THE USE OF MODERATED REGRESSION IN JOB REDESIGN DECISIONS*

This paper describes the use of moderated regression in job redesign decisions. The procedures outlined are intended to augment the job redesign strategy described by Hackman and his colleagues. Different regression models based on a theory of job design are shown to lead to different job redesign decisions. Comparisons are also made between the moderated regression procedures and the more commonly used subgroups analysis.     

Accounting for Both Random Errors and Systematic Errors in Uncertainty Propagation Analysis of Computer Models Involving Experimental Measurements with Monte Carlo Methods

A Monte Carlo method is presented to study the effect of systematic and random errors on computer models mainly dealing with experimental data. It is a common assumption in this type of models (linear and nonlinear regression, and nonregression computer models) involving experimental measurements that the error sources are mainly random and independent with no constant background errors (systematic errors). However, from comparisons of different experimental data sources evidence is often found of significant bias or calibration errors. The uncertainty analysis approach presented in this work is based on the analysis of cumulative probability distributions for output variables of the models involved taking into account the effect of both types of errors. The probability distributions are obtained by performing Monte Carlo simulation coupled with appropriate definitions for the random and systematic errors. The main objectives are to detect the error source with stochastic dominance on the uncertainty propagation and the combined effect on output variables of the models. The results from the case studies analyzed show that the approach is able to distinguish which error type has a more significant effect on the performance of the model. Also, it was found that systematic or calibration errors, if present, cannot be neglected in uncertainty analysis of models dependent on experimental measurements such as chemical and physical properties. The approach can be used to facilitate decision making in fields related to safety factors selection, modeling, experimental data measurement, and experimental design.     

GRAPHICAL REPRESENTATION OF THE EFFECTS OF MULTICOLLINEARITY

This paper considers the effects of multicollinearity on the multiple coefficient of determination and on the standard errors of the estimated regression coefficients. Limiting results are well known for these problems, and some particular cases have been investigated. However, the limiting cases do not explain the nonmonotonic relationships that have been observed in practice. This paper explores these relationships and attempts to provide a unifying graphical explanation that includes the previous results as special cases. Implications for stepwise regression and for the interpretation of any multiple regression involving multicollinear independent variables are discussed.     

What to Learn from Near Misses: An Inductive Learning Approach to Credit Risk Assessment

This paper presents a new dimension of inductive learning for credit risk analysis based on the specific impact of Type I and Type II credit errors on the accuracy of the learning process. A Dynamic Updating Process is proposed to refine the credit granting decision over time and therefore improve the accuracy of the learning process. The new dimension is tested on credit files of small Belgian businesses. Results indicate an improvement of the learning process in terms of predictive accuracy, stability, and conceptual validity of the final decision tree.     

Measuring the quality of economic forecasts

Most evaluations of forecasts have hitherto been primarily based on measure of statistical accuracy. Attention is drawn to the need for concentrating instead on the forecast components bearing on managerial decisions and on the economic effects of such decisions. The method suggested highlights the changing locus of the most influential errors over various time horizons. It also takes account of the economic consequences of under-estimates and over-estimates in forecasting, as well as of the penalties of belated correcting actions. Finally, some broader implications of this approach are discussed.     

The Effects of Publication Selection on Test Probabilities and Estimator Distributions

Editorial decisions, based in part on reported hypothesis test results, affect the probabilities associated with those results: the probabilities of Type I and Type II errors thus become different for readers than for authors. The distributions of published parameter estimates are similarly affected. A framework for studying the consequences of test-based information filtering is developed and illustrative examples are provided. The examples indicate that filtering can markedly distort the power functions of hypothesis tests and can induce large estimator biases and increases in mean square error. It is argued that test-based filtering is relevant not only to journal publication but to other forms of information dissemination as well.     

Setting Tolerable Misstatements When Auditing Aggregated Accounts*

Generally accepted auditing standards require auditors to plan audits of clients' account balances. If accounts are to be sampled, then part of this planning must include setting the tolerable misstatement for each account or class of transactions to be sampled. Although classical sampling approaches provide certain advantages, they have not been widely used because they are viewed as complex and difficult to implement. We present a remedy to these difficulties in an efficient, easily implemented optimal solution method for the problem of setting tolerable misstatements given constraints on tolerable misstatements for individual account balances as well as the overall audit. Further, our method suggests when the materialities of certain accounts or the materiality of the overall audit are irrelevant to the problem. Several example auditing problems demonstrate both our solution approach and the settings in which our approach provides a more effective or more efficient sampling plan than that provided by monetary unit sampling.     

ENROLLMENT FORECASTING: USE OF A MULTIPLE-METHOD MODEL FOR PLANNING AND BUDGETING*

A multiple-method approach is presented for making enrollment projections at the departmental, school, and university level. The various methods utilized are based on the application of a cohort model (macro level) and regression techniques (micro level). The model demonstrates its flexibility in selecting one of a number of regression techniques according to a comparison approach using historical and projected results. This multiple-method approach has proved to be highly accurate and thus useful in allocating and redistributing resources to the operating units of a university. Previous efforts have relied almost solely on a single regression technique for projections and allocation purposes rather than a method based on a measure of flexibility, adaptability, and prediction confirmation.     

Modeling Growth of Clostridium perfringens in Pea Soup During Cooling

Clostridium perfringens is a pathogen that mainly causes food poisoning outbreaks when large quantities of food are prepared. Therefore, a model was developed to predict the effect of different cooling procedures on the growth of this pathogen during cooling of food: Dutch pea soup. First, a growth rate model based on interpretable parameters was used to predict growth during linear cooling of pea soup. Second, a temperature model for cooling pea soup was constructed by fitting the model to experimental data published earlier. This cooling model was used to estimate the effect of various cooling environments on average cooling times, taking into account the effect of stirring and product volume. The growth model systematically overestimated growth of C. perfringens during cooling in air, but this effect was limited to less than 0.5 log N/ml and this was considered to be acceptable for practical purposes. It was demonstrated that the growth model for C. perfringens combined with the cooling model for pea soup could be used to sufficiently predict growth of C. perfringens in different volume sizes of pea soup during cooling in air as well as the effect of stirring, different cooling temperatures, and various cooling environments on the growth of C. perfringens in pea soup. Although fine-tuning may be needed to eliminate inaccuracies, it was concluded that the combined model could be a useful tool for designing good manufacturing practices (GMP) procedures.     

AN EVALUATION OF FORECAST ERROR IN MASTER PRODUCTION SCHEDULING FOR MATERIAL REQUIREMENTS PLANNING SYSTEMS.

Typical forecast-error measures such as mean squared error, mean absolute deviation and bias generally are accepted indicators of forecasting performance. However, the eventual cost impact of forecast errors on system performance and the degree to which cost consequences are explained by typical error measures have not been studied thoroughly. The present paper demonstrates that these typical error measures often are not good predictors of cost consequences in material requirements planning (MRP) settings. MRP systems rely directly on the master production schedule (MPS) to specify gross requirements. These MRP environments receive forecast errors indirectly when the errors create inaccuracies in the MPS. Our study results suggest that within MRP environments the predictive capabilities of forecast-error measures are contingent on the lot-sizing rule and the product components structure When forecast errors and MRP system costs are coanalyzed, bias emerges as having reasonable predictive ability. In further investigations of bias, loss functions are evaluated to explain the MRP cost consequences of forecast errors. Estimating the loss functions of forecast errors through regression analysis demonstrates the superiority of loss functions as measures over typical forecast error measures in the MPS.     

Concepts,Theory, and Techniques ROBUST METHODS: AN ALTERNATIVE APPROACH FOR ANALYZING DATA SETS CONTAINING INFLUENTIAL DATA POINTS

Recent advances in statistical estimation theory have resulted in the development of new procedures, called robust methods, that can be used to estimate the coefficients of a regression model. Because such methods take into account the impact of discrepant data points during the initial estimation process, they offer a number of advantages over ordinary least squares and other analytical procedures (such as the analysis of outliers or regression diagnostics). This paper describes the robust method of analysis and illustrates its potential usefulness by applying the technique to two data sets. The first application uses artificial data; the second uses a data set analyzed previously by Tufte [15] and, more recently, by Chatterjee and Wiseman [6].     

Correcting the Errors: Volatility Forecast Evaluation Using High‐Frequency Data and Realized Volatilities

We develop general model‐free adjustment procedures for the calculation of unbiased volatility loss functions based on practically feasible realized volatility benchmarks. The procedures, which exploit recent nonparametric asymptotic distributional results, are both easy‐to‐implement and highly accurate in empirically realistic situations. We also illustrate that properly accounting for the measurement errors in the volatility forecast evaluations reported in the existing literature can result in markedly higher estimates for the true degree of return volatility predictability.     

USING L2 ESTIMATION FOR L1 ESTIMATORS: AN APPLICATION TO THE SINGLE-INDEX MODEL

The bootstrap method is used to compute the standard error of regression parameters when the data are non-Gaussian distributed. Simulation results with L₁ and L₂ norms for various degrees of “non-Gaussianess” are provided. The computationally efficient L₂ norm, based on the bootstrap method, provides a good approximation to the L₁ norm. The methodology is illustrated with daily security return data. The results show that decisions can be reversed when the ordinary least-squares estimate of standard errors is used with non-Gaussian data.     

Cost/quality tradeoffs for control procedures in information systems

An analytical framework is presented to study the cost/benefit tradeoffs of alternative internal control scenarios designed to ensure quality of outputs from information systems. A mathematical model is used to evaluate the impact that various alternatives have on system performance and cost. The model can be used to compare the benefits gained from enhanced processing with those arising from more effective internal control procedures. Model variables under the control of system designers include placement and effectiveness of internal control procedures, cost and quality of processing activities, both manual and computerized, and cost and quality of correction procedures. The model incorporates a penalty cost incurred by failure to detect and correct errors. The analytical process is illustrated through application to a specific information system.     

Health Risks and Air Pollution — Error Analysis for a Cross-Sectional Mortality Study

An attempt is made to analyze in quantitative terms the uncertainties in multiple regression estimates of the effects of air pollution on death rates. A range of factors--statistical fluctuations in numbers of deaths, differences in local age distribution, differences in smoking habits, errors in estimated pollution levels, migration, and variability of the characterization of socioeconomic effects--are assessed as potential sources of error. Both the precision and the robustness of the regression calculation are shown to be poor. Examples and illustrative calculations are given based on a study of U. K. death rates around the 1971 Census.     

Identification and Estimation of Regression Models with Misclassification

This paper studies the problem of identification and estimation in nonparametric regression models with a misclassified binary regressor where the measurement error may be correlated with the regressors. We show that the regression function is nonparametrically identified in the presence of an additional random variable that is correlated with the unobserved true underlying variable but unrelated to the measurement error. Identification for semiparametric and parametric regression functions follows straightforwardly from the basic identification result. We propose a kernel estimator based on the identification strategy, derive its large sample properties, and discuss alternative estimation procedures. We also propose a test for misclassification in the model based on an exclusion restriction that is straightforward to implement.     

加权复合分位数回归方法在动态VaR风险度量中的应用

风险价值(VaR)因为简单直观,成为了当今国际上最主流的风险度量方法之一,而基于时间序列自回归(AR)模型来计算无条件风险度量值在实业界有广泛应用。本文基于分位数回归理论对AR模型提出了一个估计方法--加权复合分位数回归(WCQR)估计,该方法可以充分利用多个分位数信息提高参数估计的效率,并且对于不同的分位数回归赋予不同的权重,使得估计更加有效,文中给出了该估计的渐近正态性质。有限样本的数值模拟表明,当残差服从非正态分布时,WCQR估计的的统计性质接近于极大似然估计,而该估计是不需要知道残差分布的,因此,所提出的WCQR估计更加具有竞争力。此方法在预测资产收益的VaR动态风险时有较好的应用,我们将所提出的理论分析了我国九只封闭式基金,实证分析发现,结合WCQR方法求得的VaR风险与用非参数方法求得的VaR风险非常接近,而结合WCQR方法可以计算动态的VaR风险值和预测资产收益的VaR风险值。     

Simultaneous production and logistics operations planning in semicontinuous food industries

The production and logistics operations planning in real-life single- or multi-site semicontinuous food industries is addressed in this work. A discrete/continuous-time mixed integer programming model, based on the definition of families of products, is developed for the problem in question. A remarkable feature of the proposed approach is that in the production planning problem timing and sequencing decisions are taken for product families rather than for products. However, material balances are realized for every specific product, thus permitting the detailed optimization of production, inventory, and transportation costs. Changeovers are also explicitly taken into account and optimized. Moreover, alternative transportation modes are considered for the delivery of final products from production sites to distribution centers. The efficiency and the applicability of the proposed approach is demonstrated by solving to optimality two industrial-size case studies, for an emerging real-life Greek dairy industry.     

Shift Restrictions and Semiparametric Estimation in Ordered Response Models

We develop a √n‐consistent and asymptotically normal estimator of the parameters (regression coefficients and threshold points) of a semiparametric ordered response model under the assumption of independence of errors and regressors. The independence assumption implies shift restrictions allowing identification of threshold points up to location and scale. The estimator is useful in various applications, particularly in new product demand forecasting from survey data subject to systematic misreporting. We apply the estimator to assess exaggeration bias in survey data on demand for a new telecommunications service.     

EXTERNALITIES IN PROGRAM EVALUATION: THE IMPACT OF A WOMEN'S EMPOWERMENT PROGRAM ON IMMUNIZATION

Impact evaluations of development programs usually do not explicitly take into account externalities on non‐participants. Based on a unique dataset we estimate the direct as well as the spillover effects of Mahila Samakhya, a women's empowerment program in India, on child immunization. The survey covers both participants and non‐participants in program villages, as well as respondents in control villages. We account for participation selection bias using instrumental variables. We propose a direct test for the exclusion restriction using the control villages. We find both direct effects on the immunization rates of participants’ children and significant spillovers on immunization rates of non‐participants’ children. The impact of interventions might be substantially underestimated if such external effects were not taken into account. Alternative estimation methods such as propensity score matching and a regression discontinuity approach yield comparable results.