-
作者：BARNARD, GA
摘要：The note points up the areas of agreement and disagreement between the 2 papers on 2 X 2 tables by Barnard and by Pearson in the same issue of Biometrika.
-
作者：PEARSON, ES
摘要：In an effort to end the controversy which has raged since 1941, Pearson, following Barnard, has differentiated between 3 distinct situations in which data may be classed in a 2 X 2 table. The 3 situations produce tables in which (a) both margins are fixed, (b) one is fixed and (c) neither is fixed. Corresponding to the first, the appropriate test is Fisher''s exact distribution. Corresponding to the others, Pearson suggests a compound probability consisting of the probability of the observed m...
-
作者：WISHART, J
摘要：Explicit expressions for the exact cumulants of Fisher''s z-distribution are given. Cumulants for the logarithmic x 2-distribution are first found (the same result was published by Bartlett and Kendall in 1946) and the cumulants of 2 obtained by considering the z-distribution as the difference between two logarithmic x 2-distributions. Cumulants of the logarithmic t-distribution are obtained by specializing the z-distribution.
-
作者：WELCH, BL
摘要：A population parameter [eta] is estimated by a statistic [gamma], normally distributed about [eta] with variance [image], where the [sigma]i are positive numbers and the [sigma]2 are unknown variances. If independent estimates of these variances are provided by the observed data, it is possible to make probability statements about [gamma] similar to those Student gave for the mean of a single sample. The form of the exact solution is given and a series solution is developed. The series converg...
-
作者：BARNARD, GA
摘要：In considering a significance level as a probability, it is necessary to isolate a class of events having the given probability. Some of the difficulties in this isolation and identification are pointed out in connection with a particular problem.
-
作者：SILLITTO, GP
摘要：The distribution of [tau] previously given by Kendall covers the situation in which neither ranking contains ties. This paper supplements Kendall''s work for the case in which one (only) of the rankings contains ties. Formulae are given for the maximum score S (in Kendall''s sense) and for the variance of the distribution of scores. A table is given of the probability distribution of S for up to 10 ranked pairs and all numbers of paired and triplet rankings.
-
作者：PLACKETT, RL
摘要：Upper limits for the ratio of mean range to standard deviation in samples of n from any population are given by [image]. For large n, the upper limit is approx. [image].
-
作者：BARNARD, GA
摘要：Various types of expt. giving rise to results in the form of a 2 X 2 table are discussed and a test is developed for one type. The type considered is that in which samples of m and n are drawn from 2 populations and each element classified in one of 2 categories. The hypothesis tested is that the proportions in one category are the same in both populations. In this situation there is no clear-cut ordering of the sample points. An ordering is developed on the 3 criteria of convexity, sytn-metry...
-
作者：GARWOOD, F
摘要：The general problem in the mathematical study of bombing consists of estimating the overall effect on a given objective. The problem is further restricted to calculation of the mean and variance of the portion of the objective area covered by destruction areas dropped in random fashion on a section of a plane including the objective area in its interior. For the areas considered (rectangles and circles), calculation of the mean presents no great difficulty. Methods, due to H. E. Robbins and to...
-
作者：LORD, E
摘要：The t-test compares the deviation of a sample mean from the assumed population value with a root-mean-square estimate of the standard error of that difference in samples from a normal population. For routine testing of exptl. data of various types, particularly data for control of quality of manufactured products, it is desirable to have a test of this nature which requires less computation. The ratio of the observed range to the known mean range in samples from a normal population with unit s...
