Zero-Altered and other Regression Models for Count Data with Added Zeros

By D. Heilbron. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Dan Hall
2. Hui Liu
3. Felix Famoye
4. Alan Agresti
5. Yongyi Min
6. Maria Iannario
7. Jay M. Ver Hoef
8. Zheng-Yu Zhang

Citations in the corpus

• A Bayesian model for repeated measures zero-inflated count data with application to outpatient psychiatric service use (2010, neutral): “Heilbron (1994) developed a ‘zero-altered’ hurdle model—also called a ‘compatible’ model—to test for zero inflation.”
• Growth Curve Models for Zero-Inflated Count Data: An Application to Smoking Behavior (2007, background): “As a result, various methodological approaches have been developed to deal with detecting and modeling overdispersion in count data (Cameron & Trivedi, 1986, 1990; Dean & Lawless, 1989; Greene, 1994; Hausman et al., 1984; Heilbron, 1994; King, 1989; Lambert, 1992; Mullahy, 1997).”
• Bayesian Analysis of Zero-Inflated Distributions (2003, methodology): “282 Rodrigues This type of model has been used for count data with excess of zeros in many practical problems like manufacturing defects (Lambert, 1992), counting disease lesions on plants (Rideout et al., 1998) and sexual behavior (Heilbron, 1994).”
• Modelling Uncertainty and Overdispersion in Ordinal Data (2014, background): “Otherwise, if atypical categories are present, as modelled by zero-inflated (Lambert, 1992; Johnson et al., 1992) or hurdle models (Heilbron, 1994; Mullahy, 1986; King, 1998) for count data, for instance, the extra-variability may be well accounted by inserting a shelter effect (Iannario, 2012a).”
• Two alternative evaluation metrics to replace the true skill statistic in the assessment of species distribution models (2019, background): “Whereas both true and false absences can lead to ‘zero-inflated’ datasets (Heilbron 1994) that violate statistical assumptions, the latter are also a source of uncertainty in parameter estimates as artefactual signals (e.g. sampling bias, probability of detection) confounding estimates of…”
• Random effect models for repeated measures of zero-inflated count data (2005, methodology): “In the univariate response case, Heilbron (1994) defined a type of hurdle model as a compatible two-part model.”
• Zero‐Inflated Poisson and Binomial Regression with Random Effects: A Case Study (2000, methodology): “Mixtures of the Poisson distribution with zero have been considered by several authors, including Yip (1988), who described a number of insects per leaf application of such a mixture, and Lambert (1992) and Heilbron (1989, 1994), who presented regression models for counts based on mixtures of zero a”
• Gender Differences in Factors Contributing to Institutional Misconduct (2008, methodology): “Following the guidance provided by Heilbron (1994), models were estimated for the current study using zero-inflated Poisson regression (ZIP) to address the presence of overdispersion in the officially recorded incidences of inmate infractions and to account for the occurrence of excess zero counts…”
• Zero-Inflated Generalized Poisson Regression Model with an Application to Domestic Violence Data (2021, methodology): “Other models in the literature include the hurdle model (Mullahy, 1986), the two-part model (Heibron, 1994), and the semi-parametric model (Gurmu, 1997).”
• What does a zero mean? Understanding false, random and structural zeros in ecology (2019, background): “…exhibit two characteristics: first, a variance significantly larger than the mean, known as overdis‐ persion (Bliss & Fisher, 1953); and second, an excess of zero values in comparison to those expected from a classical count probability distribution, known as zero inflation (Heilbron, 1994).”
• Models for count data with many zeros (1998, methodology): “Heilbron (1994) and Mullahy (1986) discuss other compatible models.”
• Functional Form and Heterogeneity in Models for Count Data (2007, background): “The second is the conventional ZIP model proposed by Lambert (1992), Heilbron (1994) and Greene (1994).”
• Models for count data with endogenous participation (2009, methodology): “The second is the conventional ZIP model proposed by Lambert (1992), Heilbron(1994) andGreene(1994).ThePoissonmodelisnotnestedintheZIP model; there is no parametric restriction on the ZIP model that produces the Poisson specification.”