Drawbacks to integer scoring for ordered categorical data.

A Ivanova; V W Berger

doi:10.1111/j.0006-341x.2001.00567.x

Home/Archive/PubMed

PubMed · 11414585

Drawbacks to integer scoring for ordered categorical data.

A Ivanova·V W Berger

View original source ↗Export metadata ↓

Abstract

Linear rank tests are widely used when testing for independence against stochastic order in a 2 x J contingency table with two treatments and J ordered outcome levels. For this purpose, numerical scores are assigned, possibly by default, to the J outcome levels. When the choice of scores is not apparent, integer (equally spaced) scores are often assigned. We show that this practice generally leads to unnecessarily conservative tests. The use of slightly perturbed scores will result in a less conservative and uniformly more powerful test.

Explore related subjects

Biometry Humans Models, Statistical Probability Reproducibility of Results Stochastic Processes Survival Analysis Journal Article

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

A Ivanova, V W Berger. 2001. Drawbacks to integer scoring for ordered categorical data.. https://doi.org/10.1111/j.0006-341x.2001.00567.x

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2003

Size-dependent ESS sex allocation in wind-pollinated cosexual plants: fecundity vs. stature effects.

To theoretically investigate the single and compound effects of relative fecundity and relative stature of plants on size-dependent sex allocation (SDS) in wind-pollinated cosexual species, we developed a game model and analysed ESS sex allocation of large and small plants having totally or partially different reproductive resources and different pollen and seed dispersal areas in a population. We found that e.g. when both sized plants have large pollen dispersal areas relative to their seed dispersal areas, which plants are male-biased is largely determined by relative fecundity (t) and relative size of seed dispersal area (k) of the large plants to the small plants: If t >k, large plants tend to be more male-biased even if relative size of pollen dispersal area of large to small plants (l) is smaller than k. If t<k, large plants tend to be more female-biased even if with very large l, e.g. infinity l. Comprehensively, neither very large relative size of pollen to seed dispersal area, nor great expansion of pollen dispersal area with plant size, such as previous models assumed, is important for the increasing maleness with plant size in our model.

Biometry↗

PubMed2003

[Common biostatistical errors in clinical studies].

Roughly half to two third of all published biomedical studies that use statistical methods contain unacceptable errors. The present article points at common errors that may be avoided without requiring profound statistical knowledge. These errors mainly concern the minimal number of patients and sample size (statistical power), agreement between aim and conclusion, distribution of data as well as description of location and variability of data. An analysis of 150 papers in the New England Journal of Medicine and in Circulation demonstrates that these errors can also commonly be found in respected journals after statistical peer review. Editors of biomedical journals could reduce the problem by means of statistical guidelines.

Biometry↗

PubMed2003

How to decide whether small samples comply with an equidistribution.

The decision whether a measured distribution complies with an equidistribution is a central element of many biostatistical methods. High throughput differential expression measurements, for instance, necessitate to judge possible over-representation of genes. The reliability of this judgement, however, is strongly affected when rarely expressed genes are pooled. We propose a method that can be applied to frequency ranked distributions and that yields a simple but efficient criterion to assess the hypothesis of equiprobable expression levels. By applying our technique to surrogate data we exemplify how the decision criterion can differentiate between a true equidistribution and a triangular distribution. The distinction succeeds even for small sample sizes where standard tests of significance (e.g. chi(2)) fail. Our method will have a major impact on several problems of computational biology where rare events baffle a reliable assessment of frequency distributions. The program package is available upon request from the authors.

Biometry↗