Time Series Analysis for Physiological and Endocrinological Data: A Practical Guide

The study of endocrinology provides insights into the upstream drivers of behavior and physiology of wild and captive populations to both natural and anthropogenic stressors. Most studies of wildlife endocrinology rely on single samples from multiple individuals to understand differences between different demographic states or populations. However, neither stressors nor hormone secretion is static, fluctuating over time within and between individuals. The use of nontraditional sample types that accumulate hormone concentrations through time, or repeated sampling of individuals through time can elucidate natural hormone fluctuations and the influence of stressors over time. With studies containing a temporal component come numerous challenges when interpreting and analyzing the data due to the inherent correlation between data points. Time series analysis is a group of statistical methods that were developed to analyze data collected with a temporal component. Here, we present a review of classic time series analysis methods, describing how they can be used with endocrinology data while providing worked examples and R code. By integrating time series analysis into endocrinology studies, researchers can get a better understanding of the temporal and individual variation in specific hormones. Through highlighting these tools and how they can be applied, we hope that they can be more readily available to all members of the endocrinology field to further understand wildlife endocrinology.