Bias in Filipino Newspapers?

Newspaper Sentiment Analysis of the 2017 Battle of Marawi

Dexter Dave R. Valdeavilla and Maria Teresa R. Pulido

Department of Physics, Mapúa University, Intramuros, Manila City, 1002 Philippines

Keywords: Media, Large-scale Information Systems and Applications, Data Processing, Artificial Intelligence, Big Data

Algorithm, Social Science and Implications for Big Data, Data Analytics, Opinion Mining, Analytics as a

Service (AaaS).

Abstract: Newspapers provide factual reports on current events. However, news media has been shown to be

ideologically biased, often negatively shaping the readers' point of view. News on controversial issues makes

the bias of the newspaper or its writers more visible. This study aims to measure the objectivity of newspapers

by classifying news articles from three newspaper agencies covering the 2017 Battle of Marawi in Southern

Philippines. We used Aylien Sentiment Analysis Tool to detect the bias or polarity in each news article

(whether positive, negative or neutral). Negative articles on Marawi dominated the three broadsheets (45.1%

to 59.9%) while the neutral articles were the least frequent (16.1% to 21.2%). These results indicate that

newspapers apply unequal space on the different sides of an issue, which may lead to unbalanced reporting.

We also note that despite the varying number of total articles, the three papers applied the same proportion of

positive, negative and neutral articles, which may imply collusion. The emergence of Big Data greatly

increases the speed of gathering news articles on any given issue, while the Internet of Things enables readers

and journalists to measure the objectivity of the news.

1 INTRODUCTION

Bias in news media is an inherent, persistent flaw

(Efron, 1971; Anand, et al, 2007). Bias often causes

a sharp increase in political polarization,

misunderstanding, and conflict on issues (Park et al.,

2009). For example, the New York Times has been

observed to cover news events distinct from other

US newspapers (Zelizer, 2002). Major national

news organizations in the US have been seen to

present the same liberal-leaning slant (Sutter, 2000).

News accounts sometimes misrepresent certain

events, such as declining news coverage for protests

that are in reality gaining momentum (Oliver and

Maney, 2000). On the other hand, another study

using an “objective” testing of newspaper articles on

one issue yielded little evidence of partisan media

bias (Niven, 2003).

Figure 1 shows the various causes and forms of

bias (Park et al., 2009) in the media, such as

newspapers. A newspaper may contain selection

bias (which affects the amount of coverage given to

an event), and description bias (which affects the

accuracy of the coverage) (Earl et al., 2004). Bias

may be due to the preferences of the stockholders

and advertisers of a newspaper (Herman and

Chomsky, 1988), as well as competition with rival

newspapers (Ellman and Germano, 2008).

Meanwhile, readers are aware of the inherent

bias of newspapers (Anand, et al., 2007; Chiang and

Knight, 2008) and take this into account when

consuming information. It is also possible that the

measuring tools or the researchers themselves are

the ones to introduce bias in the study. Therefore,

we used a sentiment analysis software to automate

the measurement of the polarity of a news article and

to avoid personal bias from the researchers.

Sentiment analysis tools, such as those found in the

programming languages Python and R, use natural

language processing capabilities to determine the

polarity of a selection of text. Such tools indicate

whether the expression of words being used in a text

are positive, negative, or a neutral way.

408

Valdeavilla, D. and Pulido, M.

Bias in Filipino Newspapers? Newspaper Sentiment Analysis of the 2017 Battle of Marawi.

DOI: 10.5220/0007752104080413

In Proceedings of the 4th International Conference on Internet of Things, Big Data and Security (IoTBDS 2019), pages 408-413

ISBN: 978-989-758-369-8

Figure 1: Causes and forms of media bias.

Figure 2: A screenshot of an article in Google Spreadsheet using the Aylien Sentiment Analysis tool.

Bias in Filipino Newspapers? Newspaper Sentiment Analysis of the 2017 Battle of Marawi

409

In this work, we aim to measure the amount of

bias contained in Philippine broadsheets or

newspapers. To limit our dataset, we only used news

articles on a particular event with a specific time

range. We chose the 2017 Battle of Marawi due to its

national significance and inherent polarity. The

researchers have personally observed that Filipinos

generally support our armed forces against rebel

groups, however there is disagreement on whether to

implement Martial Law in Mindanao.

The Battle of Marawi City is considered the most

significant terror event in Southeast Asia in the last

15 years (Widiasari, 2018). The conflict lasted for

five months, and resulted in the mass displacement of

civilians, the widespread destruction of civilian

infrastructure, and the loss of civilian lives (Amnesty

International, 2017). Media frames or biases for the

Battle of Marawi include: humanitarianism, hope,

resilience, violence, propaganda, and criticism

(Widiasari, 2018). We gathered news articles about

the battle from major Philippine broadsheets and used

sentiment analysis software to determine the

objectivity of the news articles while excluding the

researchers’ own subjective judgment.

2 METHODOLOGY

We gathered news articles relating to the Battle of

Marawi from the websites of the three major national

newspapers: Manila Bulletin

(https://www.mb.com.ph), Philippine Daily Inquirer

(https://inquirer.com.ph), and Philippine Star

(https://www.philstar.com). We limited our search to

articles within the duration of the battle: May 23,

2017 (when war was declared) to October 22, 2017

(when government troops declared victory over the

rebels). We copied the text of each article onto

Google Sheets with the add-on Aylien Sentiment

Analysis tool (Figure 2).

The polarity of all articles were then collected and

analyzed for possible trends. We note that the time

range for the event can be divided almost equally into

five months. This is advantageous for our research, as

we can investigate trends for five different months,

and we can compare the trend for bigger and smaller

time scales.

3 RESULTS AND DISCUSSION

Table 1 lists the number of articles released by each

news agency regarding the Battle of Marawi. The

three newspapers varied in the number of articles they

released, with the PDI consistently having the highest

number of articles on Marawi. Also, the number of

articles decreased from Month 1 to Month 3, then

increased again by Month 4 when the government

steadily regained territory from the rebel forces.

Table 1: Number of news articles on the Battle of Marawi for each news agency and per month.

Month

(2017)

Manila Bulletin

(MB)

Philippine Daily

Inquirer (PDI)

Philippine Star

(PS)

TOTAL

(M1)

May 23

– Jun 21

61 193 137

391

(M2)

Jun 22

– Jul 22

54 133 103

290

(M3)

Jul 23

– Aug 22

41 54 53

148

(M4)

Aug 23

– Sep 22

50 78 59

187

(M5)

Sep 23

– Oct 23

65 69 35

169

TOTAL 271 527 387 1,185

IoTBDS 2019 - 4th International Conference on Internet of Things, Big Data and Security

410

Figure 3: The number of articles classified per month

and per newspaper.

Figure 4: The proportion of articles classified per month

and per newspaper.

Bias in Filipino Newspapers? Newspaper Sentiment Analysis of the 2017 Battle of Marawi

411

Figure 5: Average monthly distribution of positive, neutral and negative articles per news agency. The Battle of Marawi was

covered for 5 months.

For every month of the coverage on Marawi,

majority of the news was negative, followed by

positive, with minority of the news items evaluated as

neutral (Figure 3). When we normalized the values

shown in Figure 4 with respect to the total number of

articles published by each news agency, we found

highly similar proportions of positive, neutral and

negative articles (Figure 4).

Figure 5 summarizes the results. For all three

papers, a significant proportion (78% to 84%) of the

news on Marawi were found to be polarized (either

positive or negative). Majority of the articles were

negative for all three broadsheets (45.1% to 59.9%),

while a distinctly small minority of articles were

neutral (16.1% to 21.2%).

The results highly suggest the presence of bias,

particularly anti-administration bias, with regards to

the Battle of Marawi. In addition, seeing the similar

distributions of negative, neutral, and positive news

suggest collusion among the three news agencies. For

more complete results, we can use other sentiment

analysis software as well as construct our own.

The Battle of Marawi is admittedly a controversial

issue. We can also measure the polarity of news on

other issues, especially those universally regarded as

“good”, “bad” or “neutral”, to serve as a baseline for

studies such as this one.

4 CONCLUSIONS

We have studied the news coverage of the Battle of

Marawi, and have seen the prevalence of bias in the

large proportion of news with non-neutral polarity.

Moreover, the three newspaper agencies show highly

similar proportions of positive, neutral, and negative

polarities, suggesting a collusion among newspapers.

We also demonstrated the use of the Aylien

Sentiment Analysis tool to determine the polarity of a

text without inserting the researcher’s own biases or

personal interpretation.

For further study, we recommend the use of other

sentiment analysis software to compare with the

existing results. We can also construct our own

sentiment analysis software, which can be easily

trained with the massive amounts of news articles that

are generated everyday. We can also extend this work

by using the same methods and tools on other topics.

In particular, we can use universal “good”, “bad” and

“neutral” news as baselines for measuring the polarity

of an article. We can also use other news sources

aside from the three studied in this paper.

The emergence of Big Data greatly increases the

speed and convenience of gathering news articles on

any given issue. Meanwhile, the Internet of Things

enables readers to measure the objectivity of a

particular news article or source, and serves as a

check to journalists regarding their work.

ACKNOWLEDGEMENTS

We thank Mr. Carl Vincent M. Espinosa for the

invaluable technical support, the Mapúa University

Yuchengco Innovation Center for the resources in

preparing this manuscript, and our colleagues and

loved ones for their support. We also thank the

organizers of the IoTBDS 2019 Conference for

accepting this work and for the financial support.

IoTBDS 2019 - 4th International Conference on Internet of Things, Big Data and Security

412

REFERENCES

Amnesty International, 2017. [Online] The Battle of

Marawi: Death and Destruction in the Philippines.

Available from: https://www.amnesty.org/download/

Documents/ASA3574272017ENGLISH.PDF

[Accessed 20 January 2019].

Anand, B., Di Tella, R. and Galetovic, A., 2007.

Information or opinion? Media bias as product

differentiation. Journal of Economics & Management

Strategy, 16(3), pp.635-682.

Chiang, C.F. and Knight, B., 2011. Media bias and

influence: Evidence from newspaper endorsements.

The Review of Economic Studies, 78(3), pp.795-820.

Herman, E.S. and Chomsky, N. 1988. Manufacturing

Consent: the Political Economy of the Mass Media.

New York: Pantheon Books.

Earl, J., et al., 2004. The use of newspaper data in the study

of collective action. Annu. Rev. Sociol., 30, pp.65-80.

Efron, E., 1971. The News Twisters. Los Angeles: Nash

Publishing.

Ellman, M. and Germano, F., 2009. What do the papers

sell? A model of advertising and media bias. The

Economic Journal, 119(537), pp.680-704.

Niven, D., 2003. Objective evidence on media bias:

Newspaper coverage of congressional party switchers.

Journalism & Mass Communication Quarterly, 80(2),

pp.311-326.

Oliver, P. E. and Maney, G. M., 2000. Political processes

and local newspaper coverage of protest events: From

selection bias to triadic interactions. American Journal

of Sociology, 106(2), pp.463-505.

Park, S., et al., 2009, April. NewsCube: delivering multiple

aspects of news to mitigate media bias. In Proceedings

of the SIGCHI Conference on Human Factors in

Computing Systems (pp. 443-452). ACM.

Sutter, D., 2000. Can the Media be so Liberal-The

Economics of Media Bias. Cato J., 20, p.431.

Widiasari, N., 2018. Framing the War: The Marawi Siege

as Seen Through Television Documentaries. Asian

Politics & Policy, 10(1), pp.149-154.

Zelizer, B., Park, D. and Gudelunas, D., 2002. How bias

shapes the news: Challenging The New York Times'

status as a newspaper of record on the Middle East.

Journalism, 3(3), pp.283-307.

Bias in Filipino Newspapers? Newspaper Sentiment Analysis of the 2017 Battle of Marawi

413