Technical Translation: Usability Strategies for Translating Technical Documentation

Observational Evaluation 
Observing users performing tasks they would normally perform as part of 
their work is one of the most useful ways of collecting data about what us-
ers do when they use an interface and how they react to the interface. Users 
can be observed in a specially built usability lab (Preece 1993:112; Schnei-
derman 1998:128-130) or in the normal environment where the users 
work. The latter scenario is often referred to as a field study (Preece 
1994:602). Observation can take a variety of forms and there are several 
methods for recording the information. There are two fundamental forms 
of observation: 
direct observation 
and 
indirect observation
.
Direct Observation 
Direct observation involves users performing tasks while an observer is pre-
sent in the same room. The observer watches the user and makes notes, 
times actions or performs some other function such as asking questions etc. 
While this approach is valued for its informality and immediacy, there are a 
number of problems associated with direct observation. The first and per-
haps most important issue with regard to the validity of the test results is 
that the presence of an observer can have a significant impact on the users’ 
performance. This may be attributed to users believing that their perform-
ance is under constant scrutiny and that they need to “impress” the ob-
server. The overall result is that users may actually perform better under test 
conditions than they would under normal working conditions, simply be-
cause they are trying much harder, a phenomenon known as the 
Haw-
thorne effect
(Preece 1994:617; Faulkner 1998:122). 
A second problem with direct observation is that the quality of informa-
tion gathered and indeed the completeness of the information relies on the 
ability of the observer to correctly interpret what is happening and then to 
write accurate and useful notes. Furthermore, if the observer misses some-
thing, there is no way of capturing the lost data or referring back to what 
happened – the experiments are single-pass occurrences and the only record 
181

---

Assessing Usability 
of events is the notes taken by the observer. Of course, using more than 
one observer may yield more comprehensive and complete data and also 
even counteract any possible biases associated with a single observer. How-
ever, if the presence of just one observer can affect a user’s performance and 
distort the results of the test – the effects of several watchful observers could 
have disastrous consequences for the validity of the experiment. 
Both of these problems, i.e. the Hawthorne effect and the unreliabil-
ity/incompleteness of notes, can be counteracted through the use of indi-
rect observation techniques (Faulkner 1998:122).