Seven main themes could be distinguished from the data.
1. Presence of the teacher
The presence and participation of the teacher during the sessions was felt to be of added value for several reasons. First, a number of adolescents saw the teacher as a familiar figure and confidant, which made them more inclined to participate. For example, it was noticed that adolescents first whispered an answer to the teacher before mentioning it in front of the whole group, probably to be sure their answer was meaningful. Second, the teachers acted as translators if the question was not entirely clear to the adolescent. Third, they knew the context of the adolescent better than the PI did. Consequently, either questions asked by the researcher or answers given by the adolescents could be contextualised by the teacher so both sides understood each other sufficiently. Moreover, tips could be given to the PI towards adapting questioning.
“Within co-creation, the teacher also participates: writes something down now and then (e.g. on a large sheet), participates in discussions, ... You notice that pupils are at ease as a result. The teacher's input can also help to put things in context: e.g. what they mainly do at school, why they do not like certain things (e.g. swimming).” (PI, session 2, group A)
Fourth, teachers were able to give input from their experiences in physical education lessons. This helped because the teachers thought about the project from a different view, providing a fresh perspective to the topic.
“The presence of K. (teacher) during the sessions is also a great asset to this group. K. first lets the group answer by itself, but now and then he completes the questions or challenges the group to think 'differently'. E.g.: if everyone puts a green sticker on an intervention goal, he can give an example of a red sticker, which makes the group reflect on it.” (PI, session 3, group A)
Moreover, the teachers were able to contribute to the structure of the sessions, as they knew the group better. For example, each session was planned to begin with a reiteration of the project's purpose, where the session was specifically situated within the process and what the agreements were within a session (see ‘Procedure’). In consultation with one of the two teachers after the fourth session, it was decided not to do this anymore because she noticed adolescents quickly lost their focus because of it.
2. Importance of building rapport
The first session was a real search: not knowing the group, not knowing the teacher, not knowing the context adolescents live in, etc. Throughout the sessions, the researcher got to know both the group and the teacher better and gradually started to build rapport with the participants. After a few sessions, it was for example noticeable that during breaks, conversations were held that were often outside the scope of the project. Building that bond is necessary to get as close as possible to the perception of the adolescents and to give them the confidence to share their opinion. Throughout several sessions, the researchers could observe that the participants felt increasingly at ease, for example by sharing their openness about more difficult topics such as their home situations.
“By creating more trust, the participants were more open about their experiences and feelings. This is something that has to grow throughout the process.” (masters student, session 2, group B)
Moreover, a relationship of trust not only seems important between the participants and the researcher, but also among participants themselves.
“They trusted each other and everyone could speak their minds. They knew a lot about each other and picked up on it.” (masters student, session 2, group A)
3. Co-decision making power
An unprecedented step to take was daring to let go control, and give the co-creators the autonomy and ownership to steer the process. If the PI was in doubt or lost as to which methods worked best for the group, or how best to address the participants, the participants themselves were asked how they thought they could best be approached. The participants were co-creators, which means that they are on the same level as the researcher. They only perceived the theme from a different background or perspective. Moreover, by means of the process evaluation forms that the adolescents completed, the researchers were able to check how the participants had actually felt during the different sessions, so that appropriate adjustments could be made. These forms gave the participants a voice to guide the course of events.
“The main conclusion is that the majority of the adolescents did not understand everything that was being said. In the next sessions, I will have to pay attention to speaking more slowly, more clearly, more concretely.” (PI, session 1, group B)
“Very positive process evaluation. Almost all students gave a green thumbs up to all statements. The arrangement of three separate groups with three facilitators, without any feedback to the whole class, seemed to suit this group best.” (PI, session 6, group B)
4. Group dynamics
The co-creation sessions took place in two different groups, whereby it was immediately apparent that both groups were different and consequently required a different approach, even though the sessions for both were initially set up in a similar way.
In both groups, it was striking that the more articulate individuals quickly took the lead, and the more silent adolescents often just followed them. However, both groups were different in that aspect. Adolescents from the older group (17-22y) seemed to be stronger in having their own opinions and expressing them, even if these did not entirely correspond to the opinions of the more expressive participants or the majority. The youngest group (14-15y) seemed to struggle more with this. Most of them had a withdrawn attitude. It could be inferred that there was a lot of compliance in the input they provided; not compliance towards the researcher(s), but towards their more outspoken peers leading to little interaction between them. The teacher of the youngest group explained that this class group was actually a combination of three different very small class groups, who did not always attend lessons together. The oldest group, on the other hand, had been in the same class for years and seemed to respect each other's opinions despite their individual differences. The different group dynamics that prevailed are closely related to the previous paragraph that indicates that there should be an atmosphere of trust among the participants in order to make co-creation a success.
“In the exercise where they had to work together [ranking the intervention goals], there was very little cooperation, but rather everyone working individually. The opinion of the three outspoken girls was followed the most. I don't know if the rest of the group agreed with everything, as there was no discussion. When I asked if everyone agreed or if someone would put an intervention goal in a different place, there was again no response.” (PI, session 3, group B)
To counteract these groups dynamics, the decision was made to divide the youngest group into their usual smaller class groups as much as possible during the course of the sessions. This way, everyone would be motivated to speak up and express their opinions honestly.
“There was good interaction and the students cooperated well. The students dared to speak their mind honestly. This went better because of the smaller groups.” (master student, session 2, group B)
“In the small groups, they feel at ease and dare to talk more. When this has to be shared with the larger group, it became quiet again.” (PI, session 4, group B)
This shows that researchers have to adapt to the specific group they are facing, and dare to abandon their preparation of the session if it is noticed that something is not working.
5. Adapted questioning matters
How questions were asked proved to have an impact in this target group. First, it was important to present the questions as concretely as possible, with many (visual) examples for clarification. We made a lot of effort to use language appropriate to the target group and to use other, more easy words (i.e., no jargon) to explain concepts if it was felt that the participants did not understand everything (e.g. when there was no response to questions). However, it was not always easy to assess what participants did or did not understand. For example, due to the division in smaller groups, some participants dared to tell the researcher for the first time that they did not understand while they did not do this, or hardly at all, in (earlier) group discussions.
“Words that I have used several times over the course of the sessions are actually unclear to the adolescents. For example, the word 'motivate'. A. did not understand what I actually meant by this, even though I think I used this word in every previous session. Somehow, this made me a little insecure about the previous sessions, because I have the idea that they did not understand everything I asked them then either.” (PI, session 5, group B)
Therefore, during breaks, the teacher was consulted whether the questions were clear enough for the group. This shows a way of sharing power, as participants will not feel agency if fancy language is used or if they are treated like they have no experience or knowledge. Nevertheless, it was still experienced difficult to find a balance between asking a concrete question and not wanting to fill in too much for the adolescents because often they really needed a choice (i.e., ‘do you prefer this or do you prefer that?’). Furthermore, adolescents who were more in the background were more encouraged to take part in the conversation when questions were asked more individually.
“The whole class took part in the group discussion. If the somewhat quieter participants were asked a question directly, they also spoke openly about their experiences.” (master student, session 1, group A)
It was also important to keep the adolescents’ context in mind when asking questions. Many adolescents from this target group come from a more socially disadvantaged home situation, where not both parents are present. When asking a question, it was then important to talk about 'parents' or ‘at home’, rather than 'mum' or 'dad' specifically, as this could hurt the feelings of young people where 'mum' or 'dad' was not present, leading to resistance in the co-creation sessions.
6. Influence of the co-creative methods
Adolescents in both groups were hesitant when being asked a general question in front of the whole group, for example 'which apps do you use and what do you like/dislike about them'. With such a general question, it usually remained silent in which a wait-and-see attitude was adopted. It was remarkable that the adolescents became more relaxed when offered a very concrete exercise (e.g., statements or test apps on an iPad) and visual materials they could interact with (e.g., cards with pictures or posters they could vote on). Getting creative was an entry point for further discussion: in the statement exercise, adolescents could indicate why they were standing on one side of the room and not on the other side; using the activity cards, they could explain why they had chosen a particular card, or why they had voted for a particular poster; when testing different apps on the iPads, adolescents could give their input on that basis.
“H. (teacher) indicated that using the cards was a good method for the adolescents, that she was stunned by some of the young people's cooperation and that she will certainly use this way of working in her lessons in the future.” (PI, session 2, group B)
Moreover, creative/interactive methods were also a way of getting less talkative adolescents to integrate their opinions as well. For example, in the poster exercise where adolescents had to vote on intervention goals, verbally expressing an opinion was difficult for some, but through the sticker-method they could at least indicate whether they agreed the intervention goal should be integrated (green) or not (red).
Some co-creative methods were more successful than others. The session with the visual cards showing examples of barriers and facilitators provided considerable input due to the comprehensibility for the participants. Both teachers were enthusiastic about this method and indicated that it was also interesting for them to discover which methods did or did not work well with their students. One teacher indicated that this was a very good way to better understand young people with ID’s thoughts concerning movement.
“Splitting the group in two and making the assignment very visual (with cards: both activity cards and barriers/facilitators) really helped the young people. In both exercises, the groups participated well and a lot of interesting information came out.” (PI, session 2)
In the session on intervention goals, goals were displayed visually per poster, and adolescents could vote with stickers. This session was pleasant because of the interaction and active participation of the adolescents. When placing a sticker, some adolescents already explained more about why they placed a certain sticker without even being asked. Ranking the intervention goals, however, seemed to be difficult for both groups, but for a different reason. In the youngest group, there was no cooperation: more outspoken participants worked individually, and the others watched. The oldest group could not make a choice which goal they considered more important, r and lumped everything together. Only two of the 16 goals were clearly chosen as less relevant. It could have worked better if the larger group was divided into smaller groups for this exercise, increasing discussion.
Getting to test apps on an iPad was a success. The adolescents were curious about the apps, so they often enthusiastically tried out different things. By allowing adolescents to test apps and think aloud, it was possible to learn first-hand what adolescents find important.
A technique that worked less well was the story-telling during the fourth session. Whereas the youngest group was involved in the stories: “do you really know this person, where does he live?”, the oldest group did not react to this. However, it became apparent that this method was flawed due to the abstract nature of the task. Adolescents were asked to think about something that was not there yet, or without knowing what it might look like, leading to the fact that they did not really understand what the exercise was about or what was expected of them. In contrast, the previous sessions included concrete exercises, with clear direction, and adolescents were much more interactive. Due to the struggle with this abstractness, it was decided to skip the BCTs that did not have specific examples (e.g., valued self-identity, identity associated with changed behaviour), as it was noted that there was little or no response. In the last session, we did also not proceed with asking adolescents to think about a possible intervention on their own in small groups, but instead for the researcher to present an intervention idea that was based on the input adolescents provided in previous sessions and ask for their feedback.
“This session is difficult for the adolescents because they have little idea of what an app can do to meet their chosen intervention goals.” (teacher, session 4, group B)
“The process evaluation forms mainly showed that the majority of adolescents did not understand everything that was being said. A possible explanation could be that these two rather theoretical sessions (3 and 4) are rather abstract, and that it is difficult for young people to imagine an app that does not actually exist yet.” (PI, session 4, group B)
Furthermore, showing examples on a PowerPoint slide seemed to hinder the imagination of the participants. Using this method, participants were already being pushed in a direction somehow. This became clear with a PowerPoint slide showing possible incentives. Participants gave their opinion on incentives that were presented, but no other input, apart from the slide, was given. There might have been other ideas if this had been handled differently.
Tasks where participants had to write down certain things also hampered the process. A method in which they can work creatively or interactively suits them better.
“Part of the exercise was for the groups to write down what they had come up with. It struck me that writing this down was already a barrier in itself. The young people preferred to remember what they wanted to say, rather than writing it down.” (PI, session 4, group A)
Lastly, the setting in which the co-creation takes place can play a role as well. For example, during the second session in the oldest group, the weather was very nice, so it was decided to go outside with a group. By not being in the classroom for once, a much more informal setting was created which may have positively influenced the discussion dynamics.
7. Required characteristics of co-creation researchers
By conducting these sessions, it was noted that being a co-creative researcher requires enthusiasm, patience, flexibility and openness. First of all, it was important to remain enthusiastic, even if there was little initial response of the participants. Participants cannot be expected to share stories enthusiastically if the researcher him-/herself is not enthusiastic in front of the group either; enthusiasm ignites. Moreover, patience was needed. If the group did not understand the researcher or question, it was key to keep searching for ways where a similar language could be found that maximised cooperation, for example by using creativity. This required energy, but there was no point in giving up if something did not go as expected. However, when the co-creative process did not run smoothly, it was tried to keep a positive attitude: colleagues reminded the PI that no answer is also an answer. This shows that it is helpful to share uncertainties with the research team during the process. Furthermore, despite a good preparation and a clear structure at the beginning of a session, a lot of flexibility was needed from the PI. Usually, a session did not go as planned at all. For example, teachers had to reschedule sessions for various reasons (e.g., excursions, info days at school, teachers were not present themselves due to private reasons, etc.), or there was not enough time to discuss everything that was planned, or vice versa there was too much time left without having prepared another exercise.
“The statements ran out pretty quickly. I asked a number of other questions, but felt less prepared.” (PI, session 1, group B)
Moreover, it was noticed that certain methods did not work for one group, whereas they worked for the other group, or that it was just not working for both groups. A session had to be prepared on fairly short notice (i.e., one week) based on the input from the previous session(s). In the context of flexibility, it was important to have a back-up plan ready if it was felt that something was not working, or if there were time constraints. Having a back-up plan ready, made researchers feel more confident.
“Co-creation is also about letting go of control and initial plans. It is searching for what works best for the group you have in front of you.” (PI, session 4, group B)
Lastly, it was important to be open to what interests the participants. This required adaptability to letting go of what the PI had in mind based on the literature. Researcher’s expertise has to be brought into a co-creation process, but a researcher should not let his/her interpretation on what adolescents say be completely guided by what has already been read or shown in literature. Moreover, in the context of openness, the PI also shared something about herself from time to time. This helped to build rapport with the participants, as we may not expect participants to show their vulnerable side if we ourselves do not share our own thoughts and feelings on the topic. Of course, a balance has to be found for the researcher in how far things are shared.