Residential refrigerators are the largest domestic use of electricity in the US and most developed countries and, as such, have become a target for efficiency improvements. Laboratory tests of energy use are typically used to measure a refrigerator＇s energy consumption. Three different test procedures are widely used: the US Department of Energy, the International Standards Organization test, and the Japanese Industrial Standard. The features of the tests are compared and conversion factors reviewed. The DOE test is likely to yield the highest consumption of the three procedures. A critical factor is the relationship between a refrigerator＇s laboratory-based energy use and its consumption in a kitchen. The DOE test is the most carefully validated with field data. The DOE test, on average, over-predicts actual consumption in US homes by about 10%, but there is wide variation in field use for identical units. The factors affecting variation in energy use are reviewed. The ambient temperature in the kitchen is by far the most significant factor, while door openings and humidity are relatively minor. Field measurements suggest that maintenance measures, such as coil cleaning and gasket replacement, save little energy. Replacement of old refrigerators with new, efficient models often cuts refrigerator energy use by 60%. Significant reductions in refrigerator electricity use can be expected during the next decade in the US and Europe.