本研究旨在探究地方輔導團成員與國小現場教師如何在不同場域中形成素養題，並比較其命題實作差異。自十二年國民基本教育課程綱要實施以來，教育行政機關持續透過示例、研習與資源平台推動素養導向評量，但第一線教師對素養題的判準、教學範圍及評分方式仍存有疑慮。因此，本研究將焦點由題目結果前移至形成歷程，分析素養理念如何在互動、協商與修正中被重新塑形。 本研究採質性研究取向，以Michael Bratman的共享能動性為分析視角，將互鎖意圖、子計畫契合與相互回應轉化為指向性、相容性與因應性三項向度。研究參與者包括地方輔導團成員10名與國小現場教師6名，資料以半結構式訪談為主，並輔以政策文件、研習資料、評量示例、段考試卷及試題修訂版本。 本研究結果如下： 一、地方輔導團的素養示例研發，主要發生於相對專業化且責任風險較低的示例研發場域。團員的共同活動較能持續圍繞素養判準的釐清、學科本質的維持、學生認知起點的掌握、情境功能的檢視及題目品質的修正，使不同命題構想逐步形成可供國小教師參照的專業示例。 二、國小現場教師的素養命題則嵌入高責任密度的正式段考制度。段考試卷必須由同年級教師共同使用、共同評分並共同承擔，因此素養題能否進入試卷，不只取決於命題教師的理念與能力，也須同時回應教學範圍、作答時間、試卷篇幅、班際公平、答案明確性、評分一致性、同儕關係、家長反應及考後責任等條件。 三、兩個場域皆可能形成共享能動性，但呈現出不同的運作樣態。在指向性上，地方輔導團較能將素養示例的專業深化置於共同活動核心，國小現場教師則較穩定地共同指向段考試卷的完成與運作；在相容性上，地方輔導團較有機會整合學科本質、學生起點、情境功能與題目結構，國小現場教師則須優先使素養設計與時間、公平、評分及共同責任形成可運作的關係；在因應性上，地方輔導團的回應較常導向專業修題與示例深化，國小現場教師的回應則較容易在多重風險下導向題目明確化、安全化或其他補償性調整。 四、正式段考中的共同活動不只受到直接參與命題與審題者的影響。家長期待、行政規範、政策要求與考後責任雖未必直接進入命題討論，仍會透過教師的預期、風險判斷與責任考量形成「間接在場」，改變共同目標的排序、命題條件的安排及後續因應行動。因此，現場教師將題目調整為較明確、較封閉或較低風險的形式，並不必然代表其理念不足或能力不足，而可能是外部責任環境牽引下所形成的情境判斷。 五、本研究進一步提出「不便」，作為理解現場教師素養命題困難的一種去汙名化解釋。所謂「不便」，並非指教師單純覺得麻煩，也非指教師缺乏專業能力，而是指教師即使理解素養命題理念、具有實踐意願及一定的命題能力，其個人意圖仍可能因未能與共同段考的責任結構充分接合，而難以安心提出、穩定保留並共同承擔素養題。 依據上述結果，本研究指出，教育行政主管機關與地方輔導團在推動素養命題時，不宜僅著重於個別教師的理念宣導與命題技術增能，也須協助學校建立共同評量目標與命題語言，將評量討論前移至學期初或單元共備階段，並透過學生作答證據、評分規準、低威脅的共同修題及責任共擔機制，使素養命題由個別教師的命題嘗試，逐步轉化為學年教師可共同理解、修正與承擔的評量實踐。綜合而言，素養命題的落地不是政策理念的直接移植，也不只是教師個人能力的展現，而是一種在專業判斷、制度條件、人際互動與責任關係中持續協商的集體實踐。;This study examines how competency-based items are developed by members of local compulsory education advisory groups and elementary school teachers across settings and compares their item-development practices. Since the implementation of Taiwan’s 12-Year Basic Education Curriculum Guidelines， educational authorities have promoted competency-based assessment through sample items， teacher workshops， item selection programs， and resource platforms. However， frontline teachers continue to face uncertainty concerning item criteria， contextualization， curriculum scope， and the scoring of open-ended items. Accordingly， this study shifts attention from final item formats to the processes through which competency-based principles are translated， negotiated， and reshaped. Adopting a qualitative approach， this study draws on Michael Bratman’s concept of shared agency. Interlocking intentions， meshing subplans， and mutual responsiveness were contextualized as three analytical dimensions: orientation， compatibility， and adaptivity. These dimensions examine joint focus， the coordination of item-development considerations with institutional conditions， and how contextual signals shape adjustments. Participants included 10 advisory group members and 6 elementary school teachers. Data were collected through semi-structured interviews and supplemented by policy documents， workshop materials， assessment exemplars， periodic examination papers， and item-revision records. The findings are as follows. First， advisory groups develop sample competency-based items in settings with lower accountability risks. Their negotiations focus on competency criteria， subject-specific principles， students’ cognitive starting points， contextual functions， and item quality. Second， elementary school teachers develop items within formal periodic-examination settings marked by concentrated shared responsibility. Because examinations are jointly used， scored， and accounted for， teachers must consider curriculum scope， testing time， cross-class equity， scoring consistency， collegial relationships， parental responses， and post-examination accountability. Third， shared agency operates differently across the two settings. Advisory groups are oriented more toward refining competency-based exemplars， whereas elementary school teachers prioritize the stable completion of grade-level examinations. Advisory groups are also better positioned to integrate disciplinary principles， student readiness， contextual functions， and item structures， while school-based feedback more often leads to greater explicitness， closure， or risk reduction. Fourth， parental expectations， administrative regulations， policy requirements， and anticipated accountability may influence item development even when not directly represented in discussions. Through teachers’ expectations and risk judgments， these external forces indirectly shape priorities and revisions. Thus， more explicit， closed-ended， or lower-risk formats do not necessarily reflect inadequate understanding or competence， but may represent contextual judgments formed under shared responsibility. Fifth， this study offers a destigmatizing interpretation of teachers’ difficulties by framing them as contextually constrained practice rather than reluctance or professional inadequacy. Even when teachers understand competency-based principles， possess relevant expertise， and are willing to apply them， the shared-responsibility structure of periodic examinations may still make it difficult for them to confidently propose， retain， and collectively assume responsibility for such items. The study suggests moving beyond individual technical training by establishing shared assessment goals， collaborative item development， collective revision， and shared responsibility. Schools should move assessment discussions to early-semester or unit-planning stages and use student-response evidence， scoring rubrics， and a low-threat revision culture. Competency-based item development is best understood as a collective practice negotiated through professional judgment， institutional conditions， interpersonal relationships， and accountability structures.
