Measurement error in regression analysis

Consider the linear regression model Y = Xθ+ ε where Y denotes a vector of n observations on the dependent variable, X is a known matrix, θ is a vector of parameters to be estimated and e is a random vector of uncorrelated errors. If X'X is nearly singular, that is if the smallest characteristic root of X'X s small then a small perurbation in the elements of X, such as due to measurement errors, induces considerable variation in the least squares estimate of θ. In this paper we examine for the asymptotic case when n is large the effect of perturbation with regard to the bias and mean squared error of the estimate.     

Implementing a Multidimensional Poverty Measure Using Mixed Methods and a Participatory Framework

Recently, there have been advances in the development of multidimensional poverty measures. Work is needed however on how to implement such measures. This paper deals with the process of selecting dimensions and setting weights in multidimensional poverty measurement using qualitative and quantitative methods in a participatory framework. We estimate the multidimensional poverty measures developed by Alkire and Foster for a particular group: persons with psychiatric diagnoses in the United States. To select relevant dimensions and their relative ordering, two discussion groups are convened: one consisting of persons with lived-experience expertise and the other consisting of people with mental health service provision or research expertise. Several methods are used to convert dimension rankings into weights. The selection and ordering of dimensions differed between the two discussion groups, as did the resulting poverty measures. For instance, the poverty headcount using the dimensions and weights of the ‘lived experience’ group ranged from 20.61 to 26.96% as compared to a range of 18.62–33.19% using those of the ‘provider/researcher’ group. One of the main results of this study is that the Alkire Foster method is sensitive to the selection of dimensions and the methods used to derive rankings and weights. It points toward the limitation of relying exclusively on small scale qualitative methods for the selection and ranking of dimensions. In addition, the participatory framework used in this study was found to be essential in interpreting results, in particular with respect to the limitations of the data set in measuring relevant dimensions.     

International-domestic linkages in a developing-country context: the case of the Rohingyas in Bangladesh

ABSTRACT

Since 1978, the Rohingya have been fleeing Myanmar and taking refuge in Bangladesh. The state of Bangladesh is not a signatory to the Geneva Convention and does not recognize refugee rights, but the initial experiences with the Rohingya refugee population led the government to create a temporary and ad hoc domestic policy advisory and refugee management system, which eventually became highly politicized. There was also some degree of slow “externalization” of policy advice through the involvement of international organizations from 2006–2007 onward, mainly through the participation of the United Nations High Commissioner for Refugees (UNHCR) and International Organization for Migration (IOM). Over 2017–2018, there was a massive influx of refugees from Myanmar to Bangladesh. The domestic advisory and refugee management system lacked the capacity to manage the crisis and had to quickly and greatly externalize policy advice and refugee management. The UNHCR and IOM came in with a host of international organizational networks and coordinated with each other and the state through a multi-sectoral approach to manage the crisis. This externalization led to the systematization and institutionalization of the state’s domestic advisory system. However the effect of externalization on politicization is equivocal; on the one hand it decreased politicization of the domestic policy advisory system, but on the other hand, it created new levels of politicization.     

Human Capital Attainment and Gender Empowerment: The Kerala Paradox*

Objectives. The State of Kerala in southwestern India shows the highest gender development index in the country. Despite having the highest literacy rate and educational achievements of women, the suicide rate among females is very high, and domestic violence committed against women is increasing every year in Kerala. This article attempts to analyze the demographic, social, and cultural changes that are occurring in order to understand this apparent paradox of high human capital attainment and high violence and suicides experienced by women in Kerala. Methods. The study uses data from the Census of India (1991, 2001) , National Crime Records Bureau (1998–2000), and Reproductive and Child Health ( District Level House Hold Survey—2001 ) to analyze trends in crime, violence, and suicide rates, and demographic profiles of the different districts in Kerala. Results. High educational attainment has fostered new aspirations and attitudes among women in Kerala. Yet, societal and cultural norms still dictate that women should be subservient to men both at home and in the labor market. This imbalance often contributes to family violence and suicides in Kerala. Thus, beneath the veil of development lie some disturbing social, cultural, and economic issues that may be contributing to high rates of suicide and family violence in Kerala. Conclusions. This study demonstrates that high educational attainment alone will not promote gender empowerment unless the social and cultural fabric of a country or state ensures equality of women in all areas of life.     

Generalization of the immigration and the stable population model

Building upon an idea presented by Espenshade, Bouvier and Arthur (1982) about the eventual stationarity of a population subjected to a net reproduction rate of less than one and a constant stream of immigration, this note reports the consequences upon the birth trajectory for all possible values of the net reproduction rate. It has been shown that when the net reproduction rate is equal to one, the number of births increases linearly with time. As may be expected, the increase is exponential for values of net reproduction rate greater than one.     

Randomized Response Techniques Using Maximum Likelihood Estimator

The randomized response (RR) technique with two decks of cards proposed by Odumade and Singh (2009) can always be made more efficient than the RR techniques proposed by Warner (1965), Mangat and Singh (1990), and Mangat (1994) by adjusting the proportion of cards in the decks. Arnab et al. (2012) generalized Odumade and Singh strategy (2009) for complex survey designs and wider class of estimators. In this paper improvement of Arnab et al. (2012) estimator has been made by using maximum likelihood method.     

Estimation of a finite population total in two-stage and stratified sampling under superpopulation models

Optimal sampling strategies which minimise the expected mean square error for a linear design as well as model-design unbiased estimators for a finite population total for two-stage and stratified sampling are obtained under different superpopu1ation models     

Invisible or Mainstream? Disability in Surveys and Censuses in Low- and Middle-Income Countries

Social Indicators Research - This paper examines to what extent household surveys and censuses in low- and middle-income countries include disability questions and the types of questions under use....     

Optional randomized response techniques for quantitative characteristics

Gupta et al. and Huang considered optional randomized response techniques where the probability of choosing the randomized (or direct) response is fixed for all the respondents. In this paper the assumption of the constant probability of choosing the option has been relaxed by dividing respondents into two groups: one group provides direct response and the second a randomized response. The method of estimation of the population mean and variances under the modified assumption are obtained. Relative efficiencies of the proposed techniques are compared theoretically and empirically.