Ratio Type Transformed Estimators for Estimating Population Mean Using Models for Count Data with Applications to COVID-19 Incidence

Ratio type estimators’ use in count data analysis using the traditional regression coefficient based on the ordinary least square regression may not be appropriate so another method should be recruited. The number of COVID-19 cases are count data, which is the mainstay of this paper’s application as the COVID-19 pandemic has affected people’s life and way of living around the world. Estimating COVID-19 incidence can benefit in policy planning and preparing for a new wave of the virus. In this study, new ratio type estimators by transforming an auxiliary variable for estimating population mean using count data models; Poisson regression and negative binomial regression models are used to invent the alternative estimators under simple random sampling without replacement. The bias and mean square error are investigated. Simulation studies and applications to COVID-19 incidence assessed the performance of the estimators. The results revealed that the proposed estimators gave a higher percentage relative efficiency compared to the existing estimators which gave the most highest percentage relative efficiency. The proposed transformed estimators are more powerful in practice where they have more percentage relative efficiency more than the existing ones in the applications to COVID-19 incidence.