Abdominal Cavity Compliance: A Participant More in the Building Up of Visceral Obesity

give very accurate predicted values of the outcome variable.In other words, there is absolutely no relationship between the intercept values and the predicted values for two regression lines.We agree with Yajnik et al. that a good model should be based on the functional relationship between various predictors and the outcome, and the selection of variables.A statistical model is an abstract of the realistic phenomenon present in the sample on which it has been developed.The primary objective of a statistical model (equation) is to perform well outside the data set on which it has been developed.Even a good equation may not perform well if the characteristics of the two populations (on which developed and applied) are different or if the predictor variables have been measured differently or have a measurement error. DIsclOsuREThe authors declared no conflict of interest.