Meta-analysis and Random Effect Analysis

Old and New Style Random Effect Analysis
  • Ton J. Cleophas
  • Aeilko H. Zwinderman


Heterogeneity is probably the largest pitfall of meta-analyses. A random effect analytic model assumes, that heterogeneity is due to some unexpected subgroup effect rather than a residual effect, and uses a separate random variable for the purpose. Examples of fixed and random effect analyses are given both from binary and continuous outcome meta-analysis data. Within a single study heterogeneity may very well be residual, but between the overall effects of studies within a meta-analysis this is virtually never so, and it is virtually always caused by some subgroup effect. Therefore, fixed effect heterogeneity tests are slightly inappropriate. However, random effect heterogeneity tests lack power. Novel methodologies are being developed, and will be reviewed in this and many subsequent chapters of this edition.


Random Effect Random Effect Model Fixed Effect Model Effect Test Pool Odds Ratio 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. More background, theoretical and mathematical information of meta-analyses is given in Statistics applied to clinical studies 5th edition, Chaps. 32–34 and 48, Springer Heidelberg Germany, 2012, from the same authors.Google Scholar

Copyright information

© Springer International Publishing Switzerland 2017

Authors and Affiliations

  • Ton J. Cleophas
    • 1
  • Aeilko H. Zwinderman
    • 2
  1. 1.Department of MedicineAlbert Schweitzer HospitalSliedrechtThe Netherlands
  2. 2.Department of Epidemiology and BiostatisticsAcademic Medical CenterAmsterdamThe Netherlands

Personalised recommendations