Abstarct: High-stakes English proficiency exams for PhD students in Iran mainly focus on vocabulary, grammar, and reading. However, little is known about how test takers actually experience these exams or how their experiences can inform more valid and fair assessment practices. This study explored that issue through an exploratory sequential mixed-methods design focused on the English Proficiency Test (EPT) of Islamic Azad University. Phase 1 used a transcendental phenomenological approach with purposive sampling of recent EPT takers. Data were collected through in-depth interviews, reflective diaries, and artifact elicitation to identify key features of their testing experiences. The analysis (including epoche, horizonalization, and imaginative variation) revealed five main themes: (1) perceived lack of relevance between test content and doctoral communication needs; (2) reliance on memorization-baxsed and compensatory strategies; (3) anxiety and pressure caused by unclear scoring and frequent retake cycles; (4) mixed feelings about the exam’s learning value, combining reports of vocabulary improvement with perceptions of narrow study focus; and (5) concerns about cost and unequal access to preparation opportunities. Phase 2 transformed these qualitative findings into a pilot EPT Perception Questionnaire, which was administered to a larger group of doctoral candidates. Exploratory factor analysis confirmed a multifactor structure corresponding to the qualitative themes and showed satisfactory reliability for all scales. Correlation and regression analyses revealed theoretically consistent patterns: greater perceived transparency was lixnked with lower anxiety and higher fairness perceptions; perceived content relevance predicted broader strategy use and stronger beliefs in test usefulness; both factors were associated with higher pass rates after controlling for background variables. Comparisons with other tests (MSRT, TOLIMO, TOEFL, and IELTS) indicated that Iranian local tests emphasize a narrower construct and stronger washback toward discrete-point study, while international tests demonstrate broader construct coverage and clearer interpretive frxameworks. The discussion proposes several policy and design recommendations: expanding the EPT to include writing and speaking in modular formats, publishing clear test specifications and interpretive arguments, using results diagnostically with remedial support for low scorers, and reducing financial and informational inequities in access. This study offers a validated instrument for ongoing monitoring, provides a model for turning lived experiences into measurable constructs, and positions the EPT as a program that can be refined to promote more positive educational outcomes while maintaining feasibility. Limitations include its cross-sectional design and lack of external performance measures; future research should pursue longitudinal validation and task-design experiments.