Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Multifactor assessment of bias in Swedish news articles
Stockholm University, Faculty of Social Sciences, Department of Computer and Systems Sciences.
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

It can be difficult to identify bias in a text. Meanwhile, a large part of the Swedish population reads their news online and according to Statistics Sweden (SCB), only 1/3 verifies information they find false or questionable online (SCB, 2021). Although methods for identifying indicators of bias exists, there are currently no tools available for Swedish texts providing the same analysis to an end user. This study targeted the problem of identifying bias in news articles and designing a tool with this purpose by addressing the research goal: develop a tool that can assist users in identifying bias in news. This goal was approached by using Design Science Research (DSR) where experiments were used as a method to demonstrate and evaluate the developed artifact. DSR provided a framework for structuring the design work into different phases which made it a suitable choice for this study. The artifact was developed and different indicators of bias implemented as functions using dictionary analysis method and language model based classifiers. An evaluation was performed where the artifact got to analyse news, political and biased texts. The results suggests the artifact can identify the targeted indicators of bias although the political language function had poor accuracy. It was concluded that the artifact to some extent managed to solve the problem of identifying bias in text. However, evidence from previous research suggests the analysis could have been more accurate if metadata was included in the analysis. Furthermore, the generalizability of the study was limited as the data collections used were relatively narrow. It was proposed future studies continue to develop the different analysis functions for the indicators, primarily by using larger data sets.

Place, publisher, year, edition, pages
2024.
Keywords [en]
Bias, text analysis, machine learning, dictionary analysis, othering, disastrous language, political bias, BERT, LLM, junk news
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:su:diva-242774OAI: oai:DiVA.org:su-242774DiVA, id: diva2:1955706
Available from: 2025-04-30 Created: 2025-04-30

Open Access in DiVA

fulltext(1349 kB)14 downloads
File information
File name FULLTEXT01.pdfFile size 1349 kBChecksum SHA-512
0e8081683b40c869148e40da4c2a34306fc604dad03d0e2f253359546aca5499a474fc26127d036041ccdffe0bb81264b7338a7902dede0c24f999e5146381b6
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Pettersson, Linda
By organisation
Department of Computer and Systems Sciences
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 14 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 13 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf