Towards validity for a formative assessment for language-specific program tracing skills

Towards validity for a formative assessment for language-specific program tracing skills
复制标题

针对特定语言的程序追踪技能的形成性评估的有效性

DOI:
10.1145/3364510.3364525
复制
发表时间:
2019
期刊:
ACM Koli Calling International Conference on Computing Education
影响因子:
--
通讯作者:
Ko, Amy J.
Ko, Amy J.
中科院分区:
--
文献类型:
--
作者:
Nelson, Greg L.;Hu, Andrew;Xie, Benjamin;Ko, Amy J.

文献摘要

参考文献

被引文献

相似文献

形成性评估可以对学习产生积极的影响,但很少存在计算,即使是基本技能,如程序跟踪。相反,教师往往依赖过于宽泛的测试问题,缺乏衡量早期学习所需的诊断粒度。我们遵循Kane的评估有效性框架,设计了一个JavaScript程序跟踪的形成性评估,开发了“针对特定用途的有效性论证”。“这包括:1)细粒度的评分模型,以指导实践,2)项目设计,以测试我们的细粒度模型的一部分,具有低混杂引起的方差,3)覆盖测试设计,从项目空间中采样,并覆盖评分模型,以及4)形成性使用的有效性的可行性论证(可以瞄准和改善学习)。我们贡献了凯恩框架的精华,该框架适用于计算教育,以及凯恩框架在程序跟踪形成性评估中的新颖应用,重点关注评分、概括和使用。我们的应用程序还有助于一种新的方式建模可能的概念的编程语言的语义建模流行的组合物的控制流和数据流图和通过它们的路径,一个过程中产生的测试项目,并尽量减少项目混淆的原则。
Formative assessments can have positive effects on learning, but few exist for computing, even for basic skills such as program tracing. Instead, teachers often rely on overly broad test questions that lack the diagnostic granularity needed to measure early learning. We followed Kane's framework for assessment validity to design a formative assessment of JavaScript program tracing, developing "an argument for effectiveness for a specific use." This included: 1) a fine-grained scoring model to guide practice, 2) item design to test parts of our fine-grained model with low confound-caused variance, 3) a covering test design that samples from a space of items and covers the scoring model, and 4) a feasibility argument for effectiveness for formative use (can target and improve learning). We contribute a distillation of Kane's framework situated for computing education, and a novel application of Kane's framework to formative assessment of program tracing, focusing on scoring, generalization, and use. Our application also contributes a novel way of modeling possible conceptions of a programming language's semantics by modeling prevalent compositions of control flow and data flow graphs and the paths through them, a process for generating test items, and principles for minimizing item confounds.
学校及其他领域计算机科学评估的新视野:利用 ViVA 平台
DOI: 10.1145/2858796.2858801
发表时间: 2015
期刊: Proceedings of the 2015 ITiCSE on Working Group Reports
影响因子: --
作者:
D. Giordano;F. Maiorana;A. Csizmadia;S. Marsden;Charles Riedesel;Shitanshu Mishra;Lina Vinikiene
通讯作者: Lina Vinikiene
独立于语言的 CS1 知识评估的项目反应理论评估
DOI: 10.1145/3287324.3287370
发表时间: 2019
期刊: ACM Technical Symposium on Computer Science Education
影响因子: --
作者:
Xie, Benjamin;Davidson, Matthew J.;Li, Min;Ko, Andrew J.
通讯作者: Ko, Andrew J.
探索学生自我评价在入门编程中的价值
DOI: 10.1145/3291279.3339407
发表时间: 2019
期刊: Proceedings of the 2019 ACM Conference on International Computing Education Research
影响因子: --
作者:
Rodrigo Duran;Jan Rybicki;Juha Sorva;Arto Hellas
通讯作者: Arto Hellas
DOI: 10.1145/3013499.3013500
发表时间: 2017
期刊: ACM SIGCSE Bull.
影响因子: --
作者:
Andrew Luxton;Andrew Petersen
通讯作者: Andrew Petersen
学生在在线编程入门课程中的体验和评估使用
DOI: 10.1109/latice.2017.13
发表时间: 2017
期刊: 2017 International Conference on Learning and Teaching in Computing and Engineering (LaTICE)
影响因子: --
作者:
Emma Riese
通讯作者: Emma Riese