**3. Methodology**

As the theoretical and empirical state of knowledge about artistic behaviour in non-artistic settings was low, the present study pursued a descriptive-explorative objective. The qualitative research approach allowed hypotheses on the role of artists in arts-based interventions to be developed [59].

The study is based on a case the above-described approach of Art Hacking was rolled out into. In other words, a given concept was applied to the case according to the above-mentioned sequences. The starting point for the arts-based idea-finding process was an unsolved problem that a big provider of child day-care establishments introduced: How can you encourage parents to engage in voluntary activities in the establishments and have them participate in joint efforts for child education in big cities? As four artists were working simultaneously on the issue, the case promised rich and concise information (homogenous sampling) [60].

24 people joined the five-day workshop, which took place in Berlin, Germany. In each of the six-person groups that were working on the issue in parallel, one or two staff members from the central administration, two (leading) educators from day-care centres, one or two parents, one business student and one artist were always, continuously participating.

Artist 1 is a painter. Visual artist 2 has a focus on sculpture and installation art. Artist 3 is trained as an actress, mime and dancer. Artist 4 is working in improvisation theatre. All of them have worked in their artistic profession for more than ten years and possess appropriate experience and expertise. Only artist 4 is a parent who has experience with a day-care centre.

Before the workshop started, the artists had been asked to participate actively but they had not received beforehand any information about the five stages of Art Hacking, process details or specific tasks. They knew the arts-based philosophy of the approach and the issue in broad outline but not any work step or assignment given to the participants. Their role in the format was vaguely described to them as facilitators and reduced to a "Just be yourself!"

The research team consisted of the author who moderated the workshop, an assistant that supported the event and deliberately observed the happenings and the above-mentioned students. These were actively working on the issue and their assignment as observers. In order to enhance the necessary objectivity during the participatory observation, the author stayed out of the collection of data and the assistant was not involved in the work process of the participants other than setting its frame. With the exception of two students, all members of the research team had some experience with the format and other participating artists, which favoured the awareness of interesting behavioural aspects.

For structuring the observation and analysing the collected data, the critical incident technique after Flanagan [61] was used. Originally, the critical incident technique was designed for creating job profiles but it can be applied flexibly to other social-science aspects as well. The essential methodical element is a semi-structured questioning technique: Test persons report on situations in which individual behaviour had any positive or negative effects on the main goal of the activity in question. Thereby, key behaviours are identified and categorized [61].

Accordingly, both human behaviour and its significance for other actors involved were considered during the arts-based event. The objective of the analysis were the approaches the artists applied during the event and the role they played in their working groups. The artists' expert status is proved by their professional self-conception and experience.

With regard to the requirements for defining critical incidents (see [62]) 17 longer sequences in the workshop format were determined as relevant situations. Each of the chosen sequences was related to an explorative and/or creative assignment for either the small groups or the plenum, which was supposed to promote solutions to the issue. It was assumed that artistic behaviour patterns of interest would show at these points in particular.

The rule of not letting the participants know anything about the process sequence (see Chapter 2.2) had to be eased in favour of participant observation. With the critical incidents both the assistant and the student observers had a rough outline of the workshop at hand. Thus, they knew about the consecutive stages of Art Hacking but were asked not to share this information with other participants. The students received a short questionnaire for capturing artistic approaches and positions:


The participating observers were asked to outline the situation as well as to give a detailed description of behaviour and its effects on process and fellow players for every critical incident. Immediately after the workshop, an in-depth group discussion with the observers took place, methodically derived from a behavioural event interview [63].

The minutes of all five observers and the content of the group discussion were supplemented by anonymous, short written statements of non-artistic participants. Near the end of the workshop, these were requested to name situations and related activities of the artists that were striking from their point of view, as they differed from attitudes of other participants and/or promoted the process in a special way.

The written material underwent a thematic content analysis in order to categorize the observations made at the critical incidents. Categories are "groups of work behaviours that share some common theme" [62] (p. 1128). In order to objectify the findings, the observations were broken down independently by two persons at first, namely the author and her assistant. Starting from that, a pattern of main and subcategories [61] emerged.
