|Abstract or Summary
- Calls continually are made to provide economic assessments of research program
achievements and efficiency. Yet little effort has been given to develop an assessment
framework that would focus on the research discovery itself, treating the research
manager as a producer and the research technology as a knowledge production function.
The present dissertation develops such a framework and uses it, with a variety of
analytical approaches, to evaluate a two-phase international aquacultural research
program consisting of 55 distinct studies.
A Bayesian knowledge measure is developed for this purpose, allowing close
examination of each of two knowledge creation pathways – the extent of new findings
(mean surprise) and the extent of uncertainty reduction (precision). Factors affecting
each of these two pathways are estimated in decomposed form, their total effects on
knowledge achievement then combined to form an aggregate knowledge production
Team workload, education level, and scientist travel distance strongly affect
knowledge creation as postulated, although exhibiting varying effect magnitudes and
significances across the two program phases. A research study's analytical approach
significantly affects its knowledge acquisition pathways, accounting partially for the
newness of its scientific discoveries. Survey studies tend, in contrast, to have greater
potential for new findings, but yield greater uncertainty than do experimental studies.
In each of the two program phases, fish market trading and water quality are, in my
output-elasticities-based approach, respectively the least productive topic area and
research-outcome dimension. Asian researchers appear – compared to their colleagues in
South America and Africa – to achieve the highest predictive precision but the least mean
surprise, probably because of the greater maturity of their projects. In both program
phases, estimated output elasticities imply increasing knowledge returns to scale,
although the elasticities decline from 3.52 in Phase I to 1.07 in Phase II.
The dual cost function approach provides indirect insight into the program
manager’s investment decisions and to the returns to knowledge output, complementing
the primal approach. In my cost-based approach, knowledge cost elasticities are below
unity, estimated at 0.49 in Phase II and 0.37 in Phase I, consistent with the increasing
returns to scale found in the output-elasticities-based approach. Given the increasing
returns to scale estimated with both approaches, the aquacultural program appears to have
a substantial incentive to enlarge its knowledge investments. Also consistent with duality,
the least-output-productive fish-trade topic area, water-quality outcome, and Asian
research are found in my cost analysis to be the most cost-consuming.
The technical efficiencies of the aquacultural program's individual studies are also
examined, relative to both one another and to their own potentially best practices. The
examination is conducted using, successively, the Farrell input technical efficiency
measure and the directional sum-distance measure. Results are consistent across these
two efficiency instruments, confirming the conclusions about output and cost elasticities
in the previous chapters and providing a completeness to the overall research evaluation.