Multivariate Policy Brief

Introduction

Disinformation is a growing global phenomenon spurred on by political and economic actors and is spread mainly through social media and has serious real-world consequences. Many speculate that online disinformation has had a significant impact on election results in the US, Great Britain, Brazil, and numerous other countries. Online disinformation is considered a crisis by many, and while a considerable amount of research has been conducted into the factors determining one’s belief or disbelief of a fake news story, little research has been done concerning the factors that make one more or less likely to be concerned about what is real and fake online. Political ideology is one factor that may explain acceptance of disinformation, but has not been as to one’s concern about distinguishing truth online. Similarly, one’s exposure to online disinformation may impact their concern the phenomenon generally. The implications of people’s concern and exposure could have implications for policy to address the spread of online disinformation, though if concern varies by political ideology, it may complicate the possibility to act.

Literature Review

Disinformation is incredibly pervasive online. Alcott & Gentzkow (2017) investigated the role of social media in circulating fake news stories. They surveyed 1,200 individuals post-US presidential election, utilized web browsing data, and analysis of a database of 156 election-related news stories categorized as false by leading fact-checking websites leading up to the general election. Their research found that leading up to the 2016 US Presidential Elections, false stories favoring Trump were shared a total of 30 million times on Facebook, while those favoring Clinton were shared 8 million times. It also found that the average American saw one to several fake news stories in the months around the election, and just over half of those who recalled seeing a fake news story believed it. People are much more likely to believe stories that favor their preferred candidate, especially if they have ideologically segregated social media networks.

Worldwide, billions of people are exposed to online disinformation. A new report found that 70 countries have engaged in coordinated disinformation campaigns on social media in 2019, most often in support of ruling politicians or political parties. They utilize disinformation in three distinct ways: “to suppress fundamental human rights, discredit political opponents, and drown out dissenting opinions” (Bradshaw & Hanson 2019, i). They also engage in behaviors to encourage support for their political activities, distract from political scandals, and drive existing societal divisions for their political gain, all of which may be highly concerning for citizens, depending on their political allegiance and whether they have been exposed to such disinformation.

While online disinformation is becoming more widespread across the world, concern about what is real and fake online varies widely from country to country. Nic et. al., (2019) wrote a report based on the data analyzed later in this brief. According to the report, Brazil has the highest proportion of respondents that were very concerned about what is real and fake on the internet (85%), compared to Dutch respondents, of which only 30% were very concerned. There is generally a gap of exposure to and concern about online disinformation - most notably for stories completely made up for political or commercial reasons - 32% of those concerned had not been exposed in the last week.

According to Nielsen & Graves (2017), the difference between fake news and news is a “spectrum” rather than clear distinction between real and fake. They found that people use examples of poor journalism and propaganda as examples of fake news more frequently than false information that masquerades as a news report. They conducted 8 focus groups and a survey of online news users to understand audience perspectives on fake news from Finland, US, UK, and Spain in the first half of 2017.

Pennycock & Rand (2019) posed the research questions “why do people believe blatantly inaccurate news headlines” and “Do we use our reasoning abilities to convince ourselves that statements that align with our ideology are true, or does reasoning allow us to effectively differentiate fake from real regardless of political ideology?” They believed what determines response relies on one of two potential explanations: either cognitively based or based on partisanship or confirmation bias. Their solution was that people believe fake news not because of bias, but because of a lack of critical thinking.

Methodology

The 2018 Reuters Digital News Report comes from an annual online survey administered by YouGov, a survey company, sponsored by the Reuters Institute for the Study of Journalism at the University of Oxford to 75,215 online news consumers in 38 countries. The survey focuses on trust, misinformation and technology use trends, among other trends. The survey was nationally representative based on age, gender, region, and education. It excluded anyone who said that they had not consumed any news in the last month.

Overview

Political Ideology

Respondents across 38 countries were asked their political ideology, ranked from Very Left-Wing to Very Right Wing. 1 = Very left-wing, 2 = Fairly left-wing, 3 = Slightly left-of-center, 4 = Center, 5 = Slightly right of center, 6 = Fairly right-wing, 7 = Very right-wing

0	1
1742	930
6977	2686
8748	3093
13985	4942
7806	2634
5539	2146
1485	810
9739	1953

Min.	1st Qu.	Median	Mean	3rd Qu.	Max.
1	3	4	3.898	5	7

1.462

The distribution of political ideology globally fits a normal distribution, skewing only slightly to the left.

Concern about False Information Online

The increasing sophistication of propaganda and disinformation in the global age is somewhat concerning to respondents globally. Regarding the statement “Thinking about online news, I am concerned about what is real and what is fake on the internet”, respondents had the option to respond:

1 = Strongly Disagree, 2=Tend to Disagree, 3=Neither Agree nor Disagree, 4=Tend to Agree, 5=Strongly Agree.

1	2	3	4	5
14130	46536	107388	151722	61362

Min.	1st Qu.	Median	Mean	3rd Qu.	Max.
1	3	4	3.524	4	5

1.019

Respondents clearly demonstrate some concern over being able to distinguish truth from fact online.

Exposure to Disinformation

0=Respondent had not seen any form of online disinformation in the last week, 1=Respondent had seen at least one form of online disinformation in the last week.

0	1
249284	131854

About 1/3 of respondents reported seeing at least one form of online disinformation in the last week, which includes poor journalism (including clickbait), stories where facts are twisted to push a particular agenda, stories that are completely made up for political or commercial reasons, or headlines that look like news but are advertisements, satire, or the use of the term “fake news” by politicians to discredit news media they don’t like.

Correlation

-0.01744

The correlation for the variables is very weakly negative. Based on this weak correlation between political ideology and concern about online disinformation, it can’t be concluded that there is a strong relationship between the two factors.The analysis signals that respondents who identify as more right wing or very right wing are slightly less likely to be concerned about what is real and fake on the internet compared to individuals who identify as more left to very left wing individuals.

Multiple Linear Regression

Both of the relationships between concern about online disinformation and political ideology and concern and an individual having seen a form of disinformation in the last week are highly significant, with low p values. Looking at the coefficients for the independent variables, it’s clear that neither factor significantly predicts an individual’s level of concern. However, seeing disinformation in the last week contributes much more to an individual’s concern than their political ideology.

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	3.498	0.004845	721.9	0
mvdataframe1$ALL.COUNTRIES.Q1F	-0.0112	0.001124	-9.961	2.284e-23
sawfakenews	0.2006	0.003454	58.08	0

Fitting linear model: mvdataframe1\(ALL.COUNTRIES.Q_FAKE_NEWS_1 ~ mvdataframe1\)ALL.COUNTRIES.Q1F + sawfakenews

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
381138	1.014	0.009075	0.00907

With a value of .009075, R-squared for the model demonstrates that the model is very weak to explain an individual’s level of concern about what is real and fake on the internet. Therefore, there are other significant factors that predict concern that are not included in the model.

Conclusions

While both analyses yield relatively weak conclusions, the significance of both relationships indicate that concern about online disinformation is political and may have political responses and that exposure to online disinformation is highly correlated with concern. In light of the fact that over three times the quantity of fake news stories were spread in favor of Donald Trump leading up to the 2016 presidential elections compared to stories shared in favor of Hilary Clinton, this data would indicate that those who may be more likely to support Trump because of ideological affiliation may also be slightly less concerned about disinformation. It’s hard to draw conclusions as to whether individuals are less or more concerned about disinformation in general due to political affiliation or whether or not the disinformation portrays their party favorably. The fact that the data is representative of 38 countries may dilute the strength of the relationship, as the party in power and party more likely to be the subject of disinformation varies from country to country. Ultimately, these findings leave more questions than answers regarding individuals attitudes towards online disinformatin. Further research should disaggregate results by country to determine if there are clearer trends. Investigating whether age, income, educational attainment, and overall trust in the news are factors that could have significant bearing on both one’s exposure to online disinformation and their level of concern towards it.

Citations

Allcott, H., & Gentzkow, M. (2017). Social media and fake news in the 2016 election. Journal of economic perspectives, 31(2). 211-36.

Bradshaw, S., & Howard, P. N. (2019). The global disinformation order: 2019 global inventory of organised social media manipulation. Working Paper 2019.3. Retrieved from The Computational Propaganda Project, Oxford University website: https://comprop. oii. ox. ac. uk/research/cybertroops2019.

Nic, N., Fletcher, R., Kalogeropoulos, A., Levy, D. A., & Nielsen, R. K. (2018). Reuters Institute Digital News Report 2018. Reuters Institute for the Study of Journalism.

Nielsen, R. K., & Graves, L. (2017, October). “News you don’t believe”: Audience perspectives on fake news (Factsheet).

Pennycock, G., & Rand, D. G. (2019). Lazy, not biased: Susceptibility to partisan fake news is better explained by lack of reasoning than by motivated reasoning. Cognition, 188, 39-50.