Current Index to Statistics


Theses and Dissertations in the CIS Database

A Proposal for Including Theses and Dissertations in the CIS Database
Ronald Thisted, 16 March 1995

CIS does not have regular access to PhD and Masters theses, so systematically abstracting this information through our regular procedures would be haphazard at best.

I propose that we have departments of statistics and related disciplines submit lists of theses to us in machine-readable form, in a format that we specify. These files submitted by departments would either be accepted (essentially as is) or rejected (if any editorial work would be required at all to put it into the common format). Each year, departments would send us either (a) an incremental list to be added to the list from previous years, or (b) a replacement list for all of their graduates to date.

We would process the submitted data, merge the files from different departments, and index it, producing a file compatible in format with the existing records and inverted indexes to them. The dissertation records would probably not be made available in the form submitted by the departments.

Our proposed format is for departments to create an ASCII file, which they could transmit to us either via electronic mail, ftp, or floppy disk, with each thesis encoded in a refer-like format with the following markup tags:

Only relevant lines should be included; each field should appear exactly once per thesis. Distinct publications should be preceded by an empty line. Any of the lines marked with an asterisk (*) above that would be the same for all entries in a file may be entered just once, at the beginning of the file, following a line with the contents %H Common Information
and followed by an empty line. The availability information is optional, but if present should be one of the following possibilities: Not available, Contact department, Contact author, or University Microfilms (or other provider agency). Mailing addresses, telephone numbers, and the like should not be placed in the thesis record, as they may become out of date rapidly, and we have no good means of keeping them up to date.

Here is a sample entry: 

%A Thisted, Ronald A
%T Ridge Regression, Minimax Estimation, and Empirical Bayes Methods
%R PhD Thesis
%E Efron, Bradley
%I Department of Statistics, Stanford University
%C Stanford, California USA
%D 1977
%L English
%1 Xerox University Microfilms, Ann Arbor, Michigan
%K Shrinkage; Stein-James Estimator
Capitalization would follow the ``every significant word'' rule. Key words should not duplicate those in the title. Author's name should be in ``inverted order,'' that is, with family name first, followed by a comma, followed by other parts of the name.

In addition, each department submitting a file of theses would have to provide an ascii file, updated annually, indicating the name and contact information of a person to whom inquiries concerning the accuracy or completeness of information in that department's entries can be addressed. This information will be published on the CD-ROM as part of the documentation for thesis records.

Only departments or university divisions may submit material to be included; entries submitted by individuals will be declined.

Comments and submissions should be addressed to editor@statindex.org

Home | Technical Reports | Technical Documents

Last updated: 02/13/2004    [Send a question or comment]