Objective:To design a more accurate regression equation by grouping different heights (short

medium

tall).Methods: Totally 206 adult females aged between 18.1 and 35.1 years were enrolled. The body height

tibia length and ulna length were measured by standard anthropometric techniques. The subjects were randomly divided into group 1 (n=103) and group 2 (n=103). General regression equation and grouping regression equation were set up to estimate the height.The differences between the actual and estimated heights were evaluated using the two independent samples t-test.Results:The grouping regression equation using ulna length as the independent variable yielded more accurate results while that using tibia length

ulna length and tibia and ulna length as the independent variable respectively did not have obvious advantages. Conclusion: Results shows that the grouping regression equation using ulna length as the independent variable give more accurate height estimates for those who are too short or too tall.