The inadequacy of unidimensional plotting along a continuum (in this case the diagonal of a symmetric matrix) inevitably would make "buffer" elements appear non-conformist when in fact they may be part of an interconnected pattern.

No data are excluded on subjective or arbitrary grounds. No one piece of data is more highly valued than another. The consequences of this policy have to be accepted, even if they prove awkward.

Statements that will hold good for all time are difficult to obtain in archaeology. The most that can be done at any one time is to report on the current state of knowledge.

That so far the material has been dealt with in a rather subjective way provokes the question whether a means can be found of handling it objectively. [...] This chapter considers the applicability of the statistical tests employed by Wilson and the general problem whether the Linear B data are suited to statistical analysis.