Inspite of the problems of sample disturbance, if 1) e or E vs log p curve is interpreted, reasonably with some corrections for sample disturbance, 2) if the sample disturbance is not too much 3) the field exploration correctly located all the consolidating strata, and 4) if the deposit is not very soft, computed settlement will usually be reliable to about 10 to 15% of measured values.
In soft stratified soil deposits, where the settlements are large, the reliability of settlement estimate becomes highly inaccurate. The computed settlement will be either too low or too high. Computed to measured settlement ratios may be in the range of 0.57 to 3.1. Finite difference and finite element techniques may be used to model settlement and time rates. But these sophisticated methods are not in vogue due to 1) settlement computations are sensitive to several parameters (C, etc.) as the classical methods, i.e. output is no better than the input, 2) necessity for computer access, and 3) lack of general familiarity with computer technology.
More valves of C, for a stratified deposits are needed to produce a statistically reliable "parameter". A minimum of two independently determined values of the compression index should be determined which are in agreement before a 'value' is selected for stratum.