249
IV. Summary
In spite of substantial effort in agricultural data collection, review of the available summaries
and reports suggest significant problems with quality. As stated earlier, it is difficult to review data
series after a period of time has elapsed unless there are technical notes that document the survey
procedures including problems that may have affected the quality of data collected. Unfortunately, few
such notes exist in either the MAFF or the CSO publications. However, the quality of Zambia's
agricultural data may be improved in the future in a number of ways. The most important steps are
to ensure the elimination of all unnecessary questions from the two main CSO and MAFF surveys and
then to concentrate on the development and maintenance of solid data collection techniques. This will
not be an easy task. It will take the courage of policymakers and survey managers alike to make
commitments to data quality rather than data quantity.
The most expeditious way of beginning to change the Zambian data system is to employ a
team of outside evaluators to review the system in detail. Once the evaluation is complete, and a plan
of work is agreed upon by the appropriate governmental units, the government should request the aid
of donors in mounting a sustained program of survey evaluation and improvement. Following are four
alternative levels of technical support donors should consider providing to the Zambian agricultural
data system. Each alternative offers unique benefits. Assisting Zambia with a combination of these
inputs would be most effective.
(1) Short-term trip(s) with the specific task of designing and organizing surveys to collect data for
LTC research projects might be considered as a minimal level of support. Working on
specific LTC research questions, short-term technical assistance (STTA) would collaborate
with CSO statisticians in designing appropriate samples. STTA would explore the use of CSO
sampling frame to facilitate probability samples. Areas of possible data-collection collaboration
could be considered with the MAFF and/or the CSO field personnel. Any collaboration would
be accompanied by in-country training provided by the STTA. The number of trips provided
in this category would be totally dependent on LTC research needs.
Benefits: Professionally designed surveys resulting in high-quality research data. Collaboration
and a small amount of training provided for the MAFF and the CSO staff.
(2) One or two short-term trips to search for existing data that would be helpful to the LTC
researchers would probably produce information that was previously overlooked. Data series
or administrative records very likely exist which have potentially valuable information for
researchers. These may include information from governmental or parastatal units (e.g.,
mortgage listings, title registrations, and the like), or listings currently being compiled for the
CSO sampling frames. The search for these information sources would take some time and
would depend on STTA's ability to convince government officials of the value of cooperating
in the effort.
Benefits: Possible discovery of information that could be made available to the project and
others at a relatively low cost.
(3) STTA would develop an ongoing relationship with CSO and MAFF to assist in providing
incremental improvements to the agricultural data system. Following a system review, the
STTA would work to encourage changes that appear to be most effective in improving data