Count data, representing the number of occurrences of an event within a specific time or space, are prevalent across various disciplines, including ecology, epidemiology, economics, and social sciences. Unlike continuous data, count data are non-negative integers and often exhibit characteristics such as overdispersion (variance exceeding the mean) or zero-inflation (an excess of zero counts). Consequently, standard regression models designed for continuous outcomes are often inappropriate for analyzing count data. In this talk, we introduce the fundamental principles of count data analysis, highlighting the Poisson and negative binomial regression models as primary analytical tools. It discusses the assumptions underlying these models, methods for assessing model fit, and strategies for addressing common issues like overdispersion and zero-inflation through extensions such as the zero-inflated Poisson and zero-inflated negative binomial models. Furthermore, this talk emphasizes the importance of appropriate model selection and interpretation in drawing meaningful inferences from count data. By employing suitable statistical techniques, researchers can effectively model and understand the factors influencing the frequency of events in diverse real-world applications.
Zoon link: https://uoc-gr.zoom.us/j/88659969718?pwd=g6bjYPDCuUQo1bzVxjjbgQL4xFN1f3.1
