本論文探討董事會成員的教育背景對企業層面氣候變遷暴露程度的影響，研究對象為2002年至2020年間多國上市企業，將BoardEx的董事教育資料與Sautner等人（2023）所建立的氣候變遷暴露指標相匹配，後者涵蓋整體暴露程度及其機會、法規與實體三個子維度。採用產業乘以年度固定效果的OLS迴歸，結果揭示出一個理論一致的學科專屬模式。具有較高比例博士與碩士學歷董事的董事會，其氣候暴露程度顯著較高，尤其集中於機會子維度，與其研究導向的分析能力相符。具有醫學背景的董事與所有暴露變數均呈現顯著且穩健的負向關聯，反映其集中於低碳排放醫療產業的結構性特徵。具有法律背景的董事則專一且顯著地預測法規暴露與風險框架型氣候論述，而非機會或實體面向，此一學科專屬結果在所有穩健性設定中均成立。整體而言，碩士學歷比例與法律法規管道的結論，對替代性CCExposure指標、剔除金融業樣本及被解釋變數取對數等處理方式均具有穩健性。;This thesis examines whether the educational background of corporate board directors is associated with firm-level climate change exposure. Drawing on a large international panel of publicly listed firms over the period 2002 to 2020， it matches BoardEx director education records with the CCExposure measures developed by Sautner et al. (2023)， which decompose aggregate climate attention in earnings conference calls into opportunity， regulatory， and physical sub-dimensions. OLS regressions with Industry-by-Year fixed effects reveal a theoretically coherent， discipline-specific pattern. Boards with higher proportions of doctoral and master’s-degree directors exhibit significantly greater climate exposure， concentrated in the opportunity sub-dimension， consistent with their research-oriented analytical capacity. Directors with a medical background are consistently and strongly negatively associated with all CCExposure dimensions， reflecting their structural concentration in low-emission healthcare industries. Directors with a legal background positively and specifically predict regulatory climate exposure and risk-framed climate discourse， but not the opportunity or physical dimensions， a pattern that holds across all robustness specifications. Findings for the law-to-regulatory-exposure channel extend to risk-framed climate discourse across multiple robustness specifications. Directors with a business background show only a marginal and inconsistent negative association that does not survive alternative specifications. The primary findings for Masters_Ratio and the law-regulatory channel are robust to alternative CCExposure measures from Sautner et al. (2023)， to exclusion of financial sector firms， and to log-transformation of the dependent variable.
