摘要:

A course in writing and critical appraisal of medical papers that uses statistics is described, and its relationship to the goal of better integrating scientific writing into the statistics curriculum is discussed. It is concluded that writing should play an increased role in statistical education and that this can best be accomplished by distributing exercises in writing and critical appraisal throughout the curriculum. Writing assignments, such as simulated practice in producing components of scientific papers and grants, should reflect students' likely uses of statistics.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476037

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

The problem of constructing a confidence region for simultaneously estimatingp,p≥ 2, linear regression parameters for which confidence statements can be made on the individual parameters is revisited. Here, an intercept may be included among thepparemeters. The technique is due to Working and Hotelling (1929) and Scheffè (1959) and uses thepseparate projections of the exact (1 – α) 100% confidence ellipsoid (ellipse ifp= 2) to give confidence intervals for each regression parameter. The Cartesian product of thesepconfidence intervals gives ap-dimensional rectangle that contains the confidence ellipsoid and hence has a joint confidence coefficient of at least (1–α). A simple calculus proof is given to determine these projections. The projection procedure is compared with the Bonferroni procedure for this case.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476038

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

The moment-generating function of the minimum of bivariate normal random variables is derived through a lemma that can be used to evaluate certain integrals that involve the normal distribution. As an illustration of the value of the moment-generating function, the mean and mean square are deduced, and they prove to be quite complicated expressions in the general case considered here.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476039

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

LetT= (T1,…,Tk;k≥ 2) be a minimal sufficient statistic for ak-parameter natural exponential model. Consider a partition ofTinto (T1,T2), whereT1= (T1,…,Tr) andT2= (Tr+1,…,Tk; 1 <r≤k). It is shown that cumulants of the conditional distribution ofT1, givenT2=t2, can be computed through the marginal distribution ofT2and the norming constant that makes the model a probability model. The results are illustrated by a few examples.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476040

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476041

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Factorial experiments are done to measure the average effects of varying each of several factors (experimental conditions, independent variables) over fixed ranges of the other factors. Factors at two levels only are considered here. The system under study is assumed to permit the variation of only one factor per trial. Up to four factors may be of interest. The sequences proposed, of sizes 7–9, are easy to remember, simple to analyze, and straightforward to augment. Each result is compared with its immediate predecessor or successor only. Thus the three-factor one-at-a-time factorial follows the sequence. (1)1,a,ab,abc,bc,c, (1)2, where (1) means all factors at their low levels;ameans 1 factor,A, at its high level;abmeansAandBat high levels; and so forth.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476042

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Cumulative probabilities for the binomial, Poisson and chi-square distributions can be formidable to calculate because of the size of the components. A simple code is given that utilizes the ratio of successive terms and thus avoids most of the computational difficulties.

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476043

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476045

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476046

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Explanations are offered for some of the idiosyncrasies evident in computer output of sums of squares of unbalanced data described by Dallal (1992).

ISSN:0003-1305    

DOI:10.1080/00031305.1994.10476047

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor