Social identity theory

Identity has been defined as one’s “core sense of self”[19, p. 405] and as one’s “internal self-constructed, dynamic organization of drives, abilities, beliefs, and individual history” [20, p. 159]. According to Stets and Burke [21], self-identity or person identity can be defined as “the categorization of the self as a unique entity, distinct from other individuals” (p. 228). Although self-identity often is referred to as a single entity, according to Korte [22], self-identity includes both core concepts of self, which are more enduring, as well as peripheral concepts of self, which are more fluid and dynamic. Patterson and colleagues [23] describe the ongoing nature of self-identity development, which changes in relation to group identities or social identities. Thus, self-identity is dynamic and reflexive as individuals take an ongoing active role in identity formation as they strive to associate with groups they perceive as most desirable [24]. Further, individuals may engage in identity switching, drawing on a particular group identity that may be activated at different times and in different contexts, and thus, made more salient [25].

Social identity theory [24] provides a strong theoretical framework for this research because it acknowledges how individuals’ identification and association with social groups [24, 26] drive and motivate behavior [27]. Social identity theory recognizes that individuals simultaneously identify with multiple social groups, which are selected based on which identities are perceived to be most desirable and most important for maintaining a positive sense of self [24, 28]. Social identity theory explains that “individuals tend to favour those whom they see as members of their own group (‘in-group’) over those they see as belonging to another group (‘out-group’)” [29, p. 318]. Identity is a complex construct because identities are multi-dimensional; individuals simultaneously associate with multiple different social groups [18, 24]. Social identities may include gender, race, ethnicity, sexuality, ability, family roles, political affiliation, occupation, religion, as well as other social group associations. Overall, social identity theory is helpful in understanding how social identities influence people’s motivations for attending to information, including media content, best aligned with their desired social identities.

Individuals actively enact identities that are most salient for a particular context or situation [29, 30]. Individuals are not bound to one group identity, and they may move from one to another more salient or desirable group identity depending on context [30]. For example, if a woman identifies as a ‘mother,’ ‘cyclist,’ and ‘scientist,’ that woman may predominately identify as a ‘scientist’ in the lab among colleagues, a ‘mother’ when surrounded by family and children, and a ‘cyclist’ when performing in a race. Previous studies have considered how underserved science communication audiences’ attitudes toward science communication and emotions correspond to feelings of inclusion in science communication [31], but little research has examined how underserved audiences’ social identities might influence inclusion in science communication. Women’s inclusion in science communication–or whether or not women perceive science communication and science media as something relevant for themselves [31, 32]–may be tied to their social identities. Thus, it is important to consider whether the framing and design used in science communication offers points of access aligned with women’s most central social identities. Subsequently, this study used social identity theory as a framework to better understand how women’s most central social identities and STEM (science, technology, engineering, and mathematics) identity affect their initial interest and science engagement with online science media.

Conceptualizing science engagement

Science engagement has significance in multiple disciplines; thus, the extant literature has been vast and conceptualizations often disparate [13]. In the educational sciences, science engagement has typically been used to predict student interest and achievement in science classes [33–35]. Scholars in this area have defined science engagement as consisting of behavioral, emotional, and cognitive factors that may include or be synonymous with interest, enjoyment, motivation, and achievement in science activities in academic contexts [33, 34, 36, 37]. In the communication discipline, science engagement has been defined as communication about scientific or technical subjects that involves scientists, stakeholders, and citizens as well as science outreach focused on promoting, among other attitudes, the public understanding of science by various communicators (e.g., scientists, museum professionals, media professionals, etc., [38]). Some communication scholars have extended extant conceptualizations, suggesting that science engagement should extend to involvement in science policy issues [39] and decision making [40]. For example, research on public engagement with science communication in Japan focused on engagement as related to science policy [41]. Similarly, research in Europe segmented audiences into clusters that varied by level of participation in science: disengaged, aware, invested, and proactive [42]. Conceptualizations of public science engagement in the U.S. originated from conceptualizations of scientific literacy [43], which also considered public involvement in science policy, when identifying these audiences or publics: nonattentive, attentive, policy leaders, decision makers. The focus on advancing scientific literacy and public understanding of science often is reflected in current definitions of public engagement, with a shift from a “deficit model” to a “dialogue model” of communication between science experts and public nonexperts [38]. In the U.S., science engagement has been defined more recently as “intentional, meaningful interactions that provide opportunities for mutual learning between scientists and members of the public” [44]. More broadly, science engagement has been conceptualized to include public “attitudes toward science or desire to participate in future science activities such as attending a science fair or museum” [5, p. 12]. In prior work conducted by researchers in collaboration with KQED science communicators, science engagement was determined based on assessments of participants’ interest, surprise, attention, and how informative they found a particular YouTube video [6]. For the present study, we conceptualized science engagement as participants’ initial emotional responses to as well as their expression of behavioral intention to take future action (i.e., watch a video) when viewing promotional information (thumbnails and episode titles) for online science and nature videos.

Gender and science identity

Prior research has described science engagement as highly gendered and linked to traditional, cultural views that historically have associated science as masculine, thus creating “a tension between science and femininity” [45 p. 147] and further supporting masculine defaults, a form of cultural bias that associates specific characteristics or behaviors with the male gender role [46]. Gender disparities in science engagement may subconsciously be connected with long-held, gendered perceptions of science and STEM subjects (science, technology, engineering, mathematics) often developed during childhood from gendered stereotypes prevalent in, and conveyed widely through, popular culture and media [47]. Prior research found that when traditional gendered stereotypes of science endorse views of science as more appropriate topics and careers for men [48], women who associate science with masculinity and identify as feminine, may experience incompatibility with their identity as a woman and identity as a scientist [49]. Indeed, consideration of identity has illuminated how cultural production and reproduction of gender and cultural representations of gender and science affect interest and persistence in science [50].

Conceptualizations of a science or STEM identity have evolved over time, emerging from an early conceptualization of identity as being recognized “as a certain ‘kind of person’” [51, p. 99]. Early conceptual frameworks described science or STEM identity as one’s representation of one’s self as a scientist or STEM professional [52; 53]. Science identity has also been described as viewing one’s self, more generally, as “a science person” [18]. Recent conceptualizations of science identity focus on identity “as a process of becoming” [50, p. 11], thus rejecting static notions of science identity and acknowledging the potential for promoting and advancing identity work [50]. These conceptualizations, then, highlight the importance of considering identity as essential for inclusive science communication. However, studies focused on gender and science identity have mostly centered on science students [18, 54, 55] or professional women scientists [49, 52, 56–58] with fewer studies focused on women who are nonexperts in science [59]. There is a significant gap in research on how (nonexpert) women’s identities, including their science identity, influence their engagement with science media content. Understanding the influence of women’s identities on intention to engage with science media content is important for determining which features of science media content are most likely to promote interest for women viewers. Prior research has noted that science communication design should recognize that target audiences differ, and, subsequently, science communicators should use different aims, messages, and channels [2].

Current study

This study extends prior research by focusing on science engagement as related to interest in science communication [60], focusing on characteristics of promotions for online science videos intended to spark interest and drive audiences to that content. Specifically, we operationalized anticipated science engagement by asking participants whether they would select to watch any of several listed episodes based on the promotional information (i.e., the thumbnail and episode title). We had two research questions. First, to examine women’s preferences for Deep Look YouTube video promotions (thumbnails and titles) and obtain baseline data about their preferences, we asked:

1. RQ1: What are women’s preferences for Deep Look video thumbnails and titles? That is, (a) which Deep Look video thumbnails and titles are most and least preferred by women; (b) what are the reasons for these preferences; and (c) how might these preferences vary among women based on their science curiosity and science identity?

Second, to determine how women’s most central social identities influence their science engagement intentions and to determine which identities they cited as related to their preferences for Deep Look YouTube video promotions, we asked:

1. RQ2: How are women’s preferences for Deep Look YouTube video thumbnails and titles associated with their central social identities? That is, (a) which social identities do women report as their most central identities; (b) which Deep Look video titles and thumbnails are most and least compatible with their central social identities; and (c) what reasons did women provide to describe why their Deep Look video thumbnails and titles preferences were either compatible or incompatible with their most central identities?

To investigate these questions, we used an explanatory sequential (quantitative to qualitative) mixed-methods approach that included two studies: a national survey and semi-structured interviews. Explanatory sequential design can be used to “explain initial quantitative results” or to “use quantitative results about participant characteristics to guide purposeful sampling for a qualitative phase” [61 p. 77].

Method

Procedure.

The online survey included open- and close-ended questions related to featured thumbnails and titles for 12 Deep Look YouTube videos (see Fig 1). Video thumbnails and titles were selected in consultation with Deep Look producers based on audience preference metrics and variations in design and content related to (1) color(s) of the image in the thumbnail; (2) habitat of the animal/insect (land, water, or air); and (3) YouTube data reporting the number of female viewers for each Deep Look episode (e.g., the episodes watched most and least by Deep Look’s existing female audiences). The researchers determined final selection of videos.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Deep Look YouTube video thumbnails and titles selected for the survey.

Participants read the following instructions, “Below are images and titles from YouTube videos, please click on the image(s) of videos you would watch. You may select as many as you would like”.

https://doi.org/10.1371/journal.pone.0308558.g001

Note that although we chose the videos based, in part, on the viewership statistics of Deep Look’s existing female audience, we did not expect these preferences to mirror the preferences of a national sample of women. This is because we do not believe that Deep Look’s current female audience is representative of all women or of all potential female viewers. The ultimate goal of research aiming to broaden engagement with this and similar science media (among women) is to have an audience that is representative of all women.

Although the project team, which included Deep Look producers and KQED Science executives, proposed the focus on the role of women’s identities in engagement, the KQED team members (including the Deep Look producers) were not involved in data collection, analysis, or interpretation of research findings. The median survey completion time was 19 minutes.

Measures. For preliminary analyses and statistics related to scale construction and evaluation (for the following measures), please see the S2 Appendix.

Finding 1: Women generally preferred the kitten and butterfly episode thumbnails and titles, stating they preferred videos that are “interesting” and “pleasant”

Our first set of research questions asked about women’s preferences for the 12 thumbnails and titles based on their selection, ranking, and rating of those thumbnails and titles. First, the “Kitten” (61%) and “Butterfly” (61%) episode thumbnails and titles were those most frequently selected by participants. In contrast, “Spider” (8%) and “Hairworm” (7%) were the least frequently selected. Second, the kitten and butterfly episodes also had the highest rankings, and the two lowest ranked episodes were the lice and hairworm videos. See Table 1. Third, participants ratings of their most and least preferred videos along the 10 bipolar adjective scales suggested that women preferred videos that they felt were "interesting," "pleasant," or "colorful." Women did not prefer videos that they felt were "disgusting," "unpleasant," and "ugly."

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Measures of central tendency for each episode’s ranking scores.

Higher ranking videos have ranks closer to 1 and lower ranking videos have ranks closer to 12. In this table, the videos are sorted by their median ranking. Percent of viewers that are women from YouTube data are also shown.

https://doi.org/10.1371/journal.pone.0308558.t001

These results are consistent with the NLP findings. The NLP analysis found that participants were most likely to describe their most preferred thumbnails using words or phrases like attractive/colorful, cute, or interesting/curious. Participants were most likely to describe their most preferred titles as interesting, sparking their curiosity, and being informative. Overall, the top four terms mentioned as reasons for participants’ most preferred thumbnails and titles were: attractive/colorful, interesting/curiosity, cute, and informative. Participants were most likely to describe their least preferred thumbnails using words or phrases like disgusting or by stating that they did not like that animal or wildlife shown. Participants were most likely to describe their least preferred titles using words or phrases like disgusting and uninteresting. Overall, the top four terms participants mentioned as reasons for their least preferred thumbnails and titles were: disgusting, uninteresting, don’t like the animal or wildlife, and fearful (See S2 Appendix).

It is worth noting that the kitten episode, though the most popular one in this study, was one of the least viewed on the Deep Look channel at the start of this project and was not popular among women: YouTube reported that only 26% of the 214,542 total viewers of the episode were women. Similarly, YouTube reported that only 18% of the 371,616 total viewers of the butterfly episode were women. Of the 12 episodes featured in this study, the lice episode was the most popular among women on the Deep Look channel: YouTube reported that 57% of the lice episode’s 4,147,989 total viewers were women. See Table 1. We explore the individual differences in audience preferences that help to explain this finding in finding 3 below.

Finding 2: Few women anticipate engaging with any of the episodes

At the end of the survey, participants were asked whether they would like to view any of the videos featured in the survey and only 39% (n = 754 participants) said yes (61%, or 1,185 participants, said no, and 1 did not answer). This lack of interest is, perhaps, not surprising given the lack of engagement and the gender gap seen in the YouTube metrics and prior research [7]. Natural language processing findings revealed that participants who said “yes” most frequently cited expecting the videos to be interesting/curious (n = 777 mentions), informative (n = 106 mentions), and attractive/colorful (n = 62 mentions). Participants who said “no” most frequently reported that they expected the videos to be uninteresting (n = 58 mentions) and disgusting/gross (n = 37 mentions). Finding 3 discusses potential individual differences in women’s (lack of) anticipated engagement with the videos.

Finding 3: Science curiosity predicts anticipated engagement with the episodes, generally, and science identity predicts anticipated engagement with (and higher ranking of) the less popular videos

Logistic regression analyses were conducted to examine the role of science curiosity and science identity for the “watch any episode” item and for the two most frequently selected and two least frequently selected videos. Then multiple regression analyses were conducted to examine the role of these variables in predicting the preference rankings of the top two and bottom two ranked episodes.

First, we found that both science curiosity and science identity predicted participants saying “yes” when asked whether they would want to watch any video, but science curiosity was the stronger of the two predictors. When holding science identity constant, the odds participants said “yes” increased by 90% for each unit increase in science curiosity (b = 0.64, p < .001, exp(b) = 1.90). In contrast, when holding science curiosity constant, the odds of saying “yes” increased by 17% for each unit increase in science identity score (b = 0.16, p = .002, exp(b) = 1.17).

Next, we predicted selecting the two most frequently and two least frequently selected episodes. Holding science identity constant, for each unit increase in science curiosity, the odds of participants selecting the kitten episode increased by 25% (b = 0.22, exp(b) = 1.25, p < .001), the odds of participants selecting the butterfly episode increased by 68% (b = 0.49, exp(b) = 1.68, p < .001), the odds of selecting the spider episode increased by 78% (b = 0.58, exp(b) = 1.78, p < .001), and the odds of selecting the hairworm episode increased by 32% (b = 0.31, exp(b) = 1.36, p = .006). When holding science curiosity constant, for each unit increase in science identity, the odds of selecting the unpopular hairworm episode increased by 23% (b = 0.21, exp(b) = 1.23, p = .020). The odds of selecting the spider episode increased by 15%, but this effect did not meet the threshold for statistical significance (b = 0.14, exp(b) = 1.15, p = .099), and the odds of selecting the butterfly episode did not change (b = -0.04, exp(b) = 0.96, p = .444). Interestingly, holding science curiosity constant, for each unit increase in science identity score, the odds of selecting the kitten video decreased by 10%, b = -0.10, exp(b) = 0.90, p = .038. See Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

Probability of selecting to watch the kitten, butterfly, spider, and hairworm videos conditional on science curiosity (left) and science identity (right).

https://doi.org/10.1371/journal.pone.0308558.g002

To examine whether participants’ rankings for these videos were conditional on science curiosity and science identity scores, each ranking was multiplied by -1 so that rankings closer to 1 (-1) were greater in value than the lower rankings (e.g., -12).

We found that greater science curiosity is associated with higher rankings of the butterfly episode (b = 0.26, p < .001, r = 0.06), lower rankings of the lice (b = -0.45, p < .001, r = -0.09) and hairworms episodes (b = -0.49, p < .001, r = -0.11), and there was no significant relationship with the kitten episode (b = 0.10, p = .237, r = -0.03). Greater science identity score, on the other hand, is associated with lower rankings for the kitten (b = -0.49, p < .001, r = -0.15) and butterfly (b = -0.19, p < .001, r = -0.05) episodes and higher rankings of the lice (b = 0.35, p < .001, r = 0.06) and hairworm (b = 0.33, p < .001, r = 0.05) episodes. See Fig 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

Rankings (preference scores) for the kitten, butterfly, spider, and hairworm videos conditional on science curiosity (left) and science identity (right). Better rankings (greater preferences) have ranking scores closer to 1 and worse rankings (lesser preferences) have scores closer to 12. Note that we multiplied the rankings by -1 so that scores closer to 1 (-1) are greater in value than scores with lower rankings (e.g., closer to 12, -12). To avoid confusion, the negative symbols were removed from the y-axis of the figures.

https://doi.org/10.1371/journal.pone.0308558.g003

Finding 4. Women’s most central identities tended to be those describing their relationships to others and ‘colorful’, ‘attractive’, and ‘interesting’ were the descriptors of the episodes seen as most compatible with their identities

Our second set of research questions focused on which identities women reported as their most central identities. Identities were assessed in order to compare women’s preferences for thumbnails and titles with their most central identities identified by NLP. Women’s most central identities tended to be those describing their relationships to others: mother/grandmother (n = 797 mentions), friend (n = 679 mentions), and spouse/partner (n = 375 mentions). This is consistent with other research [67]. See Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Term frequency for most central identities.

This measures the number of mentions as one of the top four identities and as one of the top two identities.

https://doi.org/10.1371/journal.pone.0308558.t002

Findings for videos that were most compatible with participants’ top four reported identities are presented in Table 3. Participants were most likely to describe videos to be good matches with their self-reported central identities when they perceived the thumbnail or titles to be colorful (1,137), attractive (682) and/or informative (642). Findings for videos that were least compatible with their top four reported identities are presented in Table 4. Participants were most likely to describe videos to be bad matches with their self-reported central identities for videos with thumbnails and titles that were perceived to be disgusting (1,136) and/or fearful (273).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Term frequency for reasons for good match between most preferred videos (thumbnails and titles) and identities.

https://doi.org/10.1371/journal.pone.0308558.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Term frequency for reasons for mismatch between least preferred videos (thumbnails and titles) and identities.

https://doi.org/10.1371/journal.pone.0308558.t004

Method

Finding 5: Most interview participants considered themselves to have a science identity or to be someone interested in science

The majority of interview participants, 16 out of 24, considered themselves to have a science identity or to be someone who is interested in science.

I am really interested about astronomy, psychology, and maybe biology. It makes me wonder. It makes me curious and, I like to read interesting stuff about those topics. (Science Curious; Central Identities: Woman, Mother, Wife, Daughter)

Only five interview participants did not consider themselves to have any science identity. These interview participants were from across the four groups, but only two were from the “Science Curious” group. Three interview participants described themselves as being interested in science but did not think of themselves specifically as “a science person” (Science Indifferent; Central Identities: Black, Woman, African-American, Writer) or as “science-oriented academically” (Science Indifferent; Central Identities: Afro-American, Woman, Wife, Singer). These findings are interesting to consider further because of the overall low Science Identity scores reported for survey participants.

Interview participants who reported a science identity cited various reasons for their interest in science. Several interview participants pointed to relational watching of science television programming with spouses.

We watch documentaries in our house. My husband and I (Science Open; Central Identities: Writer, Friend, Christian, Creator)

We’ve always loved documentaries, wildlife documentaries, jungle animals, big cats, dolphins, you know, under-the-sea kind of documentaries. We’ll watch a lot of those. (Science Indifferent; Central Identities: Christian, Wife, Mother, Grandmother)

Interview participants also described an interest in collaborative science activities and projects with children. One interview participant described the importance of science on YouTube for teaching young children about science.

I have younger kids, so I’m kind of trying to get them into doing different things. Just the other day …, what was it, the Alka Seltzer in the soda bottles, and they had fun with that. They were watching YouTube videos and wanted to see how to do it themselves. Just smaller things that I’m kind of trying to get them more interested in. So different activities and things that they can do. (Science Uninterested; Central Identities: Human, Woman, Mother, Grandmother)

Another interview participant, although “Science Uninterested,” commented in great detail about how YouTube videos served as a teaching tool and resource for information about nature for teaching kids about science related to their personal experiences.

We kind of got into [science] because we had a bee hive, … somewhere outside of the house. And the kids are kind of scared to go outside. So, we started looking up videos on YouTube about the bees and how they make their hives, and the things they do to survive. So, they [the kids] are kind of interested in that, but it was more interesting to me, I think, just because I didn’t really know much of it. Kind of learning and explaining it to them as I’m going and listening and watching the videos too. So, things like that. There’s other [insects] like ants and things like that they’ve been finding outside. “Oh, I’m going to kill them,” or “I’m going to squish them.” I said, “No, no, no. I’m going to show you something. Here, this is how they make their homes,” and things like that. So, it’s more or less educational review as well because even though I’m sitting there with them and explaining it to them, I’m kind of giving them more of the insight as I’m watching it because I know of [some things about them]. (Science Uninterested; Central Identities: Human, Woman, Mother, Grandmother)

One interview participant stressed the importance of expressing a science identity because of wanting to serve as a role model for a daughter who might face racial discrimination in science.

One of the things I had to teach my children was, well, they didn’t grow up in the hood … I did … but if you have an interest in those type of subjects [i.e., science], you’re considered a nerd or you will be bullied. So, I had to kind of walk them through. My youngest daughter’s kind oflike me, she also likes science, and I had to walk her through and teach them to have pride in what they want to do and[to] just ignore people. (Science Curious; Central Identities: Data Driven, Intelligence Seeker, Happy, content Woman, Proud Black Woman)

Another interview participant noted an interest in science because of a personal need for health information to help a neighbor.

…. It’s just like something that’s happened the last few years… I’ve just, like, had this really deep thing about science. I look up a lot of medical stuff that is more science-related and I just enjoy it. So, that’s been [a] more recent thing for me. I had a brain injury…11 years ago… and so I’ve had ongoing issues [with sleep]… . The whys behind… why we don’t sleep? And some of it’s because I know people who have been sick, like a neighbor of ours, and some of it’s just, I don’t know, I read about something and then I get interested and I have to find out more about it. Yeah. So just very curious. I am a very curious person. (Science Open; Central Identities: Writer, Friend, Christian, Creator)

Other participants described interest in science because of an interest in the environment and as an extension of their personal observations of the natural world.

Well, a lot of it for me is environmental. I’m very big on clean energy and I actually had a friend [who] helped with a company trying to build a solar dish, a solar panel, and I actually got an MBA in environmental sciences. I just think the subject is awesome. You know, clean energy is important to me, especially living in a state like Nevada. So that’s been a strong pull for me. And actually, I’ve been doing a lot of research for a year or so about the virus, with the COVID virus, where it may come from, cures, how it might have been enhanced. That’s been so much of an interest for me. I just was actually thinking about maybe actually getting a Ph.D. in neuroscience because I find myself fascinated every day, I find myself interested in what’s going on trying to sift through the misinformation. Very interesting. I also like astrology [sic]. One of my daughters got me more into astrology. I’m fascinated by … Mars, I watched Rover go through the rocks on Mars, and NASA, NASA cam, and it’s just amazing. Well, yeah, I’m more scientifically inclined as I’ve gotten older. (Science Curious; Central Identities: Data Driven, Intelligence Seeker, Happy, Content Woman, Proud Black Woman)

Even several “Science Indifferent” or “Science Uninterested” interview participants described personal interest in science, which was a clear expression of science identity:

Well, I’d like to see how things how things evolve and what the reasoning is behind you know animals or plants and the climate, the earth, I’m interested in those type of things. (Science Indifferent; Central Identities: Mother, Spouse, Grandmother, Sister)

I have always been interested in learning. I have a higher IQ. Had advanced classes all my life in school, and I just, I love learning new things and how things work, and how they’re done, and how they’re made. New technology…. How it affects the environment. How any science affects the environment. (Science Uninterested; Central Identities: Lover, Fighter, Animal Activist, Environmentalist)

Various topics in science actually catch my attention, like the mosquito thing, and then now we have the cicadas here, so I’m very interested. Like, [why do] we have cicadas in the ground for so long? And they just come out. And I read [about] them and like the vaccine. (Science Indifferent; Central Identities: Female, Asian, Mother, Engineer)

A couple interview participants described professional connections to science, such as working as a research coordinator for clinical trials (Science Open; Central Identities: Woman, Singer, Philanthropist, Project Manager) and studying occupational therapy in graduate school (Science Open; Central Identities: Dog [Owner], Friend, Daughter, Student).

Finding 6: Most participants see science engagement and having a science identity as compatible with their most central identities

The majority of interview participants (18 out of 24) considered their most central identities to be compatible with a science identity. Several interview participants noted how relational identities motivated their interest in science or nature, in general, or were linked to interest in specific Deep Look thumbnails and titles. Two interview participants described how interest in science was connected to their maternal identity as a mother or grandmother. For one interview participant, science engagement is seen as important for knowing how to promote good health for children.

In terms of mother, I think every mother is interested and concerned for scientific work that’s being done related to various children’s illnesses. And I know that there’s been a lot of work done by a number of the medical organizations regarding various diseases and the science related to how … various diseases affect children, and they’re working on solutions and there were trials. And so, I definitely think mother relates in that respect. (Science Curious; Central Identities: Bird, Actress, Mother, Humorist)

For another, science engagement is seen as important for teaching grandchildren about science.

It could be that [I’m] interested in science as a grandmother. I have grandchildren [who] ask questions. So, we have a lake place. So, they ask a lot of questions about the different things they find up at the lake. They ask about the turtles, and they ask about the fish, and so… the Grandmother part [of my identity] really relates to the science for that. (Science Uninterested; Central Identities: Female, Mother, Grandmother, Friend)

Even interview participants who fell into the “Science Uninterested” quartile, like the above example, described connections with science content and their identities as mothers:

It touches on that motherly instinct for me. (Science Uninterested; Central Identities: Female, Mother, Leader, Friend)

One participant, who was an engineer, directly discussed compatibility between her identities as a mother and as an engineer, although she did not perceive these identities to be compatible for most women:

I can be [a mother and a scientist] yes, correct, but not the general population. (Science Indifferent; Central Identities: Female, Asian, Mother, Engineer)

Finding 7. Some participants believe a science identity ought to be held only by those who are working scientists

While most interview participants viewed central identities as compatible with a science identity, seven of 24 interview participants disagreed. Some distinguished between having a scientist identity and being a scientist, noting how a scientist identity and science engagement are relegated for those with an academic background in science.

In general, no, when I think when you asked [if] we [were] interested in science, I think of more of that as more of an academic kind of track and I don’t think my [identities] cover any of that. (Science Indifferent; Central Identities: Christian, Wife, Mother, Grandmother)

I kind of feel like they’re separate. I guess independent of the rest of it. I guess it’s just or should I say a piece of it. (Science Open; Central Identities: Woman, Singer, Philanthropist, Project Manager)

Interestingly, although one participant fell into the “Science Open,” quartile of science curiosity scores, this participant described an overall lack of interest in science and incompatibility with a science identity.

Anything to do with science as far as me, it’s just, I think a lack of interest in general. (Science Open; Central Identities: Woman, Spouse, Mother, Friend)

Some interview participants described, more generally, how social norms promote incompatibility between a science identity and other social identities, citing both sexism and racism.

Nigerian women, in particular, haven’t really been encouraged to being a scientist beyond maybe being a doctor. (Science Uninterested; Central Identities: Friend, Woman, Nigerian, Sister)

Women [and] womanhood have been neglected by science. I feel like the reason why science doesn’t seem appealing is because there are so many things that a woman goes through or doesn’t go through. … They’re just not catered to enough. (Science Uninterested; Central Identities: Friend, Woman, Nigerian, Sister)

Limitations and directions for duture research

Despite the contributions of these research findings, there are some limitations, particularly the extent to which these findings can be generalized. First, the current study only focused on videos produced by the YouTube series, Deep Look. This study could be replicated using other educational, digital science video series (e.g., Physics Girl, It’s Okay to Be Smart) to determine whether, and if so, to what extent the findings are generalizable. Further research with larger samples of science media products and a more in-depth analysis of product characteristics would be needed to provide more concrete advice for science communicators. Second, our survey sample did not meet national representativeness because the quota for Hispanic/Latina participants was unable to be met during the project timeframe and the average age of the sample (M = 57.81, SD = 16.9) was higher than that of the general US population (M = 38.9; United States Census Bureau data). That said, research could also consider oversampling from groups who are typically underrepresented in science engagement (e.g., women of color) to be able to examine intersectional differences with more statistical confidence. Additionally, while we found common themes among the 24 interview participants, these women represented around 1% of the total sample studied. Given additional resources and time, conducting additional interviews may have allowed for a more nuanced understanding of women’s preferences for science YouTube videos. Another limitation is the science identity scale has been used primarily to assess STEM students’ identity and is not a perfect fit for assessing science identity among non-experts. However, the scale has been validated, widely used, and helps assess non-expert participants’ perceptions of science identity. A new scale to capture and examine non-expert science identity could be a goal of future research.

Despite these limitations, the current study provided important foundational knowledge to better understand the role of gender, specifically, and identity, more broadly, in science engagement. Future research should consider other science topics on other YouTube videos. In addition, future research should directly compare the preferences of men and women related to their identities. Future research could also utilize eye-tracking technology to better understand attraction to science video thumbnail images and titles and influence on women’s intentions to engage. Finally, researchers should consider cross-media engagement with science videos on other social media platforms (e.g., Facebook, Instagram, TikTok) to increase engagement with younger audiences. Research on Instagram that focused on women posters who identify with STEM topics, found that often, women are not only challenging gender stereotypes, but also serving as role models for young viewers who are interested in these topics [58, 74].