This paper aims to develop a practical way of correcting for sample selection bias in UK regional house price indices. Such indices are invariably based on transactions data (such as Land Registry data on house sales). Unfortunately, properties that transact in any given period are unlikely to be a random sample of all properties, either in terms of location or type. Consequently, such indices may not give a reliable measure of the rate of increase in the value of the housing stock, a variable that is crucial to our understanding of wealth inequality and a range of housing policy decisions. While correction methods have been developed in the US, there is currently no UK equivalent, partly because valuation records are not kept on the whole stock. This project explores the possibility of developing a reliable correction procedure using data readily available in the UK.