TY - JOUR

T1 - Analyzing spatially distributed binary data using independent-block estimating equations

AU - Oman, Samuel D.

AU - Landsman, Victoria

AU - Carmel, Yohay

AU - Kadmon, Ronen

PY - 2007/9

Y1 - 2007/9

N2 - We estimate the relation between binary responses and corresponding covariate vectors, both observed over a large spatial lattice. We assume a hierarchical generalized linear model with probit link function, partition the lattice into blocks, and adopt the working assumption of independence between the blocks to obtain an easily solved estimating equation. Standard errors are obtained using the "sandwich" estimator together with window subsampling (Sherman, 1996, Journal of the Royal Statistical Society, Series B 58, 509-523). We apply this to a large data set describing long-term vegetation growth, together with two other approximate-likelihood approaches: pairwise composite likelihood (CL) and estimation under a working assumption of independence. The independence and CL methods give similar point estimates and standard errors, while the independent-block approach gives considerably smaller standard errors, as well as more easily interpretable point estimates. We present numerical evidence suggesting this increased efficiency may hold more generally.

AB - We estimate the relation between binary responses and corresponding covariate vectors, both observed over a large spatial lattice. We assume a hierarchical generalized linear model with probit link function, partition the lattice into blocks, and adopt the working assumption of independence between the blocks to obtain an easily solved estimating equation. Standard errors are obtained using the "sandwich" estimator together with window subsampling (Sherman, 1996, Journal of the Royal Statistical Society, Series B 58, 509-523). We apply this to a large data set describing long-term vegetation growth, together with two other approximate-likelihood approaches: pairwise composite likelihood (CL) and estimation under a working assumption of independence. The independence and CL methods give similar point estimates and standard errors, while the independent-block approach gives considerably smaller standard errors, as well as more easily interpretable point estimates. We present numerical evidence suggesting this increased efficiency may hold more generally.

KW - Binary variables

KW - Composite likelihood

KW - Generalized estimating equations

KW - Probit regression

KW - Spatial dependence

UR - http://www.scopus.com/inward/record.url?scp=34247269589&partnerID=8YFLogxK

U2 - 10.1111/j.1541-0420.2007.00754.x

DO - 10.1111/j.1541-0420.2007.00754.x

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

C2 - 17489971

AN - SCOPUS:34247269589

SN - 0006-341X

VL - 63

SP - 892

EP - 900

JO - Biometrics

JF - Biometrics

IS - 3

ER -