TY -的T1 -作者:摩根富林明——欧洲Respiratory Journal JO - Eur Respir J SP - 694 LP - 694 DO - 10.1183/09031936.00086510 VL - 36 IS - 3 AU - Miller, M.R. AU - Pedersen, O.F. Y1 - 2010/09/01 UR - //www.qdcxjkg.com/content/36/3/694.abstract N2 - From the authors:We thank the authors for raising two questions concerning our recent paper that proposed alternative ways for expressing forced expiratory volume in 1 s (FEV1) patient data 1. First was about the possible adverse effect on our conclusions from the statistical modelling for deriving the prediction equations we used. Our work used established equations supplied from other sources. The prerequisites for linear regression are that: 1) the dependent variable, which in our case is FEV1, is normally distributed with respect to the independent or predictor variables; 2) any errors in the dependent variable are not related to the independent variables; 3) the scatter of the dependent variable is uniform and independent of the predictor variables (i.e. the data are homoscedastic); and 4) the relationship is linear.The overwhelming majority of reference equations have found that the scatter of normal lung function data does not violate homoscedasticity, … ER -