Abstract
For the test of curve fitting, an indicator g was proposed, which is calculated_??_, where fi is the observed frequency, ei is the theoretical frequency calculated from the model distribution, and k is the number of classes within the normal range. Decisions are made as follows;“good” for g less than 1.18,“acceptable” for g between 1.18 and 2.00, and “dissociate” for g larger than 2.00. The fittness of Usui-Kawai's distribution model to the patients' data disribution of 22 laboratory tests were tested using this indicator, many of them showed good or acceptable fitting except very skewed distribution such as GPT,γ-GTP and TTT tests.
