Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If you need an accessible version of this item please contact JSTOR User Support

Multiple-Bias Modelling for Analysis of Observational Data

Sander Greenland
Journal of the Royal Statistical Society. Series A (Statistics in Society)
Vol. 168, No. 2 (2005), pp. 267-306
Published by: Wiley for the Royal Statistical Society
Stable URL: http://www.jstor.org/stable/3559963
Page Count: 40
  • Read Online (Free)
  • Download ($29.00)
  • Cite this Item
If you need an accessible version of this item please contact JSTOR User Support
Multiple-Bias Modelling for Analysis of Observational Data
Preview not available

Abstract

Conventional analytic results do not reflect any source of uncertainty other than random error, and as a result readers must rely on informal judgments regarding the effect of possible biases. When standard errors are small these judgments often fail to capture sources of uncertainty and their interactions adequately. Multiple-bias models provide alternatives that allow one systematically to integrate major sources of uncertainty, and thus to provide better input to research planning and policy analysis. Typically, the bias parameters in the model are not identified by the analysis data and so the results depend completely on priors for those parameters. A Bayesian analysis is then natural, but several alternatives based on sensitivity analysis have appeared in the risk assessment and epidemiologic literature. Under some circumstances these methods approximate a Bayesian analysis and can be modified to do so even better. These points are illustrated with a pooled analysis of case-control studies of residential magnetic field exposure and childhood leukaemia, which highlights the diminishing value of conventional studies conducted after the early 1990s. It is argued that multiple-bias modelling should become part of the core training of anyone who will be entrusted with the analysis of observational data, and should become standard procedure when random error is not the only important source of uncertainty (as in meta-analysis and pooled analysis).

Page Thumbnails

  • Thumbnail: Page 
[267]
    [267]
  • Thumbnail: Page 
268
    268
  • Thumbnail: Page 
269
    269
  • Thumbnail: Page 
270
    270
  • Thumbnail: Page 
271
    271
  • Thumbnail: Page 
272
    272
  • Thumbnail: Page 
273
    273
  • Thumbnail: Page 
274
    274
  • Thumbnail: Page 
275
    275
  • Thumbnail: Page 
276
    276
  • Thumbnail: Page 
277
    277
  • Thumbnail: Page 
278
    278
  • Thumbnail: Page 
279
    279
  • Thumbnail: Page 
280
    280
  • Thumbnail: Page 
281
    281
  • Thumbnail: Page 
282
    282
  • Thumbnail: Page 
283
    283
  • Thumbnail: Page 
284
    284
  • Thumbnail: Page 
285
    285
  • Thumbnail: Page 
286
    286
  • Thumbnail: Page 
287
    287
  • Thumbnail: Page 
288
    288
  • Thumbnail: Page 
289
    289
  • Thumbnail: Page 
290
    290
  • Thumbnail: Page 
291
    291
  • Thumbnail: Page 
292
    292
  • Thumbnail: Page 
293
    293
  • Thumbnail: Page 
294
    294
  • Thumbnail: Page 
295
    295
  • Thumbnail: Page 
296
    296
  • Thumbnail: Page 
297
    297
  • Thumbnail: Page 
298
    298
  • Thumbnail: Page 
299
    299
  • Thumbnail: Page 
300
    300
  • Thumbnail: Page 
301
    301
  • Thumbnail: Page 
302
    302
  • Thumbnail: Page 
303
    303
  • Thumbnail: Page 
304
    304
  • Thumbnail: Page 
305
    305
  • Thumbnail: Page 
306
    306