Polyamory in the News Corpus (POLYNEWS)

Data description The POLYNEWS corpus contains metadata and vocabulary counts for all press news articles mentioning polyamory published in English-language sources between 1994 and 2023, both included. Data format: Annotated (.csv)Type of data: Table Data collection This dataset comprises worldwide English-language press news mentioning polyamory. The data were collected in early 2024 to cover the 1994–2023 period. The data for this study were collected from the English-language news sources covered by the database Factiva (Dow Jonex & Co., 2025), using the following parameters: parameter value Search query la=en AND rst=tnwp AND (polyamory OR polyamorous OR polyamorist OR poly-amory OR poly-amorous OR poly-amorist) Date All Dates Source All Sources Author All Authors Company All Companies Subject All Subjects Industry All Industries Region All Regions Language English Duplicates Off The search yielded 9228 results as of 1 January 2024. Data source location http://www.factiva.com/ Data accessibility Repository name: POLYNEWSData identification number: 10.5281/zenodo.14989797 Data format The dataset is supplied as three files: polynews_articles.csv, containing article-level data polynews_publications.csv, containing publication-level data polynews_terms.csv, containing term-level data Article data The format of the article data file is as follows: field content type id unique ID assigned by Factiva to the article string headline text of the article headline string authors list of authors, as automatically parsed Python list authors_clean list of authors, manually cleaned up Python list date time and date of publication date in ISO8601 format publication_id unique ID assigned by Factiva to the publication where the article appeared string section original description of the section where the article was published string extra_columns additional unparsed data in the Factiva record Python list wordcount length of the article in words as reported by Factiva integer wordcount_actual length of the article in words as measured by tokenizer integer frequencies list of terms in the article with their frequencies Python Counter Publication data The format of the publication data file is as follows: field content type publication_id unique ID assigned by Factiva to the publication string wikidata_id unique ID assigned by Wikidata to the publication string publication_title full title, uniquely disambiguated if necessary string country country of publication string medium medium of publication string: print, online issn ISSN for the publication, where available string frequency frequency of appearance string type broad classification of newspaper type string: broadsheet, tabloid, local, business Term data The format of the term data file is as follows: field content type term word type, as parsed by the Penn Treebank tokeniser string count number of occurrences of term in the corpus integer log_ratio binary log of the ratio of relative frequencies of the term against the SIBOL corpus floating point number log_likelihood Dunning's log likelihood of the term against the SIBOL corpus floating point number