Food for thought: Immersive storyworlds as a way into scientific meaning-making

2.1 Immersive storyworlds

Many short-term science events involve forms of storytelling—whether in theater, podcasts, films, or readings—to communicate both the products and processes of science (Dahlstrom, 2014; Martinez-Conde & Macknik, 2017). Recently the field has addressed how storytelling can be an important tool for engaging diverse and non-science affiliated audiences (Joubert et al., 2019). Its utility for science communication includes addressing multiple dimensions of learning (visual, auditory, and embodied), embedding science within a narrative structure to aid with meaning-making and memory, and creating a programming template or format that science communication professionals can adapt and expand to new topics or audiences (Finkler & Leon, 2019; Martinez-Conde & Macknik, 2017).

While traditional storytelling typically consists of a linear narrative arc with a central protagonist, there is a growing form of story-based leisure time activity called immersive storytelling. In immersive storyworlds participants become actors in a storyline, which is usually characterized by branching plot lines that allow for varying degrees of improvisation and authorship by participants (see Bevan et al., 2021). Immersive storyworlds leverage and compete with purely digital media, and include multimedia experiences using augmented reality such as Pokemon Go; physical narrative experiences such as Escape Rooms or Punchdrunk Productions; and “transmedia” experiences such as the Marvel Cinematic Universe and affiliated spin-offs, World of Warcraft, and Comic Con.

While the term “immersive” is contested in the literature (Gröppel-Wegener & Kidd, 2019), the general notion of becoming a character or taking on a role within a fictitious world appears to be a central characteristic of immersive storyworlds (Warren, 2017). As noted by Bevan et al., 2021, communications scholar Henry Jenkins distinguishes between interactive and participatory storyworlds, with the first offering choices within a participation structure (such as many video games) and the latter offering more opportunities to fundamentally change the nature of the experience. In both cases, he describes transmedia immersive storyworlds as leveraging multiple different modalities and media, unfolding over time in serial form, and entailing performance on the part of participants (Jenkins, 2006, 2009). In a discussion of the neurological and cognitive dimensions of immersive experiences, Martinez (2014, p. 111) notes that storyworlds may generate “attention priming, empathic attachment, and emotional involvement in narratives.” Scholars also note that role-playing within storyworlds can be rewarding for both recognizing existing expertise and knowledge, thus legitimating one's value in a communal activity, and for providing opportunities for accessing new insights and skills (Finnegan, 1997; Hergenrader, 2020; McAdams, 1993).

The aesthetically and emotionally saturated immersive storyworlds of Guerilla Science, the program we report on here, scaffold elements of theater, experience design, and the visual arts into rich narratives that physically place participants in relationship to science. In taking up roles within the storyworld, participants must engage in scientific practices and conversations. In the case of Guerilla Science, participants role-play being patrons at a diner that serves exclusively a range of roasted insects (see Figure 2). Taking the plunge, embracing the role, and constructing the story through active participation involving eating the insects opens up opportunities for developing a sense of belonging in sustained science conversations.

3.1 Affect and belonging in science

Belonging has been described as a fundamental need by scholars Baumeister and Leary (1995). Belonging has been linked to a range of positive outcomes, both academic and social–emotional (see Allen et al., 2021 for an overview). Scholars have referenced belonging in terms of the “belongingness hypothesis” (Baumeister & Leary, 1995) and “a sense of belonging” (e.g., Hurtado & Carter, 1997; Trujillo & Tanner, 2014). Others have described it as “fitting in” (Rosenberg & McCullough, 1981 cited in Brockman, 2021; Walton & Brady, 2017). Recognition, a key definitional component of identity (e.g., Carlone et al., 2015) is also seen as a key feature of belonging (Dou & Cian, 2021; Trujillo & Tanner, 2014; Walton & Brady, 2017). Others have described belonging as involving a feeling of importance within a social setting such as school (e.g., Finn, 1989). It has also been described as having a sense of membership (Hurtado & Carter, 1997). Most of these studies have occurred in the context of schooling, and thus are able to observe and consider belonging as it evolves over time within more or less stable social contexts. Yet scholars have also noted the importance of opportunities for belonging as requiring both frequency and duration, and as being nurtured across settings including in casual out-of-school (e.g., family, peer) conversations and activities (Dou & Cian, 2021; Hurtado & Carter, 1997). Such views reflect an ecological approach to understanding how one develops, learns, and becomes across time and setting (Barron, 2006).

There has also been important work on designing social settings to support a sense of belonging (Calabrese Barton et al., 2021; Ong et al., 2018; Varelas et al., 2021). Walton and Brady (2017) describe how small social cues can shape initial reactions to whether a person feels that they belong or not. Creating welcoming and inviting spaces and relationships is key, as well as explicit moments of recognition that a person is present and “fits” in. Settings can be designed to support a sense of togetherness or camaraderie. Researchers have explored how such settings can create “counterspaces” (Ong et al., 2018) for those marginalized from science, contributing to their motivation to persist and commit to the social context (Brockman, 2021; Trujillo & Tanner, 2014).

Designing spaces for belonging thus blends strategies for supporting cognitive, embodied, as well as affective dimensions of participation in the activities of the social space. For example, research has noted the importance of affect in motivating and sustaining engagement with science (Martin et al., 2016; McKinnon & Vos, 2015), including how it is inextricably intertwined with the disciplinary practices of doing science (Jaber & Hammer, 2016). Over the past two decades, a growing body of sociological research has documented the role of emotion in science classrooms, focusing, in particular, on gestural, physical, and vocalized forms of emotional engagement (Bellocchi et al., 2017), and in particular laughter (Morrison, 2008; Roth et al., 2011; Scott, 2011). Scholars have noted the ways in which such emotions are both markers of and resources for deepening participation in learning (Bolen, 1999; Linnenbrink-Garcia et al., 2016). Importantly, emotions are explored not as individual traits but as emergent characteristics within social settings (Roth et al., 2011). Davis & Bellochi (2018, p. 366) argue that, in social contexts, emotions, in the form of feelings, constitute a “we-relationship,” suggesting a sense of shared perspective, community, and belonging. Roth et al. (2011) argue that laughter, in group contexts, is a mutually performative act that forges feelings of solidarity. Solidarity in turn reinforces a sense of belonging.

Research has documented how science programs designed for individuals from communities historically excluded from science demand particular attention to building a sense of solidarity and belonging (Calabrese Barton & Tan, 2020; Leggett-Robinson et al., 2018; Nasir et al., 2014). Belonging is constructed dialectically through both individual activity—actively participating, contributing, and leveraging one's own experiences in an activity—and group activity—building on and developing shared purpose, language, routines, and a growing sense of identity (Heath & McLaughlin, 1993; Lave & Wenger, 1991; Nasir & Cooks, 2009). Thus, science education programs seeking to enhance a sense of belonging for individuals who do not, for different reasons, affiliate with science, seek to ensure that the lived experiences and cultural practices of individuals—their histories and their selves— are positioned as the means for successful participation in science learning, and not as barriers that must be overcome (Bang et al., 2013; Nasir et al., 2014).

The emerging field of “inclusive science communication” builds on this research in its calls for science communication to more intentionally design for people who do not already identify, and may not feel a sense of belonging, in science (Canfield et al., 2020; Massarani & Merzagora, 2014; Polk & Diver, 2020). This literature suggests that a sense of belonging can be promoted through a range of pedagogical tactics such as using asset-based framing, integrating everyday science practices and language into the engagement, and framing science around issues of interest and importance to the target audience (Canfield et al., 2020; Gunter et al., 2021). Our study brings the importance of belonging into tension with short-term, context-dependent interactions in which a sense of belonging may be ephemeral and tied to belonging in the moment itself (rather than to a disciplinary practice or social construct such as “Science”). Like Holland et al. (1998), we propose that such momentary, even serendipitous, experiences are important windows into spaces not typically traveled that can provide, over time and space, the social means for constructing interests and identities.

3.2 Carnival and belonging in immersive storyworlds

Atkinson (2011) has described carnival as an experience that leans into “that which is not yet,” what Holland et al. (1998) would call “a figured world.” In many ways, carnival represents proto-historical instantiations of participatory transmedia experiences. For example, carnival has dimensions of seriality, multimodality, and performativity. However, contemporary transmedia storyworlds are often described as deeply saturated across time and space as afforded in particular by digital technologies in ways that carnival and the immersive science program we describe here are not designed to offer. Where transmedia events are generally co-constructed by and for self-avowed fans or aficionados, the form of carnivalesque immersive storyworlds we describe here are more ephemeral, designed to waylay the person who bought a ticket to a music festival or a country fair, and thus may not be actively seeking to engage with science.

Bakhtinian theories of carnival have a rich history as a research lens. In addition to their use in understanding the contemporary festival and its position as an emancipatory space for learning and development (Anderton, 2018; Bennett & Woodward, 2016), as noted by Bevan et al., 2021, they have been used to study language learning in elementary classrooms (Iddings & McCafferty, 2007; Lensmire, 1994; Manyak, 2001), early childhood learning (Tallant, 2015), second language learning (Ahn, 2016), drama instruction (Tam, 2010), learning in museums (Jun & Lee, 2014), and science learning in the classroom (Roth et al., 2011). Scholars note how, within the play of carnival, participants take on roles, routines, and speech patterns that invert power structures (e.g., students playing the role of teacher) and allow for the development of hybrid positionalities with respect to the subject matter (e.g., language, science, and drama) that open up opportunities for learning (Iddings & McCafferty, 2007; Tam, 2018). Manyak (2001) reported how negotiating such role-playing created a sense of agency among students as they leveraged their experiences to reproduce official forms of authority (such as voicing the role of a teacher).

Importantly, in carnival, such roleplaying differs from that of traditional theater, as carnival makes no distinction between performer and audience. “Footlights would destroy a carnival. As the absence of footlights would destroy a theatrical performance,” wrote Bakhtin (1984, p. 7). “Carnival is not a spectacle seen by the people; they live in it, and everyone participates because its very idea embraces all the people.” Carnival is constructed through acts of participation and belonging. Some researchers have referred to this phenomenon as “living the carnival” (Vigmo & Lantz-Andersson, 2014). In living the carnival, participants enter a figured world where, with irony or humor, they collectively challenge or even transgress social norms, often through embracing the grotesque, the blasphemous, or the obscene. In educational contexts, as described above, the grotesque is characterized by settings that are organized orthogonally to archetypal classrooms, where the context is deliberately ridiculous, chaotic, or perhaps unguided. Carnival is fully dialogic, emerging from social interactions that leverage diverse perspectives to construct the communal play and its meaning. Researchers argue that the dialogic nature of carnival creates opportunities for new ways of understanding, where participants, at an individual as well as a collective level, internalize, integrate, and relate new ideas and perspectives into their own evolving worldviews and understandings (Morson, 2004).

Through this Bakhtinian lens, the central conjecture of our study is that (1) the Guerilla Science storyworld—constituted by the theatrical roles, routines, props, and the central transgressive narrative conceit itself—creates an emotionally-saturated context that invites and scaffolds participation, where (2) participants' embodied selves and prior histories are positioned as resources for successful participation in roleplay within the storyworld, which in turn (3) enhances a sense of belonging in the storyworld that motivates and sustains ongoing participation and scientific conversation. These three dimensions of our central conjecture guide the coding and analysis reported below.

4.1 The Entomophatron

In the summer of 2018, we observed Guerilla Science's Entomophatron at a country fair 90 minutes north of New York City. Each year, audience members—a mix of families, teens, and adults made up of local small town tradespeople and professionals, weekenders from New York City, and seasonal agricultural workers—come to the fair to see local livestock, ride roller coasters, hear country western music, and eat carnival fare like deep fried blooming onions, hot dogs, cotton candy, shaved ice, and the like. In 2018, in one corner of the fair they encountered a booth fitted out as a 1950s American diner, with a linoleum counter and four vinyl covered stools (see Figure 3). Participants were often lured to the diner counter by an old fashioned popcorn maker that was producing free bags of popcorn, which they were then told had been seasoned with agave worm salt. Male and female wait staff, all wearing pink and black polka dotted dresses, following the Blondie comic strip motif, beckoned customers to sit down on a stool and take a look at the menu. The diner had three counter stools, but was able to accommodate crowds seated and standing of up to about 10.

The food at the Entomophatron was free, visitors were assured. A glance at the menu revealed descriptions of four different courses with enticing names; but the fine print revealed that each involved dried and roasted insects—crickets, mealworms, and ants (see Figure 4). These included, in the order listed: A blind taste test involving a cricket versus a bean nacho chip; the Bugtucky Derby, which was not a meal but a race between giant Madagascar Hissing Cockroaches; whole roasted mealworms; and The Chef's Surprise of “ants-on-a-log.” Additionally, almost always participants were offered a roasted whole cricket, and frequently science communicators offered participants a palmful of roasted whole ants. Typically, the interactions between the waitress/waiter and the diner patrons moved from initial expressions of surprise or apprehension into questions and dialogue about food sustainability and the role of insect protein in the human diet, historically, culturally, and in the future.

4.2 Data collection

A primary principle of our RPP was that the data collection would not disrupt the participant experience in the immersive storyworld. We, therefore, attached a handheld video camera and directional microphone to a corner of the diner booth. The camera was angled to pick up primarily hand gestures and dialogue; most people's faces were only partially visible in the recordings. When the diner was crowded, some participants were entirely off camera. A sign, graphically treated to fit within the aesthetic of the diner, was posted at the booth notifying participants that video was being recorded for a research study.

Over 3 days, we video-recorded a total of 48 social units (families, couples, and friend groups), involving a total of 134 individuals, participating in the activity. Each interaction, from when participants first came to the counter to when they walked away (which averaged almost 12 min [SD 7.8]), constitutes a “clip.” When social units overlapped in time (e.g., two families sat at the counter at the same time) but did not interact with one another, we coded the same video data as separate clips. To identify clips that offered rich glimpses of the performative and dialogic nature of the experience (Derry et al., 2010), we initially coded the 48 clips for instances where: (1) An insect was ingested; (2) a participant expressed prior awareness of insects as a human food source; (3) a participant shared some aspect of themselves (diet, allergies, and prior experiences); (4) there was a verbal Q&A exchange between science communicator and participant; or (5) the science communicator shared factual information. When at least one person in a social unit met three or more of the initial criteria, we subjected those clips to secondary coding. The 31 clips we identified through this process were almost twice as long, on average, as those that did not meet our criteria for secondary coding. The 12 clips that we did not subject to secondary coding, and thus that are not included in the analysis discussed below, documented people observing and listening, but speaking little or not at all (as can be seen in Table 1), and often not eating an insect. Therefore, though they may or may not have listened with interest, and walked away with new insights or not, our non-intrusive data collection methods do not allow us to make any meaning of their participation.

The rest of this article refers only to the 31 clips we subjected to secondary coding. Although we did not collect any demographic data from participants, the 102 participants in the transcripts subjected to secondary coding appeared to be mostly in their 30s–60s, with just under one quarter appearing to be teens or children, and five individuals appearing to be in their 20s. They overwhelmingly presented as white, as is representative of the region. Participants appeared to be evenly split between female and male.

4.3 Secondary data coding

The 31 clips, constituting a total of 363 video minutes, were transcribed into 31 Word documents. The research team combined a thematic content analysis of the transcripts with theorized constructs identified in the literature cited above to iteratively develop a list of tentative codes. Because our RPP involved a mixed team of researchers and practitioners engaged in collaborative analysis, it was important that our codes reflected both theoretical and practical perspectives about what constituted meaningful themes or constructs. As Xu and Zammit (2020) note, “Informed by grounded theory, codes can come from the data itself (inductive coding) as well as particular theoretical or epistemological positions (deductive coding; DeCuir-Gunby et al., 2011).” They note that this blend of inductive and deductive coding provides “a balanced, comprehensive view of the data, instead of purely relying on the frequency of codes decontextualized from their context” (p. 3). In our project, this process initially yielded four categories (theater/storyworld, emotional response, science ideas, and scientific epistemic practices). To address issues of trustworthiness (Lincoln & Guba, 1986), we standardized coding across the three coders by developing a codebook that included definitions and examples. Each transcript was coded by two coders. We reconciled questions and inconsistencies through discussion, achieving over time a shared understanding and agreement across coders. We then reviewed and consolidated the codes to create a new set of four top-level codes, with a total of 20 sub-codes and definitions (see Table 2).

Because faces and bodies were often off-screen, for reasons described above, we did not formally code for gestures. However, when the transcript contained words that could be interpreted in more than one way—such as “Really?” or “No, no, no,” where these utterances could have been stated with either apprehension or humor—coders reviewed the video data to attend to vocal intonation, or facial or bodily expressions, to clarify meanings. This meant that we brought our own cultural perspectives to bear on recognizing and coding emotions, which limits the claims this study can make about the particular dynamics of emotion and engagement in the storyworld. Finally, as described in the next section, to address confirmability (Lincoln & Guba, 1986), or lack of researcher bias, we ran descriptive analytics of frequencies and patterns of codes. That is, each of the transcripts had several more or less dramatic aha-type moments that shaped our emerging understanding of the patterns among the constructs. Running the descriptive analytics allowed us to check the frequency of these patterns across all lines of the transcripts, including the much smaller moments in the interactions.

The sub-codes for Theater represent elements that could be leveraged to create and sustain the storyworld. For example, the central conceit, meaning the underlying fictitious premise, of the insect diner could be asserted verbally when a waitress asked people to come join them at the counter, or reasserted, in the course of an interaction, by the waiter making claims about “the culinary standards” of the establishment. A participant, playing the role of diner patron, could formally request the next item on the menu. Both wait staff and customers could use the menu as a prop for sustaining participation by offering or inquiring about an item on the menu. Routines were organized around stages in the menu, such as introducing a blind taste test or contrasting the nutritional value of gummy worms compared to mealworms. Both science communicators and participants could reassert their roles of waitress or customer to initiate new phases of conversation or participation.

The sub-codes for Emotion represent, broadly, the range of reactions to the Entomophatron experience evidenced in the video recordings. Sub-codes for Belonging, conceptualized as having a sense of acceptance, value, and respect (Goodenow, 1993), were evidenced in the ways in which participants participated at (a) an individual level, for example, by sharing aspects of themselves or their histories, and (b) a group level, for example through encouraging or daring one another in the shared activity at the diner.

Science codes include the nature of the topics of scientific conversation that emerged in the data, whether expressed by the science communicator or the participant, as well as the participant's practices of science as defined by McNeill et al. (2015) in their parsing of the nine science practices identified by the U.S. National Academy of Sciences (2012). We coded for investigation when participants (a) engaged in explorations of the materiality (taste, texture) of insects or (b) asked factual questions about the insects (e.g., edible insect production and preparation). We coded for scientific sense-making when participants (a) verbally contrasted or discussed tastes and textures, in other words made meaning of their embodied experiences at the counter, or (b) discussed the possible benefits of increased insect consumption in Western diets. We found no examples of scientific critiquing.

Because we were coding the nature of the interaction and conversation, where participants and science communicators all took on improvisational roles, we coded utterances by all parties. The purpose of our analysis is not to assign a particular valence to one side or the other but to understand the nature of the interaction as a jointly produced social interaction in the context of science. The vast majority (78%) of utterances coded as scientific facts or ideas were made, as would be expected, by the science communicator, whose role in “the play” was to introduce these ideas to participants. The majority (67%) of utterances or actions coded as scientific practices, when people were initiating an investigation or beginning the process of sense-making, were made by participants, in their roles as “diner customer.”

4.4 Frequencies and patterns

The central conjecture of our study is that the storyworld creates an aesthetically and emotionally saturated context that affords opportunities for belonging in the science-rich discursive setting that is the storyworld, which in turn can support sustained science conversations among participants and science communicators. To enhance trustworthiness (Lincoln & Guba, 1986), we explored whether this relationship among constructs occurred with any regularity. That is: When passages were coded as Theater, did passages coded as Emotion follow? When clips contained a large number of instances that were coded as Belonging, were there more utterances coded as Scientific Practices, Ideas, or Facts? Additionally, we conducted a statistical analysis of the codes to determine whether there were any patterns confirming or disconfirming our conjectures.

Across the 31 clips, there were 1402 coded utterances. We ran descriptive analytics of code frequencies, both in terms of numbers of instances and in terms of percentage of total conversation, within interactions and also analyzed by duration of interactions. We first conducted a correlation analysis across the 31 clips (using the Pearson product–moment correlation coefficient) and found there to be moderate to strong correlations between all instances of the codes (see Table 3).

To explore the relationship between codes within a given conversation, we considered the adjacent two-point correlation functions for successive entries; that is, instances where two particular codes appeared immediately next to each other in the transcripts within a single conversation. We analyzed the frequencies of each pairing, e.g., the number of instances in which Theater was followed by Emotion, and so forth (see column B in Table 4). One interpretation (but see Limitations section) of the correlated two-point codes is that the former code triggers the latter. That is, for example, if the presence of Theater and Emotion are frequently correlated, asserting Theater generally triggers Emotion; this is an assumption in the practitioner literature about how stories can engage science participation (Martinez-Conde & Macknik, 2017). For the four top-level codes, this led to 16 combinations, into which we sorted our 1402 data points, while retaining statistical significance. We were not able to analyze the data using the sub-codes, as the 4000 (20 times 20) different two-point combinations precluded our results being statistically significant.

Table 4 shows the results of this analysis. We found that in 13 of the 16 combinations there was a degree of correlation between adjacent codes that could not, at the p < 0.05 level, be explained by chance alone. The associated Chi-squared analysis supporting this, under two distinct interpretations, is shown in Table 4. Our results were robust, whether or not we assume that, in the first instance (column C), the presenter steers the conversation with a script, and in the second instance (column D) the conversation is more reciprocal and free flowing. In reality, as the excerpts below will show, the interactions were somewhere in the middle: Presenters typically followed a common sequence of activities and deployed variations of set roles, routines, and science ideas and facts that they introduced at key moments in the interactions, regardless of when these moments arose. (An example is at whatever time they introduced the taste test they referred to “specimen A and specimen B.”)

We also segmented our results by conversation-duration using time windows of <5 minutes, 5–10 minutes, and >10 minutes to see how the length of the conversation related to its structure. We found no significant difference in the structure between these windows, except the unsurprising result that longer conversations contained more instances, but not a greater percentage, of certain codes. We found our results to be robust for other similar time windows. In the next section we provide a qualitative analysis of the patterns that emerged in this statistical analysis.

5.1 Storyworlds inviting and scaffolding participation

In this section we provide examples of how the theatrical elements of the storyworld—the central fictional conceit or premise, the roles, routines, and props—were leveraged to initiate and scaffold participation. By “routines” we mean the ways in which science communicators call on a number of activities or discursive assertions within an improvisational framework, in response to the moment. For example, the routine of the blind taste test is usually but not always offered first. A discussion of cricket farming often but not always ensues, and waiters typically share a set number of facts at these moments, sometimes more and sometimes less. In these examples, participants' emotional responses to the storyline—apprehension, disgust, bravado, and humor—are also highly salient. Our conjecture that the theatrical dimensions of the storyworld would create emotionally saturated contexts that invited participation was confirmed in our quantitative analysis which found that conversations in which Theater was asserted frequently were conversations in which Emotion was expressed frequently: The Pearson correlation coefficient = 0.71. Furthermore, we found that when a theatrical dimension of the storyworld was asserted there were 102 instances of emotions expressed immediately afterwards (47% of the time). A chi-squared test shows this result cannot be explained by chance at the p < 0.05 level.

In Clip 1A, the waiter is clearing the counter and putting away insect food items after the departure of a large crowd. A boy, about 9-years old, who we call Omar, had been hovering in the back of the large crowd for at least 10 min but has now made his way onto a stool at the counter. He is silently, patiently watching the waiter clean up when a 6-year-old girl, who we call Ella, clambers up to kneel on the diner stool and leans forward to see what is happening. She is accompanied by her mother and another woman with two children who do not come to the counter until later. The waiter acknowledges Ella's arrival and offers her a menu.

Clip 1A illustrates how the science communicator initiates the storyworld through repeatedly asserting his role as the waiter and the role of the participants as customers at the diner. The prop of the menu scaffolds the interaction, providing the young people a sense of the experience to come. In line 16, the waiter calls on one of his routines by suggesting he can start them on the menu, which will be the blind taste test of a bean and a cricket chip.

In asserting the storyworld, in many cases the initial invitation is met by reactions of playful horror, disgust, and a kind of morbid curiosity. In Clip 2A, a middle-aged male–female couple, who we call John and Joanna, and their female friend, who we call Nancy, are walking by the booth when they stop for popcorn. The waitress beckons them over to the counter.

The hesitation of the participants is noted in lines 4–7, and also in line 13. In line 15 the male participant echoes the waitress by ironically reinforcing the rules of the diner. Initial disgust or dismay was common, but it was generally playful. In our analysis of whether initial expressions of disgust versus delight correlated more strongly with a choice to participate we found no notable difference, suggesting that the storyworld successfully scaffolded emotional reluctance into full participation.

Playfulness and laughter saturated most of the interactions, as can be seen in line 15 above and also in Clip 3. In the next example, a man who has shared that he is a professional exterminator has joined the counter, where a woman and her son (about 12 years old), and an adult female friend of hers, have been participating for about 5 minutes already. As he joins, he notes the male waiters' dresses and comments that clearly he is joining “a fancy party.” The group at the counter has already completed the blind taste test and eaten a roasted cricket, and now Waiter Y offers a jar of mealworms.

In this clip the exterminator plays the role of diner patron by, in line 23, asking questions about the cholesterol of mealworms as if he was considering food options. This provides an opportunity for solidarity when a woman at the counter, in line 24, humorously plays along with cholesterol being the reason not to eat the mealworm. The roles of the two participants create an improvisational opportunity for the waiters to share, in lines 19–26, information on protein and fat levels of mealworms and crickets as compared to other kinds of food. In line 27–30, both Waiter Y and the exterminator reaffirm one another's role in the storyworld by engaging in the pretense of the diner as a fine dining establishment.

5.2 Leveraging present selves and prior histories as resources for successful participation

The dialogic nature of roleplaying affords opportunities for belonging in the science-rich discourse space of the storyworld. In their roles as diner patrons, participants are able to assert their past experiences, their likes and dislikes, and their opinions in relation to the experience and conversation at the diner. There were 213 instances where participants shared prior experiences, understandings or otherwise shared some aspect of themselves, including dietary restrictions or food likes and dislikes. We also saw many instances of in-group-ness, which we coded as camaraderie or membership within the science-rich discourse space. Overall, there were 270 instances of statements or gestures that were coded as Camaraderie in terms of urging, approving, revoicing, or otherwise indicating belonging in the play of the storyworld. Such utterances or exchanges represent about 27% of the total number of coded exchanges, and they increased in number in the longer events. Our conjecture that the theatrical or carnivalesque dimensions of the storyworld would create opportunities for belonging in the scientific discourse that constituted the storyworld was confirmed in our quantitative analysis: we found that conversations in which Theater was asserted frequently were conversations in which Belonging was expressed frequently (Pearson correlation coefficient = 0.75). The number of adjacent two-point correlation instances, that is when an occurrence of Theater was immediately followed by a sense of belonging, was 48. However, this two-point correlation result does not pass the Chi-squared test at the 5% confidence level, suggesting that if the assertion of a storyworld does lead to a sense of belonging, our data do not suggest that it does so instantaneously.

In the next example we draw again from Clip 1, just over 10 minutes into Ella and Omar's experience at the diner, to illustrate how the two children leverage their roles as patrons at the diner to assert themselves and their likes and dislikes into the dialogue with the waiter. Ella's mother and friend, with two boys about 7-year-old, are also present at this stage in the clip. We omit a couple of lines of dialogue between the friend and her boys.

Omar not only asserts his embodied self, when he says in line 121 that he is allergic to nothing, but as shown in line 132, he also asserts his emotional bravado in eating the ants. Ella too asserts herself in line 123 as having equal qualifications for belonging as does Omar. Later, in line 135, Omar turns to Ella to both encourage her and to communicate his own in-group membership in the insect diner culture.

Clip 4A offers a different view of how the storyworld supports roles and routines that afford belonging, as the diner creates a context for a man to share part of his cultural heritage with a younger colleague.

In Clip 4A, Manuel expresses a sense of belonging (acceptance, value, respect) at the insect diner, and in the science-rich discursive space it constitutes, by repeatedly sharing his knowledge and experience of a wide variety of insects and their culinary properties. He also references his homeland in Oaxaca and shares, in line 88, his intention of educating his colleague about these valued cultural and culinary practices. Throughout, Manuel asserts his knowledge as valuable to the science communicator and in the process, which intertwines the nostalgic and also instructive stories from his personal life with his professional expertise as a chef, establishes a sense of belonging in the dialogue.

In Clip 1C we return to Ella and Omar to illustrate how group membership in the discourse space has evolved. A young boy, who we call James, of about the same age as Omar has been watching from the side for just over a minute, but rejects an invitation to try the ants-on-a-log (see Figure 5). Soon James is joined by his father and brother. The father is interested in the insects, and, at first, when his father examines a piece of the ants-on-a-log, James chants “Eat-it, Eat-it, Eat-it” and then lifts his arms in victory when his father eats the ants and shouts “Yay!” But this attitude soon changes as the interaction continues below.

The excerpt in Clip 1C illustrates how the emotional context created by the storyworld is saturated with risk, fear, disgust, bravado, and delight creating varied opportunities for belonging in the science-rich discourse space. In this example, James both expresses his disgust with his father eating ants and reveals, in line 199, that he himself has previously eaten a worm. The camaraderie expressed by Ella and Omar, in the chanting they do in line 188, echoing the chanting James had done earlier, suggests a feeling of in-group or membership within the storyworld. Ella also asserts membership in line 175 when she tries to get her mother to try the agave worm salted popcorn. Omar asserts membership in lines 179, 187, and even 197 when he scorns James's rejection of the value of eating insects. In line 200 Ella seeks to include James in the storyworld “cast of actors” by offering him popcorn.

In Clip 2B, John, Joanna, and their friend Nancy engage in various scientific sense-making activities, related to appraising and responding to the insects, while referencing their personal experiences in an ironic manner. The excerpt starts with the waitress putting away the prop of a jar of gummy worms that the science communicators commonly present as a comparative way of introducing the nutritional value of mealworms.

In Clip 2B, the excerpt shows how individuals expressed in-groupness through forms of camaraderie. In lines 108–109, two participants in the triad voice shared apprehension. In lines 111–112, the two female participants agree that the agave worm salt popcorn is good, and the waitress, in line 113, agrees, sanctioning a shared experience. In line 116, Nancy, the friend, urges John to take the plunge and eat the mealworm. Joanna reluctantly joins in, line 118. In line 122, Nancy mugs for another one of the science communicators who is taking pictures of the event. In lines 124–125, the married couple makes a joke about dead insects they see in their basement. In this context, the science communicator is able to share the idea that mealworms are raised for human consumption—that there is an agricultural enterprise around insect production.

5.3 Belonging as a motivator of sustained participation in science engagements

Our statistical analysis of the frequency of codes found that the more time participants spent at the counter the more science ideas and practices were engaged. We found that across the 31 clips, the percentages of codes were fairly consistent, with a mean of 37% (SD 0.11) of coded activities relating to science. However, the number of science coded exchanges was substantially higher for longer events, although the relative proportion of the conversation coded science remained fairly constant: 37%, 41%, and 34% in the long (>10 minutes), medium (5–10 minutes), and short (<5 minutes) conversations, respectively. Put simply: with more time, there were many more moments coded as scientific ideas or practices.

Our conjecture that developing a sense of belonging in a science-rich discursive space, afforded by the storyworld, would sustain participation and thereby contribute to increasing opportunities for science conversations was born out by our quantitative analysis, where we found that conversations in which Belonging was asserted frequently were conversations in which Science was expressed frequently: The Pearson correlation coefficient = 0.71. We found 156 and 152 instances respectively of utterances coded as Science being immediately followed by, and proceeding, utterances coded as Belonging. This corresponded to instances of Science being followed by Belonging 32% of the time, and instances of Belonging being followed by Science, 39%. Both of these results cannot be explained by chance at the p < 0.05 confidence level.

We analyzed the nature of the factual or conceptual exchanges that occurred across the different courses. The main science concepts described or discussed included: (1) the overall enterprise of farming insects for human consumption, including tastes or reactions to the taste; (2) the long-standing human practices of consuming insects either intentionally as a valued part of the diet, or unintentionally in terms of the FDA approval of specific amounts of insect content in human foodstuffs; (3) the efficiencies and sustainability effects of increasing insect consumption in the human diet; (4) facts about insects, such as their names, behaviors, and nutritional properties; and (5) features of science as an enterprise, such as taste tests, data points, the National Science Foundation's interest in supporting the Entomophatron study, and so on. By far, the majority of the science idea exchanges (41%) focused on insect facts and nutritional features. The next most common (28%) were exchanges about the enterprise of insect farming; and then (16%) discussions about historical and contemporary human practices of eating insects. There were no discernible patterns about the distribution of the conversational foci and the duration of the events.

In this section we illustrate how conversations developed to address a range of scientific ideas. Clip 5 excerpts a 23-minute interaction, illustrating how the storyworld created opportunities for individuals to ask and respond to questions. At the counter are a couple in their 30s, Greg and Lisa, a man in his 50s, Ted, and a teenager who we call Mike; Mike and Ted each arrived independently and alone to the counter. In this case, a factual question about the agave worm salted popcorn from Greg leads to a conversation in which both Lisa and Mike make observations that lead into a discussion of human cultural practices with respect to eating insects.

The diverse perspectives in Clip 5 collectively construct the science conversation that starts with facts about insect production and moves on to social issues about marketing human consumption of insects which leads to a reflection that many humans already eat insects.

In Clip 6, a man and his son, who we call Roger, walk up to the counter and are immediately asked to join the blind taste test of the cricket chip. After an initial expression of apprehension, the boy carefully participates in the investigation in a particular way—by choosing to smell the chip first—which provides an opportunity for the waiter to expound on scientific approaches to sense-making.

Clip 6 illustrates how the science communicator is able to accept, value, and respect (lines 13 and 15) Roger's approach to investigating the chips. This acknowledgement serves both to welcome the boy and his father to the group process already underway, and also shifts a power dynamic by positioning the boy as having insights into the scientific process of gathering “data points.” This in turn allows the science communicator (in lines 5, 10, and 13–15) to expand his waiter role to that of a person speaking for science. His representation of what constitutes an appropriate way to investigate is both serious (line 15) and humorous (line 5).

In Clip 4B, Manuel and Ernesto discuss ants with the waiter, however in this case it is Manuel who is sharing his understanding of insects and human food consumption.

6.1 Methodological approach and implications

Our data collection methods were designed to document and make sense of transient and fluid science communication events in ways that did not disrupt the experience for either the participants or the science communicators. The commitment to study and learn together, while not interrupting the creative fiction of the storyworld designed by Guerilla Science staff, was crucial to the success of our research-practice partnership.

Our iterative thematic analysis, generating both theorized and grounded coding schemes, was also important to our collaborative work in the research–practice partnership (Xu & Zammit, 2020). Practitioners brought particular values to their analysis—such as the importance of humor and camaraderie in the participant experience—that led us to call out Belonging more forcefully as a top-level theme in the second iteration of the codebook, identifying the four different ways belonging was manifested in the data (see Table 2), rather than a single indicator for Emotion, as it was in the first iteration of coding. This reframing of the themes enriched our joint conceptualization of the relationships among the constructs, and the nuanced ways belonging was instantiated in the data.

Additionally our use of both qualitative discourse analysis and quantitative frequency coding was done to enhance the trustworthiness of our observations and claims. While such a blend of methods may not always be called for, in our research-practice partnership it was important for double-checking our understanding of the patterns that were emerging, especially due to our diverse professional backgrounds and training. That the patterns or correlations of themes were confirmed in the frequencies increased our confidence in our analysis. In the future, this approach would benefit from working with a larger data set to allow a finer grain analysis, including correlations of sub-codes, that might support greater insights and stronger claims.

6.2 Theoretical approach and implications

To answer our research questions about how short-term immersive storyworlds operate to attract and engage non-science seeking public audiences in science conversations, including how they afford opportunities for participants to see science in relationship to themselves, to “belong” in and with science, we drew on Bakhtin's conceptualization of carnival as a temporal, communal experience, saturated aesthetically and emotionally, where performative role-playing allows opportunities for new stances and understanding; in this case with respect to science. A Bakhtinian lens on the power and potential of immersive storyworlds is, we argue, particularly well-suited for the improvisational, analog, and typically unique (one-time) experiences that science communication efforts often design for engaging non-science seeking public audiences. This is in large part because “living the carnival” requires people to draw on their own lived experiences, positioning these personal resources as the very means for participation.

The theoretical conjectures guiding this study, which emerged from the theory of action held by the program leaders participating in the RPP, suggested, first, that the aesthetic and emotionally saturated theatricality of the storyworld would invite and sustain (scaffold) participation of people who may not otherwise seek out science experiences. In other words, there is a commonsense (but untested) assumption that if the country fair had a lecture on insect agriculture it might have attracted a very different crowd. In this case people were attracted to the science experience by the spectacle and engaged by the grotesque humor. Second, that storyworld would operate to create a sense of belonging in a science-rich discourse space, through its proffer of many different roles or entry points, all of which depend on participants activating their own prior lived experience to generate dialogue, to react, and to contribute. Third, that the sense of belonging in that science-rich discourse space would sustain participation and lead to increased and deepening science conversations and practices. Below we review elements of the storyworld that both our quantitative and qualitative analyses suggest may be at play.

6.3 Summary

Our research set out to address the questions: How do immersive storyworlds operate to attract and engage non-science seeking public audiences in science conversations? In what ways do immersive storyworlds afford opportunities for participants to see science in relationship to themselves? By leveraging Bakhtinian conceptualizations of carnival, our analysis suggests that the Guerilla Science immersive storyworld engaged public audiences, who were not actively seeking science experiences, in scientific conversations by leveraging the aesthetically and emotionally saturated nature of the storyworld, enhancing and supporting participants' sense of belonging in that science-rich discourse space, and positioning participants' prior knowledge, interests, and embodied experiences as the means for successful participation in scientific dialogue. The improvisational nature of the experience, built on a set of roles and routines, framed by the conceit of the diner and supported by props, allowed participants to see science in relationship to themselves through their own agentive acts of roleplay, and connecting the everyday grotesque nature of the experience with the larger global and historical challenges of food sustainability. We posit that these findings suggest that the immersive storyworld may be a productive strategy for engaging participants who are not necessarily seeking out science opportunities to engage in science activities and conversation.

We also suggest that adopting expansive definitions of learning (e.g., Barron, 2006; Nasir et al., 2014) requires researchers to broaden conceptualizations of belonging. Belonging in science, as described in the data we provide here, can be seen in active engagement in the moment, where one's self is brought into dialogue with science and science is brought into dialogue with one's self. It is not so much about belonging “to science” as about having legitimacy, a rightful presence, or agency “in science”—of being accepted, valued, and respected in a social context constituted by social interactions organized around scientific phenomena. Everyday science, science in everyday life, is constituted of just such moments. As per Holland et al. (1998), these improvisational moments can over time create a set of heuristics that are constitutive of identity, or processes of being and becoming. To understand how belonging in science develops across the lifespan we need to better account for how these unplanned and ephemeral moments, over time, can contribute to a science-engaged public, and how science communicators can develop more inclusive, compelling, and agentive opportunities for such engagement.

6.4 Limitations and future research

This study has several limitations. Because we chose to forego gathering additional information on participants so as not to disrupt their country fair and storyworld experiences, we have no data on participants' background, including their prior experiences or interests regarding science. However, we do know that the country fair is a general audience, family/friend-oriented entertainment venue, and not a venue branded as scientific. Therefore it is perhaps less likely that participants were favorably disposed to science than they might have been at a science fair or festival (Kennedy et al., 2017). Further, as noted above, the participants at the event we documented appeared to be overwhelmingly white, thus how these approaches play out in communities of color remains the subject for a future study.

With video data as our only data source, we could not account for any cultural variations in terms of gesture or emotion. That is, we were able to code for “disgust” when people uttered “Eww” or wrinkled their noses in a common western expression of disgust. But we only coded disgust when we, from our joint perspectives of having grown up in the United States or United Kingdom, recognized it, which means that it is likely we missed some emotional expressions. We also could only code for belonging when people verbalized their thoughts on their prior experiences, opinions, or feelings of membership or camaraderie.

We note that our quantitative analysis did not prove causation, only correlations, between codes. That is, we observed that certain combinations of codes—for example, storyworld and emotion—appeared frequently together, but we did not demonstrate that the presence of one caused the other to come about. Furthermore, we did not consider the non-adjacent two-point correlations, which would have shown the occurrence of codes that were not immediately before or after each other in time, and may have revealed some deeper conversational structure such as the assertion of a storyworld at the start of an interaction leads to a discussion of science towards its end. We did however consider the correlation between codes within each given conversation. That is, how strongly the presence of, say, emotion and science correlated within a given conversation. Furthermore, because our analysis did not differentiate between utterances by waitstaff and audience members as they jointly constructed the storyworld, the waitstaff's improvisational efforts to “steer” the conversation might have influenced the results in many ways. Nevertheless, we posit that statistically confirming these correlations adds strength to the central conjectures guiding the study and underpinning the design of the program.

Finally, this article provides accounts that contribute to a theoretical basis for why and how storyworlds may be a productive form of science engagement for non-science affiliated public audiences. However, we did not compare this approach with others to see if greater engagement, richer science dialogue, or more lasting outcomes were produced in one version or another. That remains the task of future research. We hope that future studies will benefit to some degree from our work to theorize and empirically document the potential of immersive storyworlds for engaging non-science seeking publics, and more specifically how Bakhtinian conceptions of science can be used to understand participation and belonging in shorter-term, improvisational, and fleeting science-rich interactions.