QUANTITATIVE METHODS IN ARCHAEOLOGY
Anthropology 726
Exercise 9
1. First let's try a little experiment:
Using the data in PLAY4.DTA (the same data used in the previous exercise), create a
scatter plot showing the relative locations of the points. Hint: In order to facilitate
comparison in the next step, specify an aspect ratio of 1in your plot (using the aspect
option). This produces axes of equal length. Also be sure to label your points (using the
mlabel option).
Next, subject the same dataset to nonmetric multidimensional scaling in two dimensions
[mds varlist, id(varname) method(modern) loss(stress)]. How well does STATA's
multidimensional scaling algorithm reproduce the relative locations of the points in your
original scatter diagram? What are the differences between the two plots? Can you
explain these differences using your knowledge of the nonmetric multidimensional
scaling technique?
2. The second part of this exercise is designed to give you practice in using a variety of seriation
techniques. The data consist of type frequencies (percentages) at nine Late Woodland
components in the Eno River drainage of North Carolina. Site Or231H is a historic village with
European trade goods dating to about AD 1700. The rest of the components lack European trade
goods and are presumably prehistoric. The data are presented to you in two forms: (a) a bar
chart on the sheet attached showing the relative frequencies of types at the various sites; and (b)
a STATA file called ENOSITES.DTA in which the cases are sites and the variables are types.
Your job is to do the following:
Seriate the assemblages by eye using Ford's technique. (That is, cut the attached chart
into strips and rearrange strips into battleship curves.)
Now seriate the assemblages using Gelfand's method. Hint: The dissimilarity matrix can
be generated by STATA. Using the matrix dissimilarity procedure, compute a matrix of
"city-block" distances between sites. STATA calls this measure L1 (case sensitive). Be
sure to use the names option to label your rows and columns. The "city-block" or L1
metric is simply a re-scaled version of the Brainerd-Robinson coefficient, and so it
should produce the same results as the latter. After generating the matrix, you can list it
on the screen with the matrix list command; the nohalf option will list the entire
rectangular matrix. You can then cut and paste the result into your word processor.
Now seriate the assemblages using nonmetric multidimensional scaling (using mds, as
above). Use the same dissimilarity coefficient (L1) as in the previous step.
Discuss and interpret your results. (Note: You can use Kintigh's FORD.EXE program to
produce nice seriation graphs and print them using the same techniques as in previous exercises.)