Applied Work
I have spent much of my career conducting research outside conventional academic settings.
Applied problems rarely arrive as cleanly specified research questions. Measures may be imperfect, populations heterogeneous, data generated for purposes other than the research at hand, and relevant processes only partially observable. Product, organizational, field, or client constraints may limit which comparisons are possible. Decisions also have deadlines.
The research task is therefore often diagnostic before it is analytical: clarify the decision, identify what must be measured, distinguish plausible explanations, determine what the available evidence can separate, and decide whether the question requires additional data, a different comparison, an experiment, or a new measurement system.
My applied work has spanned technology, media, public health, organizational research, public affairs, and commissioned research. Across those settings, recurring problems have involved measurement, experimentation, network and organizational analysis, computational methods, and the design of research infrastructure.
Meta
Senior Behavioral Scientist · 2019–2020
Platform-scale experimentation Networks Behavioral measurement
At Meta, I worked on research involving randomized experimentation, network modeling, behavioral measurement, integrity, civic participation, economic inequality, public-health behavior, and emerging product questions.
The work included research collaborations with the National Institutes of Health and the Centers for Disease Control and Prevention during the COVID-19 pandemic.
Platform scale changed what could be observed and tested, but it did not remove familiar research problems. Large datasets can still contain weak measures, ambiguous comparisons, heterogeneous populations, dependent observations, and multiple plausible mechanisms. A substantial part of the work involved determining which of those problems could be addressed with the available design and which remained unresolved.
Roku
Senior Behavioral Scientist · 2020–2023
Measurement infrastructure Streaming behavior Multi-source data
At Roku, I worked on behavioral measurement infrastructure and experimental, survey, product, audience, organizational, and platform research across streaming environments.
The work combined behavioral and product data, surveys, qualitative evidence, audience information, and other sources to study individual and population-level patterns across tens of millions of streaming households. I also developed analytical systems, measures, taxonomies, segmentation approaches, dashboards, and recurring research workflows for questions that could not be answered from a single data source.
A recurring problem was translating constructs that are reasonably clear in theory into measures that remain interpretable when behavior is observed indirectly through platform or organizational data. Measurement was therefore part of the substantive research problem rather than a preliminary step completed before analysis.
Iron Light
Head of Behavioral Science · 2023–2026
Field and survey experiments Audience research Applied inference
From 2023 through 2026, I led behavioral science, experimental design, measurement, and applied research infrastructure at Iron Light Public Affairs.
The work included field and survey experimentation, audience and message research, behavioral measurement, computational and mixed-method analysis, evidence synthesis, and the use of large language models in research, coding, and measurement workflows. I also built reusable analytical systems and research infrastructure intended to support repeated use across client engagements.
Public-affairs and advocacy research creates a demanding inferential environment. Outcomes are multiply determined, field conditions change, interventions occur alongside other events, and stakeholders often begin with strong explanations of what produced an observed response. The research problem is to determine which explanations the available evidence can distinguish, what alternative accounts remain viable, and whether the decision requires stronger measurement or another design.
Graystone Industries
Commissioned research and systems work
Custom research systems Modeling Experimentation
I conduct commissioned research and systems work through Graystone Industries.
Projects have included experimental design, behavioral and organizational research, measurement, social-network analysis, computational modeling, simulation, AI-system evaluation, evidence synthesis, analytical pipelines, and custom research software.
Some engagements begin with an empirical question that can be addressed using existing evidence. Others require building the measurement system, experimental environment, simulation, analytical workflow, or software tool through which the question can be studied. Methods and technical implementation are selected to fit the underlying problem, the available evidence, and the decision the research is intended to inform.
The work can therefore range from a focused analysis or experiment to the design of a reusable research system. Across those forms, I retain direct involvement in question formulation, research design, implementation, analysis, validation, interpretation, and communication.