These findings suggest that both translators and post-editors became faster by the end of the training period, but the improvement was more pronounced for translators. This is not surprising, as the training program itself was (and still is to a large extent) tailored to translators’ needs, and no courses were offered in post-editing. Clearly, translators acquired skills during training that had made their work temporarily more efficient, and post-editors must have profited from the same training, too. At the same time, it can be assumed that the lack of explicit training in post-editing skills imposed limitations on their capacity to accelerate.

An alternative or complementary explanation of this finding is that the nature of the post-editing task itself imposes constraints on how much the speed of post-editing can be increased. As students’ responses to the follow-up questionnaire indicated, the greatest advantage of post-editing is that it saves typing time, and secondly, that there is a draft translation to rely on. These factors are already present at the beginning of training, too. Thus, it can be inferred that the time gain had actually occurred before training, or more exactly, out of its context. Comparing text segments and making decisions about them, however, are cognitive processes that must always be performed, and it may take more time for them to become faster, as the majority of these processes are likely to be conscious. In contrast, it is possible that a large proportion of time gains in HT is due to an increase in automatic processing.

Another interesting finding concerning the temporal dimension of translation and post-editing was the reduction in standard deviations in both conditions. This can be interpreted as a homogenization of the group. In other words, students became more similar to each other in terms of delivery time. Homogenization is a phenomenon also reported by Quinci (2024) in her study with translation students.

When comparing B English and C English students, no significant differences were found: both groups showed increased processing speed. Therefore, it may be inferred that the category of the working language (i.e., B or C) is not a factor related to temporal gains.