Validity and fairness in leader assessment
A planned line of studies on whether leadership assessments predict outcomes equally well across groups, and how to make results clearer for the people who use them.
Research
Careful science on practical questions about people at work.
We run experiments and surveys, build and test assessments, and analyze data. Every study with people is approved by URI’s research ethics board (the IRB) before it begins.
01
Fair and accurate assessment
How do we know a test measures something real, and treats people fairly?
Organizations make hiring, development, and promotion decisions based on assessments, so those tools need to work. We build and evaluate them: defining what good performance looks like, writing items that capture it, and testing whether scores are reliable, valid, and fair across groups. This work grows out of Dr. Hanson's experience building assessments for military organizations, and the lab now asks the same questions in public-sector and business settings.
Current questions
- Do leader assessment scores predict later performance equally well for different groups?
- How do design and recruitment choices affect data quality and whose experiences are represented?
Did more competition close the strength gap?
An analysis of 2.7 million powerlifting results asking how much of women’s gains reflect a bigger talent pool and how much reflect real improvement.
Read the analysis →02
Leadership competencies and AI-assisted feedback
What does good leadership look like in a given role, and can AI help people develop it?
“Good leadership” looks different from one role to the next. Defining role-specific competencies first is what makes later selection and feedback defensible. We study how competency models are built and validated, and how assessment feedback can serve as a development tool, not only a gate. A newer thread asks whether large language models, paired with validated assessments, can give personalized feedback to many more people without lowering the standards that make it trustworthy.
Current questions
- Can conversational AI administer an assessment and give feedback that is accurate and useful?
- Can AI help translate competency models into practical guidance, such as resume advice for service members moving into civilian careers?
AI-assisted assessment and feedback
Building on a proof-of-concept platform, this project will test whether large language models paired with validated assessments can give accurate, personalized developmental feedback at scale.
03
Motivation, needs, and recovery
Can a good life outside work make up for needs that go unmet at work?
Self-determination theory holds that people need autonomy, competence, and relatedness. We study how meeting those needs, at work and away from it, shapes motivation, burnout, recovery, and well-being. Dr. Hanson's dissertation found that need satisfaction outside work predicted better recovery overall, but it did not make up for the specific needs that work left unmet. The lab is now testing that question day by day in a daily diary study, and extending it to work design, management practices, and recovery in demanding jobs.
Current questions
- Which features of work design best support autonomy, competence, and connection?
- How do need satisfaction and motivation relate to burnout in high-demand occupations?
Need satisfaction at work and away from it
A two-study project on whether meeting psychological needs outside work can make up for needs unmet at work. Study 1, a time-lagged survey, is being revised. Study 2, a daily diary study built with lab students, is in preparation for ethics review.
04
Training and workplace learning
What keeps attention on the task when learning happens on a screen and nobody is watching?
A growing share of workplace learning is self-paced and online, where distraction is the default. We study how attention, mindfulness, and self-regulation shape what people take away from training, and what helps people adapt under pressure. Earlier work showed that off-task attention is especially costly in self-directed learning. The lab builds on a published model of “mindful training” and tests its limits: when, and for whom, these approaches improve learning.
Current questions
- Does mindfulness help training because of the practice itself, or because of the label?
- How can informal, on-the-job learning be supported and measured?
Practice, label, or placebo? Mindfulness in training
A 2×2 experiment in an air traffic control simulation that separates the effects of actual mindfulness practice from simply calling an activity “mindfulness.”
Do women’s records catch up?
World-record histories in 28 events, modeled as catch-up curves to show how fast the performance gap closed and where it levels off.
Read the analysis →