Suchergebnisse

Longitudinal data is essential for understanding how the world around us changes. Most theories in the social sciences and elsewhere have a focus on change, be it of individuals, of countries, of organisations, or of systems, and this is reflected in the myriad of longitudinal data that are being collected using large panel surveys. This type of data collection has been made easier in the age of Big Data and with the rise of social media. Yet our measurements of the world are often imperfect, and longitudinal data is vulnerable to measurement errors which can lead to flawed and misleading conclusions. This book tackles the important issue of how to investigate change in the context of imperfect data.

"Measurement error in retrospective reports of work status has been difficult to quantify in the past. Issues of confidentiality have made access to datasets linking survey responses to a valid administrative source very problematic. This study uses a Swedish register of unemployment as a benchmark against which responses from two survey questions are compared and hence the presence of measurement error elucidated. We carry out separate analyses for the different forms that measurement error in retrospective reports of unemployment can take: miscounting of the number of spells of unemployment, mismeasuring duration in unemployment, and misdating starts of spells and misclassification of status. The prevalence of measurement error for different social categories and interview formats is also examined, leading to a better understanding of the error-generating mechanisms that interact when interviewees are asked to produce retrospective reports of past work status." (publisher's description)

Measurement error threatens the validity of survey research, especially when studying sensitive questions. Although list experiments can help discourage deliberate misreporting, they may also suffer from nonstrategic measurement error due to flawed implementation and respondents' inattention. Such error runs against the assumptions of the standard maximum likelihood regression (MLreg) estimator for list experiments and can result in misleading inferences, especially when the underlying sensitive trait is rare. We address this problem by providing new tools for diagnosing and mitigating measurement error in list experiments. First, we demonstrate that the nonlinear least squares regression (NLSreg) estimator proposed in Imai (2011) is robust to nonstrategic measurement error. Second, we offer a general model misspecification test to gauge the divergence of the MLreg and NLSreg estimates. Third, we show how to model measurement error directly, proposing new estimators that preserve the statistical efficiency of MLreg while improving robustness. Last, we revisit empirical studies shown to exhibit nonstrategic measurement error, and demonstrate that our tools readily diagnose and mitigate the bias. We conclude this article with a number of practical recommendations for applied researchers. The proposed methods are implemented through an open-source software package.

The underidentification of linear models with measurement error does not necessarily extend to panel data models, as has been shown by GAiliches and Hausman (1986). We discuss and extend some of their results for a simple case and address particular issues concerning identification and asymptotic variances.

The perpetual inventory method used for the construction of education data per country leads to systematic measurement error. This paper analyses the effect of this measurement error on GDP regressions. There is a systematic difference in the education level between census data and observations constructed from enrolment data. We discuss a methodology for correcting the measurement error. The standard attenuation bias suggests that using these corrected data would lead to a higher coefficient. Our regressions reveal the opposite. We discuss why the measurement error yields an overestimation. Our analysis contributes to an explanation of the difference between regressions based on 5 and on 10 year first-differences.

Personal construct psychologists have suggested various psychological functions explain differences in the stability of constructs. Among these functions are constellatory and loose construction. This paper argues that measurement error is a more parsimonious explanation of the differences
in construct stability reported in these studies.

The conference, 'Measurement Error: Econometrics and Practice' was recently hosted by Aston University and organised jointly by researchers from Aston University and Lund University to highlight the enormous problems caused by measurement error in Economic and Financial data which often go largely unnoticed. Thanks to the sponsorship from Eurostat, a number of distinguished researchers were invited to present keynote lectures. Professor Arnold Zellner from University of Chicago shared his knowledge on measurement error in general; Professor William Barnett from the University of Kansas gave a lecture on implications of measurement error on monetary policy, whilst Dennis Fixler shared his knowledge on how statistical agencies deal with measurement errors. This volume is the result of the selection of high-quality papers presented at the conference and is designed to draw attention to the enormous problem in econometrics of measurement error in data provided by the worlds leading statistical agencies; highlighting consequences of data error and offering solutions to deal with such problems. This volume should appeal to economists, financial analysts and practitioners interested in studying and solving economic problems and building econometric models in everyday operations