QUADRAT SAMPLING AND THE ESTIMATION OF NEYMAN TYPE A AND THOMAS DISTRIBUTIONAL PARAMETERS1

Abstract

Summary Computer simulation is employed to investigate the effects of quadrat size, sampling and estimation method, and cluster spread, density and size when estimating Neyman Type A and Thomas distributional parameters from quadrat samples. Results obtained indicate that quadrat size and cluster spread affect the estimates. Use of first order standard errors in selecting suitable quadrat sizes is discussed. Where clusters have non-zero spread, the usual situation in practice, cluster density is generally overestimated and cluster size underestimated, which suggests the need for development of modified estimation procedures.