Showing papers by "Ingram Olkin published in 1987"
TL;DR: In this paper, the authors present a model for estimating the effect size from a series of experiments using a fixed effect model and a general linear model, and combine these two models to estimate the effect magnitude.
Abstract: Preface. Introduction. Data Sets. Tests of Statistical Significance of Combined Results. Vote-Counting Methods. Estimation of a Single Effect Size: Parametric and Nonparametric Methods. Parametric Estimation of Effect Size from a Series of Experiments. Fitting Parametric Fixed Effect Models to Effect Sizes: Categorical Methods. Fitting Parametric Fixed Effect Models to Effect Sizes: General Linear Models. Random Effects Models for Effect Sizes. Multivariate Models for Effect Sizes. Combining Estimates of Correlation Coefficients. Diagnostic Procedures for Research Synthesis Models. Clustering Estimates of Effect Magnitude. Estimation of Effect Size When Not All Study Outcomes Are Observed. Meta-Analysis in the Physical and Biological Sciences. Appendix. References. Index.
7,063 citations
TL;DR: In this paper, a more efficient method is obtained by using Householder transformations, based on the product of orthogonal matrices, each of which represents an angle of rotation.
Abstract: In order to generate a random orthogonal matrix distributed according to Haar measure over the orthogonal group it is natural to start with a matrix of normal random variables and then factor it by the singular value decomposition. A more efficient method is obtained by using Householder transformations. We propose another alternative based on the product of ${{n(n - 1)}/2}$ orthogonal matrices, each of which represents an angle of rotation. Some numerical comparisons of alternative methods are made.
116 citations
TL;DR: In this article, a combination of the parametric and nonparametric estimates is used to fit a density function, where π (0 ≤ π ≤ 1) is unknown and π is estimated from the data, and its estimate is then used in as the proposed density estimate.
Abstract: One method of fitting a parametric density function f(x, θ) is first to estimate θ by maximum likelihood, say, and then to estimate f(x, θ) by On the other hand, when the parametric model does not hold, the true density f(x) may be estimated nonparametrically, as in the case of a kernel estimate The key idea proposed is to fit a combination of the parametric and nonparametric estimates, namely where π (0 ≤ π ≤ 1) is unknown The parameter π is estimated from the data, and its estimate is then used in as the proposed density estimate The main point is that we expect to be close to unity when the parametric model prevails, and close to zero when it does not hold We show that, under certain conditions, converges to the parametric density when the parametric model holds and approaches the true f(x) when the parametric model does not hold The procedure was applied to a number of actual data sets In each case the maximum likelihood estimate was readily obtained and the semiparametric density es
98 citations
34 citations
TL;DR: For a variety of loss functions, the authors showed that XD, where XD is diagonal, is a best equivariant estimator of positive definite matrices, and explicit expressions for XD are provided.
Abstract: Every positive definite matrix $\Sigma$ has a unique Cholesky decomposition $\Sigma = \theta\theta'$, where $\theta$ is lower triangular with positive diagonal elements. Suppose that $S$ has a Wishart distribution with mean $n\Sigma$ and that $S$ has the Cholesky decomposition $S = XX'$. We show, for a variety of loss functions, that $XD$, where $D$ is diagonal, is a best equivariant estimator of $\theta$. Explicit expressions for $D$ are provided.
28 citations
TL;DR: Morris Hansen as mentioned in this paper was a statistician at the United States Census Bureau and became a member of the National Academy of Sciences in 1986. The following conversation took place in late May 1986 in Washington, D. C.
Abstract: Morris Hansen was born on December 15, 1910 in Thermopolis, Wyoming. He entered the University of Wyoming in 1928. After a working leave for a year he returned to the University in 1930, and received his BS degree in accounting in 1934. After graduating, he joined the Bureau of the Census in Washington, D. C. He studied part time for several years at the graduate school of the Department of Agriculture and American University, and received his master's degree in statistics from American University in 1940; he became Chief of the Statistical Research Division at the Bureau of the Census in 1947, Assistant Director for Statistical Standards in 1949 and Associate Director for Research and Development in 1961. He retired from the Bureau of the Census in 1968, and joined Westat, Inc., then a small statistical research, consulting and service organization, as Statistical Advisor and Senior Vice President. In 1986 he became Chairman of the Board. He received an honorary Degree of Doctor of Laws from the University of Wyoming in 1959, was President of the Institute of Mathematical Statistics in 1953, President of the American Statistical Association in 1960 and President of the International Association of Survey Statisticians 19731977. He is Honorary Fellow of the Royal Statistical Society, and a member of the National Academy of Sciences. The following conversation took place in late May 1986 in Washington, D. C.
8 citations
TL;DR: Albert Bowker as discussed by the authors was the first Assistant Secretary for Postsecondary Education in the newly formed U S Department of Education and became the first Executive Vice President of the City University of New York (CUNY).
Abstract: Albert Bowker was born in Winchendon, Massachusetts, on September 8, 1919 He received a BS in Mathematics from MIT in 1941, and a PhD in Mathematical Statistics from Columbia University in 1949 He was on the Stanford faculty from 1947 to 1963, serving as founding Chairman of the Statistics Department and Dean of the Graduate Division In 1963, he became Chancellor of the City University of New York He returned to California in 1971 as Chancellor of the University of California at Berkeley In 1980 he was appointed as the first Assistant Secretary for Postsecondary Education in the newly formed U S Department of Education In 1981 he went to the University of Maryland as founding Dean of the School of Public Affairs and later became Executive Vice President In September 1986, he returned to the City University of New York, and now serves as Vice-President for Planning of its Research Foundation In 1961-1962, he was president of the Institute of Mathematical Statistics, and in 1964, president of the American Statistical Association Honors include the Frederick Douglass Award of the New York Urban League; the Medal for Distinguished Service of Teachers College, Columbia University; Shewhart Award of the American Society for Quality Control; Berkeley Citation; Distinguished Public Service Award, Department of the Navy; Order De Leopold II; and honorary degrees from the City University of New York, University of the State of New York (Regents), Brandeis University and Antioch University He has been a member of the boards of various professional and educational organizations including MIT, the University of Haifa and Bennington College The following conversation took place in his home in Washington, D C in October 1986
2 citations
01 Jan 1987
TL;DR: The overlap hypothesis as discussed by the authors assumes that there is no correlation between X 1 and the gain in height, X 2 - X 1, or that the correlation between x 1 and X 2 is the same as that between X1 and X 1 - X 2.
Abstract: For prototype growth data suppose that X t represents the height of an individual at time t. The overlap hypothesis asserts that there is no correlation between X 1 and the gain in height, X 2 - X 1, or that the correlation between X 1 and X 2 is the same as that between X 1 and X 1 - X 2. An estimate of the correlation when either of these hypotheses holds is obtained. A statistical test for the hypothesis is also developed, and indications of a multivariate version are presented.
1 citations