Klaus Moeltner

Workshop: Bayesian Estimation Methods in Meta-Analysis

• Bayesian Basics
  • Klaus Moeltner's UNR course (RECO777)
  • Lecture Notes
  • Prior vs. Posterior (generic example)
     
• Topic 1: Small Samples and Informed Priors / Models and Model-Averaging
  This module illustrates the use of 'outside-data' information on model parameters to derive refined (informed) priors.
  These can then be combined with a small meta-data set to derive efficient posteriors and Benefit-Transfer predictions.
  The module also highlights the ability of the Bayesian framework to assign "model probabilities" to competing specifications, and their use in
  formulating model-averaged predictions.
   
  • Presentation Slides
  • Moeltner, K., R. Woodward (2009).
    Meta-Functional Benefit Transfer for Wetland Valuation: Making the Most of Small Samples. Environmental and Resource Economics, 42 (1), p. 89-109.
    (revised submission -  please download final paper from publisher's web site)
  • Matlab code (Word format)
  • Data (Excel, see Matlab code for variable list & labels)
     
• Topic 2: Regressor-deficient Meta-data/ Methodological Indicators and Predictions
  This module deals with the "N-vs.-K Dilemma", which arises when different sub-sets of meta-data have different sets of regressors.  Deficient sub-sets can be used to derive informed priors, which can then be combined with the remaining (complete) meta-data. The module also addresses the issue of how to deal with study-methodological indicators when generating Benefit-Transfer predictions
   
  • Presentation slides
  • Moeltner, K., K.J. Boyle, R. W. Paterson (2007). Meta-Analysis and Benefit-Transfer for Resource Valuation: Addressing Classical Challenges with Bayesian Modeling. Journal of Environmental Economics and Management, 53 (2), p.250 – 269.
    (revised submission -  please download final paper from publisher's web site)
  • Technical appendix
  • Matlab code (Word format)
  • Data (Excel, see Matlab code for variable list and labels)
     
• Topic 3: 'Optimal Scope' of Meta-regressions
  This module shows how a broader definition of the dependent variable in the meta-model can lead to more efficient BT predictions for a given policy context. Bayesian model search techniques are used to examine the trade-offs between increased sample sites and a proliferation of nuisance parameters in models with augmented data space.
   
  • Presentation slides
  • Moeltner K., R. S. Rosenberger (2008), Meta-Regression and Benefit Transfer: Data Space, Model Space, and the Quest for 'Optimal Scope'. B.E. Journal of Economic Analysis & Policy, 8 (1), Contributions. http://www.bepress.com/bejeap/vol8/iss1/art31/
  • Matlab code
  • Data (Excel, see Matlab code for variable list and labels)
     

Colloquium: Benefit Transfer from Multiple Contingent Experiments: A Flexible Two-Step Model Combining Individual Choice Data with Community Characteristics

• Presentation slides
• Linked files:
  • Brooklyn survey page
  • Attribute space comparison
  • Econometric issues with MRMs based on CE's and simulated data
    
    All other files linked to the PowerPoint presentation are posted at http://www.ag.unr.edu/moeltner/EPA_STAR.htm.
    To activate links, simply place all linked files into the same folder as the presentation slides.
    
     
• Moeltner K., R. Johnston, R. S. Rosenberger, J. M. Duke, (Forthcoming). Benefit Transfer from Multiple Contingent Experiments: A Flexible Two-Step Model Combining Individual Choice Data with Community Characteristics. American Journal of Agricultural Economics.
  • Technical Appendix
  • Matlab code