Framing femicide

The data set contains 2,528 coded articles from three relevant news outlets in Mexico.

The newspapers in the study were chosen for their audience scope, considering they inform on general topics, and have national coverage in print and digital versions, the latter available for consultation since 2014. The newspapers that have different versions per region were discarded. Five newspapers have these characteristics in Mexico. To ensure equal representation in ideologies we selected one left-leaning, one right-leaning, and one centre-leaning example. We prioritized the ones with the most visits to their websites, as data was collected through their digital versions: (1) El Universal, first place in number of visits, centre ideology and open access to its online library; (2) Excélsior, second place in visits, left-wing ideology and open access to its online library; and (3) Reforma, third place in visits, right-wing ideology and online library available by paid subscription. 

The articles were collected using the search engine on each outlet’s website covering a period of seven semesters: from July 2014 (the date of the first government council meeting to study the possibility of declaring an ‘alert for gender violence’ due to the high rates of femicides) to December 2017 (date on which this research registered the increase of media attention stabilizes). Two filtering processes were used: entering keywords: 1) ‘femicide’ (feminicidio), from the results all articles that had the keyword in the headline, dealt with the topic in the body of the article, or reported on specific cases were added to the database. We discarded those that had the word in the body of the article, but the topic was not related. 2) ‘Dead woman’ (mujer muerta), the intention is to include femicides even if the news outlet did not describe them as such. Articles reporting women murdered by (ex)partners, family, or acquaintances, and single or mass murders of women possibly committed by organized crime groups were added to the database. Those referring to accidental deaths and that explicitly stated that the causes of death were not related to gender violence were discarded. The process produced a unit of analysis of N=2,528 articles from the three outlets. El Universal 1031, Reforma 969, and Excélsior 528.