CLUST, INDEX, THRAN, and CONID Computer Programs (ICPSR 5910)

Summary View help for Summary

This data collection contains eight files of computer program source code. All of these programs were written in standard Fortran IV for a CDC 6600 system at Indiana University for use with John Gillespie and Dina Zinnes, WORLD TRADE DATA, 1958-1968 (ICPSR 5402). The CLUST computer program (Parts 1 and 5) finds clusters around nations using a trade index. It allows input such as trade data for all nations in the system and for primary nations around which clusters are built, and produces several types of output data such as raw trade data, computed trade scores, GINI indices of the trade scores, members of each cluster and their cluster contribution scores, and nations that belong to more than one cluster and their respective cluster contribution scores. The INDEX computer program (Parts 2 and 6) takes a score that can be computed from world trade data and computes the score for all possible dyads and the GINI indices of the score for each nation, rotates the matrix and computes the GINI indexes again, finds the set of nations related to a nation on the basis of the score between them being greater than a certain value, and produces a frequency distribution of the scores. The CONID computer program (Parts 3 and 7) computes and prints the observed-expected trade scores between all possible dyads of nations, GINI indices on those scores for each score vis-a-vis the rest of the world, a GINI index on the indices of all nations in the world, and dominance scores for all dyads. The THRAN computer program (Parts 4 and 8) does a "threshold analysis" on a set of paired values using the correlation coefficient. This program breaks a set of ordered pairs that are sorted on the independent variable (X) into subsets so that the correlation coefficient in each subset is maximized.

Citation View help for Citation

Export Citation:

• RIS (generic format for RefWorks, EndNote, etc.)
• EndNote

Subject Terms View help for Subject Terms