本研究的焦點在於經濟合作暨發展組織（OECD）會員國及臺灣科學素養之比較，研究內 容包括彙整科學素養評量結果、進行科學素養成就統計分析、探究影響科學素養成就的相關 因素、評估OECD 會員國及臺灣科學素養成就的整體效率值，以提出各項教育績效評估的建 議。研究對象為參與PISA 2012評量的學生為主。本研究主要資料來源為OECD之2012年「國 際學生能力評量計畫」的調查結果，採用的研究方法為三階段包絡分析法：第一階段利用DEA 中的CCR、BCC 模式分析各種效率值，第二階段利用Tobit 迴歸去除環境因素所造成的相對 無效率，第三階段則把調整後的產出項投到模型，分析調整後的各種效率值。研究結果顯示， 經第二階段調整產出項後，第三階段分析結果顯示，各國的平均技術效率值及純技術效率值 是提升的，而規模報酬遞減的國家數量明顯增加，且與第一階段的效率值有顯著的差異。所 以各國皆面臨不同的經營環境，這些因素會干擾各國的經營效率評估，如果不事先予以消除， 將會使各效率值的評估結果產生偏誤。因此，可以發現達到技術效率、純技術效率、規模效 率值為1 的國家數量增加，而規模報酬固定的國家由30 個減至13 個，顯示此報酬狀態為該 國在投入與產出的比例上最為恰當，因此無須做任何改善與調整，規模報酬遞減的國家由5 個增加至22 個，顯示此報酬狀態為該國在投入與產出的比例上過大，因此各國政府可考慮調 整或降低投入產出的比例，避免可能因規模過度擴張導致資源浪費的情形。

This study compared science literacy between the Organization for Economic Cooperation and Development (OECD) member countries and Taiwan. To evaluate educational achievements, the study included archiving the assessment results of science literacy, statistically analyzing science literacy accomplishments, investigating the factors that influence science literacy, and assessing the overall efficiency of the science literacy accomplishments that the OECD member countries and Taiwan have achieved. The study participants were the high school students of the OECD member countries and Taiwan who had participated in the Programme for International Student Assessment (PISA) 2012. The study data source was OECD PISA 2012 results. A 3-stage Data Envelopment Analysis (DEA) was used. The 1-stage DEA was employed to analyze variety efficiency in the DEA of the CCR and BCC models. The 2-stage DEA was used to remove environmental factors caused by the relative inefficiencies in the Tobit regression analysis. The last stage placed the adjusted output into the model and analyzed the efficiency. It also revealed more decreasing returns to scale countries. Because different countries have different management environments, these factors can interfere with the efficiency assessment. If they are not eliminated in advance, a deviation may be observed in the analysis. This study indicated that the number of countries with technical efficiency, pure technical efficiency, and a scale efficiency value of 1 has increased, while the constant returns to scale countries from 30 to 13, indicating that the status of returns is the most appropriate proportion in input and output. Therefore, no improvements and adjustments were observed. A decrease was observed in returns to scale countries from 5 to 22, indicating that the status of returns is a huge proportion gap in input and output. Therefore, governments should consider adjusting or reducing the proportion of inputs and outputs, as resources may be wasted due to excessive scale expansion.