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

The effects of measurement errors on some multivariate procedures.

J L Fleiss, P E Shrout

    American Journal of Public Health
    |December 1, 1977
    PubMed
    Summary

    Random measurement errors can weaken statistical associations or even reverse their direction. This study examines these effects on multivariate procedures and proposes solutions.

    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

    ON THE METHODS AND THEORY OF CLUSTERING.

    Multivariate behavioral research·2016
    Same author

    A simple approximation for calculating sample sizes for comparing independent proportions.

    Biometrics·2015
    Same author

    Two plus two equals three? Do we need to rethink lifetime prevalence?

    Psychological medicine·2009
    Same author

    Psychometric evaluation of the Spanish version of CONNECT: a measure of continuity of care in mental health services.

    International journal of methods in psychiatric research·2007
    Same author

    Gender differences in the relationship of homelessness to symptom severity, substance abuse, and neuroleptic noncompliance in schizophrenia.

    The Journal of nervous and mental disease·2001
    Same author

    Risk factors for homelessness among indigent urban adults with no history of psychotic illness: a case-control study.

    American journal of public health·2000

    Area of Science:

    • Statistics
    • Multivariate Analysis

    Background:

    • Random errors of measurement are inherent in data collection.
    • These errors can significantly impact statistical analyses, particularly in multivariate settings.

    Purpose of the Study:

    • To investigate the impact of random measurement errors on fundamental multivariate procedures.
    • To identify how these errors affect the estimation of associations.

    Main Methods:

    • The study discusses the theoretical effects of random measurement errors.
    • It analyzes the consequences for various multivariate statistical techniques.

    Main Results:

    • Measurement errors can lead to attenuation, reducing the observed strength of associations.
    • In some cases, errors may distort the apparent direction of relationships between variables.

    Conclusions:

    • Random measurement errors pose a significant challenge in multivariate analysis.
    • Proposed solutions are discussed, highlighting the need for further research in this area.

    Related Experiment Videos