Signals of Public Opinion in Online Communication

A Comparison of Methods and Data Sources

Datos Bibliográficos

ID5742110
AutoresSandra Gonzalez-Bailon (0000-0002-8372-798X), Georgios Paltoglou, George Paltoglou (0000-0003-4300-7061)
Año2015
Volumen659
Número1
Páginas95-107
Fecha de publicación2015-05-01
Peer ReviewedSí
Open AccessSí
TipoARTICLE
RevistaThe Annals of the American Academy of Political and Social Science (JOURNAL)
Identificadores de la revistaISSN: 0002-7162 • E-ISSN: 1552-3349
EditorialSAGE Publications Inc (PUBLISHER)
DOI10.1177/0002716215569192
OpenAlexW2013413196
IdiomaEN
Citas recibidas43
Referencias citadas20

This study offers a systematic comparison of automated content analysis tools. The ability of different lexicons to correctly identify affective tone (e.g., positive vs. negative) is assessed in different social media environments. Our comparisons examine the reliability and validity of publicly available, off-the-shelf classifiers. We use datasets from a range of online sources that vary in the diversity and formality of the language used, and we apply different classifiers to extract information about the affective tone in these datasets. We first measure agreement (reliability test) and then compare their classifications with the benchmark of human coding (validity test). Our analyses show that validity and reliability vary with the formality and diversity of the text; we also show that ready-to-use methods leave much space for improvement when analyzing domain-specific content and that a machine-learning approach offers more accurate predictions across communication domains

Benchmark (surveying) · Coding (social sciences) · Formality · Information retrieval · Machine learning · Natural language processing · Reliability (semiconductor) · Statistics · Artificial Intelligence · Computational and Text Analysis Methods · Computer Science · Mathematics · Opinion Dynamics and Social Influence · Sentiment Analysis and Opinion Mining