Untangling serially dependent underreported count data for gender‐based violence

Abstract

Underreporting in gender-based violence data is a worldwide problem leading to the underestimation of the magnitude of this social and public health concern. This problem deteriorates the data quality, providing poor and biased results that lead society to misunderstand the actual scope of this domestic violence issue. The present work proposes time series models for underreported counts based on a latent integer autoregressive of order 1 time series with Poisson distributed innovations and a latent underreporting binary state, that is, a first-order Markov chain. Relevant theoretical properties of the models are derived, and the moment-based and maximum-based methods are presented for parameter estimation. The new time series models are applied to the quarterly complaints of domestic violence against women recorded in some judicial districts of Galicia (Spain) between 2007 and 2017. The models allow quantifying the degree of underreporting. A comprehensive discussion is presented, studying how the frequency and intensity of underreporting in this public health concern are related to some interesting socioeconomic and health indicators of the provinces of Galicia (Spain). KEYWORDSinteger autoregressive models, intimate partner violence, public health, state-dependent underreporting, underrecorded data 4404

Safety of journalists in Europe: Threats and best practices to tackle them

“…Moreover, one of the main issues, which was also encountered during the MPM2020 data collection, is that data on GBV against women journalists and its impact is not systematically collected. The efforts Robert Schuman Centre for Advanced Studies Working Papers to collect this data are weakened by the problem of underreporting (when official statistics do not represent true values due to being unreported by survivors of violence) (Fernández-Fontelo et al 2019;Gracia 2004;Palermo, Bleck, and Peterman 2014). The International Press Institute's report examining newsrooms best practices for addressing violence against journalists in the online environment also provided examples of female journalists who decided not to report incidents to their supervisors because they feared a lack of understanding and potential reputational harm.…”

Estimating the real burden of disease under a pandemic situation: The SARS-CoV2 case

“…However, Fernández-Fontelo et.at. in [12] introduced a version of the under-reporting scheme according to…”

“…In [11] it is proposed a simple model for integer-valued time series data that estimates the under-reporting of the human papillomavirus infections in the province of Girona from 2010 to 2014, the number of deaths attributable to a rare, aggressive tumour (pleural and peritoneal mesotheliomas) in Great Britain from 1968 to 2013, and the number of botulism cases in Canada from 1970 to 2013. The model mentioned above was extended by considering a more complex correlation structure among the time series observations in [12], where the authors studied the number of real cases of gender violence in Galicia from 2007 and 2017. The adequacy of the models in [11,12] was assessed through simulations of different scenarios that were well recovered by the estimation procedure, and as for real data, the results coincide with the expert's opinion.…”

“…The model mentioned above was extended by considering a more complex correlation structure among the time series observations in [12], where the authors studied the number of real cases of gender violence in Galicia from 2007 and 2017. The adequacy of the models in [11,12] was assessed through simulations of different scenarios that were well recovered by the estimation procedure, and as for real data, the results coincide with the expert's opinion.…”

Response to the letter of “Under‐reported data analysis with INAR‐hidden Markov chains”

“…A natural and convenient generalization arises when relaxing the assumption of independence between the states of under‐reporting. In this case, Fernández‐Fontelo et al propose a new model that considers dependence between the states of under‐reporting through a two‐state Markov chain. In this way, it is not necessary to modify the latent INAR(1) model in order to obtain more complex correlation structures, fulfilling identifiability as well.…”