The Kernel Method of Test Equating by Alina A. von Davier

By Alina A. von Davier

Kernel Equating (KE) is a strong, smooth and unified method of try out equating. it's according to a versatile relatives of equipercentile-like equating capabilities and comprises the linear equating functionality as a different case. Any equipercentile equating technique has 5 steps or components. they're: 1) pre-smoothing; 2) estimation of the score-probabilities at the objective inhabitants; three) continuization; four) computing and diagnosing the equating functionality; five) computing the traditional mistakes of equating and comparable accuracy measures. KE brings those steps jointly in an prepared complete instead of treating them as disparate problems.

KE exploits pre-smoothing through becoming log-linear types to attain information, and contains it into step five) above. KE presents new instruments for diagnosing a given equating functionality, and for evaluating or extra equating features so one can choose from them. during this e-book, KE is utilized to the 4 significant equating designs and to either Chain Equating and Post-Stratification Equating for the Non-Equivalent teams with Anchor try Design.

This e-book may be a big reference for a number of teams: (a) Statisticians and others drawn to the idea in the back of equating equipment and using model-based statistical equipment for facts smoothing in utilized paintings; (b) Practitioners who have to equate tests—including people with those tasks in trying out businesses, nation checking out companies and faculty districts; and (c) teachers in psychometric and dimension courses. The authors suppose a few familiarity with linear and equipercentile try out equating, and with matrix algebra.

Alina von Davier is an affiliate learn Scientist within the heart for Statistical concept and perform, at academic trying out provider. She has been a study collaborator on the Universities of Trier, Magdeburg, and Kiel, an assistant professor on the Politechnical college of Bucharest and a study scientist on the Institute for Psychology in Bucharest.

Paul Holland holds the Frederic M. Lord Chair in dimension and information at academic checking out provider. He held college positions within the Graduate university of schooling, collage of California, Berkeley and the Harvard division of information. he's a Fellow of the yank Statistical organization, the Institute of Mathematical facts, and the yank organization for the development of technology. he's an elected Member of the foreign Statistical Institute and a previous president of the Psychometric society. He was once presented the (AERA/ACT) E. F. Lindquist Award, in 2000, and was once exact a countrywide affiliate of the nationwide Academies of technological know-how in 2002.

Dorothy Thayer presently is a expert within the middle of Statistical concept and perform, at academic checking out provider. Her learn pursuits comprise computational and statistical method, empirical Bayes suggestions, lacking information strategies and exploratory info research recommendations.

Show description

Read or Download The Kernel Method of Test Equating PDF

Best econometrics books

A Guide to Modern Econometrics (2nd Edition)

This hugely winning textual content makes a speciality of exploring replacement recommendations, mixed with a realistic emphasis, A advisor to substitute options with the emphasis at the instinct at the back of the ways and their sensible reference, this re-creation builds at the strengths of the second one version and brings the textual content thoroughly up–to–date.

Contemporary Bayesian Econometrics and Statistics (Wiley Series in Probability and Statistics)

Instruments to enhance determination making in a less than perfect international This e-book presents readers with an intensive realizing of Bayesian research that's grounded within the thought of inference and optimum choice making. modern Bayesian Econometrics and records offers readers with cutting-edge simulation equipment and versions which are used to unravel advanced real-world difficulties.

Handbook of Financial Econometrics, Vol. 1: Tools and Techniques

This selection of unique articles-8 years within the making-shines a brilliant mild on fresh advances in monetary econometrics. From a survey of mathematical and statistical instruments for knowing nonlinear Markov tactics to an exploration of the time-series evolution of the risk-return tradeoff for inventory industry funding, famous students Yacine AГЇt-Sahalia and Lars Peter Hansen benchmark the present nation of data whereas members construct a framework for its development.

Extra resources for The Kernel Method of Test Equating

Example text

In this book we will investigate both CE and PSE methods. In this section we will discuss the assumptions that underlie each of them and the parameters of the final equating functions for both. In Chapters 3, 4, and 5 we show how to apply Kernel Equating to both of these methods. Finally, in Chapters 10 and 11 we will show how to implement CE and PSE in practice. The rest of this section is structured as follows: first, we identify the aspects of the NEAT Design that are common to both CE and PSE.

Data Collection Designs and KY = CσY1 = CσY2 = CσY . 29) The equality of σX1 and σX2 or of σY1 and σY2 follows from his constant order-effect assumption. Lord suggests simple method-of-moment estimates of all the relevant parameters and applies them to the linear equating function. Our mixture approach would produce almost the same estimated linear equating function as the one that Lord proposed (when wX = wY = 12 ) except for the standard deviation estimates, σ ˆX and σ ˆY . His are based on 2 = 12 (s2X1 + s2X2 ) and σ ˆY2 = 12 (s2Y1 + s2Y2 ).

L. 3 describes the issues that arise with an internal anchor test. We finish with a section on the special case of an EG Design that also has an anchor test. , P = Q. 1 Chain Equating (CE) Chain Equating may be the oldest method used for the NEAT Design. It is a simple extension of the ideas already present in the SG Design. In Livingston et al. (1990), CE was shown to give reasonable results compared to the other standard methods of treating the NEAT Design. Chain Equating (CE) uses a two-stage transformation of X scores into Y scores.

Download PDF sample

Rated 4.15 of 5 – based on 23 votes