Project BioEYES: Accessible Student-Driven Science for K–12 Students and Teachers

BioEYES, a nonprofit outreach program using zebrafish to excite and educate K–12 students about science and how to think and act like scientists, has been integrated into hundreds of under-resourced schools since 2002. During the week-long experiments, students raise zebrafish embryos to learn principles of development and genetics. We have analyzed 19,463 participating students’ pre- and post-tests within the program to examine their learning growth and attitude changes towards science. We found that at all grade levels, BioEYES effectively increased students’ content knowledge and produced favorable shifts in students’ attitudes about science. These outcomes were especially pronounced in younger students. Having served over 100,000 students, we find that our method for providing student-centered experiences and developing long-term partnerships with teachers is essential for the growth and sustainability of outreach and school collaborations.


Introduction
Over the last 30 years, many national reports have called for reform in kindergarten through 12th grade science, technology, engineering, and math (STEM) education [1][2][3], advocating for improvements in curricula and teacher preparedness.One mechanism for change is for schools to establish partnerships with STEM stakeholders, including universities.Although this type of collaboration is growing, research on sustaining and institutionalizing partnerships between formal (i.e., classroom-based experiences) and informal (i.e., science outreach programs) instruction is far from comprehensive [4][5][6][7].In 2002, we created BioEYES (www.bioeyes.org),a community-based, life science outreach program that uses live zebrafish to teach basic science principles, animal development, and genetics.This program provides scientific equipment for under-resourced schools, an experiment-driven curriculum, and a shared teaching experience between classroom teachers and university outreach educators.BioEYES offers its programs free to districts where substantial numbers of children live in poverty, and state assessments of science scores are low (e.g., School District of Philadelphia and Baltimore City Public Schools) [8][9][10][11].The majority of participating students (87%) were taught by BioEYES outreach educators or classroom "model teachers" that underwent extensive BioEYES training.Our analysis of over 19,000 students during the 2010-2015 school years shows that Project BioEYES alters students' attitudes towards science and their knowledge of concepts following the program.

Overview of Approach
During the week-long BioEYES experiment, students work as scientists using a student-centered approach, a key strategy that has been shown to increase learning (Fig 1 and videos at http://bioeyes.org/newsroom/videos.php) [12][13][14].Student learning must be an active process that connects to problems and situations relevant to students' lives [15].In this experiment, inner-city students investigate the inheritance of skin color.In five days, they collect zebrafish embryos and watch them transform from a single cell to a free-swimming larva with a visibly beating heart and a distinct pigmentation pattern.Elementary students learn about habitats, human and fish anatomy, DNA, and cells (S1 Doc and S2 Doc).Middle and high school students identify the phenotypes of their offspring (middle school) (S3 Doc and S4 Doc) or the genotypes of the parents (high school) (S5 Doc and S6 Doc).High-quality microscopes, adult zebrafish, attractive student journals, and classroom supplies are provided by BioEYES, tools that support learning through play, creativity, engagement, and curiosity [16].Such hands-on experiences elicit positive reactions that students are more apt to remember [17].
Critical questions that all STEM programs such as BioEYES must address are the degree to which they improve STEM knowledge and academic performance and, ultimately, whether they increase the representation of underrepresented groups in STEM professions.The challenges of such analyses include the need for a longitudinal study of academic performance over many years and the control of influential variables such as teacher quality and turnover, teacher professional development, and available classroom resources [13,18].We began such a longitudinal study, together with the Baltimore City Public Schools, in 2014 to assess the impact of BioEYES on a host of metrics that include students' subsequent academic performance in STEM disciplines.

BioEYES Increases Student Science Knowledge and Improves Their Science Attitudes
While the longitudinal study is ongoing, here we report on student learning and STEM attitudes collected from 19,463 students during the course of BioEYES programing (a one-week exposure per year, 2010-2015 school years).Students were asked to answer a set of knowledgebased questions as well as an assessment of their attitudes toward science and science careers before and after the BioEYES experience.At the elementary level, seven of the eight knowledge questions had significant positive gains (Table 1; S1 Table ).At the middle school level, eight of the nine knowledge questions had significant positive gains (Table 1; S2 Table ).Many of the middle school questions that showed the greatest gains involved concepts in genetics (Punnett squares with a percent change of 87%; phenotype with a percent change of 72%; and homozygous with a percent change of 54%).These genetic concepts are taught and tested in a majority of United States (US) middle school classrooms without a concrete demonstration of how they are practiced by researchers.BioEYES provides this important connection between information and application.
At the high school level, the questions that showed the greatest gains focused on characteristics of model organisms with a percent change of 64% and stem cells with a percent change of 56%, topics that are explained more in-depth during BioEYES than in a traditional science lesson (Table 1; S3 Table ).When high school students were asked genetics questions on the pretest, results showed a higher rate of correct responses in comparison to the middle school students.This led to a smaller percent change in pre-to-post scores (i.e., recessive genes with a percent change of 14%, Gregor Mendel with a percent change of 16%, and Punnett squares with a percent change of 14%).This reflects the incorporation of genetics in the 7th grade curriculum and then again in high school biology that may be further reinforced in both grades by the BioEYES program.In sum, significant positive gains in learning were observed at all grade   1), indicating that children learn these concepts when they are delivered through an authentic, student-driven experience like BioEYES.
Students' ideas about who scientists are, science's importance and popularity, and potential careers as scientists were assessed.The largest effect on attitudes occurred at the elementary school level, where significant positive changes were found in six out of eleven statements (Table 2; S4 Table ).Accordingly, among all grade levels, BioEYES exposure produced the strongest attitude shift regarding whether students "know what it's like to be a scientist" (S4-S6 Tables; A4).From its inception, BioEYES was cocreated by a practicing life scientist (SAF) with the explicit goal of providing a positive, enjoyable experience that reflects what many scientists experience on a good day at the bench.The second largest attitude change observed in elementary and high school grades (third largest in middle school grades) involved students' increased agreement with the statement that "science is becoming more popular than it used to be" (S4-S6 Tables; A9).We speculate that this reflects that their personal experience with BioEYES was fun and engaging, and that they observed the same in their classmates.It is likely no surprise that popularity is important to the social lives of students, especially in middle and high school.Science educators can thus take heart; if the classroom experience is fun and engaging, kids are likely to think science is popular, which we hypothesize will attract more students to STEM fields.Interestingly, for all grade levels BioEYES increased their ability to imagine themselves as scientists (S4-S6 Tables; A11).At the elementary level, students started with a higher degree of agreement with this question than other grade levels, leading to a smaller average change overall.These data are consistent with prior research on science attitudes describing similar declines in students' interest in science as they progress through traditional education systems [18][19][20].We did observe a significant negative attitude shift in middle school students following BioEYES for the statement "I would be interested in learning about different types of careers in science" (S5 Table; A7).We hypothesize that following BioEYES, attitude A7 was interpreted as asking them to do extra/more difficult work.Future studies will address the degree to which these attitude changes persist through their K-12 career, beyond the few days immediately following the BioEYES experience.

Impact, Recommendations, and Conclusions
Taken together, our data from thousands of children indicate that a nonprofit outreach program like BioEYES can change their attitudes about science and scientists (Table 2) in traditional classrooms and in a way likely to promote future STEM learning.We hypothesize that providing a week-long experiment (as opposed to a single science lesson) with live animals is a major factor in explaining these positive results.While we were not surprised to find that elementary students enjoyed using the microscopes we provided (Table 3), we were surprised to discover that students also liked caring for the fish, a task that involves removing fish waste from the Petri dishes.Watching the fish develop and seeing the zebrafish heartbeat were also favorites of students.These findings, too, support our claim that our use of live animals in the classroom is engaging for students.
In order to deliver this program to so many students, BioEYES staff trained approximately 1,300 teachers.Because STEM often contains complex content, we understand that teachers may feel underprepared in these areas [21].Increasing the competence of partnering teachers to deliver STEM content independently is a high priority for the BioEYES program.Our data support the emerging view that exposing teachers to guided, inquiry-based training during a three-hour professional development session and then in the classroom environment receiving support from a coteacher/expert is a "best practice" for heightening teacher ownership of the delivery [22] and for engaging students in science [23].The BioEYES professional development model provides teachers with, on average, 9-15 hours of direct collaboration per year with the program, depending on the number of classes they teach.By year three, the program is taken over by the classroom teacher without the assistance of BioEYES educators, the teachers become BioEYES "model teachers." BioEYES has had 80 model teachers per year who deliver the curricula, with program materials loaned by BioEYES staff.These teachers also help to train teachers that are new to BioEYES programming by assisting at professional development workshops held at our BioEYES Centers and by serving as the primary liaison for BioEYES at their school.Notably, the model teachers reach over a third of the number of students targeted annually.For example, in 2014-2015, 39% of the total students across all grades (4,318) were taught by BioEYES model teachers.Given limited outreach staffing and funding, the use of model teachers is a way the program can expand to serve more students without a marked increase in program costs.

Developing Outreach Efforts
We attribute the 15-year success of BioEYES to its inclusion of three major domains.Initially, BioEYES emerged from a collaboration between a biologist with an active research program (Farber; Domain 1) and a K-12 science educator (Shuda; Domain 2) who provided age-specific curriculum goals that complemented Farber's expertise.This collaboration was coupled with a substantive commitment by the scientist's host institution to support a true university/community partnership.Further, BioEYES is designed with activities (Domain 3) that are hands on, experimental, use "cool" science tools (e.g., high quality microscopes), and preferably employ live animals.The "secret sauce" of BioEYES is the overlap between these components coupled  In conclusion, data from 2010-2015 revealed that Project BioEYES is effective at increasing students' knowledge, and that changes in attitudes were more pronounced with younger students (Table 2).These data support the view that early exposure to STEM programing is most effective.Now having reached 100,000 students and 1,300 teachers in the US and Australia, these findings are particularly relevant considering that BioEYES has primarily served K-12 students from backgrounds typically underrepresented in STEM professions-low-income and minority students in public schools in large cities.Using BioEYES to reach the masses of underserved students helps spark their interest in research, which then allows them to explore college and career options in science and also in allied fields.For example, we learned of one senior in high school who was presenting research performed at Johns Hopkins University on Sudden Infant Death Syndrome.When asked in a public forum where she got her start in science, this under-represented minority woman responded that it was when BioEYES came into her 7th grade class.Opening the doors and providing a memorable K-12 science experience is our aim for the next 100,000 students.
to qualify as quality improvement by the University of Pennsylvania's Institutional Review Board.
All procedures were approved by the Carnegie Institution and University of Pennsylvania Animal Care and Use Committee (Protocol #s 142 and 804450, respectively).

Fig 1 .
Fig 1. Activities and aims of Project BioEYES.(A) Building investigative skills by using a model organism to investigate, observe, and generate hypotheses.(B and C) Working in teams, students engage in hands-on activities using microscopes and simple lab tools to care for their zebrafish larvae.(D) Students are scientists.They make qualitative observations, collect quantitative data, and engage in collaborative discourse using appropriate scientific vocabulary.Throughout this process, students analyze and interpret data, draw conclusions, and communicate results.doi:10.1371/journal.pbio.2000520.g001

Fig 2 .
Fig 2.Proven principles for creating a successful science outreach program.BioEYES began working directly with teachers, piloting the experiment in classrooms and aligning the experience with content required by local and/or national standards.We have found that building a relationship from the classroom up (i.e., teachers with the approval of their administrators, then multiple grades within a school, and finally a district-wide roll out) enables the program to prove its feasibility, curriculum alignment, and teacher enthusiasm-all required for sustainability.Moreover, K-12 outreach

Table 1 . Overall content knowledge change following BioEYES exposure. Students Assessed n % Correct Pre % Correct Post Difference Percent Change p-value (Adjusted) Significance
Data were processed by summing pre-and post-test values for each student across all test questions.Significance is estimated using a two-sided paired ttest and family-wise error rate-corrected using the Bonferroni correction.doi:10.1371/journal.pbio.2000520.t001levels (Table

Table 2 . Overall attitudes change following BioEYES exposure. Students Assessed n Average Pre Average Post Average Change Net Likert Point Change p-value (adjusted) Significance
Data were processed by summing pre-and post-test values for each student across all test questions, except for question A3.Net Likert Point Change is the difference between the sum of all Likert scale values on pre-and post-tests.Significance is estimated using a two-sided Wilcoxon signed-rank test and family-wise error rate-corrected using the Bonferroni correction.