Patterns of error in survey based estimates of consumption and their implications for energy consumption
It is well known that survey based estimates of consumption in general and of energy and water in particular are prone to recall or other forms of error. However little is known about the distribution of this error nor its correlates. This is unfortunate as a range of policy and strategy decisions have used such data to produce, for example, estimates of income and price elasticity or reported energy savings after efficiency interventions. If there is substantial systematic error in the reporting of consumption then estimates of elasticity or efficiency savings will be unreliable.
This project will use a range of linked household and individual level datasets which include reported and observed energy consumption in order to understand the patterns of energy consumption reporting error. It will use this analysis to provide guidance on the calculation of corrected consumption estimates in surveys as well as in commercial or policy oriented research. It will then use the observed consumption data to analyse the factors affecting observed rather than reported energy consumption.
The project will help to understand the patterns of energy consumption and reporting error. The work will improve the data quality of collection energy data and will help in using this data for public policy.
Data will be used from the Office for National Statistics, the Institute of Social and Economic Research and the Department of Energy & Climate Change
Prof Gabrielle Durrant, University of Southampton
Vivian So, University of Southampton
Dr Ben Anderson, University of Southampton