When in a linear GMM model nuisance parameters are eliminated by multiplying the moment conditions by a projection matrix, the covariance matrix of the model, the inverse of which is typically used to construct an efficient GMM estimator, turns out to be singular and thus cannot be inverted. However, one can show that the generalized inverse can be used instead to produce an efficient estimator. Various other matrices in place of the projection matrix do the same job, i.e., they eliminate the nuisance parameters. The relations between those matrices with respect to the efficiency of the resulting estimators are investigated. | English |

When rounded data are used in place of the true values to compute the variance of a variable or a regression line, the results will be distorted. Under suitable smoothness conditions on the distribution of the variable(s) involved, this bias, however, can be corrected with very high precision by using the well-known Sheppard’s correction. In this paper, Sheppard’s correction is generalized to cover more general forms of rounding procedures than just simple rounding, viz., probabilistic rounding, which includes asymmetric rounding and mixture rounding. | English |

Microaggregation by individual ranking (IR) is an important technique for masking confidential data. While being a successful method for controlling the disclosure risk of observations, IR is also known for its favorable property of having a relatively small effect on the results of statistical analyses. In this paper we conduct a detailed theoretical analysis on the estimation of arbitrary moments from a data set that has been anonymized by means of the IR method. We show that classical moment estimators remain both consistent and asymptotically normal under relatively weak assumptions. This theory provides the justification for applying standard statistical estimation techniques to the anonymized data without having to correct for a possible bias caused by anonymization. | English |

In a multivariate mean-variance model, the class of linear score (LS) estimators based on an unbiased linear estimating function is introduced. A special member of this class is the (extended) quasi-score (QS) estimator. It is ``extended'' in the sense that it comprises the parameters describing the distribution of the regressor variables. It is shown that QS is (asymptotically) most efficient within the class of LS estimators. An application is the multivariate measurement error model, where the parameters describing the regressor distribution are nuisance parameters. A special case is the zero-inflated Poisson model with measurement errors, which can be treated within this framework. | English |

A problem statistical offices are increasingly faced with is guaranteeing confidentiality when releasing microdata sets. One method to provide safe microdata to is to reduce the information content of a data set by means of masking procedures. A widely discussed masking procedure is microaggregation, a technique where observations are grouped and replaced with their corresponding group means. However, while reducing the disclosure risk of a data file, microaggregation also affects the results of statistical analyses. The paper deals with the impact of microaggregation on a simple linear model. We show that parameter estimates are biased if the dependent variable is used to group the data. It turns out that the bias of the slope parameter estimate is a non-monotonic function of this parameter. By means of this non-monotonic relationship we develop a method for consistently estimating the model parameters. | English |

A problem statistical offices are increasingly faced with is guaranteeing confidentiality when releasing microdata sets. One method to provide safe microdata is to reduce the information content of a data set by means of masking procedures. A widely discussed masking procedure is microaggregation, a technique where observations are grouped and replaced with their corresponding group means. However, while reducing the disclosure risk of a data file, microaggregation also affects the results of statistical analyses. We focus on the effect of microaggregation on a simple linear model. In a previous paper we have shown how to correct for the aggregation bias of the naive least-squares estimator that occurs when the dependent variable is used to group the data. The present paper deals with the asymptotic variance of the corrected least-squares estimator and with the asymptotic variance of the naive least-squares estimator when either the dependent variable or the regressor is used to group the data. We derive asymptotic confidence intervals for the slope parameter. Furthermore, we show how to test for the significance of the slope parameter by analyzing the effect of microaggregation on the asymptotic power function of the naive t-test. | English |

This paper grew out of a lecture presented at the 54th Session of the International Statistical Institute in Berlin, August 13 - 20, 2003, Schneeweiss (2003). It intends not only to outline the eventful life of Abraham Wald (1902 - 1950) in Austria and in the United States but also to present his extensive scientific work. In particular, the two main subjects, where he earned most of his fame, are outline: Statistical Decision Theory and Sequential Analysis. In addition, emphasis is laid on his contributions to Econometrics and related fields. | English |

Microaggregation is a set of procedures that distort empirical data in order to guarantee the factual anonymity of the data. At the same time the information content of data sets should not be reduced too much and should still be useful for scientific research. This paper investigates the effect of microaggregation on the estimation of a linear regression by ordinary least squares. It studies, by way of an extensive simulation experiment, the bias of the slope parameter estimator induced by various microaggregation techniques. Some microaggregation procedures lead to consistent estimates while others imply an asymptotic bias for the estimator. | English |

We consider a polynomial regression model, where the covariate is measured with Gaussian errors. The measurement error variance is supposed to be known. The covariate is normally distributed with known mean and variance. Quasi Score (QS) and Corrected Score (CS) are two consistent estimation methods, where the first makes use of the distribution of the covariate (structural method), while the latter does not (functional method). It may therefore be surmised that the former method is (asymptotically) more efficient than the latter one. This can, indeed, be proved for the regression parameters. We do this by introducing a third, so-called Simple Score (SS), estimator, the efficiency of which turns out to be intermediate between QS and CS. When one includes structural and functional estimators for the variance of the error in the equation, SS is still more efficient than CS. When the mean and variance of the covariate are not known and have to be estimated as well, one can still maintain that QS is more efficient than SS for the regression parameters. | English |

The asymptotic covariance matrices of the corrected score, the quasi score, and the simple score estimators of a polynomial measurement error model have been derived in the literature. Here some alternative formulas are presented, which might lead to an easier computation of these matrices. In particular, new properties of the variables $t_r$ and $\mu_r$ that constitute the estimators are derived. In addition, the term in the formula for the covariance matrix of the quasi score estimator stemming from the estimation of nuisance parameters is evaluated. The same is done for the log-linear Poisson measurement error model. In the polynomial case, it is shown that the simple score and the quasi score estimators are not always more efficient than the corrected score estimator if the nuisance parameters have to be estimated. | English |

We compare two consistent estimators of the parameter vector beta of a general exponential family measurement error model with respect to their relative efficiency. The quasi score (QS) estimator uses the distribution of the regressor, the corrected score (CS) estimator does not make use of this distribution and is therefore more robust. However, if the regressor distribution is known, QS is asymptotically more efficient than CS. In some cases it is, in fact, even strictly more efficient, in the sense that the difference of the asymptotic covariance matrices of CS and QS is positive definite. | English |

A measurement error model is a regression model with (substantial) measurement errors in the variables. Disregarding these measurement errors in estimating the regression parameters results in asymptotically biased estimators. Several methods have been proposed to eliminate, or at least to reduce, this bias, and the relative efficiency and robustness of these methods have been compared. The paper gives an account of these endeavors. In another context, when data are of a categorical nature, classification errors play a similar role as measurement errors in continuous data. The paper also reviews some recent advances in this field. | English |

Microaggregation is one of the most frequently applied statistical disclosure control techniques for continuous data. The basic principle of microaggregation is to group the observations in a data set and to replace them by their corresponding group means. However, while reducing the disclosure risk of data files, the technique also affects the results of statistical analyses. The paper deals with the impact of microaggregation on a linear model in continuous variables. We show that parameter estimates are biased if the dependent variable is used to form the groups. Using this result, we develop a consistent estimator that removes the aggregation bias. Moreover, we derive the asymptotic covariance matrix of the corrected least squares estimator. | English |

We prove that the quasi-score estimator in a mean-variance model is optimal in the class of (unbiased) linear score estimators, in the sense that the difference of the asymptotic covariance matrices of the linear score and quasi-score estimator is positive semi-definite. We also give conditions under which this difference is zero or under which it is positive definite. This result can be applied to measurement error models where it implies that the quasi-score estimator is asymptotically more efficient than the corrected score estimator. | English |

If rounded data are used in estimating moments and regression coefficients, the estimates are typically more or less biased. The purpose of the paper is to study the bias inducing effect of rounding, which is also seen when population moments instead of their estimates are considered. Under appropriate conditions this effect can be approximately specified by versions of Sheppard's correction formula. We discuss the conditions under which these approximations are valid. We also investigate the efficiency loss that comes along with rounding. The rounding error, which corresponds to the measurement error of a measurement error model, has a marginal distribution which can be approximated by the uniform distribution. We generalize the concept of simple rounding to that of asymmetric rounding and study its effect on the mean and variance of a distribution under similar circumstances as with simple rounding. | English |

The paper is a survey of recent investigations by the authors and others into the relative efficiencies of structural and functional estimators of the regression parameters in a measurement error model. While structural methods, in particular the quasi-score (QS) method, take advantage of the knowledge of the regressor distribution (if available), functional methods, in particular the corrected score (CS) method, discards such knowledge and works even if such knowledge is not available. Among other results, it has been shown that QS is more efficient than CS as long as the regressor distribution is completely known. However, if nuisance parameters in the regressor distribution have to be estimated, this is no more true in general. But by modifying the QS method, the adverse effect of the nuisance parameters can be overcome. For small measurement errors, the efficiencies of QS and CS become almost indistinguishable, whether nuisance parameters are present or not. QS is (asymptotically) biased if the regressor distribution has been misspecified, while CS is always consistent and thus more robust than QS. | English |

The present article considers the problem of consistent estimation in measurement error models. A linear relation with not necessarily normally distributed measurement errors is considered. Three possible estimators which are constructed as different combinations of the estimators arising from direct and inverse regression are considered. The efficiency properties of these three estimators are derived and analyzed. The effect of non-normally distributed measurement errors is analyzed. A Monte-Carlo experiment is conducted to study the performance of these estimators in finite samples and the effect of a non-normal distribution of the measurement errors. | English |

We consider a regression of $y$ on $x$ given by a pair of mean and variance functions with a parameter vector $\theta$ to be estimated that also appears in the distribution of the regressor variable $x$. The estimation of $\theta$ is based on an extended quasi score (QS) function. We show that the QS estimator is optimal within a wide class of estimators based on linear-in-$y$ unbiased estimating functions. Of special interest is the case where the distribution of $x$ depends only on a subvector $\alpha$ of $\theta$, which may be considered a nuisance parameter. In general, $\alpha$ must be estimated simultaneously together with the rest of $\theta$, but there are cases where $\alpha$ can be pre-estimated. A major application of this model is the classical measurement error model, where the corrected score (CS) estimator is an alternative to the QS estimator. We derive conditions under which the QS estimator is strictly more efficient than the CS estimator.We also study a number of special measurement error models in greater detail. | English |

Various estimators of the reduced form of a block recursive model are investigated and compared to each other. In particular it is shown that the structural reduced form estimator, which results from estimating separately each block of the block recursive model by some efficient method and then solving the system for the endogenous variables, is more efficient than the OLS estimator of the reduced form. Other reduced form estimators derived from OLS or Two Stage LS estimators of a partially reduced form have intermediate efficiency properties. The paper has been published in Schneeweiss et al (2001), but without the appendices. | English |

The paper studies the problem of estimating the upper end point of a finite interval when the data come from a uniform distribution on this interval and are disturbed by normally distributed measurement errors with known variance. Maximum likelihood and method of moments estimators are introduced and compared to each other. | English |

In a structural error model the structural quasi score (SQS) estimator is based on the distribution of the latent regressor variable. If this distribution is misspecified the SQS estimator is (asymptotically) biased. Two types of misspecification are considered. Both assume that the statistician erroneously adopts a normal distribution as his model for the regressor distribution. In the first type of misspecification the true model consists of a mixture of normal distributions which cluster round a single normal distribution, in the second type the true distribution is a normal distribution admixed with a second normal distribution of low weight. In both cases of misspecification the bias, of course, tends to zero when the size of misspecification tends to zero. However, in the first case the bias goes to zero very fast so that small deviations from the true model lead only to a negligible bias, whereas in the second case the bias is noticeable even for small deviations from the true model. | English |

We compare the asymptotic covariance matrix of the ML estimator in a nonlinear measurement error model to the asymptotic covariance matrices of the CS and SQS estimators studied in Kukush et al (2002). For small measurement error variances they are equal up to the order of the measurement error variance and thus nearly equally efficient. | English |

Two methods of estimating the parameters of a polynomial regression with measurement errors in the regressor variable are compared to each other with respect to their relative efficiency and robustness. One of the two estimators (SLS) is valid for the structural variant of the model and uses the assumption that the true regressor variable is normally distributed, while the other one (ALS and also its small sample modification MALS) does not need any assumption on the regressor distribution. SLS turns out to react rather strongly on violations of the normality assumption as far as its bias is concerned but is quite robust with respect to its MSE. It is more efficient than ALS or MALS whenever the normality assumption holds true. | English |

A polynomial structural errors-in-variables model with normal underlying distributions is investigated. An asymptotic covariance matrix of the SLS estimator is computed, including the correcting terms which appear because in the score function the sample mean and the sample variance are plugged in. The ALS estimator is also considered, which does not need any assumption on the regressor distribution. The asymptotic covariance matrices of the two estimators are compared in border cases of small and of large errors. In both situations it turns out that under the normality assumption SLS is strictly more efficient than ALS. | English |

In a polynomial regression with measurement errors in the covariate, which is supposed to be normally distributed, one has (at least) three ways to estimate the unknown regression parameters: one can apply ordinary least squares (OLS) to the model without regard of the measurement error or one can correct for the measurement error, either by correcting the estimating equation (ALS) or by correcting the mean and variance functions of the dependent variable, which is done by conditioning on the observable, error ridden, counter part of the covariate (SLS). While OLS is biased the other two estimators are consistent. Their asymptotic covariance matrices can be compared to each other, in particular for the case of a small measurement error variance. | English |

We consider two consistent estimators for the parameters of the linear predictor in the Poisson regression model, where the covariate is measured with errors. The measurement errors are assumed to be normally distributed with known error variance sigma_u^2. The SQS estimator, based on a conditional mean-variance model, takes the distribution of the latent covariate into account, and this is here assumed to be a normal distribution. The CS estimator, based on a corrected score function, does not use the distribution of the latent covariate. Nevertheless, for small sigma_u^2, both estimators have identical asymptotic covariance matrices up to the order of sigma_u^2. We also compare the consistent estimators to the naive estimator, which is based on replacing the latent covariate with its (erroneously) measured counterpart. The naive estimator is biased, but has a smaller covariance matrix than the consistent estimators (at least up to the order of sigma_u^2.) | English |

A nonlinear structural errors-in-variables model is investigated, where the response variable has a density belonging to an exponential family and the error-prone covariate follows a Gaussian distribution. Assuming the error variance to be known, we consider two consistent estimators in addition to the naive estimator. We compare their relative efficiencies by means of their asymptotic covariance matrices for small error variances. The structural quasi score (SQS) estimator is based on a quasi score function, which is constructed from a conditional mean-variance model. Consistency and asymptotic normality of this estimator is proved. The corrected score (CS) estimator is based on an error-corrected likelihood score function. For small error variances the SQS and CS estimators are approximately equally efficient. The polynomial model and the Poisson regression model are explored in greater detail. | English |

This paper discusses point estimation of the coefficients of polynomial measurement error (errors-in-variables) models. This includes functional and structural models. The connection between these models and total least squares (TLS) is also examined. A compendium of existing as well as new results is presented. | English |

The structural variant of a regression model with measurement error is characterized by the assumption of an underlying known distribution of the latent covariate. Several estimation methods, like regression calibration or structural quasi score estimation, take this distribution into account. In the case of a polynomial regression, which is studied here, structural quasi score takes the form of structural least squares (SLS). Usually the underlying latent distribution is assumed to be the normal distribution because then the estimation methods take a particularly simple form. SLS is consistent as long as this assumption is true. The purpose of the paper is to investigate the amount of bias that results from violations of the normality assumption for the covariate distribution. Deviations from normality are introduced by switching to a mixture of normal distributions. It turns out that the bias reacts only mildly to slight deviations from normality. | English |

We consider a Poisson model, where the mean depends on certain covariates in a log-linear way with unknown regression parameters. Some or all of the covariates are measured with errors. The covariates as well as the measurement errors are both jointly normally distributed, and the error covariance matrix is supposed to be known. Three consistent estimators of the parameters - the corrected score, a structural, and the quasi-score estimators - are compared to each other with regard to their relative (asymptotic) efficiencies. The paper extends an earlier result for a scalar covariate. | English |

Nach einer kurzen Einfuehrung in die Theorie der erwartungstreuen Schaetzgleichungen fuer allgemeine Regressionsmodelle und der korrigierten Schaetzgleichungen fuer Regressionsmodelle mit fehlerbehafteten Kovariablen wird die Approximationsguete eines auf Reihenentwicklung basierenden Ansatzes von Stefanski diskutiert. | English |

This Note generalizes two estimators of the quadratic regression with measurement errors by Fuller and Wolter and Fuller to the polynomial case. | English |

An adjusted least squares estimator, introduced by Cheng and Schneeweiss (1998) for consistently estimating a polynomial regression of any degree with errors in the variables, is modified such that it shows good results in small samples without losing its asymptotic properties for large samples. Simulation studies corroborate the theoretical findings. The new method is applied to analyse a geophysical law relating the depth of earthquakes to their distance from a trench where one of the earth's plates is submerged beneath another one. | English |

Ellsberg (1961) designed a decision experiment where most people violated the axioms of rational choice. He asked people to bet on the outcome of certain random events with known and with unknown probabilities. They usually preferred to bet on events with known probabilities. It is shown that this behavior is reasonable and in accordance with the axioms of rational decision making if it is assumed that people consider bets on events that are repeatedly sampled instead of just sampled once. | English |