Appendix R-National Agricultural Statistics Service (NASS) Review

Appendix R

NASS REview

Food and Nutrition Service

AG-3198-D-10-0079: Revised OMB Package (Task 8) - Evaluation of Medicare Extra Help

Generally, the survey data will be analyzed by comparing survey data between the pilot and comparison sites.  The statistical model will use the two-tailed t-test, and the sample design targets an overall 80% response rate.  Table B2.1 shows the comparison groups with the expected number of completed interviews and the minimum detectable differences from the two-tailed t-tests.  As described in Part A, the survey data will also be analyzed targeting specific subgroups on certain types of survey questions.  Caution should be taken in sample subgroups that show low response rates, for the degree of accuracy with the descriptive and inferential statistics may be affected.  Possible subgroups include counties or regions with low response, or certain groups of the elderly who have special needs and cannot participate in over-the-phone interviews.

Thanks,

Tom Pordugal

Statistical Methods Branch

Commodity Section

USDA / NASS / SD / SMB

Phone: (202) 690-3623