职业教育规模快速发展，为推动经济社会发展提供了庞大人才队伍。一直以来，高职院校毕业生的就业形势和收入情况都是社会和政府关注的热点话题。过去人力资本模型中的能力多指向单一认知能力，对个体工资收入解释力度有限。面对复杂工作环境和技术变革，非认知能力对个体经济表现的重要性逐渐凸显，近年来将非认知能力纳入分析框架的新人力资本理论成为国内学者的重点研究方向，但较少聚焦高职院校学生群体。2019年教育部和财政部联合印发《关于实施中国特色高水平高职学校和专业建设计划的意见》指出，实现引领新时代职业教育高质量发展的目标，需要加强学生合作、创新能力以及严谨专注等品质的培养。基于以上背景，本文以2022年全国“双高”院校毕业生就业情况调查数据为研究样本。论文选取合作性、尽责性和核心自我评价等三个代表性能力来测量非认知能力，采用明瑟收入方程和工具变量法，评估了非认知能力对高职毕业生的起薪影响的直接影响，并进行了性别、家庭背景和地区等方面的异质性分析。在此基础上，构建中介效应模型，探讨职业选择（是否选择白领工作）在非认知能力影响毕业生收入中的中介作用。实证分析得到以下结论：1. 非认知能力对“双高”院校毕业生起薪有显著的积极影响。具体而言，合作性、尽责性和核心自我评价及总的非认知能力等对毕业生起薪的影响均显著为正，其中核心自我评价对起薪的提升作用最大。为了克服内生性问题，本文选取同一行业内部除个体之外的非认知能力均值作为个体非认知能力的工具变量进入回归分析，两阶段工具回归的结果与基准回归一致。三个子指标的工具变量回归结果也得到了与基准回归模型一致的结论。2. 在异质性分析方面，非认知能力对起薪的影响在不同性别、家庭背景和地区存在差异。首先，非认知能力对在不同性别毕业生之间的收入效应不存在显著差异；其次，家庭社会经济地位更高的毕业生其非认知能力对起薪的影响强度明显大于家庭背景较差的毕业生，主要表现在合作性和核心自我评价子指标。最后，西部地区毕业生的非认知能力及其三个子指标的收入效应最大，尽责性指标对毕业生起薪的影响在中部和东部地区间差异不显著。3. 职业选择在非认知能力影响起薪中起到部分中介作用。非认知能力越高，职业院校毕业生从事“白领”职业的概率越高，进而起薪越高。具体维度上，合作性越强、核心自我评价越高的个体，越容易进入“白领”职业；而尽责性对起薪发挥直接促进作用，职业选择未发挥中介效应。基于实证分析结论，本研究提出以下建议：1.“双高”院校应该将学生非认知能力水平作为绩效评价指标，完善职业教育评价体系。2.高职院校应增加非认知能力培养课程、探索相关教学形式并构建良好学校氛围。3.父母应为子女营造非认知能力健康发展的良好家庭环境。

With the rapid development of vocational education, it has provided a large talent pool for promoting economic and social development. The employment situation and income status of graduates from higher vocational colleges have always been a hot topic of concern for society and the government. In the past, the ability in human capital models mostly referred to a single cognitive skill, which had limited explanatory power for individual income. Facing complex work environments and technological changes, the importance of non-cognitive skills in individual economic performance has gradually become prominent. In recent years, the new human capital theory that incorporates non-cognitive skills into the analytical framework has become a key research direction for domestic scholars, but it has rarely focused on the student population in higher vocational colleges. In 2019, the Ministry of Education and the Ministry of Finance jointly issued the "Opinions on Implementing the Plan for the Construction of High-level Vocational Colleges and Specialties with Chinese Characteristics," pointing out that to achieve the goal of leading the high-quality development of vocational education in the new era, it is necessary to strengthen the cultivation of students' cooperation, innovation ability, and qualities such as rigor and focus.Based on the above context, this paper examines the impact of non-cognitive skills on the starting salaries of graduates from "Double-high" colleges across China in 2022. Utilizing a representative sample of these graduates' employment data, we measure non-cognitive skills through three dimensions: cooperativeness, conscientiousness, and core self-evaluation. By employing the Mincer earnings equation and instrumental variable methods, we assess the direct effects of non-cognitive skills on the starting salaries of vocational graduates, and conduct a heterogeneity analysis across gender, family background, and regions. Furthermore, an intermediary effect model is constructed to explore the mediating role of career choice (specifically, the choice of white-collar jobs) in the relationship between non-cognitive skills and graduates' income. The empirical analysis reveals the following conclusions:1. Non-cognitive skills have a significant positive impact on the starting salaries of graduates from "Double-high" colleges. Specifically, cooperativeness, conscientiousness, core self-evaluation, and the overall non-cognitive skills all positively influence starting salaries, with core self-evaluation exhibiting the strongest effect. To address potential endogeneity issues, we select the mean non-cognitive ability level of individuals within the same industry (excluding the individual in question) as an instrumental variable for individual non-cognitive skills in the regression analysis. The results of the two-stage instrumental variable regression are consistent with the baseline regression. The instrumental variable regression results for the three sub-indicators also align with the baseline regression model. 2. In terms of heterogeneity, the impact of non-cognitive skills on starting salaries differs across gender, family background, and regions. Firstly, there is no significant difference in the income effect of non-cognitive skills between male and female graduates. Secondly, graduates from higher socioeconomic backgrounds exhibit a stronger effect of non-cognitive skills on starting salaries compared to those from less advantaged backgrounds, particularly in terms of cooperativeness and core self-evaluation. Lastly, graduates from the western region exhibit the largest income effect of non-cognitive skills and its three sub-indicators, while the effect of conscientiousness on starting salaries does not significantly differ between the central and eastern regions. 3. Career choice plays a partial mediating role in the relationship between non-cognitive skills and starting salaries. Higher non-cognitive skills increase the probability of vocational graduates entering "white-collar" occupations, leading to higher starting salaries. Specifically, individuals with stronger cooperativeness and higher core self-evaluation are more likely to enter "white-collar" professions. However, conscientiousness directly contributes to starting salaries, without the mediation of career choice. Based on these findings, the following recommendations are proposed: 1. "Double-high" colleges should incorporate non-cognitive skills into their performance evaluation metrics and refine their vocational education evaluation system. 2. Higher vocational colleges should introduce curricula to cultivate non-cognitive skills, explore relevant teaching methods, and foster a favorable school environment. 3. Parents should create a healthy home environment that fosters the development of their children's non-cognitive skills.