This study describes the result of implementing the Problem List Generator (PLG), a computer-based tool designed to helpYou aren't feeling well, so you go to your doctor. A physical exam reveals nothing obvious, so blood and urine samples are sent to the lab. When the results come back, your doctor has a problem. The extent to which that problem is resolved satisfactorily (and, perhaps, your health) depends on your doctor's ability to interpret all the data that have been collected. This skill, the process of turning all the information available to the physician into a diagnosis and treatment plan, is known as clinical problem solving; the subproblem of correctly interpreting the clinical laboratory data is diagnostic problem solving.Smith and Ragan (1999) defined problem solving as "the ability to combine previously learned principles, procedures, declarative knowledge, and cognitive strategies in a unique way within a domain of content to solve previously unencountered problems" (p. 132). Jonassen (2000) suggested that problem solving involves two processes: (a) the construction of a mental model of the problem (the problem space), and (b) activity-based manipulation of the problem space. He argued that general models of problem solving have proven inadequate for dealing with the rich diversity of problems faced by learners, as manifested by the fact that, although researchers tend to value problem solving as a learning outcome, many instructional design models and teaching theories provide sparse or nonspecific guidance when it comes to instructional strategies for learning problem solving. There are many kinds of problems, varying in complexity and inex-

Teaching introductory clinical pathology to veterinary students is a challenging endeavor that requires a shift in learning strategies from rote memorization to diagnostic reasoning. Educational research has identified discrete cognitive stages required to achieve the automated, unconscious thinking process used by experts. Building on this knowledge, we developed a case-based approach to clinical pathology instruction that actively engages students in the learning process and links performance with positive reward. Simulated cases provide context and create a structure, or "schema", which enhances the learning process by enabling students to synthesize facts and link them with their causal mechanism to reach a defensible diagnostic conclusion. Web-based tools, including the "Problem List Generator" and tutorials, have been developed to facilitate this process. Through the collaborative Biomedical Informatics Research Group, we are working to further develop and evaluate Web-based instructional tools and new educational methods, to clarify the diagnostic reasoning processes used by experienced clinical pathologists, and, ultimately, to better educate our future students to be effective diagnosticians.

Three related studies replicated and extended previous work ( J.A. Danielson et al. (2003),Educational Technology Research and Development, 51(3), 63-81) involving the Diagnostic Pathfinder (dP) (previously Problem List Generator [PLG]), a cognitive tool for learning diagnostic problem solving. In studies 1 and 2, groups of 126 and 113 veterinary students, respectively, used the dP to complete case-based homework; groups of 120 and 199, respectively, used an alternative method. Students in the dP groups scored significantly higher (p = .000 and .003, respectively) on final exams than those in control groups. In the third study, 552 veterinary students responding to a questionnaire indicated that the dP's gating and data synthesis activities aided learning. The dP's feedback and requirement of completeness appear to aid learning most. AbstractThree related studies replicated and extended previous work (J.A. Danielson, et al, 2003)

First-year didactic course instructors at the University of Illinois College of Veterinary Medicine leverage earlier clinical rotation experiences with weekly "Clinical Correlations" exercises to provide early exposure to critical clinical thinking (CCT). This study evaluated the efficacy of individual and paired group exercises on CCT development. Before and after instruction, the Cornell Critical Thinking Test (Level Z) (CCTTZ) was administered. Based on the hypothesis that students with higher scores would coach lower-scoring colleagues during group exercises, heterogeneous groups with similar mean scores were established for the year. Students completed 14 individual and paired group exercises over 6 months. Exercises were designed to increase in complexity and decline in scaffolding. Seven of the exercises were cases using the Applied Learning Platform (ALP) at http://www.whenknowingmatters.com . Student analyses were scored according to a six-category critical-thinking rubric using a 5-point scale. Consistent with our hypothesis, individual and group rubric scores increased significantly, plateauing near the end of the year. Contrary to our hypothesis, mean overall CCTTZ scores did not change, but there was a small statistically significant increase in the ability to assess the validity of an argument. Student attitudes were mixed. Positive comments focused on reinforcement of prior didactic instruction, while negative comments focused on preparation time needed to conduct research on clinical concepts, and on a lack of explicit evaluation by summative examinations. Nonetheless, end-of-year GPAs correlated linearly with cumulative individual rubric scores. In summary, the value of early curriculum CCT training was confirmed when discipline-specific criteria were applied.