Ignoring a Level in a Multilevel Model: Evidence from UK Census Data
- 1 May 2001
- journal article
- Published by SAGE Publications in Environment and Planning A: Economy and Space
- Vol. 33 (5) , 941-948
- https://doi.org/10.1068/a3317
Abstract
Because of the inherent multilevel nature of census data, it is often appropriate to use multilevel models to investigate relationships between census variables. For a local population, the data available from the census allow a three-level nested model to be assumed, with an individual level (level 1), an enumeration district (ED) level (level 2), and a ward level (level 3). The consequences of ignoring one of the three levels in this model are assessed here theoretically. Empirical results, based on 1991 UK Census data, are also provided, comparing the variance components estimated from the three-level model with analyses based on models where the ED or ward level are ignored. The results show how the variation that occurs at the level not included in the models is redistributed to the other levels that the models do include.Keywords
This publication has 5 references indexed in Scilit:
- Using Census Data to Investigate the Causes of the Ecological FallacyEnvironment and Planning A: Economy and Space, 1998
- Using the 1991 Census SAR in a Multilevel Analysis of Male UnemploymentEnvironment and Planning A: Economy and Space, 1997
- Searching for the Elusive Aggregation Effect: Evidence from Statistical SimulationsEnvironment and Planning A: Economy and Space, 1995
- Ecological Fallacies and the Analysis of Areal Census DataEnvironment and Planning A: Economy and Space, 1984
- Ecological Correlations and the Behavior of IndividualsAmerican Sociological Review, 1950