“All bets are off”: Flexible engagement with music-listening technologies by autistic adults

Participants

Our study involved a convenience sample of 12 autistic adults (11 men and one woman) between 18 and 37 years of age (mean age = 25.5, SD = 5.78). As a prerequisite for entrance to the internship program, all participants self-identified as having a community diagnosis of ASD or Autistic Disorder (AD). All interns scored higher than 80 on full scale IQ as measured by Wechsler Abbreviated Scale of Intelligence. Written informed consent was obtained from all participants in accordance with UC San Diego’s Institutional Review Boards, and research protocol was overseen by UCSD’s RADLab in the Department of Neurosciences. Participants were recruited on a volunteer basis and compensated. All names that appear in this document are pseudonyms created to protect participants’ anonymity.

Semi-structured interviews

We drew from DeNora (2000) in crafting our primary research questions, which explored how autistic adults understand their experiences with music and its roles in their everyday lives. We felt that the in-depth, semi-structured interview design gave participants the most flexibility to share their musical interests in rich detail, and to relate them to their daily habits. In line with methods of qualitative data collection, we designed our primary questions and follow-up probes to be open-ended and exploratory in nature.

Our interview guide was structured in two parts, beginning with a preliminary series of simple questions about the participant’s basic occupational and living situation, and method of commute to the lab. The two main portions of the interview centered on musical preferences and daily routine (Appendix 1, Interview Guide). In the first part, our probes focused primarily on autobiographical history with music and development of taste. In the second, we asked participants about their listening devices as they walked us through a typical day.

Interviews were conducted in a soundproof office in the lab and recorded using a hand-held digital recording device. Each interview was conducted by the lead author and assisted by the same team member. Neither administering researcher had any prior relationship to any of the participants.

Transcription, coding, and analysis

We based our analytic approach off of key strategies from Grounded Theory (Glaser & Strauss, 1967) and qualitative coding methods from Bryman (2016). As such our overall process was both iterative and recursive. Debriefs after each interview allowed both of the two administering researchers to discuss the sessions and document ideas in memos. After data collection, audio recordings of the interviews were transcribed and coded in Dedoose, a data-analysis app for qualitative and mixed-methods research (“Dedoose Version 8.2.14, Web Application for Managing, Analyzing, and Presenting Qualitative and Mixed Method Research Data,” 2019). Although our ultimate code categories adhered to and extended previously reported frameworks for understanding music in everyday life, we aimed to approach our data without any preconceptions. This guided the early analytical phases, which included immersion in the data by reviewing transcripts and memos, followed by several generative rounds of initial coding, done independently by the two administering researchers (Bryman, 2016). The final code categories and codebook were confirmed by all members of the research team before the final round of coding by the lead author.

Platform approaches

We observed that music technologies emerged as a natural focus in each of our conversations. In our initial coding, we were interested in understanding the specific roles technologies play in enabling daily listening habits. Rather than coding the platforms themselves, we chose to compare similar technological functions applied across a broad range of devices and apps. This was because many online-based platforms replicate functions of other more conventional music services. Apps like YouTube and Spotify, for example, offer streamed content through stations similar to broadcast radio. Therefore, the code “Following channels” was borrowed from the concept of following a media producer, and included this analogous function across all online, cloud-based, and AM/FM platforms. Our cohort explicitly described using five primary functions for discovery and organization, represented in Table 1 and Figure 1.

OPEN IN VIEWER

Table 1.

Platform approaches: code frequency.

Code	n participants	%
Searching	8	66
Getting recommendations	7	58
Making playlists	6	50
Listening to albums	6	50
Following channels	4	33

OPEN IN VIEWER

Figure 1.

Platform Approaches Code Web.

Activities

Our cohort used music to accompany a range of work-related and leisure activities shown in Table 2 and Figure 2. They are grouped here thematically for comparative purposes.

OPEN IN VIEWER

Table 2.

Activities: code frequency.

Code	n participants	%
Commuting	10	83
Working	7	58
Gaming	5	41
Sleeping	5	41
Waking up	4	33
Coping	4	33
Playing an instrument	4	33
Creating music	4	33
Creative thinking	3	25

OPEN IN VIEWER

Figure 2.

Activities Code Web.

Genres

In an early round of coding, we selected for mentions of preferred musical genres. Table 3 shows that Rock and Metal were the most popular musical genres in our cohort. We chose to code these two together because most mentions of either were grouped in context. However, several participants made nuanced distinctions between the two. At least half of the cohort also mentioned “Pop,” “Soundtracks,” and “Decades” music. Meanwhile, the three least popular genres with only one mention each were “Christian,” “Jazz,” and “Live music.”

OPEN IN VIEWER

Table 3.

Genres: code frequency.

Code	n participants	%
Rock & Metal	8	66
Pop	7	58
Decades	6	50
Soundtracks	6	50
Electronic	5	41
Classical	3	25
Musicals	3	25
Country	2	16
Disney	2	16
Rap	2	16
Christian	1	8
Jazz	1	8
Live music	1	8

Common among nearly all of the interviews was an effort to define and to describe tastes. This often led to detailed discussions of subgenres and nuances within styles. As mentioned, “Following channels” and “Listening to albums” were methods of distinguishing between these types of genre preferences. However, sometimes these musical identifications stood on their own:

I just like old school metal more than I like newer stuff . . . I can handle a bit of the screaming with the screamo . . . I do like some Megadeath and stuff . . . (Kevin)

A lot of Soundcloud rap, you know, Smokepurpp, Lil Pump, SHORELINE MAFIA, that kinda stuff. I like more EDM stuff. I like dubstep, trap, Kpop sometimes. (Dylan)

Lots of free jazz I find I don’t necessarily care for. Still might listen to that stuff though . . . free jazz is almost like the ultimate challenge musically . . . it’s like synchronized car racing or something. (Ethan)

Intentional engagement with technology

Our cohort demonstrated a range of personal approaches to using their preferred media platforms. Specifically, we were able to group responses into five major code categories of explicitly described technological uses: “Searching,” “Getting recommendations,” “Making playlists,” “Listening to albums,” and “Following channels.” These findings are in line with the complex patterns of engagement by TD individuals made possible by new music-listening technologies (Krause et al., 2015). Mirroring these results, our convenience sample actively engaged with the discrete choices afforded by music technologies. For example, more than half of our cohort used platforms like YouTube, Spotify, and Sirius XM radio explicitly for “Getting recommendations” and “Searching.” Results in these categories indicate clear patterns of behavior directed toward proactive and intentional engagement with music discovery. Half of our participants also reported “Making playlists” based on artists, genres, and miscellaneous favorites. Personal use of this organization tool demonstrated yet another axis of decision-making embedded within music devices.

The presence of “Albums” as its own method of consumption is noteworthy. Participants in our study referenced albums to comment on artists and trends and to exemplify their musical taste. As a unified musical product, albums therefore afforded participants one specific way to describe their criteria for preferences and decision-making. Two participants also said that they listen to albums in their entirety for their storyline and the narrative experience. Like listening to albums, the four participants who described “Following channels,” did so to make clear distinctions with regard to their preferred subgenres. This shows that some autistic adults use albums and music stations to self-curate their daily listening based on specific genre criteria.

The significant volume of responses in the “Platform approaches” codes suggests that our cohort overall makes full use of these music platform functions in their everyday lives. However, many of these actions like making playlists or listening to the radio may easily be taken for granted. We speculate that music technologies may offer even deeper affordances in daily life, such as defining sonic boundaries at home or attending to sensorial needs in personal spaces. Further research is needed to understand how autistic individuals use music for environmental control.

Music as a resource

Results in “Activities” agree with the broader understanding that people consciously use music as a resource in daily life (DeNora, 2000; North et al., 2004). The two most mentioned tasks accompanied by music were “Commuting” and “Working.” In their reasons for having or not having music in these scenarios, our cohort demonstrated awareness of environmental needs necessary for accomplishing these tasks. Characterizations of music’s impact on work productivity included helping to focus, staying energized, and isolating unwanted noise. The fact that commuting and working were inherent to the structure of their daily routines as interns was a unique aspect of our convenience sample. Since each participant commuted to work, we were able to ask them specifically about their habits with regard to these tasks.

The mention of using music to manage stress and emotions (“Coping”) is significant and demonstrates an additional internal awareness of personal needs. It also confirms previous findings by Allen et al. (2009) that autistic individuals use music for its therapeutic benefits. However, given the framework of our analysis, we are unable to comment on their corollary finding that autistic individuals favored using arousal states over valence terms to describe internal responses to music, possibly due to the presence of alexithymia (Allen et al., 2013; Allen et al., 2009). Our interview guide and coding scheme selected for discrete actions with music, rather than motivations for listening to it. Therefore, it is possible that if we had probed and analyzed for responsiveness, our cohort may have demonstrated a similar tendency. We do, however, interpret the presence of “Coping” as having some indication of an awareness of music’s mood-altering potential. Without having screened for participants for alexithymia, we cannot correlate these responses to severity of symptoms.

One of our most interesting findings concerned the use of music in creative contexts. “Creating music” and “Creative thinking” responses made up 50% of our sample and represented a spectrum of clear outcomes. These ranged from composing/producing music, writing rap lyrics, participating in stage production, designing video games, and making up tabletop game stories. Music-related activities like these extend previous findings that key motivations for listening to music among a sample of autistic individuals included “aesthetic pleasure,” a positive internal response to certain musical qualities (Allen et al., 2009). In the case of our cohort, musical appreciation included a desire for proactive and creative engagement with music. Given that participants in our convenience sample worked in a technology-based creative field, we speculate that they already saw themselves as creative agents. Overall, we interpret these data as suggesting a strong connection between music-listening and the activities and priorities most relevant to an individual’s present everyday life.

These findings also have important implications for therapeutic design, specifically in light of the observed benefits of creative media collaboration by autistic individuals (Hillier et al., 2012). The presence of self-motivated creative engagement with music and music technologies in our study suggests continued potential for non-performance-based music therapies.

It has been demonstrated that autistic individuals use music to feel a sense of social connection and belonging (Allen et al., 2009). In this study, social activities and attitudes toward others were occasionally brought up in the interviews, but were not the primary subject of our analysis. One participant, for example, explicitly described wanting to be a part of a creative team that produces musicals. Since this was our only explicit mention of social belonging, we chose instead to compare it to other examples of “Creating music.” On a few occasions, participants shared rich descriptions of listening to music with family and friends, especially as it concerned boundaries, memories, and differences in taste. Thus, our focus on daily routine in the interviews lent itself most naturally to an analysis in terms of these sorts of activities accompanied or inspired by music.

Genres

Similar to findings by Hillier et al. (2016), our cohort of young adults demonstrated a clear preference for various popular music genres, as opposed to classical music. Due to the small sample size, we are limited in our ability to generalize these data to other autistic populations. Instead, we note that the presence of a range of musical styles within our cohort is indicative of their ability overall to specify and articulate the nuances of their musical preferences.

We speculate that music-listening platforms play a role in teaching its users new genre-related vocabularies and how to engage with them. For our cohort, favorite channels and albums were reliable sources for preferred music, as well as discursive tools for expressing their tastes in the interviews. Future work in autism and music technology may explore the impacts of reflexive learning through technology, and how it is used independently by autistic individuals to construct taste.

Reliability, generalizability, and comparison groups

Since we were unable to double code in the final round of coding, we cannot provide intercoder reliability with respect to our ultimate results. However, the extensive process of interpreting the data through discussion and generating categories, which did include blind double coding, was recursive and exhaustive and followed standard procedures for qualitative analysis. The fact that we were also unable to conduct follow-up interviews limited the degree to which our analysis could be confirmed or expanded.

In addition, the relatively small sample size and gender imbalance limits the degree to which our findings can be generalized to other autistic adults. Nonetheless, the makeup of our convenience sample being video game designers allowed us to examine the specific use of music for creative inspiration. Further research is needed to understand how other life-stage and occupational factors manifest in one’s daily choices of music.

The use of a comparison group in a future study would enable inquiry into differences and similarities between ASD and TD groups with respect to listening patterns. Since these behaviors appear to be closely tied to many aspects of daily routine, a study would also benefit from selecting inclusion criteria based on controlled sociological factors. For example, the daily lives of ASD and TD college commuter students might provide a rich framework for comparison. Of interest would be specific of outcomes with music relevant to student life and how they are achieved using music-listening technologies. For our cohort, exploring and selecting preferred musical genre was an important motivational factor when engaging with music technologies. A comparison group would further enable exploration of how taste and musical identity influence platform approaches across ASD and TD groups.