This article introduces an information-theoretic approach for analyzing ecological data based on Kullback–Leibler information, which extends likelihood theory and avoids the limitations of null hypothesis testing. The approach emphasizes a deliberate focus on prior scientific knowledge in developing multiple working hypotheses or models. These models can then be ranked and scaled based on their likelihood given the data. The method also incorporates model selection uncertainty into estimates of precision, making it a powerful tool for valid inference. The paper discusses the limitations of traditional frequentist statistics and the challenges of Bayesian approaches in ecology. It highlights the importance of parsimony, multiple working hypotheses, and the strength of evidence in scientific inference. The Kullback–Leibler information is introduced as a fundamental concept in information theory, and its relationship to maximum likelihood is explored. Akaike's Information Criterion (AIC) is presented as a practical method for model selection and inference, allowing for the ranking of models based on their relative likelihood. The paper also discusses the use of AIC in estimating sampling variance and the importance of model averaging in ecological studies. It emphasizes the need to distinguish between inferences based on prior considerations and those resulting from data dredging. The article provides an example of applying these methods to study age- and sex-dependent rates of tag loss in elephant seals, demonstrating the effectiveness of the information-theoretic approach in ecological research. The paper concludes by emphasizing the importance of using information-theoretic methods in ecological studies, which provide a rational alternative to traditional frequentist and Bayesian approaches. It advocates for the use of these methods in the analysis of ecological data, whether experimental or observational, and highlights the need for careful consideration of model selection and inference in scientific research.This article introduces an information-theoretic approach for analyzing ecological data based on Kullback–Leibler information, which extends likelihood theory and avoids the limitations of null hypothesis testing. The approach emphasizes a deliberate focus on prior scientific knowledge in developing multiple working hypotheses or models. These models can then be ranked and scaled based on their likelihood given the data. The method also incorporates model selection uncertainty into estimates of precision, making it a powerful tool for valid inference. The paper discusses the limitations of traditional frequentist statistics and the challenges of Bayesian approaches in ecology. It highlights the importance of parsimony, multiple working hypotheses, and the strength of evidence in scientific inference. The Kullback–Leibler information is introduced as a fundamental concept in information theory, and its relationship to maximum likelihood is explored. Akaike's Information Criterion (AIC) is presented as a practical method for model selection and inference, allowing for the ranking of models based on their relative likelihood. The paper also discusses the use of AIC in estimating sampling variance and the importance of model averaging in ecological studies. It emphasizes the need to distinguish between inferences based on prior considerations and those resulting from data dredging. The article provides an example of applying these methods to study age- and sex-dependent rates of tag loss in elephant seals, demonstrating the effectiveness of the information-theoretic approach in ecological research. The paper concludes by emphasizing the importance of using information-theoretic methods in ecological studies, which provide a rational alternative to traditional frequentist and Bayesian approaches. It advocates for the use of these methods in the analysis of ecological data, whether experimental or observational, and highlights the need for careful consideration of model selection and inference in scientific research.