A number of theoretical and experimentally recorded nose whistlers, whose initiating sferics have been identified, are used to assess the precision of the generalized linear- Q methods of whistler analysis. The calculated values of nose frequency, f nc , and nose-time delay, t nc , are compared with the observed values for different measured frequency ranges on each whistler. Large systematic errors are evident. To delineate these an Extrapolation Factor, E f , has been defined to indicate the extent of extrapolation, or interpolation, from the measured points to the nose. For long extrapolations there are very large systematic errors in f nc and t nc . When the linear- Q method is used to analyse measurements around the nose, f n , is interpolated accurately but t n is over-estimated by from 30% to 70%. The systematic errors arise because the Q ( f ) curve is nonlinear and changes shape as arbitrary sferic positions are selected. Digitizing errors give rise to only small random errors in f nc (<3%) but larger errors in t nc (<10%). The paper stresses the importance of using nose whistlers, with the causative sferic identified, for evaluating the precision of whistler analysis techniques and emphasizes the need to take into account the extent of extrapolation, or interpolation, from the measured points to the nose.