On the Estimation of Population Size from a Dependent Triple-Record System

Abstract

Summary Population size estimation based on a capture–recapture experiment under a triple‐record system is an interesting problem in various fields including epidemiology and population studies. In many real life scenarios, there is inherent dependence between capture and recapture attempts. We propose a novel model that successfully incorporates the possible dependence and the associated parameters have nice interpretations. We provide estimation methodology for the population size and the associated model parameters based on the maximum likelihood method. The model proposed is applied to analyse real data sets from public health and census coverage evaluation studies. The performance of the estimate proposed is evaluated through extensive simulation study and the results are compared with existing competitors. The results exhibit superiority of the model over the existing competitors both in real data analysis and in a simulation study.

Estimating prevalence of post-war health disorders using multiple systems data

“…Also, the lack of homogeneity in capture probabilities over the individuals induces positive dependence among the lists. Note that the presence of list-dependence causes correlation bias in the estimate of the population size obtained by assuming independence among the lists 21 , 22 . Therefore, modeling TRS data becomes more challenging if individual heterogeneity is present along with list-dependence caused by behavioural response variation.…”

Estimation of Population Size with Heterogeneous Catchability and Behavioural Dependence: Applications to Air and Water Borne Disease Surveillance

“…In this context, an ecological model M tb is frequently used to account for the time and behavioural response variation (Chao et al, 2000). This model has several practical limitations (Chatterjee and Bhuyan, 2020a). See Subsection 4.2 for more details.…”

“…In this section, we first discuss the trivariate Bernoulli model (TBM) proposed by Chatterjee and Bhuyan (2020a) for modeling capture-recapture data under TRS. The TBM is used to incorporate the inherent list dependence among the capture and recapture attempts and its parameters possess physical interpretation.…”

“…This model is useful for TRS when the individuals are equally catchable in the each list. As discussed in Chatterjee and Bhuyan (2020a), the TBM can be further reduced to submodels TBM-1 and TBM-2 if we consider α 3 = 0 and α 4 = 0, respectively. In the absence of list dependence (i.e.…”

“…where x 0 = i,j,k:ijk =000 x ijk and M l = u 1 + u 2 + ... + u l−1 denote the total number of distinct captured individuals and the number of individuals captured at least once prior to the l-th attempt, respectively (Chatterjee and Bhuyan, 2020a). In this context, Chao et al (2000) discussed various likelihood based methods for estimation of N .…”