moreover, both lies and the seeding of ignorance “may be
FAKE NEWS 5
similarly motivated by a multiplicity of factors linked together by considerations of social
power” (p. 7).
This study will reference Agnotology, theories and findings about deceit from
psychology, sociology, information management, and applied linguistics to peer into the
phenomenon of FN. The project’s purpose is to seek through texts for common FN features.
Therefore, this study will take up these two research questions:
1. What are the features of FN?
2. Do educated, adult news consumers know FN when they see it, and if so, what
features do they use to identify it?
Background
To understand FN in its broadest context, it is worth considering a number of cross
disciplinary findings about lying in general, including psychological research on how people
communicate when lying, often avoiding the first person singular, using cognitively simple words and
phrases, and conveying an over-riding sense of negativity (Newman, Berry & Richards, 2003).
Psycholinguists Arciulu, Mallard, and Villar (2010) added that deceptive discourse often reflects
a paucity of detail and the absence of fillers like um. People lie at least once a day, the authors
reported, but the ability to recognize a lie remains basically a matter of chance.
In a review of a wide range of psychological studies on deception, Adenzato and Bucciarelli
(2010) said that children begin telling simple lies by the age of three. The authors went on to
describe their own research on how children from age 4 to 11 recognize mistakes and deceit. The
study’s findings suggest that youngsters get better at spotting lies and mistakes as they get older,
and more surprisingly, that they are best at distinguishing complex, rather than simple lies.
FAKE NEWS 6
Another phenomenon that has long fascinated psychologists and neuroscientists is the
Misinformation Effect which makes people remember things that never happened when they are
exposed to falsehoods. Loftus (2005) said that a seminal study using neuroimaging technology to
track how subjects recalled an event they had viewed after getting false information about it
found that “misinformation was remembered as being part of the original event 47% of the time”
(p. 361).
Deceit research in psychology, sociology, and communications has often attempted to
analyze and categorize falsehoods by apparent intent and by types, such as lies of commission
and lies of omission. According to Merriam-Webster’s Collegiate Dictionary (2003), the verb to
lie means “to make an untrue statement with intent [my italics] to deceive.”
Intentional fabrication has long played a role in American politics (known as
mudslinging). Stories about Ulysses S. Grant’s sexual peccadillos proliferated in the 1872
presidential campaign. But driven by the Internet, FN today has “gone viral,” so to speak. It
appeared on the radar screen during the 2016 election, as stories about Clinton paying people to
protest at Trump rallies and getting support from ISIS reverberated in the echo chamber of the
Internet. Legitimate news sources had to make difficult determinations about whether to report
shocking FN falsehoods propagated on the Internet and in public discourse.
Snopes.com, Factchecker.org, and other groups attempted to keep tabs on the veracity of
news reports, and BuzzFeedNews went a step farther by tracking down websites, servers, domain
codes, and hit patterns to locate the sources of some of the most egregious FN stories. In one
case, BuzzFeedNews found hundreds of false pro-Trump stories coming from a small town in
Montenegro where people with computer skills had figured out how to make money by creating
FN websites (with cleverly-disguised URL’s) linked to advertising engines like Google AdSense
FAKE NEWS 7
(Silverman, 2016). In another case, an L.A. entrepreneur who was tracked down by journalists
said he made $20,000 to $30,000 a month writing anti-Clinton FN stories at the height of the
presidential campaign last year (Smith & Sydell, 2016). Perhaps sadder still, Justin Coler, an FN
writer interviewed by the CBS news show, 60 Minutes, said that seeing the audience for his
made-up stories explode in social media was “like an addiction” (Croxton & Gonsalves, 2017).
Facebook is not the only purveyor of FN. But it is one of the biggest. According to
Silverman (2016), two billion people logon to Facebook every month and 44% of adult
Americans get their news from it. So, when FN began making news itself last fall, commentators
started urging Facebook to do something about it. Founder Mark Zuckerberg was initially
reluctant to get involved, saying that Facebook is a tech not a media company. But both
Facebook and Google Adsense eventually instituted policies using sophisticated computer
algorithms to try to find and block FN postings, though the accuracy of the algorithm approach
has been criticized (Silverman, 2016) and FN keeps flowing, as evidenced by BuzzfeedNew’s
list of last year’s top 50 FN stories (Appendix A).
Definition of FN
The term FN is now so widely used for such a variety of phenomena that even the
producers of 60 Minutes had trouble defining it (Croxton & Gonsalves, 2017). To narrow it
down and put it in the context of this study, it is useful to begin with what it is not. Political
satire of the sort produced by Saturday Night Live and The Daily Show will not be considered
FN. Uses of the term as argumentative rebuttal and derogation in political discourse, as when
President Trump Tweeted that the mainstream media deals in FN (Grynbaum, 2017), will not
meet this paper’s FN criteria.
FAKE NEWS 8
The present study will address itself exclusively to Internet-based FN stories on political
topics, beginning with the BuzzfeedNews top 50 FN stories of 2016. Since the list’s publication,
however, some of the stories have been removed from the Internet and others are non-political,
tabloid-style reports of oddities, as in “Morgue worker arrested after giving birth to a dead man’s
baby” (Silverman, 2016). Consequently, data samples have been added to reach a total of 40
stories from websites identified by multiple sources (The New York Times, CBS News,
Scientific American, etc.) as disseminators of FN (Appendix B).
A definition that suits the study’s purposes can be induced from the headlines of those
stories and from understanding the background of FN, as rehearsed above. In this study, then, FN
can be defined as:
News reports based on provably false information, disseminated on the Internet for
political purposes.
Genre Theory
With a primary aim of identifying recognizable linguistic features in recent samples of
FN, this study fits snugly under the aegis of Genre Theory. Coming of age with the rise of
English for Specific Purposes (ESP) in the 1970’s and 1980’s, the theory looks at written and
oral texts as genres shared by communities of learners with special interests (like science and
technology). Defined early on as a communicative event with a common purpose (Swales, 1990),
genre has since then been reinterpreted through myriad lenses. Bazerman (2004) called genres
“psycho-social recognition phenomena that are parts of processes of socially organized
activities” (p. 317), and Wang (2008) said that genre study “attempts to capture how writers
achieve their social purposes by using various structural forms, constructing different focuses,
and manipulating topics and readers by using various linguistic devices” (p. 363).
FAKE NEWS 9
The theory has further evolved, with the study of genres subdividing into numerous, but
not mutually-exclusive strands. One important subdivision is Discourse Analysis, which
interprets communication “beyond the sentence,” partly inspired Grice’s Cooperative Principle.
Grice proposed four maxims for successful communication: the maxim of quality (speech based
on truth and evidence), the maxim of quantity (suitably informative speech, neither too long or
short), the maxim of relevance (on-topic speech), and the maxim of manner (clarity and
directness of speech). Dilmon (2009) said that deception obstructs all four maxims when an
interlocutor is made to believe that a deceiver is being faithful to the Cooperative Principle in
general (p. 1154).
Discourse Analysis has been termed Text Analysis when focused on the written word in
its cultural setting. As such, analysis of a written genre can move beyond semantical, syntactical,
pragmatic interpretation to contextual feature like what has been omitted, which, as we have
seen, is a factor of interest to psychological deception researchers (Bazerman & Prior, 2004).
Swales (1990) broadened Genre Analysis further by proposing that samples of a
particular genre follow a number of common steps— “moves,” he called them; ergo, Move
Theory. Millar (2011) said the moves involved in the recipe genre, for instance, include “the
name of the dish, a list of ingredients and amounts, the steps for cooking, and the number of
people the dish serves” (p. 4). Move Theory has great relevance to FN, a pseudo-form of the
news story. This well-established journalistic genre, as all novice reporters learn, covers who,
what, when, where, and how, in the first paragraph, and progresses from the most important facts
to less vital ones toward the article’s end.
Move Theory is closely related to the macro-level textual exploration of Rhetorical
Analysis. As defined by Selzer (2004), Rhetorical Analysis is “an effort to understand how
FAKE NEWS 10
people within specific social situations attempt to influence others through language” (p. 281).
Among the textual features assessed in Rhetorical Analysis are semantic connotation, hyperbole,
ethos/pathos, and epideictics (an Aristotelian form of declamation invoking concepts related to
virtue and vice).
Intertextual Analysis takes the fascinating approach of looking at a text in relation to
other texts, especially how outside sources are cited (Bazerman, 2004, p. 87). Using the
intertextual approach, Wang (2008) studied Chinese news commentary about 9/11, finding that
authors attempted to distance themselves from their sources in response to socio-cultural factors
like the role played by the media in Chinese politics (Wang, 2008).
In English language and literature, Williams (2016) postulated rules for prose writing
based on ethics. Echoing the Gricean idea that communication depends on commonly-held
principles, Williams said, “The social contract between thoughtful writers and readers implies a
goodwill exchange fair to everyone and in the long-term best interest of both” (p. 222). Stylistic
features he identified, such as abstraction, overuse of the passive voice and nominalized verbs,
oversimplification, and intended obscurity, tend to break the reader-writer contract and may
signpost deception.
On a methodological level, almost all the fields of Genre Analysis mentioned above have
tapped Corpus Linguistics to assess lying texts. Research in psychology, sociology, and
communications (Van Swol, Braun & Malhotra, 2011; Allen & Blinder, 2013; Baker,
Gabrielatos & McEnery, 2013) has also relied on corpus analysis, often focused on lexical
grammatical commonalities in deception, usually by seeking collocates. However, most studies
of this kind look at dyadic, oral forms of lies and posit that underlying tensions in the mind of the
liar reveal themselves linguistically in liar texts (Newman, Berry, & Richards, 2003). As a
FAKE NEWS 11
written form, composed for publication, FN would seem less likely to show itself in that way,
though some of the features social psychologists have studied in deceptive texts—low cognitive
complexity at the sentence level, minimal use of the first person singular pronoun, and a high
occurrence of words that have to do with negative emotions—can play a role in the textual
analysis that lies at the heart of this study (Newman, Berry & Richards, 2003).
Another textual feature often assayed through corpus analysis is total word count, easy
enough to determine digitally, but much harder to understand as a possible feature of lying texts.
Van Swol, Braun & Malhotra (2011) reported varying speculation about word count. Some
studies have found high word counts for fabrications, especially when the deceiver knows that
the listener can’t confirm or disprove what he or she hears. Other studies report that low word
counts are more common in lying texts, particularly when the liar says as little as possible to
avoid touching upon something the listener might already know.
Research Method
Research Question #1: What are the features of FN?
Bhatia (1993) outlined a genre analysis system that the present study will adopt:
1. Describe the context. 2. Conduct a literature review. 3. Define the text’s contextual framework in further detail. 4. Locate samples. 5. Analyze the samples. 6. Consult a specialist in the genre about the findings.
The author will begin inductively by qualitatively examining FN samples for
characteristics that stand out and recur. This process will be informed by the deception features
touched upon in the literature of lying reviewed above, including rhetorical moves, use of
sources, authorial distancing by avoidance of the first person singular, negative emotion words
and profanity, word count, etc. Other factors arise from FN context, for instance, disguised
FAKE NEWS 12
URL’s, links to more FN, ads appearing on the same page as the FN sample, along with
sometimes hyperbolic punctuation.
Seeking out features of FN is a way of identifying and studying it as genre clearly related
to, but not at all the same as, the news story, a well-established journalism genre. In writing news
stories, reporters follow such tried-and-true rules (or moves) as:
• The first paragraph, or lede, should include all the vital information, the who, what, when, where, and how. • The story should have the form of an “Inverted Pyramid,” with all the most important facts at the top, followed by less important detail. • Facts and figures must be substantiated. • All quotes should be verified and clearly identified. • Writers must be impartial and objective; they should not editorialize. • Advertisement must be kept separate from news.
The absence or presence of factors such as these will also play a role in my initial search for FN
features.
Once a critical mass of samples has been assessed inductively, a list of possible FN
features is expected to arise. The list will be winnowed down, prioritizing features that have
occurred in the highest frequency, and then used in the analysis of the rest of the data.
This precise approach has been used in Discourse Analysis. Barton (2004) described it as
“qualitative and inductive, with basic quantitative verification” (p. 63), a process that involves
identifying features by qualitatively scrutinizing samples, confirming their occurrence over a
corpus of texts, analyzing their occurrence quantitatively, and finally drawing thematic
conclusions.
Research Question #2: Do educated, adult readers know FN when they see it? What
strategies do they use to identify it?
Brief surveys (Appendix C) will be given to the author’s Facebook friends and to
students at the Middlebury Institute of International Studies at Monterey. In order to maximize
FAKE NEWS 13
data, the surveys will be easy-to-take, beginning with a brief FN text, the number one FN news
story of 2016, “Obama signs executive order banning the Pledge of Allegiance in schools
nationwide” (Silverman, 2016). Respondents will be asked to say whether it is true or false.
Next, they will be asked whether they can recognize FN, using a 5-point Likert scale (1 never to
5 always). Finally, they will be invited to list and describe the features they use to identify FN
stories. While participants will constitute a convenience sample, they are expected to be savvy
Internet media users. The responses will undergo quantitative and qualitative analysis, hopefully
providing human confirmation of the study results.
Results
Research Question #1: What are the features of FN?
Identifying features to use in FN analysis was a dynamic process. Some items that at first
seemed important (for instance, dated quotes and non-functioning links) proved less significant
the more samples I assessed. I looked for features intrinsically by studying 10 FN texts and
extrinsically by checking to see if features mentioned in studies and articles were evident in
them. Ultimately, I arrived at 22 features in 4 categories:
Table 1. FN feature categories
Then I looked for these features in a total of 40 FN samples, including 25 articles from
Buzzfeed’s top 50 FN stories and 15 from other FN sources.
1. Word and Sentence—involving lexis and superlative/hyperbole 2. Contextual—features arising from the context 3. Intertextual—pertaining to quotations and the use of sources 4. Strategic—macro-level “moves” or rhetorical approaches
FAKE NEWS 14
Table 2. Frequency of FN features _____________________________________________________________________________________
FN Features Frequency/ Sample/Description Number of Samples in which the Feature Occurred ______________________________________________________________________________
1. Negative words 40 kill, prison, bigot 2. Hyperbole 32 greatest children out there 3. False quotes 32 Fabricated source or words 4. Unattributed sources 29 experts, sources, data 5. News-like detail 28 Facts, figures, names, dates 6. Passive voice 26 were taken, was found 7. Editorializing 26 he admitted, surprisingly 8. Negative reference, insult 26 (9 Islam) Abusive characterization 9. Extended quotes 22 Article is basically one quote 10. Famous names, sources 21 Michael Jordan, Fox News 11. News format 19 Looks like a news story 12. Sex words, phrases 18 masturbation, hot girls 13. Red herring 16 Distracting argument 14. Platitudes 16 In God We Trust 15. Oblique defamation 15 Triangular derision 16. Name-calling 14 Crooked Hillary 17. Poor copyediting 13 Spelling, grammar 18. Idiomatic expressions 11 wanna, gonna 19. Unbiased appearance 10 Looks fair, 2 sides 20. Extreme punctuation 9 Caps, exclamation points 21. Profanity 9 hell, fart 22. False Dilemma 3 Two choices only ______________________________________________________________________________
FAKE NEWS 15
Table 3. Description of top 10 FN features
1. Negative words evoke negative emotions or reactions in readers: • “Individuals who violate this order can face fines of up to $10,000 and up to one year in federal prison.”— Jimmy Rustling, abcnews.com • “People all over the country have been flooding my office with calls, telling my staff of horror stories about being harassed and intimidated by poll workers,” Obama told reporters. 2. Hyperbole/superlative involve clearly exaggerated statements or claims. • “The Pledge of Allegiance helps our children be the greatest children out there, the best there is, greater than all the other children in this world.”—Trump • “I personally went to two Donald Trump rallies and I can say with 100% certainty that none of the protesters were getting paid.”—Jimmy Rustling, abcnews.com 3. False quotes are made-up statements, either obviously false or shown false in a Google search. • “I cannot allow him [Trump] into this chair, with his finger so close to the button. The power would go to his head immediately.”—Obama • “Hillary Clinton looks like someone took a leprechaun, put make-up on it, and told him to accept money from whomever it wants if that get him the presidency.”—Pence 4. Unattributed sources are quotes, facts, and statements that lack specific source citation. • “With such few members, but billions in real estate holdings, the Church of Scientology has been called ‘the most famous small business in the world.’”—abcnews.com • “Christian Times Newspaper has learned that Guccifer, the Romanian hacker currently being held on charges for hacking Hillary Clinton’s personal email server, has been found in his Virginia jail cell, dead of an apparent suicide.”—Dave Hodges 5. New-like detail is an FN feature that shows itself in blatantly false stories, made to look true by inclusion of facts, figures, and the kind of specific detail that is a critical component of journalism. • “The liberal Democrats are openly hostile to Christians,” said Graham in an interview on WEHW radio. • “A new survey released today from the Make Michigan Great Again PAC shows that residents across the state rated Flint tap water higher on the ‘Trustworthiness’ scale than Democratic nominee for president, Hillary Rodham Clinton.”—nationalreport.net 6. Use of the passive voice obscures the doer of the action, i.e. the subject of the verb. • “Many have even said that they were flat out denied entry into the voting booths…”—usanews.com • “She [Beyoncé] was seen storming off stage and leaving the venue in a limousine shortly thereafter.”— thelastlineofdefense.org 7. Editorializing has to do with expressing an opinion, sometimes covertly. • “Barack Obama has sensationally told CNN’s Wolf Blitzer that he will not vacate the Oval office if Donald J. Trump is elected.”—burrandstreetjournal.com • “And vote many times he did.”—thelastlineofdefense.org 8. Negative reference to a minority or special interest group, or insult to an individual. • “…women are still devoid of a soul but have been shown to possess qualities common to the mammal species.”—worldnewsdailyreport.com • “…do we really need to listen to this [Beyoncé] ghetto trash music?”—thelastlineofdefense.org 9. Extended quotes occur in stories that are basically one extended quotation, or reference to a quotation. • “Actor Bill Murray Announces 2016 Presidential Run”—abcnews.com • “Trump Claims America Should Never Have Given Canada Its Independence”—burrardstreetjournal.com 10. Famous names, sources are extraneous references to famous people or news sources. • “[George] Soros-funded left wing groups…”—dangerandplay.com • “…by making short jabs at the likes of Bill O’Reilly…”—theearthchild.co.za
FAKE NEWS 16
Table 4. Feature frequency by type
Word and Sentence 1. Negative words 2. Hyperbole 3. Passive voice 4. Sex words 5. Idioms 6. Profanity
Contextual 1. News-like detail 2. Negative reference 3. Famous names 4. News format 5. Platitudes 6. Name-calling 7. Poor copyediting 8. Extreme punctuation
Intertextual 1. False quote 2. Attribution 3. Extended quote
Strategic 1. Editorializing 2. Red herring 3. Oblique 4. Unbiased 5. False dilemma
Total word counts were tabulated for each sample, but ultimately seemed of uncertain
value as an FN indicator. I compared the average story length of 499 words found in the samples
to the length of two front-page New York Times stories from April 28, 2017, both of which had
word counts of just over 1,000. This suggests that FN stories tend to be shorter than real ones,
but further research would be needed to draw any firm conclusions about the relationship
between length and FN.
Table 5. FN samples word count
It was possible to distinguish the political leaning of 31 out of 40 samples: 20 Anti
Democratic (Obama and Clinton), 2 Pro-Democratic (Obama and Clinton), 6 Pro-Trump, and 3
Anti-Trump. The remaining 9 samples had to do with religious, health, race, and women’s
issues.
Mean words—499
Median words—414
Highest—1,739 words
Lowest—38 words
FAKE NEWS 17
Table 6. FN samples political leaning by top feature in each category
Leaning Negative words News detail False quote Editorializing
Anti-Democratic (Obama, Clinton) 22 samples total
Yes in 20 samples (1-31 instances per sample) 91%
Yes all Yes in 19 samples Yes in 17 samples
Pro-Democratic (Obama, Clinton) 2 samples total
Yes in 1 sample (5 instances) 50%
Yes all Yes in 2 samples Yes in 1 sample
Pro-Trump 6 samples total
Yes in 6 samples (1-28 instances per sample) 100%
Yes all Yes in 7 samples Yes in 5 samples
Anti-Trump 3 samples total
Yes in 3 samples (3-7 instances per sample) 100%
Yes all Yes in 2 samples Yes in 2 samples
Two additional factors were noted but not specifically addressed in the FN sample
assessment, the first URL’s. Given that these are created by FN purveyors, it’s no surprise that
they can be misleading. Many are reputable-sounding, for instance http://empireherald.com,
http://worldnewsdailyreport.net, and http://nationalreport.net. Others are extremely close to
established news sources; compare the FN site http://abcnews.com.co to the real ABC News site
http://abcnews.go.com. Others are clearly satirical or doubtful, as in http://tmzhiphop.com and
http://straightstoned.com.
Finally, the look and content of the website surrounding a piece of FN can be telling.
Almost all FN samples in this study had links for additional FN headlines, some potentially
credible, others of a ridiculously vulgar tabloid sort. In many cases, when you click such a link,
you’re led eventually to an advertisement for pornography or super-vitamins. The products alone
could make a reader suspicious. However, keeping news separate from paid advertising is an
FAKE NEWS 18
inviolable principle of traditional journalism, so the blurring of the line between ads and news in
these samples can be seen as another FN marker.
Figure 1. FN ads and links
Research Question #2: Do educated, adult readers know FN when they see it? What
strategies do they use to identify it?
Survey I, sent to my Facebook friends, reported findings from 27 respondents, almost all
female and over 61. This group was highly savvy about identifying FN.
(1 Very likely true—5 Absolutely untrue) Figure 2. Survey I respondent ability to identify an FN article
FAKE NEWS 19
(1 Never—5 Always) Figure 3. Survey I respondent self-assessment about ability to decide if an article is true or false
Table 7. Survey I respondent FN features
• Look for a byline and I check the Internet address. • Source (is it biased), quotes from knowledgeable people, inflammatory language. • Reputable media outlet and author. Credible spokespersons. Reasonable content and facts. More facts than opinion. However, a lot of phony press is out there and sometimes it's important to check articles on Snopes or FactCheck.com. • Citing to various sources; common sense • I get most of my news from sources I know and trust to have high standards of responsibility and factchecking. If I don't know the source, I evaluate the quality of the writing and the author's ability for logical thought. I compare the claims being made with knowledge I already have. The quotes used in a fake news story are often the biggest giveaway, as in the story below. The Trump quote in particular is an obvious parody. Capturing a specific public figure's syntax is the hardest thing to fake. If I'm ever not sure of an article's validity after reading it carefully, I check with Snopes or a news source I trust. • Look for the media resource • Source. Emotion. Grammar.
Then I aimed the survey at a younger demographic by giving it to students at the Middlebury
Institute of International Studies in Monterey (MIIS), CA. Survey II reflects the opinions of 35
MIIS respondents, 23 females and 12 males. They were almost all age 21-40, except for two age
41-60. Findings in this younger demographic also showed sharp skills in recognizing FN.
FAKE NEWS 20
(1 Very likely true—5 Absolutely untrue) Figure 4. Survey II respondent ability to identify an FN article
(1 Never—5 Always) Figure 5. Survey II respondent self-assessment about ability to decide if an article is true or false
FAKE NEWS 21
Table 8. Survey II respondent FN features
• I look for reputable sources that I know (NY Times, BBC, etc.) • I look for other sources posting the same story. • News source; prior knowledge/context; are there other articles that confirm it (preferably from verified sources)? If there is a study or survey referenced? I either look at the original myself or take the article with a huge grain of salt. • Publications with a strong editorial reputation. Recognizable URL. "About Us" section for sources I'm not familiar with. Red flags include sensationalist headlines, or headlines that demonize a person or group. • I might look for news reported by news agencies from different countries especially regarding certain transnational controversial issues.\ I consider the logic, legality of the claims in the article. • I consider the logic, legality of the claims in the article. • I check the source--is it someone's blog or a reputable source like WaPo, the NYT, CNN, The Economist, The Hill, etc.? I look at the language--if it's too strong, the article is likely either false or heavily biased. If I still have questions, I search for reputable news sources that report on the same story. • I read sources that have a history of telling complete stories and that provide argumentation from all sides. I read a variety of different sources, left- and right-leaning, and sources at different levels, state, national, and international. I look for sponsorship or sources of funding, if possible/applicable. If something "seems" untrue, I try to verify it. • I look for multiple reputable news agencies and verified sources. I also do resources on factors that stand out to me in the news • Double check different sources for the same info. I do not trust any news agency • I look for information I can confirm using other sources, how it interacts with my understanding of the facts. • Source • Grammar and spelling matter. If there are errors, believability goes down. If it's big news I usually check 2 other stories. • The issue being discussed is highly unlikely to happen. • I read news from other sources to compare information and talk to people I trust. • Source, talk to people and get their opinions • Spelling errors, contradiction to previous knowledge • I look for references, the author, cite of publication, and Snopes • The source (if it's from Fox News or another dubious source--religious, etc.) I usually doubt it. If it comes from a name I trust (Reuters, NY Times, etc.) I tend to trust it. • Other articles about the same issues, read about the author and the publisher • You need to look at the source. The source of the source. The author. The methodology used. When it was written. What kind of article it is. Opinion paper, research paper. • Who posted/published it. Why. When was it posted or up-dated? • Check the website it's on; double check for a site description listing articles as satire or not; check sources • References, citations, quotes, publisher • Authoritative source. Opinions from third party. • Not sure; I don't generally come across much fake news, so far as I can tell. • Sources, does it come from an authoritative source? Discussion with friends. Google to search if there are similar content online. • I look for strong personal opinions/biases, or absolute statements (untrue). Inclusion of data, links to reputable sources/being from a reputable source, look of bias (true). • I check the source--is it someone's blog or a reputable source like WaPo, the NYT, CNN, The Economist, The Hill, etc.? I look at the language--if it's too strong, the article is likely either false or heavily biased. If I still have questions, I search for reputable news sources that report on the same story.
FAKE NEWS 22
Interpretation and Conclusions
FN is widely-discussed and news consumers like the ones surveyed in this study have
strategies for recognizing false reporting when they see it. Moreover, Facebook, one of the major
FN sharing websites, recently gave users a mechanism for marking false news posts, along with
a list of generalized tips for spotting them (Price, 2017). This study, however, has attempted to
find empirical evidence about FN’s nature, yielding a list of 22 common characteristics that were
tracked in 40 samples. To recap, the 10 most frequently-occurring FN features identified in this
study were:
1. Negative words 2. Hyperbole/superlatives 3. False quotes 4. Unattributed sources 5. New-like detail 6. Passive voice 7. Editorializing 8. Negative reference, insult 9. Extended quotes 10. Famous names, sources
Numbers 1 (Negative words) and 6 (Passive voice) were predicted by psychological
studies on deception mentioned earlier (Newman, Berry & Richards, 2003). Their occurrence in
study samples was consistent and marked, with negative words especially abundant in Anti
Democratic and Pro- and Anti-Trump samples. However, further research is needed to compare
the occurrence of negatives lexical items to positive ones in FN, and to compare negativity in
false news to negativity in bona fide news. For example, two sample articles from the front page
of the New York Times on April 28, 2017, showed relatively high negative word occurrence,
especially in a story about the possibility of war with North Korea (suggesting that the subject of
a story might have as much to do with negative word frequency as deceptive intent).
FAKE NEWS 23
The use of the passive voice in FN samples tended to obscure the doer of an action,
allowing for unprovable statements like “…they were denied entry,” leaving uncertainty about
who wouldn’t let them in. Without a clear agent for the action, a reader can’t even begin to
assess the validity of a statement. Journalists are taught to avoid passives, not just because they
can be confusing, but because they tend to weaken the force of the prose.
While numbers 2 (Hyperbole/superlatives), 8 (Negative reference, insult), and 10
(Famous names, sources) may seem unsurprising, these features weren’t identified in deception
literature. All three of them go against fundamental journalistic principles, in fact may be
libelous. But anything goes on the Internet and in social media, it seems.
Number 6 (News-like detail) is less a feature of FN than a method used by writers to
disguise FN as trustworthy journalism, similar to the disguised URL’s noted earlier. The same
effort to disguise can be seen in feature numbers 11 (News format) and 19 (Unbiased
appearance). When laid out with a headline, byline, and dateline, like a print newspaper article,
with supporting facts and figures and the semblance several different viewpoints, an FN item can
all too easily seem real.
The presence of all three Intertextual features—3 (False quotes), 4 (Unattributed
sources), and 9 (Extended quotes)—in top frequency positions reflects the way a news article
interacts with other written or spoken texts made by experts, involved parties, or witnesses, i.e.,
quotes. The news story genre has strict conventions about the use of other peoples’ words,
observations, and ideas, for instance, that the writer must quote and attribute anything he or she
didn’t observe at first hand. At the same time, quotes are thought to lend color and credibility,
with one quote per paragraph a common reporter’s rule of thumb. Only on rare occasions do
FAKE NEWS 24
editors allow quotes from generic un-named sources (a spokesperson, expert, scientific report,
etc.)
Based on the loose handling of quotes observed in study samples, it could be claimed that
FN is an altogether different genre of communication than real news. Unattributed or
generically-sourced quotes occurred commonly in study samples, and several FN stories were
really nothing more than extended false quotes, as in “Everyone Says They Want to Go to Africa
Or Mexico, Well Here’s Your Chance! Make America Great Again and Go Back to Your
Country for Free—Trump.” The samples were also replete with information reported at second
hand, unattributed, such as “Obama Passes Law for Grandparents to Get All their Grandchildren
Every Weekend.” Even more often, samples brazenly included quotes that readers didn’t even
have to Google to know untrue, as in Trump calling Obama “an illegitimate Muslim traitor.”
Survey respondents reported the necessity of validating an article by checking others news
websites for corroboration, but were far less focused on the need for skepticism when internal
quotes are mishandled.
Only one FN feature from the Strategic category appeared on the top 10 list:
editorializing (number 9), which means to blur the border between fact and opinion, a
journalistic taboo. However, frequency may not be the best way to judge strategy in writing
because each story generally has only one or two basic “moves,” and it takes concerted effort to
analyze and identify them, which the casual news reader may be uninclined to do.
Two of the Strategic features found in FN samples (Red herring and False dilemma)
commonly appear on lists of logical fallacies, which teachers warn writing students to avoid.
Logical fallacies have been defined by Weber and Brizee (2013) as “common errors in reasoning
that will undermine the logic of your argument, either illegitimate arguments or irrelevant points,
FAKE NEWS 25
identified because they lack evidence that supports their claims.” They typically include “straw
man” arguments, or oversimplifying a viewpoint in order to attack it; “ad hominem,” which
means to assail a person’s character rather than the view he or she espouses; and “either-or”
simplification of a far more complicated issue. In the world of FN, however, strategies for
avoiding composition problems are turned inside-out and used to create the problems they were
intended to avoid.
Taken in sum, the FN features identified and assessed for frequency in this study describe
a sub-genre of Internet-based news stories, though journalists and experienced news consumers
would strenuously resist its admission into the journalism category. Perhaps, FN is more a sub
genre of deception or of writing, in general. In any event, using the study’s 22 FN features to
define and categorize a genre may be less important, on a pedagogical level, than teaching the
features to learners to sharpen their reading, writing, and critical thinking skills.
Surveys conducted for this study show that highly-educated people, both young and old,
are perfectly able to separate true from false news. But it should be recalled that American
Internet news consumers, at large, believed FN stories to be accurate 75% of the time during the
2016 elections (Silverman, 2016). And much more concerning are the findings of a recent
Stanford University study showing that 82% of middle school students couldn’t tell the
difference between an ad labeled “sponsored content” on a website and news, and 40% of the
young people studied decided the area around Fukushima, Japan, was toxic because of a false
headline and a photo of a deformed daisy (Shellenberger, 2016).
The danger, according to K-12 educators, is that teenagers whose only exposure to news
is through their smartphones grow up lacking the critical ability to determine what is true and
what is not (Barron, 2017). Silverman, Buzzfeed media editor, said that “we need our schools to
FAKE NEWS 26
start thinking about how we integrate more media literacy and critical thinking education so that
people can make better judgments for themselves” (Davies, 2016).
Fortunately, many educational institutions and organizations have responded to the
problem of media illiteracy in an age of FN, including Stanford History Education Group, which
offers a free, downloadable social studies curriculum that teaches teenagers how to assess the
trustworthiness of sources; the education technology corporation Newsela; and Stony Brook
University’s Center for Media Literacy. Their modules on media literacy align well with
established standards, such as the State of California’s Common Core objectives for English
language arts and literacy in history/social studies (Appendix D).
The present study suggests a number of teaching methods that can be added to media
literary curriculums, not least the list of FN features identified through sample analysis, which
can serve as a useful tool for helping learners recognize doubtful news and sources. Students
could be asked to follow simplified genre analysis procedures to deduce their own lists of FN
characteristics from samples, in the process sharpening critical thinking and reading skills.
It is possible, as well, to incorporate FN recognition in general writing classes by having
students consider how deceptive news stories confound composition taboos against red herrings,
editorializing, oversimplification, and other logical fallacies. Lessons about correct use of
sources can involve examples of some of the ways FN mishandles them. Finally, learners can
and should be exposed to fundamental ethical issues that frame the relationship between honest
writer and reader. As William (2016) elegantly said, “The social contract between thoughtful
writers and readers implies a goodwill exchange fair to everyone and in the long-term best
interest of both” (p. 222).
FAKE NEWS 27
References
Adenzato, M. & Bucciarelli, M. (2010). Recognition of mistakes and deceits in communicative interactions. Journal of Pragmatics, 40(4), 608-629.
Allen, W. & Blinder, S. (2013). Migration in the news: Portrayals of immigrants, migrants, asylum seekers, and refugees in national British newspapers, 2010 to 2012. Retrieve from: http://www.migrationobservatory.ox.ac.uk/wp-content/uploads/2016/04/ReportMigration_News.pdf
Baker, P., Gabrielatos, C. & McEnery, T. (2012). Sketching Muslims: A corpus driven analysis of representations around the word “Muslim” in the British press 1998-2009. Applied Linguistics, 34(3), 255-278.
Balmas, M. (2012). When fake news becomes real: Combined exposure to multiple news sources and political attitudes of inefficacy, alienation, and cynicism. Communication Research, 41(3), 430-454.
Barron, J. (2017). In an era of fake news, teaching students to parse fact from fiction. Retrieved from: https://www.nytimes.com/2017/03/20/nyregion/fake-news-brooklyn-middle- school.html?_r=0
Bazerman, C. (2004). Speech acts, genres, and activity systems: How texts organize activity and people. C. Bazerman & P. Prior (Eds.)., What writing does and how it does it: An introduction to analyzing texts and textual practices (309-339). Mahwah, NJ: Lawrence Erlbaum.
Bazerman, C. & Prior, P. (Eds.). (2004). What writing does and how it does it: An introduction to analyzing texts and textual practices (p. 2). Mahwah, NJ: Lawrence Erlbaum.
Bedford, D. (2010). Agnotology as a teaching tool: Learning climate science by studying misinformation. Journal of Geography, 109(4), 1-9.
Belcher, D., Johns, A.M. & Paltridge, B. (2011). New directions in English for specific purposes research. Ann Arbor, MI: University of Michigan Press.
Bhatia, V. (1993). Analyzing genre: Language use in professional settings. New York, NY: Longman.
Bhatia, V. (2009). Intertextual patterns in English legal discourse. In D. Belcher (Ed.)., English for specific purposes in theory and practice (pp. 186-204). Ann Arbor, MI: University of Michigan Press.
FAKE NEWS 28
California State Board of Education (2013). Common Core state standards: English language arts & literacy in history/social studies, science, and technical subjects. Retrieved from: http://www.cde.ca.gov/be/st/ss/documents/finalelaccssstandards.pdf
Croissant, J.L. (2014). Agnotology: Ignorance and absence or toward a sociology of things that aren’t there. Social Epistemology, 28(1), 1-19.
Croxton, W., & Gonsalves, R.C. (2017). What’s “fake news?” 60 Minutes producers investigate. Retrieved from: http://www.cbsnews.com/news/whats-fake-news-60-minutes-producersinvestigate/
Davies, D. (2016). Fake new expert on how false stories spread and why people believe them. Retrieved from: http://www.npr.org/2016/12/14/505547295/fake-news-expert-on-howfalse-stories-spread-and-why-people-believe-them
Dilmon, R. (2009). Between thinking and speaking—linguistic tools for detecting a fabrication. Journal of Pragmatics, 41(6), 1152-1170.
Grynbaum, M. (2017, February 17). Trump calls the news media the ‘enemy of the American people.’ The New York Times. Retrieved from https://www.nytimes.com/2017/02/17/business/trump-calls-the-news-media-the-enemy- of-the-people.html?_r=0
Hernandez, J. (2017, February 3). China’s response to reports of torture: Fake news. The New York Times. Retrieved from https://www.nytimes.com/2017/03/03/world/asia/chinafake-news-xie.html?_r=0
Huckin, T. (2004). Content analysis: What texts talk about. What writing does and how it does it: An introduction to analyzing texts and textual practices (p. 13). Mahwah, NJ: Lawrence Erlbaum.
Jackson, C. (2016). Ipsos/BuzzFeed poll – fake news. Retrieved from http://ipsos-na.com/news-polls/pressrelease.aspx?id=7497
Lie. (2003). Retrieved March 7, from https://www.merriam-webster.com/dictionary/lie
Loftus, E.F. (2005). Planting misinformation in the human mind: A 30-year investigation of the malleability of memory. Learning & Memory, 12(4), 361-366.
Millar, D. (2011). Promoting genre awareness in the EFL classroom. English Teacher Forum 49(2), 4.
(Newman, L., Berry, D., & Richards, J. (2003). Lying words: Predicting deception from linguistic styles. PSPB, 29(5), 665-675.
FAKE NEWS 29
Obama, B. (2017). Read the full transcript of President Obama’s farewell speech. The Los Angeles Times. Retrieved from http://www.latimes.com/politics/la-pol-obama-farewell- speech-transcript-20170110-story.html
Paltridge, B. & Starfield, S. (2013). The handbook of English for specific purposes. Malden, MA: John Wiley & Sons.
Pogue, D. (2017) How to stamp out fake news. Scientific American, 316(2), 24.
Price, R. (2017). Facebook will teach you how to spot fake news. Business Insider. Retrieved from: http://www.businessinsider.com/facebook-how-to-spot-fake-news-2017-4
Selzer, J. (2004). Rhetorical analysis: Understanding how texts persuade readers. C. Bazerman & P. Prior (Eds.)., What writing does and how it does it: An introduction to analyzing texts and textual practices (309-339). Mahwah, NJ: Lawrence Erlbaum.
Shellenberger, S. (2016). Most students don’t know when news is fake, Stanford study finds. Retrieved from: https://www.wsj.com/articles/most-students-dont-know-when-news-is- fake-stanford-study-finds-1479752576
Silverman, C. (2016). Here are 50 of the biggest fake news hits on Facebook from 2016. Retrieved from https://www.buzzfeed.com/craigsilverman/top-fake-news-of- 2016?utm_term=.siz2aOvQo#.uiE6kGA9y
Smith, R., & Sydell, L. (2016). Planet Money finding the fake news king. Retrieved from: http://www.npr.org/sections/money/2016/12/02/504155809/episode-739-finding-thefake-news-king
Swales, J. (1990). Genre analysis: English in academic and research settings. Cambridge, UK: Cambridge University Press.
Tarone, E., Dwyer, S., Gillette, S., & Ickes, V. (1981). On the use of the passive in two astrophysics journal papers. English for Specific Purposes, 1(1), 123-140.
Van Swol, L, Braun, M. & Malhotra, D. (2011). Evidence for the Pinocchio Effect: Linguistic differences between lies, deception by omissions, and truths. Discourse Processes, 49(2), 1-37.
Wang, W. (2008). Intertextual aspects of Chinese newspaper commentaries on the events of 9/11. Discourse Studies, 10(3), 361-381).
Weber, R. & Brizee, A. (2013. Logical fallacies. Owl Purdue Online Writing Lab. Retrieved from: https://owl.english.purdue.edu/owl/resource/659/3/
Williams, J. (2016). Style: Lessons in clarity and grace. New York, NY: Pearson.
FAKE NEWS 30