瓜地馬拉原住民與非原住民之間教育落差的演變： Oaxaca-Blinder 分解模型比較分析

Abstract

本研究深入探討了瓜地馬拉原住民與非原住民之間教育差距在過去二十年（2006-2023年）間的演變。研究採用 Oaxaca-Blinder 分解模型，將學歷年限的差異分解為由可觀察特徵（稟賦）解釋的部分和無法解釋的部分。本研究使用了兩種模型：基準模型（考慮了年齡、性別和地理位置等基本人口統計與社會經濟因素）和擴展模型（在上述因素之外納入了父母教育程度）。 分析揭示，在兩個時期和兩種模式下，非原住民人口的學歷年限均較高，顯示出顯著但持續存在的差距。一個關鍵發現是，當擴展模型納入父母教育程度時，差距中可解釋部分的比例顯著增加，這突顯了人力資本代際傳遞在教育不平等持續存在中的深遠作用。例如，2006年，總體差距中可解釋部分從基準模型的44.2%顯著擴大至擴展模型的81.6%。儘管總體差距隨時間略有縮小，但無法解釋的部分仍然顯著，尤其是在2023年（基準模型為69.0%，擴展模型為33.2%）。這表明，諸如歧視、教育系統的文化相關性或教育的差異回報等不可觀察因素，在造成這些差異方面日益關鍵。本研究進一步按地理位置（農村/城市）和性別對結果進行了細分，持續證明父母教育在所有亞群體中均具有更高的解釋力，同時也揭示了持續存在的無法解釋部分的獨特構成模式。最終，這些結果強調了採取綜合政策干預措施的必要性，這些措施不僅要擴大教育機會，還要解決結構性不平等並推動文化相關的教育，以真正消除瓜地馬拉的教育差距。

This study conducts a deep examination of the evolution of the educational gaps between indigenous and non-indigenous population in Guatemala over two decades (2006-2023). Using the Oaxaca-Blinder decomposition model, the research breaks down the disparities in years of schooling into explained and unexplained components, by observable characteristics (endowments) and an unexplained part. Two models were used, the baseline model, which considered demographic and socioeconomic fundamental factors (age, sex, and geographic location), and an extended model, that incorporated additional factors to the variables. The analysis reveals an important but persistent gap in both periods and modes, where the non-indigenous population achieve more years of schooling. A critical finding in the substantial increase in the proportion of the gap, when the parental education is included in the extended model, this highlights the profound role of the intergenerational transmission of the human capital in the perpetration of the education inequality. An example, in 2006, the explained component in the overall gap significantly expanded from 44.2% in the Baseline to 81.6% in the Extended model. While the total gap shows a slight reduction over time, the unexplained component remains significant, particularly in 2023 (69.0% in the Baseline, 33.2% in the Extended). This suggest that unobservable factors, such as discrimination, the cultural relevance of the educational system or the differential returns of education, are everyday more crucial in the generation of these disparities. The research further disaggregates finding by geographic location (rural/urban) and by gender, consistently demonstrating the increased explanatory power of parental education across all subgroups, while also revealing distinctive patterns in the composition of the persistence unexplained portion. Ultimately, these results emphasize the need of integral politics intervention, that not only expand the access to education, but that also address structural inequalities and promote culturally relevant education to truly eliminate educational gaps in Guatemala.