Some Approximate Evidence Factors in Observational Studies

An observational study or nonrandomized experiment has exact evidence factors if it permits several statistically independent tests of the same null hypothesis H0 of no treatment effect, where these several tests depend upon different assumptions about bias from nonrandom treatment assignment. In an observational study, we are typically uncertain about what assumptions truly describe treatment assignment. If independent tests each reject H0, and if each of these tests is valid under assumptions about treatment assignment that would invalidate the others, then rejection of H0 does not depend upon the truth of any one of these assumptions. Although exact evidence factors do exist, the requirement of exact independence may limit the class of test statistics in ways that reduce Pitman efficiency and design sensitivity, so one might wish to be free of these limitations. A much larger class of statistics permits approximate evidence factors which are not independent, but which preserve some of the important consequences of independence. Each of the several tests may be subjected to a sensitivity analysis, and the tests may be combined to yield a single inference. Results are proved for a large class of statistics and they are illustrated by following a suggestion of Maritz which uses the randomization distribution of tests based on statistics that are equated to zero in defining Huber’s m-estimates. A study of damage to DNA from occupational exposures to chromium is used as an example.

[1]  T. J. Terpstra,et al.  The asymptotic normality and consistency of kendall's test against trend, when ties are present in one ranking , 1952 .

[2]  E. C. Hammond,et al.  Smoking in Relation to Mortality and Morbidity. , 1965 .

[3]  W. Brannath,et al.  Recursive Combination Tests , 2002 .

[4]  Issei Fujishiro,et al.  The elements of graphing data , 2005, The Visual Computer.

[5]  E. C. Hammond,et al.  Smoking and lung cancer: recent evidence and a discussion of some questions. , 1959, Journal of the National Cancer Institute.

[6]  Y. Benjamini,et al.  Screening for Partial Conjunction Hypotheses , 2008, Biometrics.

[7]  Govind S. Mudholkar,et al.  A class of tests for equality of ordered means , 1989 .

[8]  D. Rubin,et al.  Assessing Sensitivity to an Unobserved Binary Covariate in an Observational Study with Binary Outcome , 1983 .

[9]  T D WILSON,et al.  Smoking and Lung Cancer , 1960, Journal of the Irish Medical Association.

[10]  Martin Hellmich,et al.  Multiple testing in adaptive designs–a review , 2004 .

[11]  J. Robins,et al.  Sensitivity Analysis for Selection bias and unmeasured Confounding in missing Data and Causal inference models , 2000 .

[12]  B. L. Welch.  ON THE z-TEST IN RANDOMIZED BLOCKS AND LATIN SQUARES , 1937 .

[13]  R. Hogg.  Iterated Tests of the Equality of Several Distributions , 1962 .

[14]  J. Marden,et al.  Use of Nested Orthogonal Contrasts in Analyzing Rank Data , 1992 .

[15]  G. W. Imbens.  Sensitivity to Exogeneity Assumptions in Program Evaluation , 2003 .

[16]  E. C. Hammond,et al.  SMOKING IN RELATION TO MORTALITY AND MORBIDITY. FINDINGS IN FIRST THIRTY-FOUR MONTHS OF FOLLOW-UP IN A PROSPECTIVE STUDY STARTED IN 1959. , 1964, Journal of the National Cancer Institute.

[17]  D. Rubin.  [On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9.] Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies , 1990 .

[18]  Terence P. Speed,et al.  Introductory Remarks on Neyman (1923) , 1990 .

[19]  M. Fligner,et al.  Multistage Ranking Models , 1988 .

[20]  Abba M Krieger,et al.  Causal conclusions are most sensitive to unobserved binary covariates , 2006, Statistics in medicine.

[21]  W. G. Cochran.  The comparison of percentages in matched samples. , 1950, Biometrika.

[22]  T. DiPrete,et al.  7. Assessing Bias in the Estimation of Causal Effects: Rosenbaum Bounds on Matching Estimators and Instrumental Variables Estimation with Imperfect Instruments , 2004 .

[23]  M. Friedman.  The Use of Ranks to Avoid the Assumption of Normality Implicit in the Analysis of Variance , 1937 .

[24]  P. Rosenbaum.  Evidence factors in observational studies , 2010 .

[25]  Frederick R. Forst,et al.  On robust estimation of the location parameter , 1980 .

[26]  D. Rubin.  Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[27]  H. O. Lancaster,et al.  The derivation and partition of chi2 in certain discrete distributions. , 1949, Biometrika.

[28]  Arthur Cohen,et al.  Asymptotically Optimal Methods of Combining Tests , 1979 .

[29]  P. Rosenbaum,et al.  Amplification of Sensitivity Analysis in Matched Observational Studies , 2009, Journal of the American Statistical Association.

[30]  S. Marcus,et al.  Using Omitted Variable Bias to Assess Uncertainty in the Estimation of an AIDS Education Treatment Effect , 1997 .

[31]  P. Rosenbaum.  Sensitivity analysis for matching with multiple controls , 1988 .

[32]  Sebastian Schneeweiss,et al.  Selective prescribing led to overestimation of the benefits of lipid-lowering drugs. , 2006, Journal of clinical epidemiology.

[33]  J. I.  The Design of Experiments , 1936, Nature.

[34]  J. S. Maritz,et al.  A note on exact robust confidence intervals for location , 1979 .

[35]  K E Warner,et al.  Smoking and lung cancer: an overview. , 1984, Cancer research.

[36]  M. Lipsitch,et al.  Negative Controls: A Tool for Detecting Confounding and Bias in Observational Studies , 2010, Epidemiology.

[37]  Some Nonparametric Tests of Randomness , 1974 .

[38]  N. S. Weiss.  Inferring causal relationships: elaboration of the criterion of "dose-response". , 1981, American journal of epidemiology.

[39]  Joseph L. Gastwirth,et al.  Asymptotic separability in sensitivity analysis , 2000 .

[40]  Paul R. Rosenbaum,et al.  Design Sensitivity and Efficiency in Observational Studies , 2010 .

[41]  J. L. Hodges,et al.  Rank Methods for Combination of Independent Experiments in Analysis of Variance , 1962 .

[42]  Dylan S. Small,et al.  Sensitivity Analysis for Instrumental Variables Regression With Overidentifying Restrictions , 2007 .

[43]  Paul R Rosenbaum,et al.  Sensitivity Analysis for m‐Estimates, Tests, and Confidence Intervals in Matched Observational Studies , 2007, Biometrics.

[44]  R. Kronmal,et al.  Assessing the sensitivity of regression results to unmeasured confounders in observational studies. , 1998, Biometrics.

[45]  T. Speed,et al.  On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9 , 1990 .

[46]  Ji-liang He,et al.  Investigating DNA damage in tannery workers occupationally exposed to trivalent chromium using comet assay. , 2008, Mutation research.

[47]  J. Copas,et al.  Local sensitivity approximations for selectivity bias , 2001 .

[48]  R. Tice,et al.  A simple technique for quantitation of low levels of DNA damage in individual cells. , 1988, Experimental cell research.

[49]  E. Stuart,et al.  Impacts of age of onset of substance use disorders on risk of adult incarceration among disadvantaged urban youth: a propensity score matching approach. , 2008, Drug and alcohol dependence.

[50]  G. Gadbury.  Randomization Inference and Bias of Standard Errors , 2001 .

[51]  A. Aakvik,et al.  Bounding a Matching Estimator: The Case of a Norwegian Training Program , 2001 .

[52]  Takashi Yanagawa,et al.  Case-control studies: Assessing the effect of a confounding factor , 1984 .

[53]  Paul R. Rosenbaum,et al.  Sensitivity Analysis for Equivalence and Difference in an Observational Study of Neonatal Intensive Care Units , 2009 .