A Two-Step Approach to Analyze Satisfaction Data

In this paper a two-step procedure based on Nonlinear Principal Component Analysis (NLPCA) and Multilevel models (MLM) for the analysis of satisfaction data is proposed. The basic hypothesis is that observed ordinal variables describe different aspects of a latent continuous variable, which depends on covariates connected with individual and contextual features. NLPCA is used to measure the level of a latent variable and MLM is adopted for detecting individual and environmental determinants of the level. This approach is suggested to analyze users’ satisfaction. In fact, NLPCA is used to create a synthetic continuous measure of satisfaction (first step) and MLM are used to detect the role of external (individual or environmental) variables that can affect the level itself (second step). The proposed two-step procedure is applied to the Eurobarometer survey data about opinion of European citizens on services of general interest (SGI) aiming to evaluate and compare the opinion about SGI in different countries. The focus is on overall level of satisfaction about four major public services: fixed telephone, electricity supply, postal and rail services. The item analyzed, which are named manifest variables, are: access easiness, price, quality, information clarity and contract fairness, as reported in the 2002 Eurobarometer survey. In the first step these variables are used to set up the synthetic indicator (the overall level) of satisfaction and, in the second step, a MLM is used to test the impact of some explanatory variables on this satisfaction.     

Investigating the European perception of food using moments obtained from non-symmetrical correspondence analysis

The perception of food in Europe has been a topic of research for many years due to its importance in better understanding the role of food in helping to define the culture of a country. It is also important from a marketing perspective for identifying how consumers relate to food. Recently, this topic was discussed by Guerrero et al. (2010) who used a graphical statistical technique called correspondence analysis to identify the association between the countries that participated in the study and words that were linked with “Traditional” food. This paper explores the use of non-symmetrical correspondence analysis and provides an interpretation of the configuration of points in the graphical display in terms of its first four moments. In particular, we will focus on the skewness and kurtosis of such a configuration. Such measure's provide further detail on the nature of the association between the countries studied and the words linked with “Traditional” food.     

Temporal disaggregation and restricted forecasting of multiple population time series

This article presents some applications of time-series procedures to solve two typical problems that arise when analyzing demographic information in developing countries: (1) unavailability of annual time series of population growth rates (PGRs) and their corresponding population time series and (2) inappropriately defined population growth goals in official population programs. These problems are considered as situations that require combining information of population time series. Firstly, we suggest the use of temporal disaggregation techniques to combine census data with vital statistics information in order to estimate annual PGRs. Secondly, we apply multiple restricted forecasting to combine the official targets on future PGRs with the disaggregated series. Then, we propose a mechanism to evaluate the compatibility of the demographic goals with the annual data. We apply the aforementioned procedures to data of the Mexico City Metropolitan Zone divided by concentric rings and conclude that the targets established in the official program are not feasible. Hence, we derive future PGRs that are both in line with the official targets and with the historical demographic behavior. We conclude that growth population programs should be based on this kind of analysis to be supported empirically. So, through specialized multivariate time-series techniques, we propose to obtain first an optimal estimate of a disaggregate vector of population time series and then, produce restricted forecasts in agreement with some data-based population policies here derived.     

Predictive event modelling in multicenter clinical trials with waiting time to response

A new analytic statistical technique for predictive event modeling in ongoing multicenter clinical trials with waiting time to response is developed. It allows for the predictive mean and predictive bounds for the number of events to be constructed over time, accounting for the newly recruited patients and patients already at risk in the trial, and for different recruitment scenarios. For modeling patient recruitment, an advanced Poisson-gamma model is used, which accounts for the variation in recruitment over time, the variation in recruitment rates between different centers and the opening or closing of some centers in the future. A few models for event appearance allowing for 'recurrence', 'death' and 'lost-to-follow-up' events and using finite Markov chains in continuous time are considered. To predict the number of future events over time for an ongoing trial at some interim time, the parameters of the recruitment and event models are estimated using current data and then the predictive recruitment rates in each center are adjusted using individual data and Bayesian re-estimation. For a typical scenario (continue to recruit during some time interval, then stop recruitment and wait until a particular number of events happens), the closed-form expressions for the predictive mean and predictive bounds of the number of events at any future time point are derived under the assumptions of Markovian behavior of the event progression. The technique is efficiently applied to modeling different scenarios for some ongoing oncology trials. Case studies are considered.     

Freezing the Master Production Schedule Under Demand Uncertainty

Managing the trade-off between achieving a stable master production schedule (MPS) and being responsive to changes in customer requirements is a difficult problem in many firms where providing a high level of customer service is viewed as an important competitive factor. One alternative for managing this trade-off is to freeze an agreed portion of the MPS. This paper investigates the impact of adjustments in the design parameters of MPS freezing methods on two performance measures (MPS lot-sizing cost and stability) under stochastic demand conditions in a rolling planning horizon environment given a service level target. Simulation experiments are reported which indicate that many of the conclusions regarding the design of MPS freezing methods obtained under deterministic demand conditions hold under stochastic demand.     

A Note on Discriminant Analysis Using LAD

Using a regression approach to discriminant analysis is often incorrect because it forces the use of a binary dependent variable which violates virtually any distributional assumption for a linear model. However, assuming a Laplace distribution in an LP framework leads to a theoretical foundation for MSD discriminant analysis.     

Artificial Intelligence and Generalized Qualitative-Response Models: An Empirical Test on Two Audit Decision-Making Domains

Machine learning methods are currently the object of considerable study by the artificial intelligence community. Research on machine learning carries implications for decision making in that it seeks computational methods that mimic input-output behaviors found in classes of decision-making examples. At the same time, research in statistics and econometrics has resulted in the development of qualitative-response models that can be applied to the same kind of problems addressed by machine-learning models—particularly those that involve a classification decision. This paper presents the theoretical structure of a generalized qualitative-response model and compares its performance to two seminal machine-learning models in two problem domains associated with audit decision making. The results suggest that the generalized qualitative-response model may be a useful alternative for certain problem domains.     

Evaluating Consumer Preferences for Existing Multiattribute Products: A Non-Metric Approach

A model evaluating consumer preferences for multiattribute products is derived. The model possesses the following features: (1) the method works for existing products; (2) the input data require only overall product rankings and attribute rankings; (3) the distribution of part worths (utility) for particular attributes is derived and used to judge the trade-offs among different attributes; (4) the procedure can be implemented with existing software; and (5) the attributes do not need to be quantifiable.     

A New Framework for Manufacturing Planning and Control Systems*

This paper presents a new framework for manufacturing planning and control systems which we call iterative manufacturing planning in continuous time (IMPICT) that appears to have several advantages over the well-known material requirements planning (MRP) framework. IMPICT explicitly considers capacity constraints and total system cost (including tardiness) to determine order sizes, order release/due dates, and operation schedules in a deterministic, multi-level, finite horizon, dynamic demand environment. Continuous time scheduling variables allow setups to be carried over from one period to the next. Three new heuristics built on the IMPICT framework are presented and tested in a simulation-based, full-factorial experiment with a wide variety of problem environments. The benchmark for the experiment was materials requirements planning with operations sequencing (MRP/OS) implemented with best-case, fixed planned lead times. The experiment showed that all three heuristics were statistically better than MRP/OS. The total cost for the order merging (OM) heuristic was 25 percent better than the total cost for MRP/OS. Computational times for OM were substantially larger than for MRP/OS; however, the computational times in the experiment suggest that OM is still computationally viable for large-scale batch manufacturing environments found in industry. IMPICT is superior to standard MRP systems because it explicitly considers capacity constraints and total system costs when it creates a materials plan. IMPICT is superior to linear programming-based approaches to finite loading and scheduling found in the literature because it allows setups to be carried over from one period to another and because it is computationally viable for realistic-sized problems.     

Understanding sibling differences in child labor

This study considers sibling differences in child labor in Nepal. The data are consistent with a model where parents care equally for all children but siblings differ in comparative advantage in household production, although parental preferences and credit constraints could also be important. Girls, especially older girls, tend to work more than their brothers. This extra work increases with the number of younger siblings and the spacing between siblings. The extra work performed by girls is such that, at modal birth spacing, the younger girl actually spends significantly more time working than her older brother.