Article

Quasi-Poisson vs. negative binomial regression: how should we model overdispersed count data?

National Marine Mammal Laboratory, Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way NE, Building 4, Seattle, Washington 98115-6349, USA.
Ecology (Impact Factor: 5.18). 12/2007; 88(11):2766-72. DOI:10.1890/07-0043.1
Source: PubMed

ABSTRACT Quasi-Poisson and negative binomial regression models have equal numbers of parameters, and either could be used for overdispersed count data. While they often give similar results, there can be striking differences in estimating the effects of covariates. We explain when and why such differences occur. The variance of a quasi-Poisson model is a linear function of the mean while the variance of a negative binomial model is a quadratic function of the mean. These variance relationships affect the weights in the iteratively weighted least-squares algorithm of fitting models to data. Because the variance is a function of the mean, large and small counts get weighted differently in quasi-Poisson and negative binomial regression. We provide an example using harbor seal counts from aerial surveys. These counts are affected by date, time of day, and time relative to low tide. We present results on a data set that showed a dramatic difference on estimating abundance of harbor seals when using quasi-Poisson vs. negative binomial regression. This difference is described and explained in light of the different weighting used in each regression method. A general understanding of weighting can help ecologists choose between these two methods.

0 0
 · 
2 Bookmarks
 · 
599 Views
  • Source
    [show abstract] [hide abstract]
    ABSTRACT: High deer populations threaten the conservation value of woodlands and grasslands, but predicting the success of deer culling, in terms of allowing vegetation to recover, is difficult. Numerical simulation modeling is one approach to gain insight into the outcomes of management scenarios. We develop a spatially explicit model to predict the responses of Betula spp. to red deer (Cervus elaphus) and land management in the Scottish Highlands. Our model integrates a Bayesian stochastic stage-based matrix model within the framework of a widely used individual-based forest simulation model, using data collected along spatial and temporal gradients in deer browsing. By initializing our model with the historical spatial locations of trees, we find that densities of juvenile trees (<3 m tall) predicted after 9-13 years closely match counts observed in the field. This is among the first tests of the accuracy of a dynamical simulation model for predicting the responses of tree regeneration to herbivores. We then test the relative importance of deer browsing, ground cover vegetation, and seed availability in facilitating landscape-level birch regeneration using simulations in which we varied these three variables. We find that deer primarily control transitions of birch to taller (>3 m) height tiers over 30 years, but regeneration also requires suitable ground cover for seedling establishment. Densities of adult seed sources did not influence regeneration, nor did an active management scenario where we altered the spatial configuration of adults by creating "woodland islets". Our results show that managers interested in maximizing tree regeneration cannot simply reduce deer densities but must also improve ground cover for seedling establishment, and the model we develop now enables managers to quantify explicitly how much both these factors need to be altered. More broadly, our findings emphasize the need for land managers to consider the impacts of large herbivores rather than their densities.
    Ecology and Evolution 07/2013; 3(7):1890-901. · 1.18 Impact Factor
  • Source
    [show abstract] [hide abstract]
    ABSTRACT: In October 2010, cholera importation in Haiti triggered an epidemic that rapidly proved to be the world's largest epidemic of the seventh cholera pandemic. To establish effective control and elimination policies, strategies rely on the analysis of cholera dynamics. In this report, we describe the spatio-temporal dynamics of cholera and the associated environmental factors. Cholera-associated morbidity and mortality data were prospectively collected at the commune level according to the World Health Organization standard definition. Attack and mortality rates were estimated and mapped to assess epidemic clusters and trends. The relationships between environmental factors were assessed at the commune level using multivariate analysis. The global attack and mortality rates were 488.9 cases/10,000 inhabitants and 6.24 deaths/10,000 inhabitants, respectively. Attack rates displayed a significantly high level of spatial heterogeneity (varying from 64.7 to 3070.9 per 10,000 inhabitants), thereby suggesting disparate outbreak processes. The epidemic course exhibited two principal outbreaks. The first outbreak (October 16, 2010-January 30, 2011) displayed a centrifugal spread of a damping wave that suddenly emerged from Mirebalais. The second outbreak began at the end of May 2011, concomitant with the onset of the rainy season, and displayed a highly fragmented epidemic pattern. Environmental factors (river and rice fields: p<0.003) played a role in disease dynamics exclusively during the early phases of the epidemic. Our findings demonstrate that the epidemic is still evolving, with a changing transmission pattern as time passes. Such an evolution could have hardly been anticipated, especially in a country struck by cholera for the first time. These results argue for the need for control measures involving intense efforts in rapid and exhaustive case tracking.
    PLoS Neglected Tropical Diseases 04/2013; 7(4):e2145. · 4.57 Impact Factor
  • [show abstract] [hide abstract]
    ABSTRACT: The current financial crisis is having a major impact on European economies, especially that of Spain. Past evidence suggests that adverse macro-economic conditions exacerbate mental illness, but evidence from the current crisis is limited. This study analyses the association between the financial crisis and suicide rates in Spain. An interrupted time-series analysis of national suicides data between 2005 and 2010 was used to establish whether there has been any deviation in the underlying trend in suicide rates associated with the financial crisis. Segmented regression with a seasonally adjusted quasi-Poisson model was used for the analysis. Stratified analyses were performed to establish whether the effect of the crisis on suicides varied by region, sex and age group. The mean monthly suicide rate in Spain during the study period was 0.61 per 100 000 with an underlying trend of a 0.3% decrease per month. We found an 8.0% increase in the suicide rate above this underlying trend since the financial crisis (95% CI: 1.009-1.156; P = 0.03); this was robust to sensitivity analysis. A control analysis showed no change in deaths from accidental falls associated with the crisis. Stratified analyses suggested that the association between the crisis and suicide rates is greatest in the Mediterranean and Northern areas, in males and amongst those of working age. The financial crisis in Spain has been associated with a relative increase in suicides. Males and those of working age may be at particular risk of suicide associated with the crisis and may benefit from targeted interventions.
    The European Journal of Public Health 06/2013; · 2.52 Impact Factor

Full-text

View
25 Downloads
Available from

Jay M. Ver Hoef