
美国将大数据应用于国际学生能力评估计划(PISA)_数据分析师
大数据是教育产业重塑商业模式,促使政府、商业组织和社会企业家通力合作将实证、创意、资源整合起来成就全民终身教育的基础。因此未来教育界的巨头将是那些能够把学术权威与信息和社交网络的协同效应结合起来的领军者。更为重要的是,这将使人们在运用大数据的基础上进行应用创新。这要求体制上的协同创新,要采取更有进取、更完善的公共政策,来改变目前教育界弊病:工业化的组织模式、官僚的和以应诉导向的工作方式和策略。
这不仅仅是增加教育透明度和公共责任的问题,甚至可以说这不是主要问题。简单地把数据公布于众不能改变学生学习,老师授课和学校运作的模式。信息公开并不能自然而然地引领我们运用大数据改革教育方法。相反这一做法经常造成民众和政府在信息的控制和所有权方面的对立情绪。
运用大数据实现教育产业转型的前提是摒弃我们社会的“只读”模式。透明和合作并举。目前的情况是,坐在大办公楼里一角的某位教育专家制定了规则,成千上百名学生和老师只能遵从,没有人知道这些决定是怎么来的。如果我们能分享数据、培育民间创新和实验、开拓创造性文化,大数据可以实现大范围的信任。难怪世界经济合作与发展组织(OECD)一项关于成人技能的最新调查显示:一个人的读写能力越好,就越容易信任他人。
协同消费就是很好的一个印证。如今,我们与陌生人共享他们的汽车,甚至是房子。协同消费使人人都可以成为小微企业家,其发展驱动力在于建立与陌生人的信任。想想我们在商业世界里的行为,我们在信任他人的基础上提供信息,心甘情愿地交出信用卡数据,和各个商业行业中可信的陌生人建立联系。教育界的数据分享离我们还非常遥远。
但是这应该是我们努力的方向。几年前我们引入了国际学生能力评估计划(PISA),一项针对各国15岁青少年可比较技能的全球调研。PISA提供了大量有关教育质量的数据。PISA计划使公共教育政策的制定更加透明、高效,帮助教育力量的分配重获平衡。在微观层面,仍存有很多质疑:老师认为这是政府又一个想控制他们的问责工具。那我们该做什么?今年我们实施了“我的PISA”项目,将PISA的分析工具分发到学校。现在每个学校可以用它与全球各地相似或完全不同的学校进行比较分析。
突然间原有的状态发生了改变;学校开始使用这些数据。例如,美国弗吉尼亚州费尔法克斯郡的十所学校的校长和老师们围绕第一份报告的结论开始了长达一年的讨论。在当地教育部门(和OECD)的帮助下,他们将开始第二轮分析,进行深入的数据挖掘,更好地了解如何相互类比,并和世界各地的其他学校进行类比。这些校长和老师不再把自己看作全球舞台上的观众,而是合作的队友。换言之,在费尔法克斯郡,大数据正在建立大范围的信任。
英语原文:
Big data is the foundation on which education can reinvent its business model and build the coalition of governments, businesses, and social entrepreneurs that can bring together the evidence, innovation and resources to make lifelong learning a reality for all. So the next educational superpower might be the one that can combine the hierarchy of institutions with the power of collaborative information flows and social networks. More than anything else, this will hinge on getting people to generate innovative applications on top of big data. It’s about the co-creation of governance, about delivering more progressive and better policies than the industrial work organisation and the bureaucratic and litigation-oriented tools and strategies that we are used to in education.
This isn’t just or even mainly about improved transparency and public accountability in education. Throwing education data into the public space does not change the ways in which students learn, teachers teach and schools operate. It does not lead to people doing anything with that data and transforming education in ways that will actually change education practice. On the contrary, it often results simply in adversarial relationships between civil society and government over the control and ownership of information.
The prerequisite for using big data as a catalyst to change education practice is to get out of the “read-only” mode of our societies. It’s about combining transparency with collaboration. The way in which educational institutions often work is that you have a single expert sitting somewhere in a corner who determines the application of rules and regulations affecting hundreds of thousands of students and teachers – and nobody can figure out how those decisions were made. Big data can lead to big trust if we make that data available, train civic innovators, experiment, create a maker culture. It is no surprise that OECD’s new Survey of Adult Skills shows that the more proficient people are in literacy, the more they trust others.
Collaborative consumption provides a great example of this. These days, people share their cars and even their apartments with strangers. Collaborative consumption has made people micro-entrepreneurs – and its driving engine is building trust between strangers. Think about it: in the business world, we have evolved from trusting people to provide information, to willingly handing over credit card data, to connecting trustworthy strangers in all sorts of marketplaces. We are light-years away from that when it comes to data about education.
But here’s how we can get a little closer. Some years ago we created PISA, a global survey that examines the skills of 15-year-olds in ways that are comparable across countries. PISA has created huge amounts of big data about the quality of schooling outcomes. PISA has also helped to change the balance of power in education by making public policy in the field of education more transparent and more efficient. At the micro-level, there were still a lot of sceptics: teachers thought this was just another accountability tool through which governments wanted to control them. So what did we do? This year we put in place a kind of “MyPISA” – PISA-type instruments that we circulated out into the field. Now every school can figure out how it compares with other schools anywhere else in the world, schools that are similar to them or schools that are very different.
Suddenly, the dynamic has changed; schools are beginning to use that data. Ten schools in Fairfax county in Virginia, for example, have started a year-long discussion among principals and teachers based on the results of the first reports. With the help of district offices (and the OECD), they will be conducting secondary analyses to dig deeper into their data and understand how their schools compare with each other and with other schools around the world. Those principals and teachers are beginning to see themselves as teammates – not just spectators – on a global playing field. In other words, in Fairfax county, big data is building big trust.
数据分析咨询请扫描二维码
若不方便扫码,搜微信号:CDAshujufenxi
Excel 数据聚类分析:从操作实践到业务价值挖掘 在数据分析场景中,聚类分析作为 “无监督分组” 的核心工具,能从杂乱数据中挖 ...
2025-09-10统计模型的核心目的:从数据解读到决策支撑的价值导向 统计模型作为数据分析的核心工具,并非简单的 “公式堆砌”,而是围绕特定 ...
2025-09-10CDA 数据分析师:商业数据分析实践的落地者与价值创造者 商业数据分析的价值,最终要在 “实践” 中体现 —— 脱离业务场景的分 ...
2025-09-10机器学习解决实际问题的核心关键:从业务到落地的全流程解析 在人工智能技术落地的浪潮中,机器学习作为核心工具,已广泛应用于 ...
2025-09-09SPSS 编码状态区域中 Unicode 的功能与价值解析 在 SPSS(Statistical Product and Service Solutions,统计产品与服务解决方案 ...
2025-09-09CDA 数据分析师:驾驭商业数据分析流程的核心力量 在商业决策从 “经验驱动” 向 “数据驱动” 转型的过程中,商业数据分析总体 ...
2025-09-09R 语言:数据科学与科研领域的核心工具及优势解析 一、引言 在数据驱动决策的时代,无论是科研人员验证实验假设(如前文中的 T ...
2025-09-08T 检验在假设检验中的应用与实践 一、引言 在科研数据分析、医学实验验证、经济指标对比等领域,常常需要判断 “样本间的差异是 ...
2025-09-08在商业竞争日益激烈的当下,“用数据说话” 已从企业的 “加分项” 变为 “生存必需”。然而,零散的数据分析无法持续为业务赋能 ...
2025-09-08随机森林算法的核心特点:原理、优势与应用解析 在机器学习领域,随机森林(Random Forest)作为集成学习(Ensemble Learning) ...
2025-09-05Excel 区域名定义:从基础到进阶的高效应用指南 在 Excel 数据处理中,频繁引用单元格区域(如A2:A100、B3:D20)不仅容易出错, ...
2025-09-05CDA 数据分析师:以六大分析方法构建数据驱动业务的核心能力 在数据驱动决策成为企业共识的当下,CDA(Certified Data Analyst) ...
2025-09-05SQL 日期截取:从基础方法到业务实战的全维度解析 在数据处理与业务分析中,日期数据是连接 “业务行为” 与 “时间维度” 的核 ...
2025-09-04在卷积神经网络(CNN)的发展历程中,解决 “梯度消失”“特征复用不足”“模型参数冗余” 一直是核心命题。2017 年提出的密集连 ...
2025-09-04CDA 数据分析师:驾驭数据范式,释放数据价值 在数字化转型浪潮席卷全球的当下,数据已成为企业核心生产要素。而 CDA(Certified ...
2025-09-04K-Means 聚类:无监督学习中数据分群的核心算法 在数据分析领域,当我们面对海量无标签数据(如用户行为记录、商品属性数据、图 ...
2025-09-03特征值、特征向量与主成分:数据降维背后的线性代数逻辑 在机器学习、数据分析与信号处理领域,“降维” 是破解高维数据复杂性的 ...
2025-09-03CDA 数据分析师与数据分析:解锁数据价值的关键 在数字经济高速发展的今天,数据已成为企业核心资产与社会发展的重要驱动力。无 ...
2025-09-03解析 loss.backward ():深度学习中梯度汇总与同步的自动触发核心 在深度学习模型训练流程中,loss.backward()是连接 “前向计算 ...
2025-09-02要解答 “画 K-S 图时横轴是等距还是等频” 的问题,需先明确 K-S 图的核心用途(检验样本分布与理论分布的一致性),再结合横轴 ...
2025-09-02