Modelling age variation in survival and reporting rates for recovery models

In this paper, we focus on models for recovery data from birds ringed as young. In some cases, it is important to be able to include in these models a degree of age variation in the reporting probability. For certain models this has been found, empirically, to result in completely flat likelihood surfaces, due to parameter redundancy. These models cannot then be fitted to the data, to produce unique parameter estimates. However, empirical evidence also exists that other models with such age variation can be fitted to data by maximum likelihood. Using the approach of Catchpole and Morgan (1996b), we can now identify which models in this area are parameter-redundant, and which are not. Models which are not parameter-redundant may still perform poorly in practice, and this is investigated through examples, involving both real and simulated data. The Akaike Information Criterion is found to select inappropriate models in a number of instances. The paper ends with guidelines for fitting models to data from birds ringed as young, when age dependence is expected in the reporting probability.     

A model of the ring-recovery reporting process for the Cape Griffon Gyps coprotheres

By their very nature, statistical models are constructed on the basis of a number of simplifying assumptions. It is usual to assume that all the individuals in a 'group' or 'cohort' have similar survival, recovery or reporting probabilities. From a number of earlier studies of the Cape Griffon Gyps coprotheres in southern Africa, it is clear that there have been many violations of these assumptions of homogeneity. To get a better understanding of the process whereby a dead ringed bird is found and reported, an analysis was undertaken of 575 recoveries from 7130 individuals ringed as nestlings. From a series of univariate generalized linear models, it was found that the proportion of ringed birds reported dead varied with the following factors: (1) ring prefix (representing different grades and thicknesses of aluminium): there was considerable variation in reporting rate between cohorts fitted with different ring prefix series used; (2) metal type: birds fitted with monel metal rings were reported at a rate twice that of those bearing aluminium rings; (3) colour rings: recoveries of birds with colour rings were much more likely to be reported than birds with only a metal ring; (4) epoch: the reporting rate has increased steadily from the 1950s through to the mid-1980s. All of these factors are confounded and so a number of multivariate generalized linear models were constructed. It was found that the variations in the cohort-specific reporting rate could be described well by a model including factors for metal-ring type and the presence or absence of colour rings. Using the tougher monel metal ring along with a set of colour rings more than doubles the reporting rate and their continued use is strongly recommended for future studies. The year-to-year variations could be accounted for by this model but the colony of ringing did not enter the model. The models used were based on two assumptions: (i) the reporting rate was constant for all individuals within a given cohort and (ii) the recoveries were complete. It is argued that the results are congruent with these assumptions. There is now a clearer model of the manner in which the ring-recovery reporting process proceeds and this has opened the way to building a more realistic statistical model to estimate survival in the Cape Griffon.     

A model of the ring-recovery reporting process for the Cape Griffon Gyps coprotheres

By their very nature, statistical models are constructed on the basis of a number of simplifying assumptions. It is usual to assume that all the individuals in a 'group' or 'cohort' have similar survival, recovery or reporting probabilities. From a number of earlier studies of the Cape Griffon Gyps coprotheres in southern Africa, it is clear that there have been many violations of these assumptions of homogeneity. To get a better understanding of the process whereby a dead ringed bird is found and reported, an analysis was undertaken of 575 recoveries from 7130 individuals ringed as nestlings. From a series of univariate generalized linear models, it was found that the proportion of ringed birds reported dead varied with the following factors: (1) ring prefix (representing different grades and thicknesses of aluminium): there was considerable variation in reporting rate between cohorts fitted with different ring prefix series used; (2) metal type: birds fitted with monel metal rings were reported at a rate twice that of those bearing aluminium rings; (3) colour rings: recoveries of birds with colour rings were much more likely to be reported than birds with only a metal ring; (4) epoch: the reporting rate has increased steadily from the 1950s through to the mid-1980s. All of these factors are confounded and so a number of multivariate generalized linear models were constructed. It was found that the variations in the cohort-specific reporting rate could be described well by a model including factors for metal-ring type and the presence or absence of colour rings. Using the tougher monel metal ring along with a set of colour rings more than doubles the reporting rate and their continued use is strongly recommended for future studies. The year-to-year variations could be accounted for by this model but the colony of ringing did not enter the model. The models used were based on two assumptions: (i) the reporting rate was constant for all individuals within a given cohort and (ii) the recoveries were complete. It is argued that the results are congruent with these assumptions. There is now a clearer model of the manner in which the ring-recovery reporting process proceeds and this has opened the way to building a more realistic statistical model to estimate survival in the Cape Griffon.     

Bias‐reduced marginal Akaike information criteria based on a Monte Carlo method for linear mixed‐effects models

In linear mixed‐effects (LME) models, if a fitted model has more random‐effect terms than the true model, a regularity condition required in the asymptotic theory may not hold. In such cases, the marginal Akaike information criterion (AIC) is positively biased for (?2) times the expected log‐likelihood. The asymptotic bias of the maximum log‐likelihood as an estimator of the expected log‐likelihood is evaluated for LME models with balanced design in the context of parameter‐constrained models. Moreover, bias‐reduced marginal AICs for LME models based on a Monte Carlo method are proposed. The performance of the proposed criteria is compared with existing criteria by using example data and by a simulation study. It was found that the bias of the proposed criteria was smaller than that of the existing marginal AIC when a larger model was fitted and that the probability of choosing a smaller model incorrectly was decreased.     

Estimating survival rates from recoveries of birds ringed as young: A case study

Models for analysis of survival rates from recoveries of birds ringed as young were evaluated for lesser snow geese ringed on the breeding grounds at La Perouse Bay, Manitoba from 1970 to 1992. Analyses of birds ringed as adults and young indicate recovery rates were lower for young than for older birds and varied independently in the two age classes. Adult survival increased over the study, while that of immatures decreased. Using only recoveries of birds ringed as young, models incorporating observed variation in recovery rates were not identifiable. Many sets of constraints could be selected to make the model identifiable, but there was no objective way to distinguish among them, and each could lead to different conclusions. The observed changes in survival rates could not be estimated. Previous analyses have shown that models for birds ringed as young can be used to test even complex variation in survival rates, provided that recovery rates do not vary with age. Unfortunately, as shown by the geese, this assumption may often be violated and can only be tested with additional data, preferably from birds recaptured or ringed as sub-adults or adults.     

Estimating survival rates from recoveries of birds ringed as young: a case study

Models for analysis of survival rates from recoveries of birds ringed as young were evaluated for lesser snow geese ringed on the breeding grounds at La Perouse Bay, Manitoba from 1970 to 1992. Analyses of birds ringed as adults and young indicate recovery rates were lower for young than for older birds and varied independently in the two age classes. Adult survival increased over the study, while that of immatures decreased. Using only recoveries of birds ringed as young, models incorporating observed variation in recovery rates were not identifiable. Many sets of constraints could be selected to make the model identifiable, but there was no objective way to distinguish among them, and each could lead to different conclusions. The observed changes in survival rates could not be estimated. Previous analyses have shown that models for birds ringed as young can be used to test even complex variation in survival rates, provided that recovery rates do not vary with age. Unfortunately, as shown by the geese, this assumption may often be violated and can only be tested with additional data, preferably from birds recaptured or ringed as sub-adults or adults.     

Bivariate Hofmann distributions

The aim of this paper is to develop some bivariate generalizations of the Hofmann distribution. The Hofmann distribution is known to give nice fits for overdispersed data sets. Two bivariate models are proposed. Recursive formulae are given for the evaluation of the probability function. Moments, conditional distributions and marginal distributions are studied. Two data sets are fitted based on the proposed models. Parameters are estimated by maximum likelihood.     

Modelling the duration of spells of withdrawal from a female dominated labour force

It is of particular importance to nurse manpower planners to determine the likelihood and time-scale of return to service of nurses who have temporarily withdrawn from the labour force. Such behaviour is common in female dominated professions and is referred to in the nursing literature as being in ‘limbo’. In this paper we develop non-parametric, parametric and network models to describe the distribution of stay in ‘limbo’ and estimate the probability of return to service. These models are fitted to data from the Northern Ireland nursing service. It is inherent in the problem that the data are statistically incomplete. This must be taken into account in the statistical methods used to estimate the various model parameters.     

Maximum likelihood estimators in finite mixture models with censored data

The consistency of estimators in finite mixture models has been discussed under the topology of the quotient space obtained by collapsing the true parameter set into a single point. In this paper, we extend the results of Cheng and Liu (2001) to give conditions under which the maximum likelihood estimator (MLE) is strongly consistent in such a sense in finite mixture models with censored data. We also show that the fitted model tends to the true model under a weak condition as the sample size tends to infinity.     

Estimating age-specific survival rates of tawny owls--recaptures versus recoveries

We compared estimates of annual survival rates of tawny owls ( Strix aluco ) ringed in southern Finland from several different sampling methods: recoveries of birds ringed as young; recaptures of birds ringed as young; recoveries of birds ringed as adults as well as young; combined recoveries and recaptures of birds ringed as young, and combined recoveries and recaptures of birds ringed as adults and young. From 1979 to 1998, 18 040 young owls were ringed, of which 983 were recaptured as breeders in subsequent years during this period, and 1764 were recovered dead at various locations. In addition, 1751 owls were ringed as adults, of which 612 were later recaptured and 199 were recovered dead. First-year survival rates estimated using only recoveries of birds ringed as young averaged 48%, while apparent survival rates estimated using only recaptures from birds ringed as young averaged 10-13%. Use of combined recapture-recovery models, or supplementary information from recoveries of birds ringed as adults, produced survival estimates of 30-37%. Survival estimates from young-recoveries-only models were biased high, because of violation of the assumption of constant recovery rates with age: birds dying in their first-year were one-third less likely to be found and reported than older birds. In contrast, recaptures-only models confounded emigration with mortality. Despite these differences in mean values, annual fluctuations in estimated first-year survival rates were similar with all models. Estimates of adult survival rates were similar with all models, while those for second-year birds were similar for all models except recaptures-only. These results highlight the potential biases associated with analysing either recaptures or recoveries alone of birds ringed as young, and the benefits of using combined data.     

Understanding and Addressing the Unbounded “Likelihood” Problem

The joint probability density function, evaluated at the observed data, is commonly used as the likelihood function to compute maximum likelihood estimates. For some models, however, there exist paths in the parameter space along which this density-approximation likelihood goes to infinity and maximum likelihood estimation breaks down. In all applications, however, observed data are really discrete due to the round-off or grouping error of measurements. The “correct likelihood” based on interval censoring can eliminate the problem of an unbounded likelihood. This article categorizes the models leading to unbounded likelihoods into three groups and illustrates the density-approximation breakdown with specific examples. Although it is usually possible to infer how given data were rounded, when this is not possible, one must choose the width for interval censoring, so we study the effect of the round-off on estimation. We also give sufficient conditions for the joint density to provide the same maximum likelihood estimate as the correct likelihood, as the round-off error goes to zero.     

Tests for symmetry of two-piece normal distribution

In this paper, classical optimum tests for symmetry of two-piece normal distribution is derived. Uniformly most powerful one-sided test for the skewness parameter is obtained when the location and scale parameters are known and is compared with sequential probability ratio test. An ad-hoc test for symmetry and likelihood ratio test for symmetry for large samples, can be found in literature for this distribution. But in this paper, we derive exact likelihood ratio test for symmetry, when location parameter is known. The exact power of the test is evaluated for different sample sizes.     

Estimating multiple-membership logit models with mixed effects: indirect inference versus data cloning

Multiple-membership logit models with random effects are models for clustered binary data, where each statistical unit can belong to more than one group. The likelihood function of these models is analytically intractable. We propose two different approaches for parameter estimation: indirect inference and data cloning (DC). The former is a non-likelihood-based method which uses an auxiliary model to select reasonable estimates. We propose an auxiliary model with the same dimension of parameter space as the target model, which is particularly convenient to reach good estimates very fast. The latter method computes maximum likelihood estimates through the posterior distribution of an adequate Bayesian model, fitted to cloned data. We implement a DC algorithm specifically for multiple-membership models. A Monte Carlo experiment compares the two methods on simulated data. For further comparison, we also report Bayesian posterior mean and Integrated Nested Laplace Approximation hybrid DC estimates. Simulations show a negligible loss of efficiency for the indirect inference estimator, compensated by a relevant computational gain. The approaches are then illustrated with two real examples on matched paired data.     

A class of observation-driven random coefficient INAR(1) processes based on negative binomial thinning

Integer-valued time series models and their applications have attracted a lot of attention over the last years. In this paper, we introduce a class of observation-driven random coefficient integer-valued autoregressive processes based on negative binomial thinning, where the autoregressive parameter depends on the observed values of the previous moment. Basic probability and statistics properties of the process are established. The unknown parameters are estimated by the conditional least squares and empirical likelihood methods. Specially, we consider three aspects of the empirical likelihood method: maximum empirical likelihood estimate, confidence region and EL test. The performance of the two estimation methods is compared through simulation studies. Finally, an application to a real data example is provided.     

Test for age-specificity in survival of the common tern

Much effort in life-history theory has been addressed to the dependence of life-history traits on age, especially the phenomenon of senescence and its evolution. Although senescent declines in survival are well documented in humans and in domestic and laboratory animals, evidence for their occurrence and importance in wild animal species remains limited and equivocal. Several recent papers have suggested that methodological issues may contribute to this problem, and have encouraged investigators to improve sampling designs and to analyse their data using recently developed approaches to modelling of capture-mark-recapture data. Here we report on a three-year, two-site, mark-recapture study of known-aged common terns (Sterna hirundo) in the north-eastern USA. The study was nested within a long-term ecological study in which large numbers of chicks had been banded in each year for > 25 years. We used a range of models to test the hypothesis of an influence of age on survival probability. We also tested for a possible influence of sex on survival. The cross-sectional design of the study (one year's parameter estimates) avoided the possible confounding of effects of age and time. The study was conducted at a time when one of the study sites was being colonized and numbers were increasing rapidly. We detected two-way movements between the sites and estimated movement probabilities in the year for which they could be modelled. We also obtained limited data on emigration from our study area to more distant sites. We found no evidence that survival depended on either sex or age, except that survival was lower among the youngest birds (ages 2-3 years). Despite the large number of birds included in the study (1599 known-aged birds, 2367 total), confidence limits on estimates of survival probability were wide, especially for the oldest age-classes, so that a slight decline in survival late in life could not have been detected. In addition, the cross-sectional design of this study meant that a decline in survival probability within individuals (actuarial senescence) could have been masked by heterogeneity in survival probability among individuals (mortality selection). This emphasizes the need for the development of modelling tools permitting separation of these two phenomena, valid under field conditions in which the recapture probabilities are less than one.     

Smooth Semi‐nonparametric Analysis for Mixture Cure Models and Its Application to Breast Cancer

Mixture cure models are widely used when a proportion of patients are cured. The proportional hazards mixture cure model and the accelerated failure time mixture cure model are the most popular models in practice. Usually the expectation–maximisation (EM) algorithm is applied to both models for parameter estimation. Bootstrap methods are used for variance estimation. In this paper we propose a smooth semi‐nonparametric (SNP) approach in which maximum likelihood is applied directly to mixture cure models for parameter estimation. The variance can be estimated by the inverse of the second derivative of the SNP likelihood. A comprehensive simulation study indicates good performance of the proposed method. We investigate stage effects in breast cancer by applying the proposed method to breast cancer data from the South Carolina Cancer Registry.     

Modelling postfledging survival and age-specific breeding probabilities in species with delayed maturity: A case study of Roseate Terns at Falkner Island,Connecticut

We modelled postfledging survival and age-specific breeding probabilities in endangered Roseate Terns ( Sterna dougallii ) at Falkner Island, Connecticut, USA using capture-recapture data from 1988-1998 of birds ringed as chicks and as adults. While no individuals bred as 2-year-olds during this period, about three-quarters of the young that survived and returned as 3-year-olds nested, and virtually all surviving birds had begun breeding by the time they reached 5 years of age. We found no evidence of temporal variation age of first breeding of birds from different cohorts. There was significant temporal variation in the annual survival of adults and the survival over the typical 3-year maturation period of prebreeding birds, with extremely low values for both groups from the 1991 breeding season. The estimated overwinter survival rate (0.62) for adults from 1991-1992 was about three-quarters the usual rate of about 0.83, but the low survival of fledglings from 1991 resulted in less than 25% of the otherwise expected number of young from that cohort returning as breeding birds; this suggests that fledglings suffered a greater proportional decrease in survival than did adults. The survival estimates of young from 1989 and 1990 show that these cohorts were not negatively influenced by the events that decimated the young from 1991, and the young from 1992 and 1993 had above-average survival estimates. The apparent decrease since 1996 in development of fidelity of new recruits to this site is suspected to be due mainly to nocturnal disturbance and predation of chicks causing low productivity.     

The Bernstein–von Mises theorem in semiparametric competing risks models

Semiparametric Bayesian models are nowadays a popular tool in event history analysis. An important area of research concerns the investigation of frequentist properties of posterior inference. In this paper, we propose novel semiparametric Bayesian models for the analysis of competing risks data and investigate the Bernstein–von Mises theorem for differentiable functionals of model parameters. The model is specified by expressing the cause-specific hazard as the product of the conditional probability of a failure type and the overall hazard rate. We take the conditional probability as a smooth function of time and leave the cumulative overall hazard unspecified. A prior distribution is defined on the joint parameter space, which includes a beta process prior for the cumulative overall hazard. We first develop the large-sample properties of maximum likelihood estimators by giving simple sufficient conditions for them to hold. Then, we show that, under the chosen priors, the posterior distribution for any differentiable functional of interest is asymptotically equivalent to the sampling distribution derived from maximum likelihood estimation. A simulation study is provided to illustrate the coverage properties of credible intervals on cumulative incidence functions.     

Random effects regression models for count data with excess zeros in caries research

We extend the family of Poisson and negative binomial models to derive the joint distribution of clustered count outcomes with extra zeros. Two random effects models are formulated. The first model assumes a shared random effects term between the conditional probability of perfect zeros and the conditional mean of the imperfect state. The second formulation relaxes the shared random effects assumption by relating the conditional probability of perfect zeros and the conditional mean of the imperfect state to two different but correlated random effects variables. Under the conditional independence and the missing data at random assumption, a direct optimization of the marginal likelihood and an EM algorithm are proposed to fit the proposed models. Our proposed models are fitted to dental caries counts of children under the age of six in the city of Detroit.     

Focused information criteria for copulas

In this paper, we extend the focused information criterion (FIC) to copula models. Copulas are often used for applications where the joint tail behavior of the variables is of particular interest, and selecting a copula that captures this well is then essential. Traditional model selection methods such as the Akaike information criterion (AIC) and the Bayesian information criterion (BIC) aim at finding the overall best‐fitting model, which is not necessarily the one best suited for the application at hand. The FIC, on the other hand, evaluates and ranks candidate models based on the precision of their point estimates of a context‐given focus parameter. This could be any quantity of particular interest, for example, the mean, a correlation, conditional probabilities, or measures of tail dependence. We derive FIC formulae for the maximum likelihood estimator, the two‐stage maximum likelihood estimator, and the so‐called pseudo‐maximum‐likelihood (PML) estimator combined with parametric margins. Furthermore, we confirm the validity of the AIC formula for the PML estimator combined with parametric margins. To study the numerical behavior of FIC, we have carried out a simulation study, and we have also analyzed a multivariate data set pertaining to abalones. The results from the study show that the FIC successfully ranks candidate models in terms of their performance, defined as how well they estimate the focus parameter. In terms of estimation precision, FIC clearly outperforms AIC, especially when the focus parameter relates to only a specific part of the model, such as the conditional upper‐tail probability.