Search

In this study, the authors analyzed data from a sample of thirty-two middle school students from an urban school district in the southeastern United States who used MIT's App Inventor to design, create, and remix mobile apps during an afterschool program for one school year. This paper focuses on computer science learning outcomes as measured by an assessment created by the authors. Findings indicated a linear relation between the number of apps a participant created during the given afterschool program and their level of accuracy on the assessment. Findings and discussion should be of interest to both researchers and practitioners.

We investigated relationships among expressed emotions, perceived motivation, perceived emotions, and three dependent variables (i.e., behavioral engagement, cognitive engagement, and student achievement) in the context of a blended online course. We defined "expressed emotions" as emotions detected by an affective computing tool in messages that students posted to online discussion forums in a blended undergraduate writing course. The results of two-step hierarchical multivariate regressions revealed that expressed emotions differentiated positive emotions from negative emotions better than perceived emotions did. Moreover, while no significant effect emerged for perceived motivation and perceived emotions, expressed fear was a significant predictor of student achievement (i.e., final score). Although affective computing is in its infancy, our findings suggest the potential use of expressed emotions for educational research and practice.

This study, which took place during a 7-week middle school computing course, used structural equation modeling to examine the overall cumulative relationship between self-efficacy, interest, and prior computing experience and students' computer science learning outcomes. The findings indicated that 52% of the variance of student success, measured by a computational thinking quiz and rubric-based evaluations of participants' computing artifacts, was related to the aforementioned learner characteristics. These findings have implications for theory and practice and suggest that future research and instructional design practice in K-12 computing education should take these learner characteristics into account.

Background and context: The variables that researchers measure and how they measure them are central in any area of research, including computing education. Which research questions can be asked and how they are answered depends on measurement. Objective: To summarize the commonly used variables and measurements in computing education and to compare them to best practices in measurement for human-subjects research. Method: Systematic literature review analyzing 197 papers published during 2013-2017 in computing education research venues. Findings: The review illuminates common practices related to: variables measured (including learner characteristics), measurements used, and type of data analysis. The paper lists standardized measurements that were used and highlights commonly used variables for which no standardized measures exist. Implications: The paper concludes with information about best practices currently being used in the community that should be continued, as well as pointing out practices that could be improved along with recommendations for how to begin to adopt those best practices.