The following is an excerpt of the intervention methodology. For more information, please see the full text of the article on the publisher's website or through your institution's library.

Participants [...] Four hundred and twenty-nine of the 900 persons invited to attend our discussion group sessions did so, and these formed the pool of potential participants in our experiment. In the biggest towns— Cali, Florencia, and Cucuta—slightly less than half the potential participants were ex-combatants, while in the remaining places they constituted less than a third, approximately.

Procedures Once in the field, we applied a stratified, random system for assigning participants to the discussion groups, in order to ensure that we obtained balanced groups: using attendance lists drawn up and reviewed on-site by the field contractors, we were able to randomly assign a balanced number of excombatants and community members to the experiment groups. [...] A total of 174 individuals were randomly drawn from our pool of potential participants (82 ex-combatants and 92 community members) and then randomly assigned to treatment and control conditions in almost perfectly balanced groups. Remaining individuals from the pool of potential participants were assigned to nonexperiment discussion groups, which are not included in the present analysis. We then randomly assigned different conditions (perspective-giving, argumentation, and no-treatment control condition protocols) to each experiment group [...] After everyone had finished a pretest questionnaire, basic rules were explained: there would be audio-only recording, no one would have to identify himself or herself while speaking [...] the main discussion question for all groups was restated: [Verbal Stimulus 1...] From then, discussion flowed spontaneously without moderation. [...] We randomly assigned either a perspective-giving treatment (seven groups, N = 59), an argumentation condition (seven groups, N = 57), or a no-treatment control condition (seven groups, N = 58) to the 21 groups. Every group consisted of an average of eight participants [...] Perspective-giving treatment was established in instructions that were both read out aloud and handed individually to all participants on small pieces of paper and exhorted them to make references to their own personal perspective and history when justifying their proposals. This instruction was absent for the control conditions. [...]

Measures Participants were given a pretest questionnaire immediately before the discussion exercise and a posttest one immediately after it. Pretest items consisted of basic demographic information (e.g., age, gender, education level, and whether they considered themselves conflict victims), five Likerttype items on perceptions of others’ authenticity in previous day-to-day discussion experiences, and one item on the level of trust in their community. Our main dependent variable is attitude change. [...] In the pretest questionnaire, 12 interspersed Likert-type items were used for measuring attitudes to community members [...] and to ex-combatants [...]. In each case, half of the sentences were worded in positive terms and the other half in a more negative tone. Sentences aimed to capture biases in terms of who is to blame for political, economic, social, and security problems and who is contributing to solving them. [...] In addition to attitudes, the posttest questionnaire also measured discussion-quality perceptions [...] fellow-discussant authenticity perceptions [...] overall trust in participants [...] and overall trust in their community [...].