Purpose The authors developed online learning modules to train graduate and undergraduate student instructors (GUSIs) on grading and delivering feedback in quantitative disciplines. The authors report results from multiple assessments conducted during recent training events at a mid-sized, research-intensive institution and discuss implications for educational development. Design/methodology/approach Using pre/post-assessments, the author measured participants' learning gains and skill development. In Study 1, the authors measured learning gains for 109 computer science GUSIs randomly assigned to complete the modules or not. Participants who completed the modules performed significantly better on the post-assessment relative to the control group across all seven module learning objectives aligned with GUSI responsibilities. In Study 2, we iterated on both assessments and modules, replicating Study 1 results for GUSIs from other quantitative disciplines. In Study 3, the authors compared learning gains from online modules to in-person training sessions, focusing on the authentic task of providing written feedback on student work. Findings Proficiency improved equally and significantly via both training modalities. Originality/value At research-intensive universities, GUSI training can be inconsistent and difficult to scale and rarely assessed via direct measures of outcomes. To the authors’ knowledge, this is the first study to rigorously measure GUSI skill development via authentic assessment tasks such as grading student work and/or providing effective written feedback rather than simply testing knowledge. This study also addresses implications for designing and implementing effective GUSI training at scale.

Evidence-based practice in educational development includes leveraging data to iteratively refine center for teaching and learning (CTL) services. However, CTL data collection is often limited to counts and satisfaction surveys rather than direct measures of outcomes. To directly assess impacts of consultations on course and syllabus design, we analyzed 94 clients’ syllabi (32 faculty, 62 graduate students and postdocs) before and after consultations. Faculty and non-faculty clients demonstrated significant change following consultations (6% and 10% gains in syllabus rubric scores, representing 50% and 31% of possible gains and effect sizes of 0.73 and 1.04 standard deviations, respectively). We compared faculty clients to quasi-experimental control groups that did not receive consultations. Syllabi from non-clients scored lower and did not demonstrate similar changes across semesters. Attendance at a CTL seminar on course and syllabus design did not explain variation in clients’ syllabi. We discuss implications for assessment of CTL services and how we leveraged formative assessments to inform and iteratively refine our educational development practices.

This study updates and expands upon past work on how tenure-track hiring committees evaluate teaching effectiveness to provide centers for teaching and learning (CTLs) current data as they support graduate students and postdocs navigating the academic hiring process. In this study, 166 hiring committee chairs from nine academic disciplines and a variety of institution types (e.g., Baccalaureate, Master’s, and Doctoral) responded to survey questions addressing how they evaluate teaching effectiveness for tenure-track positions. Results indicate that hiring chairs across institution types and disciplines value a candidate’s teaching effectiveness. Hiring committees use a variety of documents to gather information about an applicant’s teaching, but the teaching philosophy statement is most common. The survey responses also provide greater clarity on what hiring chairs find effective in applicants’ teaching philosophy statements, diversity and inclusion statements, and presentations of student evaluations. We discuss implications of these findings for scholarship of educational development (SoED) researchers, CTL leadership engaged in strategic planning, and educational developers providing consultations and workshops for future faculty.

Copious research demonstrates the benefits of adding active learning to traditional lectures to enhance learning and reduce failure/withdrawal rates. However, many questions remain about how best to implement active learning to maximize student outcomes. This paper investigates several “second generation” questions regarding infusing active learning, via Think-Pair-Share (TPS), into a large lecture course in Computer Science. During the “Share” phase of TPS, what is the best way to debrief the associated course concepts with the entire class? Specifically, does student learning differ when instructors debrief the rationale for every answer choice (full debrief) versus only the correct answer (partial debrief)? And does the added value for student outcomes vary between tasks requiring recall versus deeper comprehension and/or application of concepts? Regardless of discipline, these questions are relevant to instructors implementing TPS with multiple-choice questions, especially in large lectures. Similar to prior research, when lectures included TPS, students performed significantly better (~13%) on corresponding exam items. However, students’ exam performance depended on both the type of debrief and exam questions. Students performed significantly better (~5%) in the full debrief condition than the partial debrief condition. Additionally, benefits of the full debrief condition were significantly stronger (~5%) for exam questions requiring deeper comprehension and/or application of underlying Computer Science processes, compared to simple recall. We discuss these results and lessons learned, providing recommendations for how best to implement TPS in large lecture courses in STEM and other disciplines.

In this study, we investigated the optimal placement of animations and practice and feedback exercises with respect to each other and to static text and graphics in an online DNA replication module. We randomly assigned students in a first-semester introductory biology course for freshmen biology majors and nonmajors to one of four online modules with animations and practice exercises (assets) either embedded with the text and images or saved until the end after all the text and images. Although we expected to find that embedding assets with text and images would improve learning outcomes, we were surprised to find that outcomes did not change with the location of these assets with respect to the text and static images. Instead, we found that colocation of exercises with animations was correlated with the biggest student learning gains, independent of where these assets appeared relative to the static text and static images. This held true both for immediate posttest scores and delayed application questions on a midterm exam. Our findings could have potential implications for how to best design online learning modules that feature multiple assets.

Pharmacy educators are improving education of professional program students by incorporating active learning techniques. Team-based learning “flips” the classroom, creating different roles for faculty and students compared to traditional lecture-based pedagogy. Implementing team-based learning on a large scale, such as across multiple semesters, introduces challenges that are distinct from implementation on a smaller scale. We describe our experience at the University of Michigan College of Pharmacy with adopting team-based learning in our curriculum. We adopted team-based learning as a unifying pedagogy across our five-semester therapeutics problem-solving course sequence. We experienced challenges distinct from those that accompany smaller scale adoption of an active learning pedagogy. Specifically, garnering faculty support, logistical issues, and implementation of the new pedagogy by faculty and students were all affected by the large scale of adoption. We share our experience with a large-scale pedagogical shift, highlighting challenges and lessons learned for other faculty in health professions education who may be interested in leveraging the benefits of active learning across several courses involving many faculty.

Change • January/February 2014 By Mary C. Wright, Timothy McKay, Chad Hershock, Kate Miller, and Jared Tritz Learning Analytics (LA) has been identified as one of the top technology trends in higher education today (Johnson et al., 2013). LA is based on the idea that datasets generated through normal administrative, teaching, or learning activities—such as registrar data or interactions with learning management systems—can be analyzed to enhance student learning, academic progress, and teaching practice. Examples of LA projects in colleges and universities include Purdue University’s “Course Signals” system, an early-alert notification for struggling students, and Austin Peay State University’s “Degree Compass,” a course recommender program based on predictive analytics.

Understanding the behavior of mixtures of species based solely on knowledge of the individual (component) species remains a big challenge for plant ecology. We used the observed outcome of two-species mixtures from a garden competition experiment with five clonal sedge species (two runners and three clumpers), and compared the data with predictions from a highly parameterized simulation model based only on monoculture data of these species. After two growing seasons (i.e., 300 growing days), overall performance of the mixtures (total biomass and ramet number of the mixtures, proportion of the biomass and ramet number of each species) was predicted rather well by the simulation model and the simulated variables were all within the 50% fit of observed values. Therefore, the single-species parameterization can capture most of the important processes that determine species behavior in the mixture and there is strong equivalence of species and a weak species-specific effect. This study demonstrates the power of modeling studies to perform virtual experiments for explicit hypothesis testing. Using this approach, we show that performance of mixtures can be realistically predicted by models parameterized and calibrated based on single species information. (C) 2011 Elsevier B.V. All rights reserved.

Although many kinds of data can be used to guide instructional consultations, research comparing the efficacy of such data is scant, especially in engineering. In this study, multiple modes of assessment were used to evaluate the impact of consultations informed by different kinds of data. This study illuminates two key aspects of instructional consultations: (1) their efficacy varies depending on the kind of data used to guide them, with student feedback from a Small Group Instructional Diagnosis (SGID) having the largest positive impact, and (2) the instructional consultant plays a key role in helping both interpret the available data and identify strategies for improvement. These findings suggest three implications for practice: (1) whenever possible, SGID‐based consultations should be offered systematically and proactively for engineering faculty, (2) data for other kinds of consultations should be tailored to the needs of the individual instructor, and (3) instructional consultants should be available to collaborate with faculty to enhance their teaching, thereby building an engineering culture that actively supports teaching and learning.

Individual traits are often assumed to be linked in a straightforward manner to plant performance and processes such as population growth, competition and community dynamics. However, because no trait functions in isolation in an organism, the effect of any one trait is likely to be at least somewhat contingent on other trait values. Thus, to the extent that the suite of trait values differs among species, the magnitude and even direction of correlation between values of any particular trait and performance is likely to differ among species. Working with a group of clonal plant species, we assessed the degree of this contingency and therefore the extent to which the assumption of simple and general linkages between traits and performance is valid. To do this, we parameterized a highly calibrated, spatially explicit, individual-based model of clonal plant population dynamics and then manipulated one trait at a time in the context of realistic values of other traits for each species. The model includes traits describing growth, resource allocation, response to competition, as well as architectural traits that determine spatial spread. The model was parameterized from a short-term (3 month) experiment and then validated with a separate, longer term (two year) experiment for six clonal wetland sedges, Carex lasiocarpa, Carex sterilis, Carex stricta, Cladium mariscoides, Scirpus acutus and Scirpus americanus. These plants all co-occur in fens in southeastern Michigan and represent a spectrum of clonal growth forms from strong clumpers to runners with long rhizomes.Varying growth, allocation and competition traits produced the largest and most uniform responses in population growth among species, while variation in architectural traits produced responses that were smaller and more variable among species. This is likely due to the fact that growth and competition traits directly affect mean ramet size and number of ramets, which are direct components of population biomass. In contrast, architectural and allocation traits determine spatial distribution of biomass; in the long run, this also affects population size, but its net effect is more likely to be mediated by other traits. Such differences in how traits affect plant performance are likely to have implications for interspecific interactions and community structure, as well as on the interpretation and usefulness of single trait optimality models.

Physiological integration and foraging behavior have both been proposed as advantages for clonal growth in heterogeneous environments. We tested three predictions concerning their short- and long-term effects on the growth of the clonal perennial sedge Schoenoplectus pungens (Pers.) Volk. ex Schinz and R. Keller: (1) growth would be greatest for clones with connected rhizomes and on heterogeneous soil, (2) clones would preferentially place biomass in the nutrient-rich patches of a spatially heterogeneous environment, and (3) physiological integration would decrease a clone’s ability to forage. We tested our predictions by growing S. pungens clones for 2 years in an experimental garden with two severing treatments (connected and severed rhizomes) crossed with two soil treatments (homogeneous and heterogeneous nutrient distribution). Severing treatments were only carried out in the first year. As predicted, severing significantly decreased total biomass and per capita growth rate in year one and individual ramet biomass both in year one and the year after severing stopped. This reduction in growth was most likely caused by severing damage, because the total biomass and growth rate in severed treatments did not vary with soil heterogeneity. Contrary to our prediction, total biomass and number of ramets were highest on homogeneous soil at the end of year two, regardless of severing treatment, possibly because ramets in heterogeneous treatments were initially planted in a nutrient-poor patch. Finally, as predicted, S. pungens concentrated ramets in the nutrient-rich patches of the heterogeneous soil treatment. This foraging behavior seemed enhanced by physiological integration in the first year, but any possible enhancement disappeared the year after severing stopped. It seems that over time, individual ramets become independent, and parent ramets respond independently to the conditions of their local microsite when producing offspring, a life-history pattern that may be the rule for clonal species with the spreading “guerrilla” growth form.

Although clonal plants comprise most of the biomass of several widespread ecosystems, including many grasslands, wetlands, and tundra, our understanding of the effects of clonal attributes on community patterns and processes is weak. Here we present the conceptual basis for experiments focused on manipulating clonal attributes in a community context to determine how clonal characteristics affect interactions among plants at both the individual and community levels. All treatments are replicated at low and high density in a community density series to compare plant responses in environments of different competitive intensity. We examine clonal integration, the sharing of resources among ramets, by severing ramets from one another and comparing their response to ramets with intact connections. Ramet aggregation, the spacing of ramets relative to each other, is investigated by comparing species that differ in their natural aggregation (either clumped growth forms, with ramets tightly packed together, or runner growth forms, with ramets loosely spread) and by planting individual ramets of all species evenly spaced throughout a mesocosm. We illustrate how to test predictions to examine the influence of these two clonal traits on competitive interactions at the individual and community levels. To evaluate the effect of clonal integration on competition, we test two predictions: at the individual level, species with greater clonal integration will be better individual-level competitors, and at the community level, competition will cause a greater change in community composition when ramets are integrated (connected) than when they are not. For aggregation we test at the individual level: clumped growth forms are better competitors than runner growth forms because of their ability to resist invasion, and at the community level: competition will have a greater effect on community structure when ramets are evenly planted. An additional prediction connects the individual- and community-level effects of competition: resistance ability better predicts the effects of competition on relative abundance in a community than does invasion ability. We discuss additional experimental design considerations as revealed by our ongoing studies. Examining how clonal attributes affect both the individual- and community-level effects of competition requires new methods and metrics such as those presented here, and is vital to understanding the role of clonality in community structure of many ecosystems.

Guided inquiry-based lab activities improve students’ recall and application of material properties compared to structured inquiry Prior research suggests that lab courses taught by following traditional, recipe-based instructions may lead to suboptimal learning outcomes when compared to inquiry-based procedures where the students must think for themselves. However, the relative impacts of different forms of inquiry-based learning are unclear. In an attempt to extend the current research, we measured student learning in a materials lab course that taught labs in two different ways. Our main objective was to test whether implementing structured or guided inquiry-based learning methods, would lead to better learning outcomes. Two inquiry-based learning methods were implemented in a junior-level, civil engineering materials lab course that featured three independent labs based on the materials of concrete, wood, and masonry. The concrete and wood labs were taught using a structured inquiry-based approach known as predict-observe-explain or POE. Students predicted the outcome of a lab exercise before following a standard, recipe-based lab protocol provided by the instructor. Afterwards, students evaluated their predictions by explaining observed data and underlying concepts of material properties. In contrast, the masonry lab was taught using a guided inquiry-based approach. Students were given an authentic engineering question: is the design of a masonry building for which they were given drawings feasible? To answer the question, they needed to decide which material properties were important and identify the experimental methods necessary to determine those properties. Because the activities related to guided inquiry were new to the students, the activities were broken into scaffolded tasks and questions. Students were assessed on all labs via a final exam at the end of the semester, allowing for a within-subjects comparison of student learning using both instructional approaches. Analysis of the data showed a significant difference in student learning for the content related to the guided inquiry-based lab compared to the content from the two structured inquiry-based labs. Relative to content from the two structured inquiry-based labs, students exhibited an 11% gain, over a full letter grade, in learning on content related to the inquiry-based lab (Cohen’s d = .86). This gain was consistent across recall- and application-based exam items. Design for implementation of inquiry-based methods can take significant time and preparation. Student development of experimental methods appears to be an effective inquiry-based method in this case, but requires a significant amount of guidance and oversight to effectively implement. In addition to improved exam performance, application of results to solve real-world problems gives the students an understanding of how their experimental work relates to their other courses and the world in general, which provides context and may increase motivation. Given the workload, a best practice may be implementing these methods incrementally rather than implementing a wholesale change in a course. The guided inquiry-based methods applied in this materials lab course can be applied in all types of classes, but methods are most easily transferable to laboratory, design, and problem-based project courses. For future iterations of this course, we are redesigning the two structured inquiry-based labs using guided inquiry approaches and will be continuing to collect data to assess their effectiveness.