We talked about the specificity of the interview questions and agreed that they were experienced as broad/general, illustrating a phenomenon that we during the process started to name “top down questioning”, more adapted to neurotypical cognitive styles. On the one hand, we felt that this made it difficult to “get at” what we were looking for — it was easy to get stuck in generalized topics like “exercise and diet.” On the other hand, we felt that keeping the questions open was good in order not to exclude anything valuable. While we were critical of the open questions and wanted more specific phrasing, we also found it difficult to suggest concrete alternatives. One of us pointed out that overly narrow, detail-oriented more in line with a bottom-up-approach commonly associated with an autistic cognitive style, questions could threaten people’s ability to freely associate and bounce ideas off each other (which was the point of the group interview), and that we might lose what we’re looking for that way as well.

The interviewer, however, felt that she found an answer to the first question by responding to question three:

Several of us suggested asking for specific examples, although we also acknowledged the risk that people might feel stuck if they have to recall memories to come up with examples. One of us suggested including two different phrasings in the same question to spark associations in more participants. We also discussed that our experience of the interview guide may not only depend on the guide itself, and therefore might not be solved simply by improving it. Some experiences are part of the interview process itself — for example, the uncertainty at the beginning and the “warm-up” time it takes before one feels comfortable.

In all of the interviews, one question naturally led to the next without the interviewer needing to steer very much. The first question, about what it feels like to feel good, was difficult to talk about without moving into the next question — that is, about other people. Two of us expressed the logical flow like this:

In exploring how autistic-friendly interviewing can capture meaningful experiences, we reflected on what flourishing might mean in autistic lives. We talked about the fact that we lack an autistic language for flourishing – naming -, which made the questions feel “neurotypically formulated”:

One of us described feeling angry at the word “good, “ that it feels bland and without meaningful content. We all agreed on the challenge of finding a better, more “autistic-adapted” word — one that hasn’t been so overused and isn’t strongly associated with neurotypical communication — but we became frustrated by how difficult it was to come up with alternative naming of our experiences. Some suggestions that came up were “a pulse that feels right” or “an energy that is positive.”

We talked about the possibility that there might not be a specific “autistic flourishing, “ but rather that flourishing is experienced similarly by all people regardless of neurotype. However, the path to it may differ between neurotypes, as one of us expressed:

Following this, we emphasized that even if the content of flourishing may be similar between neurotypes, the autistic path to flourishing should not be confused with a neurotypical one — for example, it should not turn into us being encouraged to “learn” to enjoy neurotypical social situations.

When it came to the experience of a possible “autistic flourishing” as entity or phenomenon — and whether such a thing exists — the group was divided. Several of us felt that, as humans with similar biology, the feeling of “good” should ultimately be the same for all people. At the same time, there was uncertainty in the group about whether this is actually true or if we simply haven’t yet managed to fully “grasp” what autistic flourishing really is. One of us shared that since she often feels “broken in the brain” after a day with many social interactions, her experience of feeling “good” is when her brain feels like it’s holding together. However, this might perhaps be more about the internal metaphor or image of that experience (centered, focused, calm) being different, rather than the core feeling itself.

As all of us had been involved in the construction of the interview guide, this mirrored receiving the interview guide in advance. It could be experienced as both helpful and hindering. It was helpful in allowing one to prepare and think about how to respond, but at the same time, it could feel like a limitation — having already formulated an answer in your head and becoming “stuck” in that. When arriving at the interview, the intention was then to deliver the answer you had already prepared, instead of feeling open and spontaneous together with the group. One of us described it like this:

Since we had different views on this, we discussed that sending out the interview guide in advance should be an offer rather than something that is routinely done with everyone. Alternatively, that everyone could receive the questions before the first interview session, but not for the following sessions.

Since interviews may spark thoughts that one doesn’t have time to express and process during the interview, we talked about the importance of having the opportunity to follow up with the interviewer (and possibly the rest of the group) afterwards, either verbally (with/to the interviewer) or in writing (to the interviewer or the whole group). One of us suggested that it might be good to set a time window for this:

Several of us mentioned that having a time limit for the interview felt reassuring. Even if we didn’t manage to answer all questions or say everything we may have wanted, a lengthier interview would mean a less safe participation due to risk of cognitive overload. One of us suggested the alternative that those who still have energy could continue the interview after a 15-minute break. The suggestion was to have a 45-minute interview, a 15-minute break, and then continue another 45 minutes for those who felt for it. The break would ensure some processing time and enable participants to return to the second session either with reflections from the first session — or with new thoughts.

At the end of each interview we talked about whether we felt a need for a follow-up interview to continue the conversations. However, we realized we couldn´t know this right away. After all interviews had been conducted, we arrived at not doing more interviews but encouraged everyone to add their reflective thoughts during the process of co-writing. The following dialogue about whether or not we should have more interviews took place during the first pilot interview:

As autistic people we have experiences of being perceived as “talking too much or too little”. This is commonly described in autism literature as difficulties in turn-taking in the conversation. This anxiety, of being unsure about “how much” including how detailed one´s responses should be (c.f. ‘infodumping’), and “when” one should talk was brought up as a challenge in the focus group setting- This can be referred to as part of a broader umbrella of “research anxiety” (33). One of us said:

The rest of us agreed with this. This is contrasted with a situation where one is only in conversation with the interviewer (individual interviews) — a situation with less cognitive load; less cognitive input, less multi-modal sensory and social complexity, and a different expectation in the turn-taking, including being expected to talk “a lot.” One of us pointed out that although individual interviews do provide more clarity about when to speak, they come with other disadvantages instead.

During the first interview, it was therefore suggested that in upcoming group interviews (in the main research project), the interviewer should direct questions to specific participants — “go around the group” with each question — in order to avoid uncertainty about who should speak. In the second interview, this was done, and the interviewer experienced it as helpful for everyone at the beginning. However, once the conversation got going, it instead became a barrier in communication. When someone started to speak, they stopped themselves because the interviewer hadn’t directed the question to them, leading to an unnatural interruption in the interview situation. We talked about how it’s important to agree within the group on when and how it’s okay to jump in or interrupt — and that interrupting doesn’t necessarily have to be a negative thing.

In the following we explore synchronous (such as real-time interviews) versus asynchronous (such as email interviewing or written interviews that allow for different cognitive processing times, c.f. 10, 34) data collection. We talked about how it’s challenging to come up with what you want to say during the (synchronous) interview time, and that the answers you give can end up being rather short or general because we weren’t “done thinking.” Questions during an interview may trigger a lot of thoughts that grow and take shape in one´s mind during the interview. However, they may not be ready to be expressed until at the end of the interview - but then, there’s no time left. One of us expressed it like this:

Similarly, another of us said that she may expresses herself briefly before she has had time to fully process her thoughts, and then at the end of the interview, she may feel that there was much more she would have liked to say in relation to what the others had said.

Another one of us expressed the opposite — that she was able to quickly formulate her thoughts and then became tired.

At the beginning of the interviews, the interviewer mentioned the option to write in the chat if one wanted to, but the chat was never used during the pilot interviews. Several of us suggested that it would be better if the interviewer were more concrete in their instructions about when and how the chat could be used; for example, ‘when you get an association while someone else is talking and are afraid you’ll forget it or won’t have time to say it’.

We talked about the challenge of, on the one hand, daring to associate freely and, on the other hand, staying on topic. We felt that it’s important for participants to feel free to make associations and express themselves without fear that someone else will perceive it as wrong or ‘off track’. Even though the interview focuses on a certain topic (in this case, autistic flourishing), sometimes you need to take detours to get to what you really want to say. We experienced that our mainly detail-oriented, bottom-up cognitive style involved telling a story step by step, where meaning is built gradually and the point is reached through the whole sequence. We discussed how it’s important that the interviewer guides the conversation back to the topic if we go ‘off track’, but not too abruptly, allowing for this step by step-storying. We talked about how it’s important for the interviewer to be clear at the beginning of the interview that there are no right or wrong answers, that the “range of relevance” is very broad, and that free associations are encouraged.

One of us described that the group interview felt comfortable precisely because it reduced the pressure to constantly have something to say oneself, and allowed time to process and make associations in your head while others were talking. Several of us also described that it was important to have the possibility to say “pass” if you needed more processing time, and that this option was introduced at the beginning of the interview by the interviewer.

As autistic people we have a tendency to consider questions very exactly and literally (35, p. 1134), also referred to as manifest communication. Our frustration regarding the lack of autistic languages (or pre-dominantly bottom-up naming) can be linked to this, as neurotypical language (or pre-dominantly top-down naming) is not suited for manifest communication. Our tendency to interpret things literally may have something to do with the idea that questions have one correct answer and that the “truth” is a valuable thing per se, and a sense that it is very important to get it precisely right. A problem may therefore present itself when the question in itself is difficult to understand, due to ambiguities of pre-dominantly top-down naming or if the fundamental problem underlying the question is unclear or hidden. “How do you feel?” might make autistic people think, “Why are you asking me this? Why is this important? Is it important at all? Or is it just a polite question with no deeper meaning?” The response will then depend on what the autistic person believes the purpose of the question is. Similar experiences were also described by autistic people in Finn et al.’s study (36, p. 2091), who found job interviews too challenging when the questions were not clear enough. This concept of taking questions very seriously and literally becomes important in the setting of research interviews. We believe that it is important not only to formulate the questions in a specific way but also to clarify the intention behind the questions and allowing for bottom-up step-by-step answering.

Another issue related to literal thinking and manifest communication is that there are so many levels on which to answer a question. Questions such as “what is a good life for an autistic person” or “How do you experience yourself having a good life - if you would describe it as sensations in your body?” may not lack clarity but they may be perceived as unclear regarding which theoretical level is being referred to because they are very broad (in line with a top down processing style). In line with a more detail-oriented, bottom up processing style, we may ask ourselves: “On what level should we answer? Why is this question important? What will it explain? Since autistic people often strive to answer literally we believe it is extra important to clarify at some point exactly on what level(s) of (top down) abstraction the study is aiming to explore. If the scope of research is very broad, it is important that researchers are clear about this as well. Questions that seem to hide a deeper meaning that we cannot understand may make us reluctant to answer and we tend to get stuck on definitions of words. However, when we understand what we really are “getting at”, or the concept of something, it is easier for us to relax and let go of the minuscule details and paint a more interesting greater picture. We therefore believe that, for autistic people to be able to share their experiences in an authentic way, it is fundamentally important to clarify the goal of the study, as this can help in understanding the questions despite different relationships with the individual words in the interview guide.

Hermeneutical marginalization of autistic people means that they not only struggle to describe their experiences but also to understand them themselves (37, p. 564). To investigate our own minds and bodily experiences, is difficult and intimidating. We may also have a strong urge of wanting to get to the core, the very essence, of things we investigate (including when we are being interviewed). We may be looking for something that could be unique to “us”, this group of people, who interpret ourselves as divergent from non-autistic people. Divergent in relation to people who we by definition can know only from secondhand account. How do we know in what ways we differ and in what ways we do not? Comparing oneself to someone whose internal perspective one has never experienced is a difficult, if not impossible, task. We can compare ourselves to the observed behaviors we see in non-autistic people, but we never gain full insight into the core that underlies these behaviors. We can take part in non-autistic people’ descriptions of how they function, but even then, we do not fully grasp it, as we may not always recognize the way they use words or the experiential sphere from which their descriptions originate (37, p. 566). The risks of framing responses based on comparisons with non-autistic functioning are that we may find it harder to articulate our thoughts (becoming uncertain about whether they overlap too much with non-autistic functioning) or that we may express them in an overly extreme way to clearly distinguish the autistic experience from the non-autistic one. This means that autistic voices may not be adequately represented. We also risk falling out of the neurodiversity approach and into the biomedical deficit perspective (38), which argues that neurotypicality is the human benchmark against which other neurotypes should be measured. We therefore suggest that, in interviews with autistic people, the focus should be shifted away from contrasting autism with neurotypicality and instead encourage autistic bottom-up exploration in a safer interview situation. The interviewer should inform participants that the study focuses on autistic perspectives, regardless of whether they overlap with a neurotypical perspective or not. The uniqueness of autistic thinking and perception will emerge naturally.

Furthermore, since we feel that we lack an autistic language, it may be valuable in future interviews to conduct more multifaceted interviews — focusing on communication through multiple forms of expression and documenting the full experience: not just spoken or written language, but also visual, sensory, and nonverbal ways of expressing oneself.

From a biopsychosocial perspective, we recognize that human functioning emerges from the interaction between personal, social, and environmental factors. In our case, we were discussing in groups of people (social) where most of us did not know each other in advance via a digital platform (environmental) online. This is practical and convenient (e.g., no travelling, easy to find a time when everyone can join and people can choose a place where they feel comfortable) (39, p. 1297), but it may not be the most liberating way of exploring our minds. It is known that online interviews come with certain risks, such as a sense of feeling restrained by the limited way people express themselves non-verbally during online interviews, issues with turn-taking and taking initiative to speak, and difficulties in speaking about personal matters (40, p. 6-9). In the present study, we noted a tendency to repeat and acknowledge what someone had already said, which could be due to the interview being conducted digitally, but could also be because we did not know each other and were therefore uncertain in the interaction. When people don’t know each other and the context is new, the interview can be perceived as “artificial”, which can reduce engagement during the interview (41, p. 4-5). Thus, there may have been a clash between our personal needs, other people, and the environment we were in. Regardless of the practicality of the digital setting, we wonder what it would have been like to dig deeper into the material of our collective mind in a physical meeting. To compensate for the limitations of digital platforms, we suggest conducting repeated digital group interviews. This could help participants become more comfortable with each other, the interviewer and the interview situation. The artificial interview situation becomes more of a natural social context as the participants, like other social groups, build relationships and a shared foundation to rely on. This enables deeper and more meaningful discussions (41, p. 5). To accommodate different cognitive styles and processing needs (some have a faster processing style, speak and get tired early, while others need more time to think before speaking), it can help to use a conversation format that allows for more variation. For example, starting with a bit of silence before people speak or working in pairs first can enable different cognitive styles and processing needs. It can likely have a positive effect on participants’ ability to share their thoughts when participation isn’t expected to be constant or to take a specific form.

The number of participants is also important to consider. As an interviewer, ÅH found that the optimal group size was three participants and herself/the interviewer. This group size kept the discussion lively without placing too much speaking expectations on anyone, while also meeting ÅH´s own processing needs, and enabled her to keep track of who said what and when. If the groups had been larger, it might have been difficult to fully “take in” the entire group, with attention potentially focusing on the speaker and missing what others are experiencing. We also felt that this number was optimal from the participants’ perspective, as it limits the amount of information they need to process.

The interview duration (45 minutes) were maybe too short. Several of us felt that we could have discussed longer. On the other hand, some of us appreciated that the time was not “too long” (so they didn’t get a cognitive overload). As a comparison, the non-autistic participants in Gray et al.’s study (39, p. 1295) felt that Zoom interviews should not last longer than an hour, as they similarly otherwise get a cognitive overload. Whether the time was sufficient in the present study may depend on the number of participants in the interview. The time did not feel as limited in the third interview, where there were two participants and the interviewer, compared to the first and second interviews, where there were three participants and the interviewer. Overall, however, our feeling was that the interviews did not manage to capture everything that should have been captured (there was not enough time to ask all questions in the interview guides and follow up all potentially interesting thoughts). This reinforces the argument above—that it may be a good idea to interview autistic people multiple times. This allows for deeper exploration of a topic while reducing risks of cognitive overload. Another idea could be to take a break after perhaps 40-45 minutes and then offer an additional interview session for those who have the cognitive capacity to continue. However, it is important to consider that this may not meet the needs of the autistic interviewer, as they need to remain sharp and focused the whole interview.