%0 Thesis
%9 S3
%A Ndayizeye, Oscar
%A Putro, Nur Hidayanto Pancoro Setyo
%B Penelitian dan Evaluasi Pendidikan
%D 2026
%F UNY:90799
%I Sekolah Pascasarjana
%K program evaluation, Fuzzy C-Means, TPACK, internship, I–U  collaboration
%T SMART KB-RLB Evaluation Model for Interns’ Soft Skills  Enhancement in The Industry–University Collaboration.
%U http://eprints.uny.ac.id/90799/
%X This research aimed to 1) design the SMART KB RLB internship evaluation  model susceptible to evaluate internship program soft skills enhancement in the  Industry-University collaboration, 2) prove the model's feasibility and effectiveness,  3) determine the model’s instruments' psychometric quality, 4) demonstrate interns’  soft skills enhancement in the industry–Universitas Negeri Yogyakarta collaboration,  and 5) prove the level of TPACK integration in the internship program.  This is an R&D project that merges the evaluator’s brand KB with SMART data  analysis using the Fuzzy C-Means Machine Learning algorithm with R-Reaction, L-  Learning, and B-Behaviour (henceforth, KB-RLB) to develop a higher education  internship program evaluation model. As the focus was on general soft skills  enhancement, the research took place at Yogyakarta State University, specifically in  the Faculty of Languages, Arts, and Culture (FLAC), within the English Literature and  Indonesian Literature Study Programs. The research subjects included interns,  academic supervisors, and internship coordinators (n = 120). The data collection  techniques included a questionnaire, a situational judgement test (SJT), and  documentation. RStudio version 2025.09.2+418 was used to analyse quantitatively the  psychometric validation (CTT, IRT, CFA) and fuzzy C‑means clustering to profile  interns’ mastery of soft skills, while data reduction, condensation, and verification  were used to analyse the qualitative data.  The R&D research results indicate that the SMART KB-RLB model design  1) includes the following components: reaction, learning of seven soft skills,  behaviour, TPACK integration, instruments that satisfy psychometric properties  standards, Feedback, and SMART Data Analysis through algorithmic interns’ soft  skills mastery clustering, all of which enable program recommendations. The model is  found to be 2) feasible as it can be logistically and contextually applied in a real-world  internship program soft skills enhancement evaluation without significant barriers; it  is also effective as it helped to profile interns’ soft skills mastery, it identified aspects  to optimise, is cost and time-effective, and environment-friendly as most data are  collected online and available in electronic form. In terms of the 3) model’s quality,  the psychometric property analysis indicates strong validity (Aiken’s V≈ 0.86),  moderate reliability (CR ≈ 0.70), and good CFA fit (CFI/TLI > .93; RMSEA = .035;  <0.14 <SRMR <0.08); IRT parameters indicate fair-to-good discrimination and  moderate item difficulty (b = –1.790 to 0.330). As evidence of soft skills enhancement,  4) the Fuzzy C-Means clustering identified two intern profiles: “Excellent” (46.2%,  ≥85, that is, an A) and “Moderate” (53.8% of interns, have 79≤ score <83, a B score);  communication skills emerged as the strongest predictor of supervisor and overall  interns’ scores, with time management soft skills correlating strongly with problem-  solving (r = 0.57). 5) TPACK integration is still basic as it is used for production and  communication; it is not yet at the pedagogical transformative level.