This study investigated the effects of a group of task factors on advanced ESL learners’ actual and perceived listening performance. We examined whether the speed, linguistic complexity, and explicitness of the listening text and characteristics of the text necessary for task completion influence comprehension. We also explored what textual factors learners perceived as causing difficulty. The participants were sixty-eight students who performed 18 versions of a listening task, each followed by a perception questionnaire. Nine additional students engaged in stimulated recall. The listening texts were analysed in terms of a variety of measures, utilizing Praat v5.0.25, Cohmetrix v2.0, and WebVocabProfiler v3. We used Rasch and regression analyses to estimate task difficulty and its relationship to the text characteristics. Six measures emerged as significant predictors of task difficulty, including indicators of lexical range, density, and diversity, and causal content. The stimulated recall comments were more reflective of these findings than the questionnaire responses.