I’m looking for a consultant to support the implementation of a multi-level Bayesian model with spatial structure, using Stan or brms.
Project summary:
This is part of an applied research project with a hierarchical data structure and geo-coded locations. The goal is to estimate organization- and location-level heterogeneity while adjusting for spatial correlation — across multiple outcome types.
Key modeling features:
- Data structure with multiple levels of nesting (decisions within interviews, interviews within locations, locations within organizations; multiple organizations may operate within the same spatial clusters)
- Locations are spatially clustered (with nearby but distinct points)
- Each interview contains several standardized decision tasks
- Covariates at the organization, interview, and task levels
- Outcomes include binary, ordinal (e.g., Likert-style), and continuous variables
- Interest in estimating both between- and within-organization heterogeneity in decisions (adjusting for spatial effects and other covariates), as well as modeling beliefs about other organizations’ decisions — including potential bias and additional variance relative to the other organization’s own responses
If this sounds like a fit and you’ve worked on similar models (e.g., Bayesian hierarchical modeling, spatial GPs, or CAR/ICAR priors), feel free to just send a quick note or link to your work.
Timeline
Data collection begins soon and will take about a month. I’m looking to engage someone early to plan the model structure, and then support estimation, diagnostics (e.g., R-hat, ESS, divergences), and posterior predictive checks – with flexibility as the analysis evolves.
Compensation:
This is a paid consulting role. Budget can be discussed based on experience and scope. Happy to talk through feasibility depending on fit and availability.
Tools:
Stan or brms preferred, but open to discussion.
Posting anonymously for now, as the project is part of a job market paper where sole authorship is important and work is still in development. Happy to share more context privately.
Feel free to reply here, DM me, or email: stan.for.anonymous@gmail.com