Modeling health survey data with excessive zero and <i>K</i> responses

Ting Hsiang Lin, Min‐Hsiao Tsai

Statistics in Medicine · 2012 · 26 citations · 12 references

Concepts

TL;DR

Zero‑inflated Poisson regression is widely used for data with many zeros, but existing models rely on data‑specific features and struggle when respondents cluster on identical responses, creating peaks. The authors introduce a flexible mixture model combining multinomial logistic and Poisson regression to capture excessive counts beyond zeros, including a specified integer K. The model uses a Poisson component for counts and a multinomial logistic component for excess zeros, K, and other values, demonstrated with examples involving frequent ones and sixes. The proposed zero‑ and K‑inflated models outperform standard zero‑inflated Poisson and ordinary Poisson regressions in fit.

Abstract

Zero-inflated Poisson regression is a popular tool used to analyze data with excessive zeros. Although much work has already been performed to fit zero-inflated data, most models heavily depend on special features of the individual data. To be specific, this means that there is a sizable group of respondents who endorse the same answers making the data have peaks. In this paper, we propose a new model with the flexibility to model excessive counts other than zero, and the model is a mixture of multinomial logistic and Poisson regression, in which the multinomial logistic component models the occurrence of excessive counts, including zeros, K (where K is a positive integer) and all other values. The Poisson regression component models the counts that are assumed to follow a Poisson distribution. Two examples are provided to illustrate our models when the data have counts containing many ones and sixes. As a result, the zero-inflated and K-inflated models exhibit a better fit than the zero-inflated Poisson and standard Poisson regressions.

References

12