Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Marginal versus conditional versus 'structural source' models: a rationale for an alternative to log-linear methods

R R Regal1, E B Hook

  • 1University of Minnesota, Department of Mathematics and Statistics, Duluth 55812, USA.

Statistics in Medicine
|February 17, 1998
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Lake Superior Rural Cancer Care Project, part I: an interventional trial.

Cancer practice·2002
Same author

Lake Superior Rural Cancer Care Project, part II: provider knowledge.

Cancer practice·2002
Same author

Statistical and logical considerations in evaluating the association of prenatal folic-acid supplementation with pregnancy loss.

American journal of medical genetics·2001
Same author

Design and validation of instruments to measure knowledge.

Journal of cancer education : the official journal of the American Association for Cancer Education·2001
Same author

Search for ethnic, geographic, and other factors in the epidemiology of Down syndrome in South America: analysis of data from the ECLAMC project, 1967-1997.

American journal of medical genetics·2001
Same author

Instruments to measure cancer management knowledge of rural health care providers.

Journal of cancer education : the official journal of the American Association for Cancer Education·2001

Log-linear models can bias population size estimates. Analyzing spina bifida data reveals that assuming conditional independence between data sources can be inaccurate, leading to biased results when marginal independence is present.

Area of Science:

  • Biostatistics
  • Epidemiology
  • Population Dynamics

Background:

  • Log-linear models are standard for capture-recapture data, focusing on conditional interactions between data sources.
  • Practitioners often find marginal associations more intuitive and applicable than conditional ones.

Purpose of the Study:

  • To demonstrate how assuming conditional independence can bias population estimates when sources are marginally independent.
  • To introduce 'structural source modeling' for analyzing interdependencies between information sources.

Main Methods:

  • Re-analysis of published spina bifida case data from upstate New York.
  • Development of a structural model for source interactions, implying marginal independence for two sources.
  • Derivation of population total estimates based on marginal independence.

Related Experiment Videos

Main Results:

  • The assumption of conditional independence yielded biased population estimates.
  • Estimates derived from marginal independence were larger than those from conditional dependence models.
  • The proposed structural source model highlighted marginal independence between two data sources.

Conclusions:

  • Conditional independence assumptions in log-linear models can lead to significant estimation bias.
  • Modeling potential interdependencies (structural source modeling) is recommended when underlying source relationships are understood.
  • Marginal independence-based estimates provided a more accurate population total for the spina bifida data.