Dataset Information

A global sensitivity test for evaluating statistical hypotheses with nonidentifiable models.

ABSTRACT: We consider the problem of evaluating a statistical hypothesis when some model characteristics are nonidentifiable from observed data. Such a scenario is common in meta-analysis for assessing publication bias and in longitudinal studies for evaluating a covariate effect when dropouts are likely to be nonignorable. One possible approach to this problem is to fix a minimal set of sensitivity parameters conditional upon which hypothesized parameters are identifiable. Here, we extend this idea and show how to evaluate the hypothesis of interest using an infimum statistic over the whole support of the sensitivity parameter. We characterize the limiting distribution of the statistic as a process in the sensitivity parameter, which involves a careful theoretical analysis of its behavior under model misspecification. In practice, we suggest a nonparametric bootstrap procedure to implement this infimum test as well as to construct confidence bands for simultaneous pointwise tests across all values of the sensitivity parameter, adjusting for multiple testing. The methodology's practical utility is illustrated in an analysis of a longitudinal psychiatric study.

SUBMITTER: Todem D

PROVIDER: S-EPMC3076640 | biostudies-literature | 2010 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A global sensitivity test for evaluating statistical hypotheses with nonidentifiable models.

Todem D D Fine J J Peng L L

Biometrics 20090723 2

We consider the problem of evaluating a statistical hypothesis when some model characteristics are nonidentifiable from observed data. Such a scenario is common in meta-analysis for assessing publication bias and in longitudinal studies for evaluating a covariate effect when dropouts are likely to be nonignorable. One possible approach to this problem is to fix a minimal set of sensitivity parameters conditional upon which hypothesized parameters are identifiable. Here, we extend this idea and s ...[more]

PMID: 19645705

Similar Datasets

Project description:Tissues are often made up of multiple cell-types. Blood, for example, contains many different cell-types, each with its own functional attributes and molecular signature. In humans, because of its accessibility and immune functionality, blood cells have been used as a source for RNA-based biomarkers for many diseases. Yet, the proportions of any given cell-type in the blood can vary markedly, even between normal individuals. This results in a significant loss of sensitivity in gene expression studies of blood cells and great difficulty in identifying the cellular source of any perturbations. Ideally, one would like to perform differential expression analysis between patient groups for each of the cell-types within a tissue but this is impractical and prohibitively expensive. To test the relationship between measured gene expression in mixed samples and the expression of genes in the isolated pure subsets, we begin with a situation in which all factors are known. Tissue samples from the brain, liver and lung of a single rat were analyzed using expression arrays (Affymetrix) in triplicate. Homogenates of those three tissues were then mixed together at the cRNA level. We then measured the gene expression pattern of each mixed sample. Such mixtures mimic the common scenario in which biological samples in a dataset are heterogeneous and vary in the relative frequency of the component subsets from one another. We mixed rat brain, liver and lung biospecimens derived from one animal at the cRNA homogenate level in different proportions. 3 technical replicates each. Snap frozen rat liver and brain was kept frozen while cutting it into pieces. cDNA synthesis and labeling was done with a starting amount of 1 Î¼g, using the Affymetrix Eukaryotic One-Cycle Target Hybridization. Washing, Staining and scanning protocol for Eukaryotic Cartridge Arrays with User-Prepared Buffers and Sloutions according to the technical manuals (Affymetrix GeneChip Expression, Analysis for Cartridge Arrays using the GCAS version 1.4, Affymetrix GeneChip Expression Analysis (P/N 701021,Rev. 5) , Affymetrix GeneChip Expression Wash, Stain and Scan (P/N 702731, Rev. 3) , following the manufacturer/s instructions. Data was RMA normalized.

Dataset Information

A global sensitivity test for evaluating statistical hypotheses with nonidentifiable models.

Publications

A global sensitivity test for evaluating statistical hypotheses with nonidentifiable models.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets