Integration of Deep Learning Approaches and PjBL Models: Analysis of Students' Writing Skills Cycle Towards News Publication

Abstract

This study aims to analyze the effectiveness of the integration of the Deep Learning approach and the Project-Based Learning (PjBL) model in improving the news writing ability of high school grade XI students towards digital publications. The main problem of the study is the low critical reasoning and factual verification skills of students in conventional news text learning. The method used is Classroom Action Research (CAR) model Kemmis and McTaggart in two cycles. Data were collected through intensive observation, documentation, performance tasks, and interviews, which were analyzed using the interactive models of Miles, Huberman, and Saldaña. The results showed a significant improvement in students' writing skills, with classical completeness increasing from 20% in the pre-cycle to 53.33% in Cycle I,and reaching a peak of 90% at the end of Cycle II. Qualitatively, the integration of the Deep Learning dimension through personal coaching has successfully overcome students' literacy and technical barriers. This research provides an implication that the synergy of the Deep Learning approach and the PjBL model is transformatively able to transform grade XI high school students into critical and collaborative information producers. The final results were validated by the success of the publication of student news works that met professional journalism standards (factual and fitable) on the school's digital platform. Based on these findings, the model is recommended as a strategic framework to answer the challenges of digital literacy at the high school level.