4.8 Accidental, few, present, abundant

Using a progressive ordinal classification (4.7) is in itself not essentially new in archaeology. Converting measurements to the logarithm of the values is closely related to this. Low values are emphasized somewhat more, high values somewhat less. One thousand finds is no longer one hundred times as much as ten finds, but only three times as much. Our classification is based on the same principle, on the understanding that we are dealing with discrete values, whereas in logarithms they are continuous. This was achieved by rounding off the result of the cube root to a whole number, which could of course also be done with the logarithm. The essential difference is in fact in the degree to which the influence of large numbers is reduced. In the 'raw' numbers that influence is large, in the logarithm it is less and in the cube root even more so. In the cube root the slope is at the beginning a little less steep and at the end somewhat steeper than with the (natural) logarithm. This provides a little more differentiation, both in small and in large numbers.

We feel that, for the analysis of the data from our regional archaeological investigation, the use of a progressive classification fits the best. The limitations of the data are simply too great to use the actual counts. The 'translated' table of the Michelsberg sites below (Table 11) will be analyzed using one of the graphic displays provided by Exploratory Data Analysis.