Gender stereotypes prevalent in society are considered an important explanation for gender differences and their influence on learners' self-concept, interests, and emotions (Marsh et al., 2005). Mathematics and science subjects are often associated with a masculine stereotype, while language subjects are typically perceived as having feminine connotations (Kessels et al., 2006). Girls are particularly sensitive to stereotypical environments that maintain science as a male domain (Mason et al., 1991). One way to counteract the development of stereotypes can be learning environments that are appropriate for students' needs (Armando et al., 2022). If they do not present science in stimulating, relevant, and equitable ways, only students who fit the stereotype of science will continue to enroll in it (Gardner et al., 1989; Mason & Kahle, 1989). Due to stereotypical perceptions in learning environments, the use of a well-deliberated design of pedagogical agents (PA) as a digital learning guide that supports students in their learning process aims to counteract negative effects of stereotypes (e.g., on science self-concept). The assumption is that a gender-neutral looking PA with certain pop cultural features found in students' worlds that appealed to both genders would reduce the perceived masculinity of physics as a subject and enhance the students' science self-concept. According to Fitzgerald (2018, 2019) integrating elements of popular culture, such as characteristics of superheroes, into the development of pedagogical agents can positively support student learning. As a consequence, the aim of this research project was to investigate at first how students perceive ideal and given PAs as learning guides of a digital learning environment and second to determine the influence of gender-neutral looking PAs, developed based on the results of the first study, on students' science self-concept and interest in a web-based comic (webcomic). The first study examined secondary students' perceptions (N = 128) of an ideal PA for a web-based multimedia environment for science education. To investigate possible influencing factors of learners' popular culture in selecting characteristics of their ideal PA, characteristics of well-known superheroes (human = Batgirl/Batman; human-like = Supergirl/Superman; robot = Cyklob; alien = She-Hulk/Hulk) were included in the form of categories. In addition, to examine possible effects of different agent features, we developed nine comic-based pedagogical agents that differed in gender (female, male, gender-neutral), age (young, old), fiction (real, fiction) and human-likeness (human, human-like, non-human). Its purpose was to investigate the popularity, acceptance and perceived emotions related to the presented educational offer. To identify possible predictors of the selection of PAs, in addition to gender and age, and digital media use, students' emotions (Bradley & Lang, 1994) were assessed. From the results, two pedagogical agents, “Kogni“, a gender-neutral looking superhero, and “MetaRob“, a non-human robot, were selected for embedding in a webcomic. We assumed that a gender-neutral looking PA as a superhero and a non-human agent (robot) would invalidate the masculinely charged subject of physics and empower both male and female students in their science self-concept. To investigate these assumptions, 73 students participated in Study 2 in which they were guided in a webcomic about floating and sinking through the two selected PAs. A significant increase of science self-concept from pre- to posttest was found with a medium effect regardless of the gender of the students. The results have shown that gender-neutral looking PAs can play an important role in enhancing students' self-concept in digital science education. Independent of the usual classroom and regular instruction, the design of digital media offers the possibility of creating an inclusive learning environment that can reduce these stereotypes and can make students feel empowered in their science self-concept.

The first study examined secondary students' perceptions about an ideal pedagogical agent. Overall, 128 students (gender: female = 41.4%; age: M (SD) = 12.02 (0.71)) from eight classes at four schools (85,2 % secondary school, 13,3% community school) participated. In a first step, they selected characteristics of an ideal PA by choosing from four categories: gender (female, male, non-binary, indifferent), age (child, teenager, young adult, adult, old adult), fiction (real, fiction) and human-likeness (human, human-like, non-human). The human-likeness category subdivided as follows: human, human-like superhero, robot, animal, alien, mythical creature, and others. To identify possible predictors of the students' choice of specific characteristics of an ideal PA, in addition to gender, age, and digital media use, students' emotions (valence, arousal, dominance) toward the ideal PA were assessed using the Self-Assessment Manikin scales (SAM; Bradley & Lang, 1994). In a second step, students rated nine comic-based PAs designed with the above characteristics from first rank (most popular) to ninth rank (least popular) in terms of popularity. Next, possible preferences in selecting a PA were explored by having students rate their perceived acceptability as well as their emotions toward the first, second and ninth ranked PA, also using the SAM scales. In the second study, a PA that has positive encounters with both genders was provided. Two new PAs, designed by the outcome from Study 1, were incorporated into a web-based comic (webcomic) with simulations to measure their effect on students' science self-concept and interest in science comics. Based on the results of Study 1 two PA had been designed: Kogni, a gender neutral looking superhero and MetaRob, a robot. To investigate the effect of the two PAs on science self-concept and interest, 73 secondary students (gender: female = 38.4%, divers = 5.5 %; age: M (SD) = 11.45 (0.83)) participated in a pre/post design study. Participants were given a questionnaire of scientific self-concept developed by Authors (2010). The questionnaire consists of nine items (Cronbach's alpha = 0.90; e.g.,"I am gifted in science") with a 4-point scale. A questionnaire by Authors (2022) was adapted to measure students' interest in science comic. The questionnaire consists of four items (Cronbach's alpha = 0.83; e.g., "I like reading comics.") with a 4-point scale. Furthermore, in the posttest, assessments on the PAs design were developed with a total of three items on well-being, likability, and role as a learning guide.

It was hypothesized that learners' current popular culture would have an impact on learners' conception of an ideal PA, which will be reflected in an increased selection of the superhero, robot, and alien categories. Descriptive analysis of the results showed that the male participants chose the alien and the robot mainly in the first rank, followed by the gender-neutral looking superhero, while for the girls the superhero was ranked the most first. A chi-square test showed a significant result for both ranks, first rank χ²(2) = 7.27, p = .013, V = 0.24 (girls 50% human, boys 20% human), ninth rank χ²(2) = 15.24, p < .001, V = 0.35 (girls 36.5% human, boys 68.9% human). In the space available for this proposal, only a portion of the results could be presented. For the second study no significant effect was found for interest in science comics between pretest and posttest (t(72) = 0.15, p = .881). Based on the use of gender-neutral PAs with current popular culture learning characteristics that include both genders, we hypothesize that there should be no differences in science self-concept between students' gender. However, a significant increase (t(72) = 3.46, p = .001) of science self-concept from pre- to posttest was found with a moderate effect size (d = 0.404). To draw conclusions about the relationship between the effect of PAs and self-concept, correlations were calculated with the following results: All items on the PAs correlate positively with self-concept for all learners (Pearson's Correlation, r (0.347- 0.554), p < .001). These findings apply to both PAs and show that designing PAs gender-neutral can have a positive impact on students' science self-concept of both genders and provide an opportunity to create an inclusive learning environment in which students overcome these stereotypes and feel empowered.

Armando, M., Ochs, M., & Régner, I. (2022). The Impact of Pedagogical Agents' Gender on Academic Learning: A Systematic Review. Frontiers in Artificial Intelligence, 5. Authors. (2010). [Title omitted for blind review]. Authors. (2022). [Title omitted for blind review]. Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: the Self-Assessment Manikin and the Semantic Differential. Journal of Behavior Therapy and Experimental Psychiatry, 25(1), 49–59. Dark, M. L. (2005). Using Science Fiction Movies in Introductory Physics. The Physics Teacher, 43(7), 463–465. Fitzgerald, B. W. (2018). Using superheroes such as Hawkeye, Wonder Woman and the Invisible Woman in the physics classroom. Physics Education, 53(3), 035032. Fitzgerald, B. W. (2019). Exploring the electromagnetic spectrum with superheroes. Physics Education, 54(1), 015019. Gardner, A. L., Mason, C. L., & Matyas, M. L. (1989). Equity, excellence, and "just plain good teaching." The American Biology Teacher, 57, 72-78. Kessels, U., Rau, M. & Hannover, B. (2006). What goes well with physics? Measuring and altering the image of science. British Journal of Educational Psychology, 74 (4), 761–780. Marsh, H. W., Trautwein, U., Lüdtke, O., Köller, O., & Baumert, J. (2005). Academic self-concept, interest, grades, and standardized test scores: Reciprocal effects models of causal ordering. Child Development, 76 (2), 397–416. Mason, C. L., & Kahle, J. B. (1989). Student attitudes toward science and science-related careers: A program designed to promote a stimulating gender-free learning environment. Journal for Research in Science Teaching, 26, 25-39. Mason, C. L., Kahle, J. B., & Gardner, A. L. (1991). Draw‐a‐scientist test: Future implications. School science and mathematics, 91(5), 193-198. Um, E., Plass, J. L., Hayward, E. O., & Homer, B. D. (2012). Emotional design in multimedia learning. Journal of educational psychology, 104(2), 485.