Michael Fischer Reading and Analyzing Cross-cultural Data from tables

  1) Look at the marginal totals for the row category and column category.
  These give the internal distribution of societies across these categories. This
is useful to contextualise later thoughts. For example, looking at the table
below we can see that 117 of 186 societies in the sample have the wife
residing with the husband’s group after marriage. We would therefore
expect to see some of our larger cross-comparison totals to be in the cells
in this column. Likewise for Bride-price or wealth. Indeed the cell which
includes both of these values is the largest in the table, which we would

  2) “Read” some of the results by translating the totals and cells for the table
  into expressions in your native language, e.g. English, French, Japanese
etc. You must read the table before you can say much about it.

  3) Now look at the distribution of values down the columns and across the
  rows. Compare rows. Compare columns. Do values seem to be evenly
distributed (accounting for the differences in marginal totals). Are their
any inversions (eg larger values to the left for Bride-price, to the right for


  4) The following table (which you get by checking the “Expected” box on the
  opening screen, helps to access the distribution. It relays how many more
or less societies we find in a cell than we would expect if the two categories
did not interact or relate to each other.


  5) Here we have removed columns to make the relationship more explicit. I
  did this using the merge facility, keeping the three most divergent row and
column values.

6) Another Example...