This study aims at testing the Learning-by-Exporting (LBE) on the TFP from the perspective of the evaluation literature. The focus is posed on the distribution of the outcome, the pre-entry selection bias is addressed, and both “statistical issues†and the influence of macroeconomic factors are accounted for. Basing upon a panel of Italian manufacturing firms, we design an experiment by aligning and pooling cohorts of starter, incumbent exporter and domestic firms and we further address the panel drop out. Main findings are that internationalisation has an impact on firms’ TFP, which is larger for best performing firms. Then it is shown that estimates of LBE impact are biased when i) the heterogeneous influence of macro-factors across groups and cohorts, and ii) the drop out of some firms from the panel are not accounted for.