This paper contrasts private equity data from three leading providers and draws implications for research and practice. If each data provider’s sample were a random draw from the same underlying universe, we’d expect similar messages to emerge across all data sources. This is not the case. Different providers have different mixes of funds (e.g. venture capital versus buyout), especially in periods studied by prior research. Turning to performance data, the picture is even more troubling. Performance data cover only a small fraction of funds started. Moreover, return data from the three sources often signal different performance to limited partner investors. Some of these differences are not readily explained by random variation and suggest systematic effects related to data methods and sample selection. We also review findings of past research on private equity with particular attention to the data used. Our findings highlight the need for better private equity data to improve investment decisions and enable research. We conclude with thoughts on steps toward such improved data.