Online Program

Sample Design and Multi-Method Sample Allocation
Fawzi Al Nassir, SRA International, Inc. 
Owen Hung, SRA International, Inc. 
Kimberly Renee Hylton, SRA International, inc. 
Dorothy Lucille Kester Jackman, SRA International 
*David McGrath, Defense Manpower Data Center (DMDC) 
Kimberly Williams,  
Kristin Williams, Defense Manpower Data Center 

Keywords: Power allocation, proportional allocation, industry, state, stratified sampling

The purpose of the research is to produce estimates by state and by industry based on the North American Industry Classification System (NAICS). All states and DC were included. For CEI, industry was defined by NAICS and grouped based on the distribution of Reservists in the industry resulting in 15 industry groups. Government agencies were classified into Federal and Local. The sample design was stratified sampling without replacement. Strata were defined by the crossing of 51 state and 15 industry groups, resulting in 765 strata. The sample was allocated in two stages. To assure proper representation of each state since the number of establishments greatly differs among states, Power Allocation at 0.05 power was used. Proportional Allocation was used to allocate the industry group sample within states. The Non-CEI sample was similarly allocated resulting in 102 strata of two industry groups and 51 states. An Excel based sampling tool was developed to produce the final sample allocation. The final sample was 59k EST for CEI and 20K for Non-CEI. Practical and theoretical justifications of combining allocation methods are the emphasis of this article.