Technical Translation: Usability Strategies for Translating Technical Documentation

Error Rates 
A two-sided Wilcoxon-Mann-Whitney exact test of the null hypothesis of 
no difference between the total error score for the Control group and Ex-
perimental group was also conducted. This yielded a 
p-value of 0.0079
in-
dicating that there was a statistically significant difference between the 
scores for the Control group and the Experimental group. 
We can see from Figure 7 that the Control group using the original ver-
sion of the user guide committed on average 31.9 errors in comparison to 
the Experimental group who committed 10.9 errors. This represents quite a 
significant divergence in the effectiveness of each group of users in per-
forming tasks. Indeed, looking at individual error criteria we can see some 
rather striking trends. 
Some 60% of Control group failed to complete one task (criterion 1) 
During the tasks, the Experimental group consulted the user guide 44.4% 
less than the Control group (criterion 2) 
The Experimental group remembered 84.2% more commands and icons 
than the Control group (criterion 3) 
242
ware would be used and used computers every day. While the pilot group 
pants in the main study were more experienced in the particular work 
environment, had a greater understanding of the context in which the
•
•
•
•

---

Results of Usability Evaluation
There were no observations of frustration in the Experimental group 
while in the Control group there were three individual observations of 
frustration (criterion 7) 
There were 76.9% fewer observations of confusion in the Experimental 
group (criterion 8) 
There was a single observation of satisfaction in the Experimental group 
compared with no observations in the Control group (criterion 9) 
The Control group used 58.8% more incorrect commands or input (cri-
terion 10) 
In the Control group, two participants mistakenly thought they had 
completed one or more tasks; in the Experimental group, all participants 
successfully completed the tasks before finishing work (criterion 11) 
Comparing the average total error rates for each group (Figure 8), we can 
see that the Experimental group, which used the user guide featuring Iconic 
Linkage, committed on average 10.9 errors over the course of the test. In 
comparison, the Control group, which was using a user guide without 
Iconic Linkage, committed on average 31.9 errors over the course of the 
test. To put this another way, we can say that the Experimental group 
made 65.8% fewer mistakes than the Control group. 
243
The Experimental group made 69.2% fewer incorrect icon choices than 
the Control group (criterion 4) 
The Experimental group made 73.5% fewer incorrect menu choices than 
the Control group (criterion 5) 
The Experimental group did not ask any questions during the test 
whereas the Control group asked on average 1.8 questions (criterion 6) 
•
•
•
•
•
•
•
•

---

Assessing Usability 
Figure 7: Average Usability Criteria Error Scores for Each Group in Main Study 
Figure 8: Average Group Error Rates in Main Study