A SIMPLE METHOD TO INCREASE PRECISION OR REDUCE EXPENSE IN REGRESSION EXPERIMENTS TO PREDICT THE PROPORTION OF LEAN MEAT OF CARCASSES

By a combination of complete and incomplete dissections a considerable increase of precision or reduction of expense may be obtained in the construction of a prediction formula for the proportion of lean meat of a carcass from objective carcass measurements. This approach, referred to as double-regression, is demonstrated on data involving dissection of pig carcasses in The Netherlands. Advantages of double-regression relative to ordinary regression with respect to precision and experimental cost are discussed. Expressions for optimal sample sizes at fixed total cost are derived.