Estimation of a process capability index for inverse gaussian distribution

In this article, interval estimates of Clements' process capability index are studied through bootstrapping when the underlying distribution is Inverse Gaussian. The standard bootstrap, the percentile bootstrap, and the bias-corrected percentile bootstrap confidence intervals are compared.     

A Duration‐sensitive Measure of the Unemployment Rate: Theory and Application

The measurement of unemployment, like that of poverty, involves two distinct steps: identification and aggregation. In this two‐step process, the issue of identifying the unemployed has received considerable attention but, once the unemployed have been identified, the aggregation issue has been addressed by simply ‘counting heads’: the unemployment rate is conventionally defined as the proportion of the labour force that, on a given date, is unemployed. This, in particular, leads to differences between individuals, in their unemployment experiences being ignored when the unemployment rate is being computed. This paper — predicated on the proposition that what matters to a person is not just the fact of unemployment but also its duration — proposes a methodology, derived from the measurement of income inequality, for adjusting unemployment rates so as to make them ‘duration‐sensitive’. In consequence, different values of the ‘duration‐sensitive’ rate will, depending upon the degree of inequality in the distribution of unemployment duration, and upon the extent to which society is averse to such inequality, be associated with the same value of the conventionally defined unemployment rate. A numerical example, based on published data for seven major OECD counties, illustrates the methodology.     

Severity and determinants of stunting in children under age 2 years in Odisha (India): a tribal v/s non-tribal analysis

Prevalence of stunting among India’s tribal under-five children has witnessed a 3% annual decline in the last 8 years. Cross-sectional data of 1000 children (287 tribal and 713 non-tribal) aged 0–23 months from Odisha’s Rapid Survey of Children (RSOC, 2014) was analysed to identify the predictors of stunting and severe stunting among tribal children and to suggest policy and programme implications. Results show significant determinants of childhood stunting to be birth order, maternal illiteracy and determinants for severe stunting were maternal age <18 years at marriage and <20 years at birth, and <3 antenatal clinic visits. Severe stunting in tribal children was predicted by basic causes: poverty and maternal age <20 years at first birth and age <18 years at marriage. Findings strongly suggest to accelerate efforts towards reducing childhood stunting in Odisha more specifically among tribal children for whom a special strategy is urgently warranted.     

US College Students' Reports of Receiving Health Information on College Campuses

Abstract

Institutions of higher education are in a unique position to promote healthy behaviors by providing health education to students, but little information exists about the proportion of students reached by such efforts. The authors used data from a nationally representative sample of college students to describe the extent to which students reported receiving health information from their colleges and universities, to examine the characteristics of students who received such information, and to determine specific sources of health information. Approximately three quarters of college students reported they received information on at least one health topic, and 6% received information on all of the topics examined. Those who reported receiving health information from their colleges or universities were likely to be “traditional” college students. To achieve relevant national health objectives, health educators must increase the proportion of students they reach and the number of health topics they cover.     

Exploring Multidimensional Well-Being in Switzerland: Comparing Three Synthesizing Approaches

Well-being is increasingly conceptualized and measured from a multidimensional perspective in order to provide policy makers and the general public with the necessary tools for monitoring social change. Synthetic measures of multidimensional concepts such as well-being are, therefore, needed. This paper provides a systematic comparison of three methodological approaches applicable in a top-down perspective to construct a synthetic multidimensional measure of well-being using the same theoretical framework and the same data. The three methodologies used are the confirmatory factor analysis, the Alkire and Foster counting approach and the posetic approach. The first two methodologies belong to the family of composite indicators, the most common way to synthetize multidimensional information. The third was developed in order to deal with the construction of composite indicators’ two major issues, namely weighting and aggregating. The main contribution of this paper is methodological; it aims at discussing the advantages and drawbacks of the three methodologies, and at identifying their similarities and specificities.     

A pseudo-complete sample technique for estimation from censored samples

An iterative procedure is presented whereby singly right censored samples are transformed into pseudo-complete samples. This involves the use of order statistics to calculate expected “complete” values of the censored observations. Estimates of the distribution parameters are then calculated by employing standard complete sample estimators. This procedure is applicable to various types of distributions.