| Popis: |
A validity methodology of the Gerber method through a comparison study to the Rose Gottlieb reference method is presented.The analysis of Gerber (G) and Rose Gottlieb (RG), in a total of 100 raw milk samples, with a range of 1.80 and 5.50% (m/m) of fat, were carried out according to oificial FIL-IDF 152 A: 1997 and FIL-IDF ID: 1996 rules respectively. The regression equation found (RG = 0.9998 G + 0.0278) presented a gradient not significantly different to 1.000 (0.9846 < b < 1.0149) and an intercept not significantly differing from 0 (-0.0274 < a < 0.0830). The residual standard deviation was Sy,x = 0.0853 and the accuracy of the method was therefore + 0.17g/100 g (+ 1.987 Sy,x) at present conditions. Gerber method exactuess, expressed as repeatability, presented a repeatability standard deviation of Sr= 0.0308 and a repeatability of r = 0.087 g/100 g, according to FIL-IDF 128: 1985; thus allowing to estimate both the method reproducibility (R = 2 r) and confidence Interval, including the mean precision and accuracy (random error and systematic error). On the other hand, the analysis of media values indicated the acceptance of the nule hypotesis, t0.05 = -0.1645 (P = 0.86) and a correlation coefficient of r = 0.9972.It is concluded that Gerber Method for milk, carried out according to FIL-IDF 152A: 1997 rule, with a milk volume of 10.77 ml presented acceptable validity parameters, in lineal, precision and accuracy terms. |
