The paper presents a new method for evaluating resource preference by comparing usage and availability data. Traditional methods often rely on arbitrary decisions about what is available to an animal, leading to inconsistent conclusions. The proposed method uses ranks of usage and availability to determine preference, providing comparable results whether or not questionable components are included. This method allows for statistical tests of significance and resolves inconsistencies in previous approaches. The paper also introduces a hierarchical model of selection processes, which helps resolve inconsistencies in studies of selection. This hierarchy includes first-order selection (species range), second-order selection (home range), third-order selection (habitat use within home range), and fourth-order selection (food procurement at a feeding site). This model helps clarify the relationship between different levels of selection and provides a framework for understanding preference and avoidance. The method is illustrated with an example involving wetland usage and availability data for mallards. The results show that the inclusion or exclusion of certain components can significantly affect conclusions about preference. The proposed method uses ranks to avoid absolute statements about preference and provides a more reliable way to compare preferences among components. The method is statistically robust, as it uses ranks rather than exact measurements, which are often subject to error. This approach is particularly useful when availability measurements are uncertain. The method also allows for significance tests and can be applied to real data, as demonstrated by the analysis of mallard habitat use. The paper concludes that the proposed method provides a more accurate and consistent way to evaluate resource preference by comparing usage and availability data. It emphasizes the importance of considering the hierarchical nature of selection and the potential for bias in traditional methods. The method is recommended for use in ecological studies where accurate assessment of resource preference is needed.The paper presents a new method for evaluating resource preference by comparing usage and availability data. Traditional methods often rely on arbitrary decisions about what is available to an animal, leading to inconsistent conclusions. The proposed method uses ranks of usage and availability to determine preference, providing comparable results whether or not questionable components are included. This method allows for statistical tests of significance and resolves inconsistencies in previous approaches. The paper also introduces a hierarchical model of selection processes, which helps resolve inconsistencies in studies of selection. This hierarchy includes first-order selection (species range), second-order selection (home range), third-order selection (habitat use within home range), and fourth-order selection (food procurement at a feeding site). This model helps clarify the relationship between different levels of selection and provides a framework for understanding preference and avoidance. The method is illustrated with an example involving wetland usage and availability data for mallards. The results show that the inclusion or exclusion of certain components can significantly affect conclusions about preference. The proposed method uses ranks to avoid absolute statements about preference and provides a more reliable way to compare preferences among components. The method is statistically robust, as it uses ranks rather than exact measurements, which are often subject to error. This approach is particularly useful when availability measurements are uncertain. The method also allows for significance tests and can be applied to real data, as demonstrated by the analysis of mallard habitat use. The paper concludes that the proposed method provides a more accurate and consistent way to evaluate resource preference by comparing usage and availability data. It emphasizes the importance of considering the hierarchical nature of selection and the potential for bias in traditional methods. The method is recommended for use in ecological studies where accurate assessment of resource preference is needed.