Purpose – The purpose of this paper is to demonstrate the application of robust techniques to the estimation of hedonic house price indexes. Design/methodology/approach – The authors use simulation analysis to compare an index estimated using ordinary least squares (OLS) with several indexes estimated using robust techniques. The analysis uses sales transactions data from a US city. The authors then explore how robust methods can correct for omitted variables under some circumstances and how they affect the revision problem that occurs when longitudinal hedonic indexes are updated. Findings – Robust methods can resolve missing variable problems in some circumstances and also can substantially reduce the revision problem in longitudinal hedonic indexes. Practical implications – Robust techniques may be preferable to OLS when constructing longitudinal hedonic indexes. Originality/value – This is the first paper to undertake a systematic analysis of the applicability of robust techniques in constructing hedonic house price indexes.