Assessing a Data Administrative Program - High-Performance Web Databases: Design, Development, and Deployment - page 686

A wide variety of ways to collect measurements are not included on the
survey. For example, the program that has the goal of improving data qual-
ity and the data quality metadata will simply collect the measurements that
already exist. Data administration programs that have operationally
defined all their goals in this way would be able to conduct assessment of
their program at any time.

Storing the raw data
collected in a database allows the data administration department to use it
in the future (e.g., for comparison with future assessments or for other pro-
gram uses). A program assessment is likely to be a large investment — one
that must be protected. If a survey software package is used, it will likely sup-
port the long-term storage of the survey itself, the raw measurements, and
the compiled results. Several statistical analyses can be performed on the
results. If unfavorable results occur from running the statistical analyses,
they should not be ignored. If they are accurate, no matter how unpleasant,
they are revealing something significant about the program. Data administra-
tors should take this opportunity to address them with assertive solutions.

Step 5: Collating information and calculating results.

This step is imperative. It is worse to conduct
an assessment and not report the results than to not perform an assess-
ment. If the results are not flattering, it should be used as an opportunity to
focus on how to improve things. if the results are flattering, data adminis-
trators should shout them from the rooftops and send a copy of the results
to everyone. One of the final products of the assessment should be an
updated data administration program plan. It should document the data
administration program accomplishment, provide the rationale for drop-
ping any tasks from the original plan, and forecast the strategies and tasks
for the out years.

Step 6: Reporting results.

USING RESULTS FROM OTHER SURVEYS

The ideal situation would be to conduct periodic surveys and keep the
results to compare year after year. Few data administration programs are
old enough to have had this luxury. They can simulate that situation by
using the results of other surveys. Some example surveys include:

• DAUG, “Data Administration Survey Report.”
• GUIDE “Selling Data Administration for the 80s and Beyond.”
• Center for Information System Research, MIT, Sloan School of Manage-
ment, “Managing the Data Resource: A contingency perspective.” This
paper presents case studies of 31 data management efforts in 20 differ-
ent firms.
• Database Newsletter, “Results of the 1991 Advanced Information
Resource Management Survey Part I.” The survey collected informa-
tion from US and Canadian firms.