What is the Office of Evaluation Sciences?
OES is a team of applied experts based at the General Services Administration (GSA) that combines academic expertise with experience implementing and evaluating evidence-based program improvements. The private sector quickly learns what does not work, what works, and what works most cost-effectively - government should, too. OES supports GSA’s Office of Government-wide Policy’s mission to “use policies, evidence, and analysis to help agencies drive efficiency, savings, and improved mission performance.”
What does OES do?
OES recruits top talent in diverse scientific fields such as economics, psychology, public health, and statistics from academic institutions and non-profit organizations into government to complete a one-year, in-person Fellowship. OES also accepts subject matter experts from other Federal agencies on detail to build capacity governmentwide in designing and implementing rapid, low-cost evaluations.
OES work directly with agency collaborators across government to:
- Use existing administrative data to measure outcomes that matter
- Design operationally feasible, no- or low-cost changes to improve agency outcomes using proven insights
- Integrate rigorous evaluations within current program implementation and agency constraints, to learn quickly what works and doesn’t, and what is cost-effective to improve agency outcomes
- Provide technical services and capacity building on how to implement and analyze rigorous evaluations
- Disseminate results and learnings to government and the public
What does OES look for in a potential project?
In general, a project may be a good fit if it contains:
- A clear touchpoint (e.g., a communication channel, a signup point, etc.) between the Federal agency program and an individual
- An outcome of interest depends in part on people’s actions (e.g., individuals taking up a program, making a payment, etc.)
- An outcome of interest that is reflected in data that is currently collected (or could be easily collected) by the Federal agency (relevant administrative data sets exist)
- A program population size that is large enough to be statistically and policy relevant
- The ability to assign groups of people to different versions of an intervention in order to compare outcomes and learn what works best (e.g., half of program recipients receive information presented in a standard way, and half in a modified, evidence-based improvement format)
- An agency collaborator willing to work alongside OES and to share results across government