Technical Translation: Usability Strategies for Translating Technical Documentation

Sullivan & Chapanis 1983 
Like Foss 
et al.
(1981), Sullivan and Chapanis set out to investigate ways of 
improving the usability of computer documentation by evaluating different 
versions of a user guide for a text editing software application. However, 
this particular study differs from that of 
Foss et al.
on a number of points: 
firstly, this study involves the creation of an entirely new manual from 
scratch; secondly, the study is concerned with the content and formulation 
of information in the manual as opposed to the organisation of information 
or the use of supplementary sources of information. 
According to the authors, the purpose of this study was… 
ment and, based on that experience, to propose a general methodology 
for the preparation of documentation (Sullivan and Chapanis 1983:114). 
The study consists of two broad components. The first component in-
volved producing an “improved” manual on the basis of a review of litera-
ture on writing instructions. This involved implementing style guidelines 
such as using short, active sentences, simple language, sequencing events in 
the order they are performed etc. The second stage involved conducting a 
comparative analysis of the original manual and the new version of the 
manual. Both of these stages involved an element of experimental evalua-
tion and are described in greater detail below. 
The experiment was carried out over two sessions for each subject. In the 
first session, subjects familiarised themselves with the computer using a key-
sen
to use either the old manual or the new manual. The first task was allocated 
pressurised
tasks within
205
… to improve an existing computer manual through the application of 
rules of document preparation, to measure the extent of the improve-
board learning program. In the second session, subjects were randomly cho
2.5 hours but subjects were informed that they should not feel 
because most people would be unable to complete all of the
given a 
the allotted time. When this task was completed, subjects were 

---

Assessing Usability 
for an
hour. 
During the tasks, the subjects were told that if they encountered a prob-
lem, they should first try to solve it themselves. They could only ask the 
experimenter if they really could not solve a problem themselves. If subjects 
asked general questions, the experiment referred to the manual. If subjects 
asked specific questions, the experimenter gave specific answers. 
When the tests were completed and all data had been gathered, perform-
ance was measured on the basis of the following criteria (Sullivan & Cha-
panis 1983:119): 
quality of text produced 
number of different commands used successfully 
number of questions asked 
type of questions asked 
Recognising the importance of subjective user attitudes to overall usability 
levels, the authors used a post-test questionnaire to determine subjects’ atti-
tudes to the manual, the software and the tasks. However, attitudes to the 
software and tasks were not related to which manual was used (
ibid.
:122). 
In contrast to the initial questionnaire, the one used here consisted of thir-
teen 7-point Likert scale questions. Four of these questions related to the 
manual while the remainder related to the software and the tasks. 
The authors found that the subjects who used the new manual performed 
significantly better in the tasks than the subjects who had used the old man-
ual and that the new manual group were able to use more commands suc-
cessfully without the assistance of the experimenter. They also found that 
the group using the old manual asked more than four times as many ques-
tions as the other group and that this indicated serious problems with the 
old manual. 
The questions asked by subjects during the tasks were categorised as fol-
lows:
1. manual organisation problems 
2. global misunderstanding 
3. not knowing what command to use 
4. failing to distinguish between writing and editing modes 
206
15 minute break before commencing the second task which lasted
•
•
•
•

---

Experiment to Test the Impact of Iconic Linkage
5. not knowing how to use a command or code 
6. system-dependent problems 
7. task clarification questions 
8. negative comments 
The results of the test showed that in all categories except category 4, the 
subjects who used the new manual asked far fewer questions than the other 
group. In category 4, the number of questions remained constant. 
The overall findings of the study showed that it is possible to improve 
the usability of user documentation through a combination of “local” pro-
duction rules and iterative production processes. The authors also maintain 
that the involvement of potential users in the development process is essen-
tial.