7 Conclusions and Neolithization model

7.1 Methodological conclusions

There are both theoretical and practical objections to a site-typological analysis in a regional archaeological investigation. To mention a few, the limited representativeness of surface finds, inaccuracy of chrono-typological dating, the distorting influence of post-depositional processes or the large differences in quality of documentation between amateurs and museums. Furthermore there are a large number of research factors affecting the analysis, such as the choices we made in favour of certain amateurs and museums, the use of the centre of chronological gravity, allowing a maximum of 10% of other guide artefacts on a 'clean' site or reducing the counts to progressive ordinal classes. Yet the results presented here seem to allow the conclusion that a site-typological analysis is worthwhile after all. The decision in the Meuse Valley Project to consider the nature of a site as one of three archaeological correlates for the (food) economy is certainly useful, provided the data and techniques are suitably adapted and it is accepted that the results are indicative only.

An important analytical decision is the reduction of the data to ordinal classes. The quality of the data simply does not allow an analysis of counts or percentages. A 'level of measurement' has been selected that suits the quality of the archaeological data and yet still contains sufficient information for a typological analysis. As far as we are concerned, the emphasis is on graphical analytical techniques, where the data are presented in a number of different ways so as to allow exposure of the structure in the data. In this respect our approach closely matches the principles of the Exploratory Data Analysis and its use in GIS-applications (Farley et al. 1990; Williams et al. 1990) .