|
|
Small Area Income & Poverty EstimatesModel-based Estimates for States, Counties, & School Districts |
|
Cautions about comparisons between intercensal model-based county estimates and direct survey estimates
SAIPE model-based estimates of poverty and median household income use Census 2000 and ACS direct survey estimates in their construction. Consequently, there is an almost certain positive correlation between the Census 2000 estimates, ACS direct survey estimates, and their intercensal model-based estimates, which should be represented in confidence intervals for their difference. Failure to do so will result in too many differences being considered "not significant". We currently do not have estimates of the individual correlations or advice on a general magnitude that can be assumed.
This caution does not apply to state-level comparisons between Census 2000 because the Census 2000 estimates have close to negligible sampling error.
Cautions about comparisons between SAIPE model-based estimates for different years
Comparisons between 2005 SAIPE estimates (based on ACS data) and SAIPE estimates for previous years (based on CPS ASEC data) are not advised because there is a break in the time-series due to the switch from CPS ASEC to ACS data. In future years, comparisons at the state-level estimates will be possible.
Comparisons of pre-2005 state-level estimates are possible; and 2005 estimates will be comparable with estimates produced in the future. For the time being, one can compare 2005 and 2006 ACS direct state estimates. In making these state-level comparisons, the greater precision of the direct state estimates results in a larger ratio of model-error variance to sample-error variance in the state model than that observed in the county model. Consequently, ignoring the model-error correlation across years is an unacceptable assumption and no crude approximation is given to make such comparisons. Methodology for Testing for a Rise in Child Poverty Rate [PDF 141k] describes a methodology for estimating if any states have a five percent or greater significant change in child poverty rate between two years.
There is currently not an accepted method for comparing model-based county estimates because there is a smaller ratio of model-error variance to sample-error variance.
Cautions about comparisons between model-based estimates for different states or different counties in the same year
All SAIPE model-based estimates are correlated because they depend on the same regression coefficients. Therefore, to make comparisons between states or counties, one cannot simply take the variances (implied by the confidence intervals) for the two different places and apply the usual means difference hypothesis test. However, a method for comparing counties is being developed, similar to the method currently used for comparing states. The difficultly of implementing this method at the county-levelis that the sampling and model errors are probably correlated across counties within a state.
[PDF] or 
denotes a file in Adobe’s Portable Document Format.
To view the file, you will need the
Adobe® Acrobat® Reader
available free from Adobe.