DOCSLIB.ORG

Artefact Corrections in Meta-Analysis

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Artefact Corrections in Meta-Analysis Postprint Advances in Methods and Practices in Psychological Science XX–XX Obtaining Unbiased Results in © The Author(s) 2019 Not the version of record. Meta-Analysis: The Importance of The version of record is available at DOI: XXX www.psychologicalscience.org/AMPPS Correcting for Statistical Artefacts Brenton M. Wiernik1 and Jeffrey A. Dahlke2 1Department of Psychology, University of South Florida, Tampa, FL, USA 2Human Resources Research Organization, Alexandria, VA, USA Abstract Most published meta-analyses address only artefactual variance due to sampling error and ignore the role of other statistical and psychometric artefacts, such as measurement error (due to factors including unreliability of meas- urements, group misclassification, and variable treatment strength) and selection effects (including range restriction/enhancement and collider biases). These artefacts can have severe biasing effects on the results of individual studies and meta-analyses. Failing to account for these artefacts can lead to inaccurate conclusions about the mean effect size and between-studies effect-size heterogeneity, and can influence the results of meta-regression, publication bias, and sensitivity analyses. In this paper, we provide a brief introduction to the biasing effects of measurement error and selection effects and their relevance to a variety of research designs. We describe how to estimate the effects of these artefacts in different research designs and correct for their impacts in primary studies and meta-analyses. We consider meta-analyses of correlations, observational group differences, and experimental effects. We provide R code to implement the corrections described. Keywords psychometric meta-analysis; measurement error; reliability; range restriction; range variation; selection bias; collider bias Received 11/06/18; Revision accepted 10/02/19 Meta-analysis is a critical tool for increasing the rigor of describe how to estimate the effects of these artefacts in research syntheses by increasing confidence that ap- different research designs and how to correct for their parent differences in findings across samples are not impacts in meta-analyses. We consider meta-analyses merely attributable to statistical artefacts (Schmidt, of correlations, observational group differences, and ex- 2010). However, most published meta-analyses are perimental effects. concerned only with artefactual variance due to sam- As noted above, most published meta-analyses are pling error and ignore the role of other statistical and concerned only with sampling error and do not correct psychometric artefacts, such as measurement error for other statistical artefacts. For example, of the 71 (stemming from factors including unreliability of meta-analyses published in Psychological Bulletin dur- measurements, group misclassification, and variable ing 2016–2018, only 6 made corrections for measure- treatment strength) and selection effects (including range restriction/enhancement and collider bias). In Corresponding Author: this paper, we provide a brief introduction to the bias- Brenton M. Wiernik, Department of Psychology, University of South Florida, 4202 East Fowler Avenue, ing effects of measurement error and selection effects PCD4118G, Tampa, FL 33620, USA. and their relevance to a variety of research designs. We Email: [email protected] 2 Wiernik et al. ment error and only 1 corrected for selection biases (a grouping variables, measurement error is also called similar review by Schmidt, 2010, found similarly low misclassification (Alexander et al., 2014a). rates of corrections for statistical artefacts). Corrections Measurement error can come from a variety of for measurement error are commonly applied in indus- sources, including truly random response errors (e.g., trial–organizational (I–O) psychology meta-analyses, momentary distractions, arbitrary choices between but rarely in other psychology subfields or other adjacent scale points), transcription errors, transi- disciplines (Schmidt & Hunter, 2015; Schmidt, Le, & ent/temporal effects (e.g., performing poorly on a cog- Oh, 2009). Corrections for selection effects are even nitive assessment due to fatigue on a particular day), rarer. They are typically only performed in meta-anal- environmental effects, content sampling effects (i.e., yses of personnel selection and educational admissions the specific items or content used in the measure do not research (Dahlke & Wiernik, 2019a; Sackett & Yang, function the same way as possible alternative items or 2000). Measurement error and selection artefacts have content), rater effects (e.g., differential knowledge/ severe biasing effects on the results of individual motivation or idiosyncratic beliefs across raters), and studies and meta-analyses. Failing to account for these low sensitivity or specificity of measurement instru- artefacts can lead to inaccurate conclusions about the ments, among others. These diverse sources of error mean effect size and between-studies effect-size hetero- can be grouped into four “classical” categories geneity, and can influence the results of meta-regres- (Cronbach, 1947; Schmidt et al., 2003); see Table 1 for sion, publication bias, and sensitivity analyses. When descriptions. Importantly, measurement error is not measurement error and selection effects are considered only about whether participants’ responses were cor- in meta-analyses outside of I–O psychology, they are rectly recorded, but more about the process that gener- often treated simply as indicators of general “study ated the responses and whether the responses would be quality” and used as exclusion criteria or as moderators the same if the process were repeated at different times, (Schmidt & Hunter, 2015, p. 485); both approaches are by different raters, or with a different instrument or suboptimal, as measurement error and selection effects item set (Schmidt & Hunter, 1996). For psychological have predictable effects on study results. The best way measures, random response error and transient error to handle measurement error and selection effects is to are typically the largest sources of measurement error apply statistical corrections that account for the known (Ones et al., 2016). impacts of these artefacts. Below, we describe the The amount of measurement error in a sample of impacts of measurement error and selection effects on scores is quantified using a reliability coefficient, de- primary studies and meta-analyses, as well as methods fined as the proportion of the observed-score variance to correct for these artefacts. that is consistent (i.e., believed to be “true”), or one minus the proportion of the observed-score variance Measurement Error attributable to measurement error: Measurement error is an artefact that causes observed 2 2 ����� ������ 1 ′ (i.e., measured) values to deviate from the “true” values �� 2 2 = ���� = 1 − ���� of underlying latent variables (Schmidt & Hunter, Conceptually, the reliability coefficient is the correla- 1 1996). For example, consider a political psychologist tion between two parallel measures of a construct. The assessing political orientation using a 10-item measure with items rated on a 7-point scale. A respondent might 1 In classical test theory, an individual’s “true score” is the obtain a mean score of “5” (somewhat conservative) or expected value of the individuals’ response to a measurement “3” (somewhat liberal) across the 10 items, when their process that has been repeated an infinite number of times, such that true score is in fact “4” (moderate/centrist). Measure- all errors of measurement average out to zero across observations. That it, it the “true score” is the score on the measure without any ment error is also called unreliability (Rogers et al., measurement error. The “true score” does not necessarily 2002), observational error (Walter, 1983), and infor- correspond to the individuals’ standing on the intended latent mation bias (Alexander et al., 2014a). For continuous construct—that is a question of measure validity, not reliability. If a variables, measurement error is also called (low) preci- measure has poor validity as an indicator of the intended construct, sion (Stallings & Gillmore, 1971); for dichotomous or even correlations corrected for measurement error will poorly reflect correlations with the latent construct variable (for discussions, see Borsboom, 2006; Schmidt et al., 2013). Correcting for Statistical Artefacts 3 Table 1. Four classical sources of measurement error and reliability estimators sensitive to each. Source of error Description and examples Random response error Truly random error specific to each item/response (e.g., random fluctuations in response time, momentary lapses in attention); unique to each measurement Reliability estimators: – All reliability estimators Transient error Error due the specific time or environment in which data are gathered (e.g., participant mood, environmental distractions); shared by measures completed within a meaningfully short time span Reliability estimators: – Test-retest reliability (coefficient of stability) – Delayed parallel forms reliability (coefficient of equivalence and stability) – Delayed internal consistency (e.g., test-retest alpha; Green, 2003) Content sampling error Error due to specific content used on a measure (e.g., interpretations of specific (item/test specific factor error) test items); shared by measures with the same or highly similar content Reliability estimators: – Parallel forms reliability
Load more

Recommended publications
  • An Introduction to Psychometric Theory with Applications in R
    What is psychometrics? What is R? Where did it come from, why use it? Basic statistics and graphics TOD An introduction to Psychometric Theory with applications in R William Revelle Department of Psychology Northwestern University Evanston, Illinois USA February, 2013 1 / 71 What is psychometrics? What is R? Where did it come from, why use it? Basic statistics and graphics TOD Overview 1 Overview Psychometrics and R What is Psychometrics What is R 2 Part I: an introduction to R What is R A brief example Basic steps and graphics 3 Day 1: Theory of Data, Issues in Scaling 4 Day 2: More than you ever wanted to know about correlation 5 Day 3: Dimension reduction through factor analysis, principal components analyze and cluster analysis 6 Day 4: Classical Test Theory and Item Response Theory 7 Day 5: Structural Equation Modeling and applied scale construction 2 / 71 What is psychometrics? What is R? Where did it come from, why use it? Basic statistics and graphics TOD Outline of Day 1/part 1 1 What is psychometrics? Conceptual overview Theory: the organization of Observed and Latent variables A latent variable approach to measurement Data and scaling Structural Equation Models 2 What is R? Where did it come from, why use it? Installing R on your computer and adding packages Installing and using packages Implementations of R Basic R capabilities: Calculation, Statistical tables, Graphics Data sets 3 Basic statistics and graphics 4 steps: read, explore, test, graph Basic descriptive and inferential statistics 4 TOD 3 / 71 What is psychometrics? What is R? Where did it come from, why use it? Basic statistics and graphics TOD What is psychometrics? In physical science a first essential step in the direction of learning any subject is to find principles of numerical reckoning and methods for practicably measuring some quality connected with it.
    [Show full text]
  • Fatigue and Psychological Distress in the Working Population Psychometrics, Prevalence, and Correlates
    Journal of Psychosomatic Research 52 (2002) 445–452 Fatigue and psychological distress in the working population Psychometrics, prevalence, and correlates Ute Bu¨ltmanna,*, Ijmert Kanta, Stanislav V. Kaslb, Anna J.H.M. Beurskensa, Piet A. van den Brandta aDepartment of Epidemiology, Maastricht University, P.O. Box 616, 6200 MD Maastricht, The Netherlands bDepartment of Epidemiology and Public Health, Yale University School of Medicine, New Haven, CT, USA Received 11 October 2000 Abstract Objective: The purposes of this study were: (1) to explore the distinct pattern of associations was found for fatigue vs. relationship between fatigue and psychological distress in the psychological distress with respect to demographic factors. The working population; (2) to examine associations with demographic prevalence was 22% for fatigue and 23% for psychological and health factors; and (3) to determine the prevalence of fatigue distress. Of the employees reporting fatigue, 43% had fatigue only, and psychological distress. Methods: Data were taken from whereas 57% had fatigue and psychological distress. Conclusions: 12,095 employees. Fatigue was measured with the Checklist The results indicate that fatigue and psychological distress are Individual Strength, and the General Health Questionnaire (GHQ) common in the working population. Although closely associated, was used to measure psychological distress. Results: Fatigue was there is some evidence suggesting that fatigue and psychological fairly well associated with psychological distress. A separation
    [Show full text]
  • When Psychometrics Meets Epidemiology and the Studies V1.Pdf
    When psychometrics meets epidemiology, and the studies which result! Tom Booth [email protected] Lecturer in Quantitative Research Methods, Department of Psychology, University of Edinburgh Who am I? • MSc and PhD in Organisational Psychology – ESRC AQM Scholarship • Manchester Business School, University of Manchester • Topic: Personality psychometrics • Post-doctoral Researcher • Centre for Cognitive Ageing and Cognitive Epidemiology, Department of Psychology, University of Edinburgh. • Primary Research Topic: Cognitive ageing, brain imaging. • Lecturer Quantitative Research Methods • Department of Psychology, University of Edinburgh • Primary Research Topics: Individual differences and health; cognitive ability and brain imaging; Psychometric methods and assessment. Journey of a talk…. • Psychometrics: • Performance of likert-type response scales for personality data. • Murray, Booth & Molenaar (2015) • Epidemiology: • Allostatic load • Measurement: Booth, Starr & Deary (2013); (Unpublished) • Applications: Early life adversity (Unpublished) • Further applications Journey of a talk…. • Methodological interlude: • The issue of optimal time scales. • Individual differences and health: • Personality and Physical Health (review: Murray & Booth, 2015) • Personality, health behaviours and brain integrity (Booth, Mottus et al., 2014) • Looking forward Psychometrics My spiritual home… Middle response options Strongly Agree Agree Neither Agree nor Disagree Strong Disagree Disagree 1 2 3 4 5 Strongly Agree Agree Unsure Disagree Strong Disagree
    [Show full text]
  • Reliability” Within the Qualitative Methodological Literature As a Proxy for the Concept of “Error” Within the Qualitative Research Domain
    The “Error” in Psychology: An Analysis of Quantitative and Qualitative Approaches in the Pursuit of Accuracy by Donna Tafreshi MA, Simon Fraser University, 2014 BA, Simon Fraser University, 2011 Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Department of Psychology Faculty of Arts and Social Sciences © Donna Tafreshi 2018 SIMON FRASER UNIVERSITY Fall 2018 Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation. Approval Name: Donna Tafreshi Degree: Doctor of Philosophy Title: The “Error” in Psychology: An Analysis of Quantitative and Qualitative Approaches in the Pursuit of Accuracy Examining Committee: Chair: Thomas Spalek Professor Kathleen Slaney Senior Supervisor Professor Timothy Racine Supervisor Professor Susan O’Neill Supervisor Professor Barbara Mitchell Internal Examiner Professor Departments of Sociology & Gerontology Christopher Green External Examiner Professor Department of Psychology York University Date Defended/Approved: October 12, 2018 ii Abstract The concept of “error” is central to the development and use of statistical tools in psychology. Yet, little work has focused on elucidating its conceptual meanings and the potential implications for research practice. I explore the emergence of uses of the “error” concept within the field of psychology through a historical mapping of its uses from early observational astronomy, to the study of social statistics, and subsequently to its adoption under 20th century psychometrics. In so doing, I consider the philosophical foundations on which the concepts “error” and “true score” are built and the relevance of these foundations for its usages in psychology.
    [Show full text]
  • Governing Collective Action in the Face of Observational Error
    Governing Collective Action in the Face of Observational Error Thomas Markussen, Louis Putterman and Liangjun Wang* August 2017 Abstract: We present results from a repeated public goods experiment where subjects choose by vote one of two sanctioning schemes: peer-to-peer (informal) or centralized (formal). We introduce, in some treatments, a moderate amount of noise (a 10 percent probability that a contribution is reported incorrectly) affecting either one or both sanctioning environments. We find that the institution with more accurate information is always by far the most popular, but noisy information undermines the popularity of peer-to-peer sanctions more strongly than that of centralized sanctions. This may contribute to explaining the greater reliance on centralized sanctioning institutions in complex environments. JEL codes: H41, C92, D02 Keywords: Public goods, sanctions, information, institution, voting * University of Copenhagen, Brown University, and Zhejiang University of Technology, respectively. We would like to thank The National Natural Science Foundation of China (Grant No.71473225) and The Danish Council for Independent Research (DFF|FSE) for funding. We thank Xinyi Zhang for assistance in checking translation of instructions, and Vicky Zhang and Kristin Petersmann for help in checking of proofs. We are grateful to seminar participants at University of Copenhagen, at the Seventh Biennial Social Dilemmas Conference held at University of Massachusetts, Amherst in June 2017, and to Attila Ambrus, Laura Gee and James Walker for helpful comments. 1 Introduction When and why do groups establish centralized, formal authorities to solve collective action problems? This fundamental question was central to classical social contract theory (Hobbes, Locke) and has recently been the subject of a group of experimental studies (Okada, Kosfeld and Riedl 2009, Traulsen et al.
    [Show full text]
  • The Role of Observational Errors in Optimum Interpolation Analysis :I0
    U.S. DEPARTMENT OF COMMERCE NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION NATIONAL WEATHER SERVICE NATIONAL METEOROLOGICAL CENTER OFFICE NOTE 175 The Role of Observational Errors in Optimum Interpolation Analysis :I0. K. H. Bergman Development Division APRIL 1978 This is an unreviewed manuscript, primarily intended for informal exchange of information among NMC staff members. k Iil The Role of Observational Errors in Optimum Interpolation Analysis 1. Introduction With the advent of new observing systems and the approach of the First GARP Global Experiment (FGGE), it appears desirable for data providers to have some knowledge of how their product is to be used in the statistical "optimum interpolation" analysis schemes which will operate at most of the Level III centers during FGGE. It is the hope of the author that this paper will serve as a source of information on the way observational data is handled by optimum interpolation analysis schemes. The theory of optimum interpolation analysis, especially with regard to the role of observational errors, is reviewed in Section 2. In Section 3, some comments about determining observational error characteristics are given along with examples of same. Section 4 discusses error-checking procedures which are used to identify suspect observations. These latter procedures are an integral part of any analysis scheme. 2. Optimum Interpolation Analysis The job of the analyst is to take an irregular distribution of obser- vations of variable quality and obtain the best possible estimate of a meteorological field at a regular network of grid points. The optimum interpolation analysis scheme attempts to accomplish this goal by mini- mizing the mean square interpolation error for a large ensemble of analysis situations.
    [Show full text]
  • Monthly Weather Review James E
    DEPARTMENT OF COMMERCE WEATHERBUREAU SINCLAIR WEEKS, Secretary F. W. REICHELDERFER, Chief MONTHLY WEATHER REVIEW JAMES E. CASKEY, JR., Editor Volume 83 Cloeed Febru 15, 1956 Number 12 DECEMBER 1955 Issued Mm% 15, 1956 A SIMULTANEOUS LAGRANGIAN-EULERIAN TURBULENCE EXPERIMENT F'RA.NK GIFFORD, R. U. S. Weather Bureau The Pennsylvania State University * [Manuscript received October 28, 1955; revised January 9, 1956J ABSTRACT Simultaneous measurements of turbulent vertical velocity fluctuationsat a height of 300 ft. measured by means of a fixed anemometer, small floating balloons, and airplane gust equipment at Brookhaven are presented. The re- sulting Eulerian (fixed anemometer) turbulence energy mectra are similar to the Lagrangian (balloon) spectra but " - with a displacement toward higher frequencies. 1. THE NATURE OF THEPROBLEM function, and the closely related autocorrelation function. This paper reports an attempt to make measurements of The term Eulerian, in turbulence study, impliescon- both Lagrangian and Eulerianturbulence fluctuations sideration of velocity fluctuations at a point or points simultaneously. This problem owes its importance largely fixed in space, as measured for example by one or more to the fact that experimenters usually measure the Eu- fixed anemometers or wind vanes. The term Lagrangian lerian-time form of the turbulence spectrumor correlation, implies study of the fluctuations of individual fluid par- which all 6xed measuring devices of turbulence record; cels; these are very difficult to measure. In a sense, these but theoretical developments as well as practical applica- two points of view of the turbulence phenomenon are tionsinevitably require the Eulerian-space, or the La- closely related tothe two most palpable physical manifesta- grangian form.
    [Show full text]
  • Properties and Psychometrics of a Measure of Addiction Recovery Strengths
    bs_bs_banner REVIEW Drug and Alcohol Review (March 2013), 32, 187–194 DOI: 10.1111/j.1465-3362.2012.00489.x The Assessment of Recovery Capital: Properties and psychometrics of a measure of addiction recovery strengths TEODORA GROSHKOVA1, DAVID BEST2 & WILLIAM WHITE3 1National Addiction Centre, Institute of Psychiatry, King’s College London, London, UK, 2Turning Point Drug and Alcohol Centre, Monash University, Melbourne, Australia, and 3Chestnut Health Systems, Bloomington, Illinois, USA Abstract Introduction and Aims. Sociological work on social capital and its impact on health behaviours have been translated into the addiction field in the form of ‘recovery capital’ as the construct for assessing individual progress on a recovery journey.Yet there has been little attempt to quantify recovery capital.The aim of the project was to create a scale that assessed addiction recovery capital. Design and Methods. Initial focus group work identified and tested candidate items and domains followed by data collection from multiple sources to enable psychometric assessment of a scale measuring recovery capital. Results. The scale shows moderate test–retest reliability at 1 week and acceptable concurrent validity. Principal component analysis determined single factor structure. Discussion and Conclusions. The Assessment of Recovery Capital (ARC) is a brief and easy to administer measurement of recovery capital that has acceptable psychometric properties and may be a useful complement to deficit-based assessment and outcome monitoring instruments for substance dependent individuals in and out of treatment. [Groshkova T, Best D, White W. The Assessment of Recovery Capital: Properties and psychometrics of a measure of addiction recovery strengths. Drug Alcohol Rev 2013;32:187–194] Key words: addiction, recovery capital measure, assessment, psychometrics.
    [Show full text]
  • PSY610: PSYCHOMETRICS/STATISTICS FALL, 2006 Dr. Gregory T. Smith Teaching Assistant: 105 Kastle Hall Lauren Gudonis 257-6454 L
    PSY610: PSYCHOMETRICS/STATISTICS FALL, 2006 Dr. Gregory T. Smith Teaching Assistant: 105 Kastle Hall Lauren Gudonis 257-6454 [email protected] [email protected] office: 111 J Office Hours: Monday and Wednesday 11:00 - 11:45 and by appt. Required Texts: 1. Keppel, G. (2004). Design and Analysis: A Researcher’s Handbook. (4th Edition). Upper Saddle River, N.J.: Pearson/Prentice-Hall, Inc. Course Outline 8/24 1. Introduction. Independent and dependent variables. Correlational and experimental studies. Implications for interpretation of results. Example. Randomization and Random Sampling. 8/29 2. Levels of Measurement. Examples. Relation of levels of measurement to statistical tests. Chart. Three major questions to be answered by statistical analysis: existence, nature, and magnitude of relationships between X and Y (or Y1 and Y2). Introduction to probability. 8/31 3. Probability: an intuitive approach and a formal approach. Independence and conditional probabilities. Introduction to the logic of hypothesis testing. 9/5 4. The logic of hypothesis testing. Introduction to parametric (normal distribution) techniques. One sample goodness of fit test z. Sampling distribution of the mean. Central limit theorem. Standard error of the mean. Lab #1 Introduction to lab. Probability. Definitions, examples, and exercises. Introduction to computer system. Homework #1 assigned. 9/7 5. Logic of hypothesis testing, decision rule, terms, power, Type I and Type II errors. 9/12 6. Student's t test: sampling distribution, procedures, assumptions, degrees of freedom. Unbiased and biased estimates. Directional v. nondirectional tests, confidence intervals, t and z compared. Lab #2 Review material from preceding two lectures; homework; SPSS exercises. 9/14 7.
    [Show full text]
  • Arxiv:2103.01707V1 [Astro-Ph.EP] 2 Mar 2021
    Draft version March 3, 2021 Typeset using LATEX preprint2 style in AASTeX63 Mass and density of asteroid (16) Psyche Lauri Siltala1 and Mikael Granvik1, 2 1Department of Physics, P.O. Box 64, FI-00014 University of Helsinki, Finland 2Asteroid Engineering Laboratory, Onboard Space Systems, Lule˚aUniversity of Technology, Box 848, S-98128 Kiruna, Sweden (Received October 1, 2020; Revised February 23, 2021; Accepted February 23, 2021) Submitted to ApJL ABSTRACT We apply our novel Markov-chain Monte Carlo (MCMC)-based algorithm for asteroid mass estimation to asteroid (16) Psyche, the target of NASA's eponymous Psyche mission, based on close encounters with 10 different asteroids, and obtain a mass of −11 (1:117 0:039) 10 M . We ensure that our method works as expected by applying ± × it to asteroids (1) Ceres and (4) Vesta, and find that the results are in agreement with the very accurate mass estimates for these bodies obtained by the Dawn mission. We then combine our mass estimate for Psyche with the most recent volume estimate to compute the corresponding bulk density as (3:88 0:25) g/cm3. The estimated ± bulk density rules out the possibility of Psyche being an exposed, solid iron core of a protoplanet, but is fully consistent with the recent hypothesis that ferrovolcanism would have occurred on Psyche. Keywords: to be | added later 1. INTRODUCTION and radar observations (Shepard et al. 2017). Asteroid (16) Psyche is currently of great in- There have, however, been concerns that Psy- terest to the planetary science community as che's relatively high bulk density of approxi- 3 it is the target of NASA's eponymous Psyche mately 4 g/cm is still too low to be consis- mission currently scheduled to be launched in tent with iron meteorites and that the asteroid 2022 (Elkins-Tanton et al.
    [Show full text]
  • Factor Analysis 1 Factor Analysis
    Factor analysis 1 Factor analysis Factor analysis is a statistical method used to describe variability among observed, correlated variables in terms of a potentially lower number of unobserved variables called factors. In other words, it is possible, for example, that variations in three or four observed variables mainly reflect the variations in fewer such unobserved variables. Factor analysis searches for such joint variations in response to unobserved latent variables. The observed variables are modeled as linear combinations of the potential factors, plus "error" terms. The information gained about the interdependencies between observed variables can be used later to reduce the set of variables in a dataset. Computationally this technique is equivalent to low rank approximation of the matrix of observed variables. Factor analysis originated in psychometrics, and is used in behavioral sciences, social sciences, marketing, product management, operations research, and other applied sciences that deal with large quantities of data. Factor analysis is related to principal component analysis (PCA), but the two are not identical. Latent variable models, including factor analysis, use regression modelling techniques to test hypotheses producing error terms, while PCA is a descriptive statistical technique[1]. There has been significant controversy in the field over the equivalence or otherwise of the two techniques (see exploratory factor analysis versus principal components analysis). Statistical model Definition Suppose we have a set of observable random variables, with means . Suppose for some unknown constants and unobserved random variables , where and , where , we have Here, the are independently distributed error terms with zero mean and finite variance, which may not be the same for all .
    [Show full text]
  • Error Statistics of Asteroid Optical Astrometric Observations
    Icarus 166 (2003) 248–270 www.elsevier.com/locate/icarus Error statistics of asteroid optical astrometric observations Mario Carpino,a,∗ Andrea Milani,b andStevenR.Chesleyc a Osservatorio Astronomico di Brera, via Brera, 28, 20121 Milan, Italy b Dipartimento di Matematica, Università di Pisa, Via Buonarroti, 2, 56127 Pisa, Italy c Navigation & Mission Design Section, MS 301-150, Jet Propulsion Laboratory, Pasadena, CA 91109, USA Received 29 May 2001; revised 25 October 2002 Abstract Astrometric uncertainty is a crucial component of the asteroid orbit determination process. However, in the absence of rigorous uncer- tainty information, only very crude weighting schemes are available to the orbit computer. This inevitably leads to a correspondingly crude characterization of the orbital uncertainty and in many cases to less accurate orbits. In this paper we describe a method for carefully assessing the statistical performance of the various observatories that have produced asteroid astrometry, with the ultimate goal of using this statistical characterization to improve asteroid orbit determination. We also include a detailed description of our previously unpublished automatic outlier rejection algorithm used in the orbit determination, which is an important component of the fitting process. To start this study we have determined the best fitting orbits for the first 17,349 numbered asteroids and computed the corresponding O–C astrometric residuals for all observations used in the orbit fitting. We group the residuals into roughly homogeneous bins and compute the root mean square (RMS) error and bias in declination and right ascension for each bin. Results are tabulated for each of the 77 bins containing more than 3000 observations; this comprises roughly two thirds of the data, but only 2% of the bins.
    [Show full text]