Despite the growing interest in the use of Twitter in higher education, its effects on aspects such as student engagement and learning continue to yield unclear and even contradictory results. The combination of evaluative methodologies for these different impacts contributes to confusion and bias. The aim of this study is to analyse whether the use Twitter to share and discuss current news produces engagement and its effects on learning by university students. The analysed sample is composed of 93 students earning a political science degree, split into three groups. The analysis is based on a mixed approach combining statistical and qualitative methods (content analysis and thematic coding of self-reported assessments) used to evaluate the degree of engagement, the evidence of learning and the relationship between these two dimensions. The results show a very low degree of engagement and little evidence of learning. The relationships between one dimension and another dissipate after detailed scrutiny of the statistically significant correlations found and are inflated by enthusiastic self-evaluations of the learning experience by students. The results warn about the risks for the research community of magnifying the benefits of the use of Twitter for educational purposes. The educational community should reflect on the effectiveness and efficiency of the use of Twitter for the purpose of stimulating independent, collaborative and reflective learning.