Available at

Average 
0.984 
0.990 
1.014 
0.946 
0.980 
As shown in Table 14, the overall average expansion rate in LOCOSSI is lower 
than 1 (value = 0.980), which means that interpreter trainees used the mechanism 
of compression. This corpus presents little evidence of compression, which 
confirms Chernov’s assumption (2004). However, Chernov’s theoretical framework 
should not be taken for granted as other studies did not show any evidence of 
compression (Iacovoni, 2010).
In general, interpreter trainees used fewer words in their interpretations. This is 
especially striking in STU04 and STU06’s interpretations. If we look at the table 
somewhat closer, we see that there are many differences across students and not 
so many across interpreting tasks. IN03 stands out as its STU average expansion 

---

Results and discussion 
 
page 66 
rate is the only one being higher than 1 but, in general, expansion rate scores are 
very similar across interpreting tasks.
Linking the delivery rate scores and the expansion rate scores can allow to outline 
some tendencies across students (as we said, lexical density scores are 
surprisingly similar across students). It is interesting to see that the majority of 
students who have an average delivery rate markedly higher than the IN average 
delivery rate, namely STU01, STU03, STU05, also have an average expansion 
rate higher than 1. In the same vein, the lowest delivery rate for IN01 and IN02 
was respectively STU04’s score (82.24 w/m) and STU06’s score (79.76 w/m) and 
the lowest expansion rate for IN01 and IN02 was respectively STU04’s rate (0.849) 
and STU06’s rate (0.801).
4.1.4 Interim conclusion 
Summing up the trends outlined in the previous sections, we can say that: