The problematic estimation of "imitation effects" in multilevel models
It seems plausible that a personâ€™s demographic behaviour may be influenced by that among other people in the community, for example because of an inclination to imitate. When estimating multilevel models from clustered individual data, some investigators might perhaps feel tempted to try to capture this effect by simply including on the right-hand side the average of the dependent variable, constructed by aggregation within the clusters. However, such modelling must be avoided. According to simulation experiments based on real fertility data from India, the estimated effect of this obviously endogenous variable can be very different from the true effect. Also the other community effect estimates can be strongly biased. An "imitation effect" can only be estimated under very special assumptions that in practice will be hard to defend.
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Øystein Kravdal, 2002. "Education and fertility in sub-Saharan africa: Individual and community effects," Demography, Springer, vol. 39(2), pages 233-250, May.
- repec:att:wimass:9127 is not listed on IDEAS
- Manski, Charles F, 1993.
"Identification of Endogenous Social Effects: The Reflection Problem,"
Review of Economic Studies,
Wiley Blackwell, vol. 60(3), pages 531-42, July.
When requesting a correction, please mention this item's handle: RePEc:dem:demres:v:9:y:2003:i:2. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Editorial Office)
If references are entirely missing, you can add them using this form.