Study title

Full edition for scientific use. The AUTNES dataset on party Facebook pages contains information on parties’ Facebook posts during the six weeks of election campaign for the Austrian general election in 2017. We retrieved posts for all relevant parties and their top candidates. Each post constitutes a unit of analysis. The coding procedure applies the AUTNES relational approach of recording subjects, predicates, and objects to party Facebook pages. The subject is the party or candidate that operates the Facebook page and is coded with the name (if an individual is present), organisational affiliation and appearance in the coding unit (text only, picture only, text and picture). There are two types of objects: issues and object actors. Issues are recorded by coders selecting from the AUTNES issue coding scheme the dominant policy issue and the dominant campaign issue in the coding unit. The issue predicate numerically records whether the subject’s position towards the policy issue is one of support, rejection, or conveys a neutral/ambivalent stance. Up to ten object actors are recorded from each coding unit in the same way as the subject actor, supplemented with an evaluation by the subject actor (positive, negative, or neutral), as well as references to campaign or policy issues. In addition to the basic subject–predicate–object structure, we code several additional variables: variables describing the Facebook page, the coding unit and user interactions. Variables: Variables referring to Facebook posting: URL; type of Facebook page (party or candidate page); timestamp (date when the posting was published); text of the posting; number of interactions and date of their collection; technical problems with coding; content of the posting: continuous text, picture, video; variables referring to the author of the Facebook page: author’s organisation; author’s name; mention of the author in the posting; variables referring to issues: campaign issue; predicate; policy...