Causal Reasoning from Longitudinal Data*

Abstract. This paper reviews some of the key statistical ideas that are encountered when trying to find empirical support to causal interpretations and conclusions, by applying statistical methods on experimental or observational longitudinal data. In such data, typically a collection of individuals are followed over time, then each one has registered a sequence of covariate measurements along with values of control variables that in the analysis are to be interpreted as causes, and finally the individual outcomes or responses are reported. Particular attention is given to the potentially important problem of confounding. We provide conditions under which, at least in principle, unconfounded estimation of the causal effects can be accomplished. Our approach for dealing with causal problems is entirely probabilistic, and we apply Bayesian ideas and techniques to deal with the corresponding statistical inference. In particular, we use the general framework of marked point processes for setting up the probability models, and consider posterior predictive distributions as providing the natural summary measures for assessing the causal effects. We also draw connections to relevant recent work in this area, notably to Judea Pearl's formulations based on graphical models and his calculus of so‐called do‐probabilities. Two examples illustrating different aspects of causal reasoning are discussed in detail.     

Conditional Likelihood Estimators for Hidden Markov Models and Stochastic Volatility Models

ABSTRACT.  This paper develops a new contrast process for parametric inference of general hidden Markov models, when the hidden chain has a non-compact state space. This contrast is based on the conditional likelihood approach, often used for ARCH-type models. We prove the strong consistency of the conditional likelihood estimators under appropriate conditions. The method is applied to the Kalman filter (for which this contrast and the exact likelihood lead to asymptotically equivalent estimators) and to the discretely observed stochastic volatility models.     

An empirical likelihood goodness-of-fit test for time series

Summary. Standard goodness-of-fit tests for a parametric regression model against a series of nonparametric alternatives are based on residuals arising from a fitted model. When a parametric regression model is compared with a nonparametric model, goodness-of-fit testing can be naturally approached by evaluating the likelihood of the parametric model within a nonparametric framework. We employ the empirical likelihood for an α -mixing process to formulate a test statistic that measures the goodness of fit of a parametric regression model. The technique is based on a comparison with kernel smoothing estimators. The empirical likelihood formulation of the test has two attractive features. One is its automatic consideration of the variation that is associated with the nonparametric fit due to empirical likelihood's ability to Studentize internally. The other is that the asymptotic distribution of the test statistic is free of unknown parameters, avoiding plug-in estimation. We apply the test to a discretized diffusion model which has recently been considered in financial market analysis.     

Estimation for Continuous Branching Processes

The maximum-likelihood estimator for the curved exponential family given by continuous branching processes with immigration is investigated. These processes originated from population biology but also model the dynamics of interest rates and development of the state of technology in economics. It is proved that in contrast to branching processes with discrete space and/or time the MLE gives a unified approach to the inference. In order to include singular subdomains of the parameter space we modify the MLE slightly. Consistency and asymptotic normality for the MLE are considered. Concerning the asymptotic theory of the experiments, all three properties LAQ, LAN, and LAMN occur for different submodels     

Asymptotic Theory of Outlier Detection Algorithms for Linear Time Series Regression Models

Outlier detection algorithms are intimately connected with robust statistics that down‐weight some observations to zero. We define a number of outlier detection algorithms related to the Huber‐skip and least trimmed squares estimators, including the one‐step Huber‐skip estimator and the forward search. Next, we review a recently developed asymptotic theory of these. Finally, we analyse the gauge, the fraction of wrongly detected outliers, for a number of outlier detection algorithms and establish an asymptotic normal and a Poisson theory for the gauge.     

On the asymptotic normality of the R-estimators of the slope parameters of simple linear regression models with associated errors

ABSTRACT

The purpose of this paper is to prove, under mild conditions, the asymptotic normality of the rank estimator of the slope parameter of a simple linear regression model with stationary associated errors. This result follows from a uniform linearity property for linear rank statistics that we establish under general conditions on the dependence of the errors. We prove also a tightness criterion for weighted empirical process constructed from associated triangular arrays. This criterion is needed for the proofs which are based on that of Koul [Behavior of robust estimators in the regression model with dependent errors. Ann Stat. 1977;5(4):681–699] and of Louhichi [Louhichi S. Weak convergence for empirical processes of associated sequences. Ann Inst Henri Poincaré Probabilités Statist. 2000;36(5):547–567].     

Asymptotic behaviour of proportions of observations in random regions determined by central order statistics from stationary processes

In this paper, we show that proportions of observations that fall into a random region determined by a given Borel set and a central order statistic converge almost surely, provided that the corresponding population quantile is unique. We also describe three types of possible asymptotic behaviour of these proportions in the case of non-unique population quantile. As an application of our findings we establish limiting properties of numbers of ties with a central order statistics in a discrete sample. Our results are derived not only for independent and identically distributed observations but more generally for strictly stationary and ergodic sequences of random variables.     

Policy advocacy for women's unpaid care work: comparing approaches and strategies in Nepal and Nigeria

This article focuses on policy advocacy programmes in Nepal and Nigeria, instigated by ActionAid International with local women's rights organisations and non-government organisations, and supported by the Institute of Development Studies, UK. These programmes aimed to challenge women's unequal responsibility for care work and to influence policymakers to understand the importance of providing services to support them. One of the main components of these programmes was child care, a responsibility which most women experience 24/7 for at least two decades of their lives, and which profoundly shapes their lives and opportunities. We examine the processes through which each of the country teams have engaged with public policies on this issue, focusing on the similarities and differences in each context. Although the programme of activities has been implemented in much the same way in both countries, and each chose to focus on early child-care provision as the main policy demand, the partnerships and policy processes chosen differ greatly. Specifically, we distinguish between ‘critical engagement’ (in the case of Nepal), as compared to ‘constructive engagement’ (in the case of Nigeria). The article ends with some reflections on the challenges facing the teams during their work, and the implications and lessons that can be drawn from these two case studies.     

Maternal Education and Investments in Children's Health

Maternal education differences in children's academic skills have been strongly linked to parental investment behaviors. This study extended this line of research to investigate whether these same maternal education patterns in parenting are observed among a set of parenting behaviors that are linked to young children's health. Drawing on data from the Early Childhood Longitudinal Study, Birth Cohort (n = 5,000) and longitudinal models incorporating random effects, the authors found that higher levels of maternal education were associated with more advantageous health investment behaviors at each phase of early development (9 months, 2 years, 4 years, 5 years). Moreover, these disparities were typically largest at the developmental stage when it was potentially most sensitive for children's long‐term health and development. These findings provide further evidence of a developmental gradient associated with mothers' education and new insight into the salience of mothers' education for the short‐ and long‐term health and well‐being of their children.     

A note on spectral estimation for ARMA processes

The standard frequency domain approximation to the Gaussian likelihood of a sample from an ARMA process is considered. The Newton-Raphson and Gauss-Newton numerical maximisation algorithms are evaluated for this approximate likelihood and the relationships between these algorithms and those of Akaike and Hannan explored. In particular it is shown that Hannan's method has certain computational advantages compared to the other spectral estimation methods considered