Statistical methods in epidemiology: I. Statistical errors in hypothesis testing.

A S Rigby

doi:10.3109/09638289809166071

Home/Archive/PubMed

PubMed · 9571378

Statistical methods in epidemiology: I. Statistical errors in hypothesis testing.

A S Rigby

View original source ↗Export metadata ↓

Abstract

PURPOSE: Although scientific journal editors are making use of statisticians in the review process, the quality of statistical reporting in many journals remains poor. In many cases the problem for the scientist would appear to be a lack of understanding of basic statistics. The focus of the scientist is on showing 'p < 0.05', when what is actually required is a statement about effect size and interval estimation. The aim of this paper is to show the inadequacy of reporting of results using p-values alone. This paper is the first in a series detailing common statistical methods, with a view to aiding potential authors in their statistical presentation of data. METHOD: A review of the basic hypothesis test, using examples from the author's own teaching experiences. RESULTS: Type I and type II errors are defined; the problem of multiple comparisons is highlighted; interval estimation is introduced. CONCLUSIONS: The case for considering the p-value as an error probability is made which suggests ways of improving statistical presentation and thus expediting the statistical review process.

Explore related subjects

Confidence Intervals Epidemiologic Methods Epidemiologic Studies Humans Models, Statistical Odds Ratio Probability Research Design Sensitivity and Specificity Journal Article Review

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

A S Rigby. 1998. Statistical methods in epidemiology: I. Statistical errors in hypothesis testing.. https://doi.org/10.3109/09638289809166071

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

KEEP EXPLORING

Related citations

PubMed2002

["P" as in Program Package. About torturing data and significant "fishing expeditions"].

For the study of prognostic factors, the medical researchers have access to a number of advanced statistical techniques available in standard program packages. A tradition has developed where survival or time to relapse is analysed on the basis of statistical materials with few patients but a large number of possible explanatory variables. In statistical "fishing expeditions" the p-values are used to sort out potentially useful prognostic variables. Since the number of observations is small, all relevant prognostic factors do not give statistical significance. Since a large number of variables are tested there is a considerable risk for spurious significances. It is not enough to show that a prognostic factor seems to be efficient in the patient group where it was first found. The result must be verified in further studies of independent groups of similar patients.

Confidence Intervals↗

PubMed2002

Substance abuse and the need for money management assistance among psychiatric inpatients.

Patients who mismanage their funds may benefit from financial advice, case management or the involuntary assignment of a payee who restricts direct access to funds. Data from a survey of psychiatric inpatients at four VA hospitals (N = 236) was used to evaluate the relationship between substance abuse and clinician-rated need for money management assistance. Multivariate analytic techniques were used to control for sociodemographic factors and psychopathology. Alcohol and drug use severity both were modestly associated with need for assistance. The effect of substance use severity was greater in patients who were also diagnosed with a major mental illness. Clinicians indicated that 27 patients (11% of the sample) required an involuntary payee and 21 of the 27 (78%) had a Substance Abuse diagnosis. Only drug use severity was significantly associated with need for a payee. These data describe a substantial unmet need for money management assistance in psychiatric inpatients, particularly among those with substance abuse disorders. There is a need to examine the process by which the Social Security and Veterans Benefits Administrations assign payees to determine whether patients with co-morbid substance abuse are not being assigned a payee in spite of their discernible need for one.

Confidence Intervals↗

PubMed2002

Low agreement among 24 doctors using the Neer-classification; only moderate agreement on displacement, even between specialists.

Twenty-four orthopaedic surgeons classified 42 pairs of radiographs according to the Neer system for proximal humeral fractures. Mean kappa value for inter-observer agreement was 0.27 (95% CI 0.26-0.28) with no clinically significant difference between orthopaedic residents ( n=9), fellows ( n=6) and specialists ( n=9). Mean kappa for agreement of displacement versus non-displacement was 0.41 (95% CI 0.39-0.43) overall, and 0.50 (95% CI 0.45-0.56) within the specialist group. The agreement found in our study is unsatisfactory from a clinical perspective.

Confidence Intervals↗