The materials and datasets accompanying the paper “Fostering Constructive Online News Discussions: The Role of Sender Anonymity and Message Subjectivity in Shaping Perceived Polarization, Disinhibition, and Participation Intention in a Representative Sample of Online Commenters”. In this paper we report on an experiment in which we aimed to reduce perceived polarization and increase intention to join online news discussions through manipulating sender anonymity and message subjectivity (i.e., explicit acknowledgements that a statement represents the writer’s perspective, e.g., “I think that is not true”).

Data filesDataset_raw – SPSS raw datafile Dataset_restructured_coding incl – SPSS restructured data file from variables to cases, coding of participants’ comments has been included as an additional variable Dataset_backstructured_for MEMORE – SPSS backstructured data file from cases to variables in order to conduct the mediation analysis in MEMORE Coding participant comments – Excell file with the coding of participants comments by the R script, including the manual checking SPSS Syntax – SPSS syntax with which the variables were constructed in the Dataset R Script – R script for all the analyses, except the mediation because that was conducted in SPSS Supplemental material Questionnaire Design lists of stimuli Stimuli lists (1-4) Dutch words and phrases for automated subjectivity coding

Structure data package From the raw dataset, we made the restructured dataset which also includes the calculated variables, see the SPSS Syntax. This structured dataset was the basis for the analyses in R. The backstructured dataset is based on the restructured dataset and needed for conducting the repeated measures mediation with SPSS MEMORE. The coding dataset was also analyzed in R, and provides the input for the column “CodingComments” in the restructured dataset.

Method: Survey through the LISS panel Universe: The sample consisted of 302 participants, but after removing the 8 participants that had not completed the survey, the final sample consisted of 294 participants (Mage = 54.80, SDage = 15.53, range = 17 – 88 years; 55.4% male and 44.6% female). 3.1% of the sample completed only primary education, 25.6% reported high school as their highest completed education, 31.1% had attained secondary vocational education, 25.6% finished higher professional education, and 14.7% had a University degree as their highest qualification. Notably, whereas we preselected participants on their online activity, 49.7% of the sample indicated that they do not respond to online news articles anymore, suggesting that actual participation in online discussions fluctuates over time. Of the people that do react, 54.1% also engages in discussions in online news article threads. Of those, 8.8% discusses almost never, 45% multiple times per year, 35% multiple times per month, 10% multiple times per week, and 1.3% multiple times per day. Country/Nation: The Netherlands