As, according to comments from our test persons, producing multistimulus associations is subjectively hard, in order not to overload the test takers, we divided the dataset into 40 sections, each comprising 50 items. A total of 40 test sheets were printed, each showing the test items from one section. The test takers were instructed to produce for each test item of five words the spontaneous association that first came to mind. It was also mentioned that there are no “right” or “wrong” answers, but that we were only surveying human word associations.
As an example, the list “work desk bureau secretary post” was provided, together with the response “office”. As the experiment was conducted at the University of Mainz, Faculty of Translation Studies, Linguistics and Cultural Studies (FTSK) in Germersheim (Germany), all participants were non-native English speakers. For this reason, they were asked to rank their English proficiency on the following scale: native / very good / good / satisfactory / basic knowledge / no or very little knowledge. Note, however, that because the FTSK is specialized in training translators and interpreters, of whom many focus on English, as expected the majority of students indicated “very good” English proficiency, and none indicated a proficiency below “satisfactory”. The experiments were conducted in the winter term of 2014/15 and in the summer term of 2015 in the courses given by the author. Course topics were Translation Systems, Language and Cognition, Computational Linguistics, Electronic Dictionaries, Corpus Linguistics, Machine Translation, Computer Aided Translation and Translation Memories (mostly seminars at either undergraduate or graduate level). Participation in the association test was voluntary. As the task was perceived to be tedious, many questionnaires showed some omissions. Altogether, 66 questionnaires were filled out, which implies that some of the 40 sections of the dataset were dealt with by more than one student.