Biostatistics 4:265-278 (2003)
© 2003 Oxford University Press
Insights on bias and information in group-level studies
Departments of Biostatistics and Environmental Health, Box 357232, University of Washington, Seattle, WA 98195-7232, USA sheppard{at}u.washington.edu
Ecological and aggregate data studies are examples of group-level studies. Even though the link between the predictors and outcomes is not preserved in these studies, inference about individual-level exposure effects is often a goal. The disconnection between the level of inference and the level of analysis expands the array of potential biases that can invalidate the inference from group-level studies. While several sources of bias, specifically due to measurement error and confounding, may be more complex in group-level studies, two sources of bias, cross-level and model specification bias, are a direct consequence of the disconnection. With the goal of aligning inference from individual versus group-level studies, I discuss the interplay between exposure and study design. I specify the additional assumptions necessary for valid inference, specifically that the between- and within-group exposure effects are equal. Then cross-level inference is possible. However, all the information in the group-level analysis comes from between-group comparisons. Models where the group-level analysis provides even a small percentage of information about the within-group exposure effect are most susceptible to model specification bias. Model specification bias can be even more serious when the group-level model isn't derived from an individual-level model.
Keywords: Aggregate data studies; Cross-level inference; Ecological studies; Relative risk
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. G. Kenward and J. H. Roger The use of baseline covariates in crossover studies Biostat., November 13, 2009; (2009) kxp046v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Schildcrout and P. J. Heagerty On outcome-dependent sampling designs for longitudinal binary response data with time-varying covariates Biostat., October 1, 2008; 9(4): 735 - 749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wakefield and S. J.-P. A. Haneuse Overcoming Ecologic Bias using the Two-Phase Study Design Am. J. Epidemiol., April 15, 2008; 167(8): 908 - 916. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Manca, P. C. Lambert, M. Sculpher, and N. Rice Cost-Effectiveness Analysis Using Data from Multinational Trials: The Use of Bivariate Hierarchical Modeling Med Decis Making, August 1, 2007; 27(4): 471 - 490. [Abstract] [PDF]
|
|
|
|
|
|
J. S. Schildcrout, L. Sheppard, T. Lumley, J. C. Slaughter, J. Q. Koenig, and G. G. Shapiro Ambient Air Pollution and Asthma Exacerbations in Children: An Eight-City Analysis Am. J. Epidemiol., September 15, 2006; 164(6): 505 - 517. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Ferrari, R. Kaaks, M. T. Fahey, N. Slimani, N. E. Day, G. Pera, H. C. Boshuizen, A. Roddam, H. Boeing, G. Nagel, et al. Within- and Between-Cohort Variation in Measured Macronutrient Intakes, Taking Account of Measurement Errors, in the European Prospective Investigation into Cancer and Nutrition Study Am. J. Epidemiol., October 15, 2004; 160(8): 814 - 822. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. L. Prentice, W. C. Willett, P. Greenwald, D. Alberts, L. Bernstein, N. F. Boyd, T. Byers, S. K. Clinton, G. Fraser, L. Freedman, et al. Nutrition and Physical Activity and Chronic Disease Prevention: Research Strategies and Recommendations J Natl Cancer Inst, September 1, 2004; 96(17): 1276 - 1287. [Abstract] [Full Text] [PDF]
|
|