Much debate surrounds the issue of whether allowing candidates to listen to recordings twice is more desirable in language tests than offering just one opportunity. Using regression models, this study investigates, analyses and interconnects both item difficulty and stimulus length in relation to the frequency of stimulus presentation and its effect on test results. The study involved 1,266 eighth-grade learners with perceived A2/B1 level in English as a first foreign language. The final discussion seeks to offer explanations for the results that clearly indicate the advantages of listening once; likewise, potential implications for listening tests and classroom practice are raised.