Who Are the Talented Ones——The 'Excellence Gap' Among Chinese Urban and Rural Students and Its Contributing Factors

• In Event: Education, Difference and Development (I)

Proposal

Abstract: Talent plays a crucial role in enhancing a country's technological competitiveness, and the lack of representation of talented individuals from disadvantaged backgrounds raises significant social concerns about the barriers faced by the less privileged. Therefore, analyzing the sources of talent has important practical implications for building a strong education system and promoting educational equity. This article, based on PISA 2018 data, defines talented students as individuals with academic achievements at or above the 90th percentile. Firstly, by using logit regression and a random forest model, we estimated the factors influencing an individual's likelihood of becoming a talent. Subsequently, we employs Oaxaca-Blinder decomposition to explore the origins of the excellence gap between urban and rural students. The research findings are as follows:
1.Early educational experiences, metacognitive abilities, and growth mindset significantly impact an individual's likelihood of becoming a talent. On the other hand, factors such as family socio-economic status and school faculty, which have a greater influence on academic achievements, exert a smaller impact on whether one becomes a talent.
2.Within the excellence gap between urban and rural students, school average socio-economic status, and the quality of school faculty have a more substantial effect, similar to the factors influencing academic achievements. This suggests that urban schools are more effective in converting educational resources into student learning outcomes.
In light of the pursuit of high-quality and balanced compulsory education, it is crucial to further expand the availability of high-quality pre-school education and enhance students' metacognitive abilities to promote the development of talent and ensure fairness in education.
Keywords: urban-rural gap; academic achievement; excellence gap; random forest; Oaxaca-Blinder decomposition.

 
Some findings:
[About the excellence gap] The achievement gap is one of the crucial research directions in the field of educational inequality. In academic research and policy discourse, the disparities in academic achievement among different groups, such as those based on family backgrounds, urban-rural divides, and racial differences, are often regarded as key indicators to measure educational equity. Researchers frequently employ methods of comparing group means for analysis. However, relying solely on mean-based gap comparisons can oversimplify the issue, greatly overlooking within-group distribution differences and disparities across different percentiles among groups. Therefore, in recent years, similar to the attention given to "income polarization" in income distribution research, the study of the achievement gap has also shifted its focus from mean-based disparities to exploring the "excellence gap" (or "gifted gap") based on high levels of academic achievement. Plucker et al. (2010) first introduced the concept of the excellence gap, defining it as "the difference among students at the highest levels of academic achievement."


Fig 1: Urban and Rural Distribution of Math Test Fig 2: Urban-Rural Disparities in Math at Different Percentiles.

In the distribution of mathematics test scores among urban and rural students, the mean and high-scoring urban students are higher than the rural samples, indicating that the urban environment is significantly better than the rural environment. Regarding the urban-rural mathematics disparities at different percentiles, as the percentile increases, the disparities show a clear growing trend. At the median level, the gap is approximately 20 points, but at the 90th percentile, the gap is about 30 points.

Fig 3: Random Forest Estimation of Variable Importance in Determining Talent

The estimation results of the random forest model show that the five most important variables influencing whether an individual can become a talent are as follows: comprehension and memory-related metacognitive ability, duration of pre-school education, growth mindset, summarization-related metacognitive ability, and information evaluation-related metacognitive ability.

Authors

• Yuankang Xue, East China Normal University

• Yuheng Zhang, Institute of Education, Xiamen University