Background. Authors of scientific articles in one language are often required to provide abstracts of their papers in a second language, and they use a variety of ways to achieve this. Aims. The aims of our studies were (i) to assess the effects of using native English speakers to improve non-native speakers' translations of their abstracts, and (ii) to explore the variations in edits produced by different native speakers. Methods. In Study 1, a French abstract was translated into English by two French authors. Each of these English abstracts was then edited by two native speakers of English (leading to four abstracts altogether). Textual and stylistic analyses were made of these four abstracts. In Study 2, over 200 academics rated these four abstracts on a series of semantic-differential scales to see how readers judged each abstract and what discriminations they made between them. Results. Study 1 showed how the two English editors imposed their own particular styles of writing on to the anglicized French abstracts, and how each changed the meaning in different ways. Study 2 showed that readers did in fact evaluate these pairs of English abstracts differently, indicating that the two English editors wrote in recognizably different `voices', whilst ostensibly carrying out the same task. Comparing the results of Study 1 (using textual analyses) with those of Study 2 (using readers' judgements), it appeared that improving readability and changing style were independent skills, and each had differing effects on meaning and understanding. Conclusions . These data suggest that there are benefits from using native English speakers to edit translated abstracts, but that there will never be a `one-to-one' translation from one language to another, and that different measures may reveal different effects.