Overdispersion in the Poisson Regression Model

Palmer, Abraham A.; Losilla, Josep-María; Vives, Jaume; Jiménez, Rafael Zequeira

doi:10.1027/1614-2241.3.3.89

Cited by 16 publications

(6 citation statements)

References 33 publications

(49 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We did not observe ants on sprouts wider than 7 cm ( S1 Table ) [ 98 ]. The data from plants 40c-B and 89 were outliers (observations carried out in early morning), so were removed from the PMRM analysis [ 99 ]. Type III test of fixed model in PMRM showed that the inclusion of the cladode width was negatively associated with the number of ants on sprouts ( Table 1 ), that is, ants visited narrower cladodes more frequently.…”

Section: Resultsmentioning

confidence: 99%

First description of extrafloral nectaries in Opuntia robusta (Cactaceae): Anatomy and ultrastructure

et al. 2018

View full text Add to dashboard Cite

To our knowledge, there are no studies about the structure and ecological function of extrafloral nectaries (EFNs) in Opuntia robusta. This is the first description of EFNs in O. robusta, where young spines have an interesting structure and a secreting function, which are different from EFNs described in other Cactaceae species. We used light, scanning-electron, and transmission-electron microscopy to examine morphology, anatomy, and ultrastructure of the secretory spines in areoles in female and hermaphrodite individuals of O. robusta. Young cladodes develop areoles with modified and secretory spines as EFNs only active during the early growth phase. EFNs are non-vascularized structures, with no stomata, that consist of a basal meristematic tissue, a middle elongation region, and an apical secretory cone formed by large globular epidermal cells, containing nectar and medullar elongated cells. We observed the presence of Golgi apparatus, vesicles and plastids in the medullar and sup-epidermal cells of the spine. We propose that the nectar is stored in the globular cells at the apex of the spine and secreted by breaking through the globular cells or by pores. We recorded a more frequent presence of ants on younger cladode sprouts producing young secreting spines: this result is parallel with the predictions of Optimal Defense Hypothesis, which states that younger plant organs should be better defended than older ones because their loss produces a higher fitness impairment. Although Diaz-Castelazo’s hypothesis states that a more complex structure of EFNs correlates with their lower among-organs dispersion, comparing to less complex EFNs, non-vascularized structure of EFNs in O. robusta is not associated with their higher among-organs dispersion likened to O. stricta, which produces vascularized EFNs. We provide evidence that this characteristic is not a good taxonomic feature of Opuntia genus. Moreover, the comparison of EFNs of O. robusta and O. stricta suggests that the hypothesis of Diaz-Castelazo should be revised: it is rather a rule but not a law.

show abstract

Section: Resultsmentioning

confidence: 99%

First description of extrafloral nectaries in Opuntia robusta (Cactaceae): Anatomy and ultrastructure

et al. 2018

View full text Add to dashboard Cite

show abstract

“…We performed the overdispersion test proposed by Cameron and Trivedi [2009, p. 575] and failed to reject the null hypothesis of no significant over dispersion (p value = 0.307). However, we estimate the regression using heteroskedasticity-consistent robust standard error to account for potential heterogeneity (arising, for example, from differences among the sizes of the observations, although most of the independent variables are measured in logarithmic form) that could be an additional source of overdispersion [Palmer et al, 2007].…”

Section: Econometric Methodsmentioning

confidence: 99%

Forest cover, socioeconomics, and reported flood frequency in developing countries

Ferreira¹,

Ghimire²

2012

Water Resources Research

View full text Add to dashboard Cite

[1] In this paper, we analyze the determinants of the number of large floods reported since 1990. Using the same sample of countries as Bradshaw et al. (2007), and, like them, omitting socioeconomic characteristics from the analysis, we found that a reduction in natural forest cover is associated with an increase in the reported count of large floods. This result does not hold in any of three new analyses we perform. First, we expand the sample to include all the developing countries and all countries for which data were available but were omitted in their study. Second, and more importantly, since forest management is just one possible channel through which humans can influence reported flood frequency, we account for other important human-flood interactions. People are typically responsible for deforestation, but they are also responsible for other land use changes (e.g., urbanization), for floodplain and flood emergency management, and for reporting the floods. Thus, in our analysis we account for population, urban population growth, income, and corruption. Third, we exploit the panel nature of the data to control for unobserved country and time heterogeneity. We conclude that not only is the link between forest cover and reported flood frequency at the country level not robust, it also seems to be driven by sample selection and omitted variable bias. The human impact on the reported frequency of large floods at the country level is not through deforestation.

show abstract

“…Consequently, the accuracy of parameter and standard error estimates are undermined (e.g. Palmer et al, 2007). Although alternatives such as the quasi-likelihood methods have been proposed to tackle these problems, they only serve as partial solutions, and their finite sample performance can be unsatisfactory (Nelder and Lee, 1992).…”

Section: Introductionmentioning

confidence: 99%