(Re)-Assessing the Role of Qualifying Examinations in Training Next-Generation Biostatisticians

In this presentation, I frame the biostatistics doctoral qualifying exam (QE) within both the historical context of the discipline's founding and its modern role as a curricular "gatekeeper". I present some mutually agreed upon ideas regarding the purpose of the exam(s) and then explore whether QEs in biostatistics are aligned with that purpose or should be redesigned to be better aligned as well as to support inclusive excellence and student success. I contend that as currently operationalized, these high-stakes assessments often lack clear rubrics and may function as structural barriers to student success. I also demonstrate that students may not always have clear guidance on how best to prepare for this high-stakes milestone I advocate for moving beyond "assimilationist inclusion" toward transparent assessment methods that better measure genuine research readiness. 

As graduate programs in (bio)statistics re-assess the role of qualifying examinations as milestones toward degree completion, it is important to first consider the fundamental purpose of the qualifying examinations in the students' overall training. Although different programs may arrive at slightly different answers, many would likely say that the overarching purpose of these examinations is to enhance the rigor of the training. To that end, I propose a number of suggestions toward thoughtful exam development as we enter the era of pervasive generative AI. I approach this in two pieces: (1) exam development, and (2) exam assessment. 

Doctoral training in biostatistics aims to develop independent researchers capable of advancing statistical methods and to prepare collaborative scientists to lead interdisciplinary teams in study design, data analysis, and ethical decision making. Qualifying exams are commonly used to assess theoretical mastery and applied competence before students advance to dissertation research. Yet across biostatistics programs, the structure, content, and expectations of qualifying exams vary, and it remains unclear whether current exam models accurately reflect the skills needed in contemporary methodological research or in real world collaborative practice.

This talk will examine the core competencies expected of PhD trained biostatisticians, review the predominant qualifying exam formats used across programs, and evaluate the degree to which these formats align with evolving methodological and collaborative demands of the field. I will highlight gaps between exam objectives and professional expectations, raise questions about how qualifying exams might be modernized, and invite a broader discussion on evidence informed approaches for assessing doctoral readiness in biostatistics. 

Arguably, the most important aspect of PhD training is the transition from "classroom-based learning", where material is organized and synthesized by an instructor in a structured manner and where problems posed have known solutions (indeed, there is a solution), to "self-directed learning", which requires the individual to organize and synthesize a typically large body of existing work, and then contend with problems for which there is no known solution (indeed, there may not be a solution). Key to success in this transition is a shift in mindset, the development of new time management skills, and increased comfort with uncertainty, setbacks and responsibility. Since all of these are both highly personal and cannot be taught in a classroom, PhD training programs need to provide experiential mechanisms through which students are initially exposed to what self-directed learning entails, and though which they begin to develop the corresponding skills. Simultaneously, as with all aspects of training, programs need mechanisms through which student progress can be assessed. I will argue that qualifying exams provide such a mechanism, providing a means to balance programmatic responsibilities of: (i) giving students avenues for growth and development, (ii) ensuring appropriate progress towards timely completion; and (iii) setting students up for successful long-term careers. 

A growing and diverse pipeline is bringing students into biostatistics Ph.D. programs from many different backgrounds, training experiences, and scientific interests. It is only natural that such expansion of discipline and student diversity should be accompanied by continued reconsideration of education and training, and recent experience points to qualifying exams (QEs) as warranting attention. But QEs are just one strand in a complicated web of student pipeline, admissions processes, curriculum, and departmental priorities, and as such cannot be considered in isolation. This talk defends the maintenance of QEs - even traditional, timed, written exams - but looks to address their evolving place in biostatistical training by reconsidering how the goals of the QE can be appropriately situated within the broader training process to maintain departmental standards while maximizing student success. In particular, I advocate that we elevate considerations of what we hope students to gain from the QE process to receive at least equal weight alongside considerations of what we are trying to evaluate through rigorous examination.