Comparison of ratio-synthetic, sample-size dependent and EBLUP estimators as estimators of food-animal productivity parameters.

A comparison was made of three small-area sampling methods [two traditional design-based methods (ratio-synthetic, sample-size dependent) and one model-based method (EBLUP)] in estimation of some cow and sow population productivity parameters. Performance was evaluated in estimating both farm-specific mean responses and mean animal response over all farms using sample sizes of 100 and 25. Differences in results obtained with the cow and sow data are discussed in terms of the impact of sample size and population size on sampling method. There was a tendency for the model-based method to be the best performer in situations most likely to be operational when the sampling is done as part of a food-animal monitoring scheme. The situations are identified where the sample-size-dependent method performed best.