Wealthy, capital-intensive estates are severely overrepresented in historical collections of probate inventories. This paper discusses a new methodology to neutralise the wealth bias in this important historical data source, enabling one to use probate inventories as a source of information on all households, rather than merely on the biased, probated subsection. The methodology is based on establishing the probability of being inventoried – in a certain time period and geographical area – as a function of the value of the decedent’s estate. The probability function is established by means of an iterative fitting process, using occupational information from contemporary parish registers to establish the target values. The inverse of the resulting probability function is subsequently used to reweight the probate inventory dataset. Thereby the historical process which created the inventoried household subsection in the first place is reversed, thus providing an unbiased view of the full population of households. A number of example applications demonstrate the value of the approach.