Exploring the Fascinating World of Raking Algorithms in Surveys
Delve into the intriguing realm of raking algorithms used in surveys, from their historical origins to practical applications. Discover how these algorithms align samples to control totals, reduce variance, and correct biases. Explore convergence practices, convergence improvements, and detailed examples illustrating the algorithm's utility.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
E N D
Presentation Transcript
Rakes Progress Revisited Nada Ganesh and Fritz Scheuren WSS September 24, 2014
Snacks and Knacks About the Talk Title? Application History First! Information Theory too? Raking s Roles in Surveys? Some Examples (Handout)? More Snacks and Knacks? Wraps and Conjectures!
Why Call Algorithm Raking? Maybe a Lofty Origin? Perhaps a 1951 opera by Stravinsky inspired by 1733-5 engravings by Hogarth Or maybe a Humbler Origin! The Dairy Farm Rake? What is Your Guess?
Basic Raking Algorithm Repeated Ratio-ing r of a (Weighted) sample total t to a control total T as in r=T/t Total by Total repeatedly until all the adjusted sample totals t were as close as wanted to T
Application History First! The Pre-1940 Period The 1940 Deming-Stephan AMS Paper to align Survey/Census Not Widely used in 1940 Census? Because of WWII! Regularly Forgotten/Recovered
Information Theory too? Kullback Kully and Ireland Terry s Proof of Convergence From Calculus a Bounded Monotonic Series Converges Information Theory Affirms this When Constraints are Consistent
Rakings Roles in Surveys? Aligning to Agreed upon Control totals aka Calibration -- Variance reduction -- Bias Correction MSE? Hand-waving but Maybe harmful Sometimes because unit based
More Snacks and Knacks? Convergence Practices and Possible Improvements Variable Bounds on Dimensions Iterating Worst Dimensions First Early/Differential Stopping Rules Family Harmonic Adjustments
More Application Settings Chip Alexander and CPS/ACS Survey and Administrative Linked Files (CPS/IRS/SSA) File Alignment before Raking? Reducing Impact on Linkage Errors (Maybe both Variance/Bias)
Two Examples Two Examples Geometric Family Weight Example (Ganish) Changing Convergence Algorithm to Quicken Speed (Ganish) Role for Synthetic Estimation (Phil?)
Wraps and Conjectures! A Multi-Use Tool, Like Pliers Computational Barriers long since gone More work on Diagnostics needed though More work on multiple weights
Many, Many, Many Thanks
