政府資助研發機構計畫績效評估之研究－以經濟部科技專案為例

Abstract

由於技術市場不健全、創新不足、發展需要等因素使然，不論是已開發國家或開發中國家皆試圖藉由技術政策來影響產業技術發展，以協助本國產業維持、提升競爭力；其中，以政府資助研發機構直接協助產業發展技術即為技術政策重要之一環。政府資助研發機構計畫之經費來源乃以政府公共預算浥注，受到近年來科技活動日益膨脹及各國總體預算緊縮等因素影響，要求此類計畫進行績效評估之 呼聲乃日趨高漲。為有效評估研發計畫之績效，許多學者依據研發活動的特性發展出績效評估模式，依其功能與特性可以區分為系統模式與階段模式。然因研發活動之複雜性使然，欲對其進行績效評估往往面臨許多問題，從而使得諸多評估活動與模式往往無法獲得正確的評價。 本研究透過文獻探討，先詳究績效評估之意涵，進而歸納出績效評估活動五要素：受評主體、衡量準則、衡量行動、評斷準則、評斷行動，再將文獻提及之龐雜問題歸整為受評主體問題、衡量問題、評斷問題等三大類，使績效評估面臨的問題得以條理分明，並將政府資助研發機構計畫之特性與此三大類問題一一比對，以收釐清政府資助研究機構計畫特性之效。 績效評估模式乃因應研發績效評估之困境而生，衡諸眾多模式，其內涵、目的雖然各異，本研究深加分析之後，從中推論其出共通點，從而歸納為投入、作業、產出、擴散、效益等五階段。由於績效評估模式乃屬概念模式，旨在適切表達研發活動，實際運用時必須依受評主體之特性予以調整；因此本研究將績效評估模式與政府資助研究機構計畫之特性予以連繫，以使評估模式概念足供政府資助研究機構計畫績效評估之用。 最後，本研究由歸整後之績效評估面臨問題分析出績效評估之要領，並配合與政府資助研究機構計畫特性連繫後之績效評估模式，再根據評估目的調整績效層面，歸納出多元尺度的評估方法。

Because of imperfect market for technology, insufficient innovation, and needs for development, many governments, including developed and developing countries, attempt influencing industrial technology by technology policy to assist domestic industries maintaining and elevating their competitiveness. Government-funded research institutes (GFRI), aiming to develop industrial technology directly, play an important role in technology policy. Constrained by expanding technology activities and shrinking public budget, performance evaluation of R&D project of GFRI is increasingly demanded. To evaluate effectively the performance of R&D project, many models conforming to the property of R&D project are elaborated. However, with the complexity of R&D project, evaluation would be faced with a lot of problems, and so many evaluation activities and models would not be properly justified. By literature review, this study explores the meanings of performance evaluation, and infers that evaluation activity contains five elements: entity evaluated, measurement criteria, measurement activity, judgement criteria, and judgement activity. Thence, complicated problems facing performance evaluation could be clarified by inducing all the problems to three categories: problems of entity evaluated, problems of measurement, and problems of judgement. Furthermore, to elucidate the performance evaluation of GFRI, all the three kinds of problems should be paralleled with characteristics of GFRI. Performance evaluation models in literatures are developed to solve the predicament of performance evaluation. In spite of the fact that each model is distinct from each other in connotation and purpose, some common properties could be deduced. That is, R&D activity could be divided into five stages: input, operation, output, diffusion, and effect. For practical usage, performance evaluation models, being conceptual ones, should be adjusted to characteristics of entity evaluated. To apply effectively the concept of evaluation model for GFRI evaluation, our study associates performance evaluation model with characteristics of GFRI. After inducing the recipe of performance evaluation from problems facing performance evaluation and accommodating the recipe to performance evaluation model associated with characteristics of GFRI, we conclude a multidimensional evaluation method by adjusting performance dimension according to evaluation purpose.