Area based stratified random sampling using geospatial technology in a community-based survey

BMC Public Health

Table 2 Proposed stratified sampling based on Hispanic household proportion

Strata	% Hispanic Households	Eligible Blocks	Selected Blocks	Surveys per Block	Surveys per stratum ^a	Surveys per Stratum by Hispanic Origin ^b
Strata	% Hispanic Households	Eligible Blocks	Selected Blocks	Surveys per Block	Surveys per stratum ^a	Hispanic	Non-Hispanic
Stratum 1	≤10%	268	10	6	60	5	55
Stratum 2	10–30%	101	10	6	60	20	40
Stratum 3	30–50%	54	12	4	48	36	12
Stratum 4	> 50%	32	12	4	48	48	0
Total		455^c	44		216	109	107

^aDerived by multiplying selected blocks and surveys per block
^bFor strata 1 and 2, distribution of Hispanic and non-Hispanic surveys within each block roughly equate to household proportions. Oversampling of Hispanics was planned a priori to reach recruitment goals thus proportions of Hispanic surveys in strata 3 and 4 were set higher than actual Hispanic household proportions
^cTotal of 647 blocks in Albertvile, AL with 192 blocks with zero household, leaving n = 455 eligible blocks for sampling

ISSN: 1471-2458