Using a linear mixed-effect model framework to estimate multivariate generalizability theory parameters in R.

Multivariate generalizability theory (mG-theory) is an important framework in many behavioral and educational studies, as it describes useful psychometric properties of multidimensional assessments. Nevertheless, the use of mG-theory estimation is limited due to the lack of available software for carrying out the necessary calculations: users rely heavily on independent software programs such as mGENOVA and the BUGS/JAGS suite of programs. Considering the prevalence of R software, this paper presents a solution using the glmmTMB package to accomplish the estimation task. Users adopting the proposed method may find it more convenient for conducting both applied investigation and simulation studies without the need to switch between different software programs.

The optimal number of options for multiple-choice questions on high-stakes tests: application of a revised index for detecting nonfunctional distractors.

Research suggests that the three-option format is optimal for multiple choice questions (MCQs). This conclusion is supported by numerous studies showing that most distractors (i.e., incorrect answers) are selected by so few examinees that they are essentially nonfunctional. However, nearly all studies have defined a distractor as nonfunctional if it is selected by fewer than 5% of examinees. A limitation of this definition is that the proportion of examinees available to choose a distractor depends on overall item difficulty. This is especially problematic for mastery tests, which consist of items that most examinees are expected to answer correctly. Based on the traditional definition of nonfunctional, a five-option MCQ answered correctly by greater than 90% of examinees will be constrained to have only one functional distractor. The primary purpose of the present study was to evaluate an index of nonfunctional that is sensitive to item difficulty. A secondary purpose was to extend previous research by studying distractor functionality within the context of professionally-developed credentialing tests. Data were analyzed for 840 MCQs consisting of five options per item. Results based on the traditional definition of nonfunctional were consistent with previous research indicating that most MCQs had one or two functional distractors. In contrast, the newly proposed index indicated that nearly half (47.3%) of all items had three or four functional distractors. Implications for item and test development are discussed.

A practical guide to test blueprinting.

A test blueprint describes the key elements of a test, including the content to be covered, the amount of emphasis allocated to each content area, and other important features. This article offers practical guidelines for developing test blueprints. We first discuss the role of learning outcomes and behavioral objectives in test blueprinting, and then describe a four-stage process for creating test blueprints. The steps include identifying the major knowledge and skill domains (i.e. competencies); delineating the specific assessment objectives; determining the method of assessment to address those objectives; and establishing the amount of emphasis to allocate to each knowledge or skill domain. The article refers to and provides examples of numerous test blueprints for a wide variety of knowledge and skill domains. We conclude by discussing the role of test blueprinting in test score validation, and by summarizing some of the other ways that test blueprints support instruction and assessment.

Crowdsourcing for assessment items to support adaptive learning.

PURPOSE: Adaptive learning requires frequent and valid assessments for learners to track progress against their goals. This study determined if multiple-choice questions (MCQs) "crowdsourced" from medical learners could meet the standards of many large-scale testing programs. METHODS: Users of a medical education app (Osmosis.org, Baltimore, MD) volunteered to submit case-based MCQs. Eleven volunteers were selected to submit MCQs targeted to second year medical students. Two hundred MCQs were subjected to duplicate review by a panel of internal medicine faculty who rated each item for relevance, content accuracy, and quality of response option explanations. A sample of 121 items was pretested on clinical subject exams completed by a national sample of U.S. medical students. RESULTS: Seventy-eight percent of the 200 MCQs met faculty reviewer standards based on relevance, accuracy, and quality of explanations. Of the 121 pretested MCQs, 50% met acceptable statistical criteria. The most common reasons for exclusion were that the item was too easy or had a low discrimination index. CONCLUSIONS: Crowdsourcing can efficiently yield high-quality assessment items that meet rigorous judgmental and statistical criteria. Similar models may be adopted by students and educators to augment item pools that support adaptive learning.

A Six Year Experience of a National Leadership Fellowship in Northern Ireland: Achieve, Develop, Explore Programme for Trainees (ADEPT).

Introduction: The Achieve, Develop, Explore Programme for Trainees (ADEPT) Clinical Leadership Fellowship Programme was established in response to growing recommendations to underpin healthcare reconfiguration in Northern Ireland with a collective leadership strategy. The fellowship combines a leadership development programme with a project carried out within a host organisation. With the fellowship now in its sixth year, a need was identified to assess its impact on the fellows' leadership skills, career choices, achievements, and views on both the fellowship and how to develop future leaders. Methods: Demographic data for all ADEPT fellows was held centrally through Northern Ireland Medical and Dental Training Agency (NIMDTA) and assessed anonymously. A mixed-methods questionnaire was composed using Smart Survey. Likert scale questions were designed to determine the extent to which participants believed ADEPT supported their development of strong and exemplary elements of the nine dimensions of the NHS Healthcare Leadership Model. The questionnaire was distributed electronically to all ADEPT alumni in November 2021 and remained open for 4 weeks. Results: There have been 46 ADEPT fellows to date (72% female; all fellows were white). ADEPT fellows were most commonly from Psychiatry (33%), Paediatrics (17%) and Obstetrics and Gynaecology (15%). There were 19 responses from the alumni cohort of 46 (41%). 75% of respondents reported that their project resulted in publication, presentation or award. Leadership skill development was identified as best in "Evaluating Information" and "Engaging the Team", whereas skills in "Sharing the Vision" and "Developing Capability" saw less improvement. The majority felt that the fellowship had been useful in securing their position as a consultant or general practitioner and 50% went on to pursue senior leadership positions. Conclusion: The ADEPT Clinical Leadership Fellowship delivers effective leadership training as measured by the nine domains of the NHS Healthcare Leadership Model. It provides value for host organisations through the projects undertaken and by developing doctors who are more likely to engage in future formal leadership roles. ADEPT alumni saw the value in their leadership experience and felt it should be embedded in standard postgraduate training schemes to reach a wider audience.

Measurement precision for repeat examinees on a standardized patient examination.

Examinees who initially fail and later repeat an SP-based clinical skills exam typically exhibit large score gains on their second attempt, suggesting the possibility that examinees were not well measured on one of those attempts. This study evaluates score precision for examinees who repeated an SP-based clinical skills test administered as part of the US Medical Licensing Examination sequence. Generalizability theory was used as the basis for computing conditional standard errors of measurement (SEM) for individual examinees. Conditional SEMs were computed for approximately 60,000 single-take examinees and 5,000 repeat examinees who completed the Step 2 Clinical Skills Examination(®) between 2007 and 2009. The study focused exclusively on ratings of communication and interpersonal skills. Conditional SEMs for single-take and repeat examinees were nearly indistinguishable across most of the score scale. US graduates and IMGs were measured with equal levels of precision at all score levels, as were examinees with differing levels of skill speaking English. There was no evidence that examinees with the largest score changes were measured poorly on either their first or second attempt. The large score increases for repeat examinees on this SP-based exam probably cannot be attributed to unexpectedly large errors of measurement.

Snocross "Shark-bite" Laceration.

CASE PRESENTATION: A snowmobile racer fell from his sled and was run over by another, sustaining "shark bite" to his hand and leg. He was evacuated to a trackside medical trailer where the characteristic wounds were felt to require further exploration at a hospital. DISCUSSION: "Shark bite" is a colloquial term for lacerations sustained from metal studs attached to a snowmobile's track. "Shark-bite" lacerations may be more prone to complications than other lacerations commonly sustained in motorsports events.

A Monte Carlo Study of Confidence Interval Methods for Generalizability Coefficient.

Computing confidence intervals around generalizability coefficients has long been a challenging task in generalizability theory. This is a serious practical problem because generalizability coefficients are often computed from designs where some facets have small sample sizes, and researchers have little guide regarding the trustworthiness of the coefficients. As generalizability theory can be framed to a linear mixed-effect model (LMM), bootstrap and simulation techniques from LMM paradigm can be used to construct the confidence intervals. The purpose of this research is to examine four different LMM-based methods for computing the confidence intervals that have been proposed and to determine their accuracy under six simulated conditions based on the type of test scores (normal, dichotomous, and polytomous data) and data measurement design (p×i×r and p× [i:r]). A bootstrap technique called "parametric methods with spherical random effects" consistently produced more accurate confidence intervals than the three other LMM-based methods. Furthermore, the selected technique was compared with model-based approach to investigate the performance at the levels of variance components via the second simulation study, where the numbers of examines, raters, and items were varied. We conclude with the recommendation generalizability coefficients, the confidence interval should accompany the point estimate.

Bispecific Anti-HIV Immunoadhesins That Bind Gp120 and Gp41 Have Broad and Potent HIV-Neutralizing Activity.

We have constructed bispecific immunoglobulin-like immunoadhesins that bind to both the HIV-envelope glycoproteins: gp120 and gp41. These immunoadhesins have N terminal domains of human CD4 engrafted onto the N-terminus of the heavy chain of human anti-gp41 mAb 7B2. Binding of these constructs to recombinant Env and their antiviral activities were compared to that of the parental mAbs and CD4, as well as to control mAbs. The CD4/7B2 constructs bind to both gp41 and gp140, as well as to native Env expressed on the surface of infected cells. These constructs deliver cytotoxic immunoconjugates to HIV-infected cells, but not as well as a mixture of 7B2 and sCD4, and opsonize for antibody-mediated phagocytosis. Most surprisingly, given that 7B2 neutralizes weakly, if at all, is that the chimeric CD4/7B2 immunoadhesins exhibit broad and potent neutralization of HIV, comparable to that of well-known neutralizing mAbs. These data add to the growing evidence that enhanced neutralizing activity can be obtained with bifunctional mAbs/immunoadhesins. The enhanced neutralization activity of the CD4/7B2 chimeras may result from cross-linking of the two Env subunits with subsequent inhibition of the pre-fusion conformational events that are necessary for entry.

The impact of statistical adjustment on conditional standard errors of measurement in the assessment of physician communication skills.

The use of standardized patients to assess communication skills is now an essential part of assessing a physician's readiness for practice. To improve the reliability of communication scores, it has become increasingly common in recent years to use statistical models to adjust ratings provided by standardized patients. This study employed ordinary least squares regression to adjust ratings, and then used generalizability theory to evaluate the impact of these adjustments on score reliability and the overall standard error of measurement. In addition, conditional standard errors of measurement were computed for both observed and adjusted scores to determine whether the improvements in measurement precision were uniform across the score distribution. Results indicated that measurement was generally less precise for communication ratings toward the lower end of the score distribution; and the improvement in measurement precision afforded by statistical modeling varied slightly across the score distribution such that the most improvement occurred in the upper-middle range of the score scale. Possible reasons for these patterns in measurement precision are discussed, as are the limitations of the statistical models used for adjusting performance ratings.

Developing and verifying the psychometric integrity of the certification examination for imaging informatics professionals.

The American Board of Imaging Informatics (ABII) was founded in 2005 by the Society of Imaging Informatics in Medicine (SIIM) and the American Registry of Radiologic Technologists (ARRT). ABII's mission is to enhance patient care, professionalism, and competence in imaging informatics. This is accomplished primarily through the development and administration of a certification examination. The creation of the exam has been an exercise in open community involvement with SIIM providing access to the PACS community and ARRT providing skilled psychometric support to ensure a balanced and comprehensive examination. The process to generate the exam required several years and the efforts of dozens of subject matter experts active who volunteered to submit and validate questions for the examination. This article describes the organizational and statistical processes used to generate test items, assemble test forms, set performance standards, and validate test scores.

Indices of Subscore Utility for Individuals and Subgroups Based on Multivariate Generalizability Theory.

Conventional methods for evaluating the utility of subscores rely on traditional indices of reliability and on correlations among subscores. One limitation of correlational methods is that they do not explicitly consider variation in subtest means. An exception is an index of score profile reliability designated as G , which quantifies the ratio of true score profile variance to observed score profile variance. G has been shown to be more sensitive than correlational methods to group differences in score profile utility. However, it is a group average, representing the expected value over a population of examinees. Just as score reliability varies across individuals and subgroups, one can expect that the reliability of score profiles will vary across examinees. This article proposes two conditional indices of score profile utility grounded in multivariate generalizability theory. The first is based on the ratio of observed profile variance to the profile variance that can be attributed to random error. The second quantifies the proportion of observed variability in a score profile that can be attributed to true score profile variance. The article describes the indices, illustrates their use with two empirical examples, and evaluates their properties with simulated data. The results suggest that the proposed estimators of profile error variance are consistent with the known error in simulated score profiles and that they provide information beyond that provided by traditional measures of subscore utility. The simulation study suggests that artificially large values of the indices could occur for about 5% to 8% of examinees. The article concludes by suggesting possible applications of the indices and discusses avenues for further research.

The Use of Multivariate Generalizability Theory to Evaluate the Quality of Subscores.

Conventional methods for evaluating the utility of subscores rely on reliability and correlation coefficients. However, correlations can overlook a notable source of variability: variation in subtest means/difficulties. Brennan introduced a reliability index for score profiles based on multivariate generalizability theory, designated as G , which is sensitive to variation in subtest difficulty. However, there has been little, if any, research evaluating the properties of this index. A series of simulation experiments, as well as analyses of real data, were conducted to investigate G under various conditions of subtest reliability, subtest correlations, and variability in subtest means. Three pilot studies evaluated G in the context of a single group of examinees. Results of the pilots indicated that G indices were typically low; across the 108 experimental conditions, G ranged from .23 to .86, with an overall mean of 0.63. The findings were consistent with previous research, indicating that subscores often do not have interpretive value. Importantly, there were many conditions for which the correlation-based method known as proportion reduction in mean-square error (PRMSE; Haberman, 2006) indicated that subscores were worth reporting, but for which values of G fell into the .50s, .60s, and .70s. The main study investigated G within the context of score profiles for examinee subgroups. Again, not only G indices were generally low, but it was also found that G can be sensitive to subgroup differences when PRMSE is not. Analyses of real data and subsequent discussion address how G can supplement PRMSE for characterizing the quality of subscores.

When Listening Is Better Than Reading: Performance Gains on Cardiac Auscultation Test Questions.

PURPOSE: In 2007, the United States Medical Licensing Examination embedded multimedia simulations of heart sounds into multiple-choice questions. This study investigated changes in item difficulty as determined by examinee performance over time. The data reflect outcomes obtained following initial use of multimedia items from 2007 through 2012, after which an interface change occurred. METHOD: A total of 233,157 examinees responded to 1,306 cardiology test items over the six-year period; 138 items included multimedia simulations of heart sounds, while 1,168 text-based items without multimedia served as controls. The authors compared changes in difficulty of multimedia items over time with changes in difficulty of text-based cardiology items over time. Further, they compared changes in item difficulty for both groups of items between graduates of Liaison Committee on Medical Education (LCME)-accredited and non-LCME-accredited (i.e., international) medical schools. RESULTS: Examinee performance on cardiology test items with multimedia heart sounds improved by 12.4% over the six-year period, while performance on text-based cardiology items improved by approximately 1.4%. These results were similar for graduates of LCME-accredited and non-LCME-accredited medical schools. CONCLUSIONS: Examinees' ability to interpret auscultation findings in test items that include multimedia presentations increased from 2007 to 2012.

Investigating the Performance of Second Language Medical Students on Lengthy Clinical Vignettes.

This study evaluated the extent to which medical students with limited English-language experience are differentially impacted by the additional reading load of test items consisting of long clinical vignettes. Participants included 25,012 examinees who completed Step 2 of the U.S. Medical Licensing Examination®. Test items were categorized into five levels based on the number of words per item, and examinee scores at each level were evaluated as a function of English-language experience (English as a second language [ESL] status and scores on a test of English-speaking proficiency). The longest items were more difficult than the shortest items across all examinee groups, and examinees with more English-language experience scored higher than those with less experience across all five levels of word count. The effect of primary interest-the interaction of word count with English-language experience-was statistically significant, indicating that score declines for longer items were larger for examinees with less English-language experience; however, the magnitude of this interaction effect was barely detectable (η2 = .0004, p < .001). Additional analyses supported the conclusion that the differential effect for examinees with less English-language experience was small but worthy of continued monitoring.

Using National Health Care Databases and Problem-Based Practice Analysis to Inform Integrated Curriculum Development.

One challenge when implementing case-based learning, and other approaches to contextualized learning, is determining which clinical problems to include. This article illustrates how health care utilization data, readily available from the National Center for Health Statistics (NCHS), can be incorporated into an educational needs assessment to identify medical problems physicians are likely to encounter in clinical practice. The NCHS survey data summarize patient demographics, diagnoses, and interventions for tens of thousands of patients seen in various settings, including emergency departments (EDs), clinics, and hospitals.Selected data from the National Hospital Ambulatory Medical Care Survey: Emergency Department illustrate how instructional materials can be derived from the results of such public-use health care data. Using fever as the reason for visit to the ED, the patient management path is depicted in the form of a case drill-down by exploring the most common diagnoses, blood tests, diagnostic studies, procedures, and medications associated with fever.Although these types of data are quite useful, they should not serve as the sole basis for determining which instructional cases to include. Additional sources of information should be considered to ensure the inclusion of cases that represent infrequent but high-impact problems and those that illustrate fundamental principles that generalize to other cases.

The relationship between educational preparation and performance on the ARRT examination in radiation therapy.

PURPOSE: To determine if graduates of different types of educational programs obtain similar scores on the Examination in Radiation Therapy administered by the American Registry of Radiologic Technologists. The results will help inform discussions regarding educational requirements for radiation therapists. METHODS AND MATERIALS: Test scores were obtained for 531 candidates who had taken the examination for the first time in 1997, 1998, or 1999. Candidates were divided into the following three categories, based on the type of educational program attended: hospital-based certificate, associate's degree, or bachelor's degree. To determine if test scores were related to the type of educational preparation, analyses of variance were conducted separately to test for differences in total scores and section scores, and scores on test questions intended to measure critical thinking skills. RESULTS: Candidates with an associate's degree scored slightly lower than candidates with a bachelor's degree on the total test (p < 0.10) and lower than candidates with either a certificate or bachelor's degree on Section B of the examination (Treatment Planning and Delivery, p < 0.10). Baccalaureate candidates did not obtain higher scores than those prepared in certificate programs. On critical thinking questions, candidates with certificates scored higher than those with associate's degrees (p < 0.10). Some evidence suggested that candidates with a certificate scored higher on critical thinking than those with a bachelor's degree (p < 0.10), and that candidates with a bachelor's degree scored higher than candidates with an associate's degree (p < 0.10). CONCLUSIONS: Although some of the differences in the mean test scores among the three educational groups were statistically significant, all differences were small and do not support one type of educational preparation over another.

The influence of educational preparation on evaluations of applicants for radiation therapy positions: a policy-capturing experiment.

PURPOSE: To determine whether radiation therapy department administrators prefer to hire graduates with certain types of educational preparation. The study was undertaken by the American Registry of Radiologic Technologists as part of a larger project to determine educational requirements for radiation therapists. METHODS AND MATERIALS: Forty-one department administrators evaluated applications from a pool of 984 hypothetical applicants for the position of radiation therapist. Applications were created by systematically varying eight characteristics such as years of experience, quality of educational program, and ratings from prior references. Type of educational program (baccalaureate degree, associate's degree, or hospital certificate) was of particular interest in this study. Each administrator evaluated 24 applications and assigned a rating ranging from 1 to 5 to indicate the extent to which he or she desired to interview each applicant. All ratings and applicant characteristics were coded and subjected to regression-type analyses to determine the relative importance of each applicant characteristic to administrators' decision-making policies. RESULTS: Information obtained from applicant references had the greatest impact on administrators' evaluations of applicant quality. Specifically, reference ratings of cooperation and technical skills were the two most important characteristics, followed closely by reference ratings of interpersonal skills and dependability. Quality of educational program had some influence, as did years of experience. Type of educational program had virtually no impact on interview decisions for a vast majority of the administrators. CONCLUSIONS: When making hiring decisions about hypothetical applicants, department administrators place most emphasis on evidence relating to past performance and give almost no weight to type of educational preparation. The extent to which these results generalize to actual applicants is addressed in the article.

Empirically mapping the subspecialties of cardiovascular-interventional technology.

As the practice of cardiovascular interventional technology (CVIT) has evolved over the last 50 years, so has the role of radiographers employed in this specialty. In 1991, the American Registry of Radiologic Technologists (ARRT) initiated a certification program to recognize radiologic technologists practicing in CVIT. The certification program consisted of a single examination that covered all aspects of CVIT (e.g., neurologic, cardiac, genitourinary). In 2000, the ARRT conducted a study to investigate further the nature of subspecialization occurring within CVIT. A comprehensive job analysis questionnaire was developed that consisted of 137 clinical activities organized into 19 general domains of practice. The questionnaire was completed by a national sample of 848 radiologic technologists working in CVIT, who indicated the frequency with which they performed each of the 137 activities. Responses were subjected to cluster analysis to classify technologists into homogeneous groups corresponding to different CVIT subspecialties. Results indicated that CVIT consists of two major subspecialties: one corresponding to cardiac procedures and one corresponding to procedures involving organ systems other than the heart. Other smaller subspecialties also emerged from the cluster analysis. A multidimensional scaling of the profiles suggested that CVIT subspecialization can be explained by two dimensions: (1) whether the procedures are diagnostic or interventional and (2) the type of organ system involved. The findings are discussed in terms of their implications for education, certification, and performance evaluation.