Optimizing Survey Data: Factor Reduction with SPSS in STEM Education

Abstract

Survey instruments are foundational to many disciplines, educational, social and behavior research. Yet their reliability and use are often compromised by excessive item length, respondent fatigue, and analytical complexity. This study demonstrates the application of Principal Component Analysis (PCA) within SPSS to optimize 5-point Likert-scale instrument measuring perceptions of STEM (Science, Technology, Engineering, and Mathematics) education among Albanian university students. The study sample contains 120 respondents who completed STEM-program secondary education. Several statistical measures are implemented, Kaiser-Meyer-Olkin (KMO) and Bartlett’s tests, followed by Varimax-rotated PCA to determine optimal number of factors without compromising data reliability. Results yielded a robust four-component solution—Cognitive & Career Development, Curricular & Pedagogical Quality, Institutional Support, and Stakeholder Engagement—accounting for 82.1% of total variance. Findings reveal that while Albanian students recognize STEM’s cognitive and career benefits, they report significant deficits in institutional resources and teacher preparedness. Methodologically, this study offers a useful model for survey optimization, enhancing reliable results without sacrificing validity. The approach is generalizable across educational and social science contexts where efficient, psychometrically sound instruments are needed for evidence-based policy and practice.

Received: 10 July 2025 / Accepted: 27 August 2025 / Published: 25 September 2025