This paper presents the results of a small-scale study into advanced trainee interpreters’ performance in tasks which involve consecutive interpretation of openly evaluative texts, with particular focus on the use of agentless structures and nominalizations by male and female subjects. It seeks answers to the following questions: i) Is the interpreter’s involvement in the ongoing discourse a factor that may elicit agentless structures in the output? ii) Does the preference for such constructions seem to be related to the gender of the interpreter? The analysis is based on 40 interpretations of four formal addresses, of which two express criticism and the other two praise of the audience. One text in each set is addressed to students of English at the University of Silesia, a group to which the trainee interpreters belong and with which they identify. The results indicate that while there is no substantial difference in the use of agentless structures in contexts which involve identification of the interpreter with the ultimate receiver and in contexts which preclude identification, nominalizations tend to be used slightly more frequently in the former set of circumstances. It also appears that female interpreters are more likely to use nominalizations in texts which express open evaluation of the audience with which they identify, irrespective of the direction of valuation.