Loss of resilience in population numbers in response to environmental perturbations may be predicted with statistical metrics called early warning signals (EWS) that are derived from abundance time series. These signals, however, have been shown to have limited success, leading to the development of trait‐based EWS that are based on information collected from phenotypic traits such as body size. Experimental work assessing the efficacy of EWS under varying ecological and environmental factors are rare. In addition, disentangling how such warning signals are affected under varying ecological and environmental factors is key to their application in biological conservation. Here, we experimentally test how different rates of environmental forcing (i.e. warming) and varying ecological factors (i.e. habitat quality and phenotypic diversity) affected population stability and predictive power of early warning signals of population collapse. We analyzed population density and body size time series data from three phenotypically different populations of a protozoan ciliate Askenasia volvox in two levels of habitat quality subjected to three different treatments of warming (i.e. no warming, fast warming and slow warming). We then evaluated how well abundance‐ and trait‐based EWS predicted population collapses under different levels of phenotypic diversity, habitat quality and warming treatments. Our results suggest that habitat quality and warming treatments had more profound effects than phenotypic diversity had on both population stability and on the performance of abundance‐based signals of population collapse. In addition, trait‐based EWS generally performed well, were reliable and more robust in forecasting population collapse than abundance‐based EWS, regardless of variation in environmental and ecological factors. Our study points towards the development of a predictive framework that includes information from phenotypic traits such as body size as an indicator of loss of resilience of ecological systems in response to environmental perturbations.