Many behavioral traits appear to be complex, with an assumed distribution in genetic liability underlying known and unknown environmental influences. Quantitative genetic methods enable more effective selection on single or multiple complex traits than selection on phenotypes because of the estimation of additive genetic variance in the population (the chief cause of resemblance of relatives and the only genetic component that can be estimated from observations on the population) and estimated breeding values of individuals. Estimation of genetic correlations between multiple traits reveals the effects selection on 1 trait will have on the breeding values of others and inform of unfavorable genetic relationships between objective traits. Improvement in multiple traits can be optimized via the use of selection indices, making the best use of all phenotypic information available to achieve specific selection objectives. Therefore, quantitative methods offer effective means of improving the accuracy of selection of behavioral traits. Quantitative genetic methods use phenotypic data on a large and representative proportion of the population, which have in livestock species typically taken the form of empirical measurement of continuous variables. Obtaining behavioral phenotypes presents challenges inherent in quantifying behaviors, but certain steps can be taken in the design of phenotypic protocols to ensure that phenotypic data are conducive to quantitative genetic analysis.