Post-2000 Estimates Planning (2KEP) How will we make estimates after the 2000 Census?
Form A Committee FSCPE Members –Ken Darga, Dean Judson, Melanie Martindale, Mohammed Shahidullah, Donna Tolson, Greg Williams Census Staff –Bill Schooling, Sam Davis, Greg Harper, Barbara van der Vate, Signe Wetrogan, Jim Fitzsimmons Other Members –Patty Becker, Meyer Zitter, Joe Salvo, Ed Spar
Hold Meetings July Ground Rules June Presentations, Discussion August Criteria, Scoring
Ground Rule Highlights State will produce and deliver estimates on a continuous, timely basis State will supply input data and documentation State method must be replicable State totals will come from Census Bureau Census Bureau will maintain estimates to be used as a “fallback”
Call for Proposals Early 2000 Received Proposals from 12 States Funding Requested from 8 States
States Submitting Estimates Alaska California Colorado Illinois Iowa Virginia North Carolina Oregon
Difference Measures Accuracy –MALPE –MAPE –MEAN LOSS (Webster’s Rule) Outliers –Top 5% Absolute Percent Error –Top 5% Webster’s Rule Loss Function
More Decision Aids Significance Tests Quality Scores
Results Some Methods did not do as well as PEB Some came out about the same Some really came out better A variety...
Results Not as well as PEB (Iowa, Colorado (some), Illinois (some), Oregon (some) Came out about the same - Illinois (some), Oregon,(Some), North Carolina Some really came out better - California, Virginia
Implications of Using Substitute Estimates Method Inconsistent Among States –Like the 1980s Results More Difficult to Explain May Not Be Able to Report Components
Component Alternatives No Components - (Boo!!!) Show Births, Deaths, “Residual” –(Like in the 1980s) Show Births, Deaths, IRS-Derived Domestic Migration, International Migration –(“But they don’t add up” - Boo Hoo) –Throw all difference into residual
Plans Reconvene 2KEP Committee in Winter –Look at Results in Depth Make Recommendations on What to Do