The BER Lab conducts research on human thinking and reasoning about foundational biological ideas and uses this work to design more effective learning environments.

Artificial Intelligence in Learning and Assessment. The BER Lab was a pioneer in the use of open-source Natural Language Processing (NLP) and Machine Learning technologies in biology assessment. The goal of this work was to advance the richness and rigor of learning measures (e.g., Nehm & Haertig 2011; Nehm, Ha, Mayfied, 2011; Moherrari et al. 2014; Ha & Nehm 2016). The ongoing expansion of the Next Generation Science Standards (National Research Council, 2014) and implementation of Vision and Change in Undergraduate Biology (AAAS, 2011) demanded that new forms of assessment be developed to measure twenty-first-century science competencies. The recognition that undergraduate science courses needed to move away from assessing massive numbers of easily assessed (but quickly forgotten) facts required that new measures of learning be developed for large-enrollment classes. Scientific practices such as explanation and argumentation, the utilization of cross-cutting concepts such as cause/effect and structure/function, and the construction of scientific models that enhance understanding required new types of assessment tools. These practices foster deeper engagement with the processes of science through debate, discussion, analysis, and critique and are therefore important learning objectives (National Research Council 2012). The challenge is that these practices require oral or written language, and because of this, time-consuming analyses and scoring of language by teachers and researchers is needed (cf. Rector et al. 2012). A decade ago the lab began exploring how the use of open-source NLP-based machine learning tools could be used to measure learning in a cost-effective way, thereby facilitating the adoption of more ambitious learning objectives in undergraduate science. The lab has been supported by multiple National Science Foundation research grants to advance understanding of science assessment using NLP-based ML tools and the learning of core biological ideas. Recently the Lab has been studying predictive learning analytics and the role it can play in engineering effective learning environments in science classrooms and the role of bias in AI systems.

Measuring faculty change in evidence-based practices in undergraduate education. Vision and Change in Undergraduate Biology (AAAS, 2011) included ambitious policy guidelines recommending evidence-based reform in the teaching, learning, and assessment of undergraduate biology. One focus of this report was the call for biology faculty to change their pedagogical practices away from didactic lecturing and toward student-centered learning. There is a strong evidence base supporting these recommendations. Student-centered pedagogical practices have been found to improve course outcomes, sense of belonging, and retention in STEM (Freeman et. al 2014). Student-centered learning has also been found to disproportionately benefit students from backgrounds that are underrepresented in science (Theobald and Freeman 2020). One instrument, known as the COPUS (The Classroom Observation Protocol for Undergraduate STEM) is a widely-used instrument designed to collect and categorize individual class-level observational data about student and instructor behaviors in undergraduate learning environments (Smith et al. 2013). Although the COPUS instrument has been widely adopted as a measure of classroom behaviors and learning environments, important methodological and conceptual questions remain about its appropriateness for measuring the magnitudes of educational reform. The lab is seeking to advance understanding of the role that the COPUS may play in the measurement of undergraduate STEM education reform by using probabilistic approaches to sampling in order to better understand how best to measure change.Measuring–and mitigating–conflict with evolution. Although personal, familial, and community conflict with evolution have been documented in the literature, these scales require conceptualization as a construct and operationalization as a measure. The Scales of Conflict with Evolution Measure (SECM) instrument was developed in response to these needs. Using a construct validity framework, the content, internal structure, convergent, and substantive validity of the SECM were evaluated using Rasch analysis, Structural Equation Modeling (SEM), and follow up questioning. The conceptual utility of the instrument was explored by examining whether it added explanatory insights into evolution acceptance above and beyond religiosity, evolution knowledge, and background variables. The SECM is an easy-to-administer instrument to measure conflict with evolution and is supported by several forms of validity evidence. The SECM has potential for facilitating measurement of evolutionary conflict in educational settings, thereby raising instructor awareness of conflict levels in students, promoting rigorous evaluations of educational interventions designed to reduce conflict, and fostering conceptual advances in the field of evolution education. Future work is needed to gather additional forms of validity evidence and to test current validity claims in additional participant samples. SECM measures should also be incorporated into more complex SEM models that treat evolution knowledge and religiosity as part of the structural paths to evolution acceptance. Such models could provide insights into the most worthwhile targets for the development of educational interventions to mitigate conflict at multiple scales.

Advancing more sophisticated conceptual models of biological thinking and reasoning. Many policy documents emphasize that student understanding of living systems requires the integration of concepts that span levels of biological organization, encompass the tree of life, and cross different fields of study (NRC, 2009; AAAS, 2011; NSF, 2019). Educational efforts to foster cognitive and practice-based competencies that align with disciplinary frameworks (such as Vision and Change) must consider what is known about student thinking about living systems rather than pieces and parts of them. Perhaps as a consequence of disciplinary isolation, remarkably little work in Biology Education Research (BER) has sought to identify common threads in the fabric of student confusion and to weave them into unified models of biological reasoning that are capable of explaining seemingly disparate educational challenges (Nehm 2019). To foster disciplinary unification and more integrative models of BER, these features should (1) span different biological subdisciplines and (2) undergird broad learning challenges about core ideas about living systems. Recently, Nehm (2019) proposed that three areas–unity and diversity; randomness, probability, and contingency; and scale, hierarchy, and emergence—are likely to be valuable ideas for the development of discipline-grounded conceptual frameworks for BER. Ongoing work is building empirical connections to this theoretical framework.

Equity, Diversity, and Inclusion in EEB. The Ecological and Evolutionary Sciences (EES) prepare students for careers in many areas of central importance to society, including environmental conservation and protection, biodiversity management, sustainability, medical genetics, and evolutionary medicine (Sbeglia and Nehm 2021). Nevertheless, the EESs appear to be one of the least diverse areas of study in terms of degree completion and career participation (ESA 2017). Although many STEM fields have been working to address patterns of underrepresentation, the EESs have only recently begun to focus on this topic (although see important work by Mead et al. 2015). Understanding the causes of race- and gender-based attrition, and addressing them early in academic pathways, could help maintain and foster a diverse talent pool for EES degrees and careers. Proposed but largely under-investigated variables that may exacerbate race-based attrition in biology in general and EES in particular include students’ knowledge of, conflict with, and acceptance of evolution. Prior work suggests (i) discussing the bounded nature of science, (ii) presenting the diversity of scientific views about religion, and (iii) introducing underrepresented scientist role models may collectively reduce conflict, increase acceptance, and foster greater understanding of the importance of EES to society. We are investigating (i) pre-course patterns of acceptance, conflict, religiosity, and evolutionary knowledge that characterize biology students in terms of racial and ethnic backgrounds and (ii) the impact that standard instruction and the addition of research-based instructional strategies alter these patterns. This work is designed to use evidence to eliminate institutional barriers to the success of all students in the ecological and evolutionary sciences. These efforts build on the lab’s 20-year history working to address inequities in educational achievement.

Interventions that address student misconceptions in biology Documenting student misconceptions and incorporating them into measurement instruments has laid important groundwork for developing and testing interventions designed to improve student learning of biology, including evolution (e.g. Nehm and Reilly 2007). Intervention studies remain comparatively rare, however, and are often small scale (e.g., one or two classes), lack robust research designs (e.g., no comparison groups, univariate designs), do not statistically control for background variables (e.g., knowledge, gender, race, EL status), utilize different instructors (many with atypical backgrounds, such as expertise in biology education; see Andrews et al. 2011), and measure change using instruments adopting single measures of learning. Ongoing work is exploring how best to address student misconceptions and the role that active learning plays in overcoming misconceptions.

Reasoning about matter and energy transformation in biological systems

GTA professional learning of equitable practices.

Untangling the casual contributors to evolution acceptance.