In search of experimental evidence on Scratch programming and students' achievements in the first-year college computing class? Consider these datasets.

This article presents datasets representing the demographics and achievements of computer science students in their first programming courses (CS1). They were collected from a research project comparing the effects of a constructionist Scratch programming and the conventional instructions on the achievements of CS1 students from selected Nigerian public colleges. The project consisted of two consecutive quasi-experiments. In both cases, we adopted a non-equivalent pretest-posttest control group design and multistage sampling. Institutions were selected following purposive sampling, and those selected were randomly assigned to the Scratch programming class (experimental) and the conventional (comparison) class. A questionnaire and pre- and post-introductory programming achievement tests were used to collect data. To strengthen the research design, we used the Coarsened Exact Matching (CEM) algorithm to create matched samples from the unmatched data obtained from both experiments. Future studies can use these data to identify the factors influencing CS1 students' performance, investigate how programming pedagogies or tools affect CS1 students' achievements in higher education, identify important trends using machine learning techniques, and address additional research ideas.

Impact of a Scratch programming intervention on student engagement in a Nigerian polytechnic first-year class: verdict from the observers.

An engaging first programming class (CS1) often inspires students' passion for computer science (CS). However, the evidence in the literature suggests that the average CS1 classes are anything but engaging for many students. The performance of CS compared to other science, technology, engineering, and mathematics (STEM) courses in international student engagement surveys seems to substantiate CS1 failure, attrition rates, and lack of diversity in most CS classes. Meanwhile, for its simplicity in introducing programming to beginners, primary and secondary schools use Scratch, an educational programming environment developed by the Massachusetts Institute of Technology, USA. For the same reason, higher institutions now include some forms of Scratch instruction in CS1. The question remains, to what extent is Scratch engaging, especially for students in higher education? This study addressed this gap by observing college computer science students exposed to a constructionist Scratch programming pedagogy. We adopted a descriptive design based on quantitative observations. To observe the class during a weekly 2-hour session, we employed five CS educators, one observer per week. Each observer, employing a 20-item observation protocol, rated the extent of affective, behavioral, and cognitive engagement of first-year CS students in a polytechnic in North Central Nigeria. Most of the students were learning to program for the first time. Analysis of the data showed a significant agreement in the ratings of the five observers for overall student engagement, although the impact was moderate. However, while agreement in their ratings for affective engagement was significant, with a large effect, there was no significant concordance in their ratings for behavioral engagement. Observers also significantly agreed in their ratings for cognitive engagement; however, the impact was moderate. These findings suggest that employing Scratch in higher education can be engaging and useful, especially for students with no prior programming experience.