Ecological Bias, Confounding, and Effect Modification
Abstract
Ecological bias is sometimes attributed to confounding by the group variable (ie the variable used to define the ecological groups), or to risk factors associated with the group variable. We show that the group variable need not be a confounder (in the strict epidemiological sense) for ecological bias to occur: effect modification can lead to profound ecological bias, whether or not the group variable or the effect modifier are independent risk factors. Furthermore, an extraneous risk factor need not be associated with the study variable at the individual level in order to produce ecological bias. Thus the conditions for the production of ecological bias by a covariate are much broader than the conditions for the production of individual-level confounding by a covariate. We also show that standardization or ecological control of variables responsible for ecological bias are generally insufficient to remove such bias.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
A potential limitation of the study is the use of aggregated randomized controlled trial data and not individual subject data, with the exception of one study.
“A potential limitation of our work is the use of aggregated RCT data and not individual subject data [113, 114], with the exception of one study [86].”
From Estimating the Effect of Aerobic Exercise Training on Novel Lipid Biomarkers: A Systematic Review and Multivariate Meta-Analysis of Randomized Controlled Trials (2023) · cites this paper as [113]
Cited by
Sorted by citations
Bringing context back into epidemiology: variables and fallacies in multilevel analysis
Diez-Roux, A. V. (1998) · American Journal of Public Health · doi:10.2105/ajph.88.2.216 · Cited by 1,317 · see citing works
Sud, S., Friedrich, J. O., Taccone, P., Polli, F., Adhikari, N. K. J., Latini, R., Pesenti, A., Guérin, C., Mancebo, J., Curley, M. A. Q., et al. (2010) · Intensive Care Medicine · doi:10.1007/s00134-009-1748-1 · Cited by 563 · see citing works
Invited commentary: ecologic studies—biases, misconceptions, and counterexamples
Greenland, S., & Robins, J. (1994) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a117069 · Cited by 494 · see citing works
Self-citationEcological effects in multi-level studies
Blakely, T. A., & Woodward, A. J. (2000) · Journal of Epidemiology and Community Health · doi:10.1136/jech.54.5.367 · Cited by 385 · see citing works
Ecological study of social fragmentation, poverty, and suicide
Whitley, E., Gunnell, D., Dorling, D., & Smith, G. D. (1999) · BMJ · doi:10.1136/bmj.319.7216.1034 · Cited by 250 · see citing works
Suicide, religion, and socioeconomic conditions. An ecological study in 26 countries, 1990.
Neeleman, J., & Lewis, G. (1999) · Journal of Epidemiology and Community Health · doi:10.1136/jech.53.4.204 · Cited by 147 · see citing works
The ISAAC Phase One Study Group, Asher, M. I., Stewart, A. W., Mallol, J., Montefort, S., Lai, C. K., Aït-Khaled, N., & Odhiambo, J. (2010) · Respiratory Research · doi:10.1186/1465-9921-11-8 · Cited by 144 · see citing works
Ethnic minority suicide: a small area geographical study in south London
NEELEMAN, J., & WESSELY, S. (1999) · Psychological Medicine · doi:10.1017/s003329179800806x · Cited by 138 · see citing works
Office-based preventive dental program and statewide trends in dental caries
Achembong, L. N., Kranz, A. M., & Rozier, R. G. (2014) · Pediatrics · doi:10.1542/peds.2013-2561 · Cited by 38 · see citing works
Wood, G., Taylor, E., Ng, V., Murrell, A., Patil, A., van der Touw, T., Wolden, M., Andronicos, N., & Smart, N. A. (2023) · Sports Medicine · doi:10.1007/s40279-023-01817-0 · Cited by 31 · see citing works
“A potential limitation of our work is the use of aggregated RCT data and not individual subject data [113, 114], with the exception of one study [86].”
BackgroundGenser, B., Teles, C. A., Barreto, M. L., & Fischer, J. E. (2015) · Environmental Health · doi:10.1186/s12940-015-0047-2 · Cited by 23 · see citing works
- Authors
- GREENLAND, S., & MORGENSTERN, H.
- Affiliations
- University of California, Los Angeles (US)
- ORCID
- GREENLAND, SANDER
- Venue
- International Journal of Epidemiology
- Publisher
- Oxford University Press (OUP)
- Published
- 1989
- Volume
- 18
- Issue
- 1
- Pages
- 269-274
- ISSN
- 0300-5771 (ISSN-L)
- 1464-3685
- OpenAlex
- W2155669879
- Cited by
- 540
- References
- 0
- Citations by year
- 2026: 24 · 2025: 17 · 2024: 12 · 2023: 19 · 2022: 20 · 2021: 27 · 2020: 17 · 2019: 7 · 2018: 13 · 2017: 9 · 2016: 12 · 2015: 16 · 2014: 18 · 2013: 16 · 2012: 16
- FWCI
- 10.60
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- International Journal of Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Keywords
- ecological bias, confounding, effect modification, risk factors, standardization, ecological control
- MeSH terms
- Data Interpretation, Statistical, Epidemiologic Methods, Humans, Risk Factors
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
