Technical Translation: Usability Strategies for Translating Technical Documentation

Approaches to Empirical Evaluation 
There are several approaches and methods for assessing usability which can 
be used. We can group these methods into two broad categories: 
analytical 
and 
empirical 
(Faulkner 1998:113). 
Analytical evaluation
, according to Preece (1993:109), uses formal or 
semi-formal methods of describing the interface in order to predict user 
performance. These methods include such strategies as GOMS (
goals
, 
op-
tasks
). Analytical evaluation is primarily a formative approach and as such is 
of little significance here. 
179
erators
, 
methods
and 
selection rules
) and KAT (
knowledge analysis of 

---

Assessing Usability 
Empirical evaluation 
as described by Faulkner (
ibid.
) includes expert 
evaluation or heuristic analysis, observational evaluation, survey evaluation 
and cognitive walk-throughs (Hill 1995:120). Within this broad category of 
empirical evaluation, we can distinguish between 
absolute
and 
comparative
studies. Absolute experiments (for example, expert or heuristic evaluation) 
involve assessing the interface or system on the basis of predefined specifica-
tions, criteria and scores. Such standards might include, for example, Niel-
other hand, involve assessing an interface and comparing it with some other 
alternative interface or version of the interface (Downton 1991:331; Faulk-
ner 1998:113). For the purposes of this research, the comparative approach 
is preferable because we are concerned with determining whether Iconic 
Linkage can improve the usability of user guides in comparison to user 
guides where Iconic Linkage is not present. 
We can further divide empirical evaluation into methods which involve 
users and methods which do not involve users. This distinction is important 
because Landauer (1995:281) maintains that, in contrast to analytical evalua-
tion, expert evaluation and cognitive walkthroughs, testing involving users 
is the best way of evaluating usability. He says “only by studying real work-
ers doing real jobs in real environments can we be sure that what we find 
out is truly relevant” (
ibid.
). With this in mind, the following pages will 
discuss only those empirical evaluation techniques which involve users, 
namely observational evaluation, survey evaluation and experimental 
evaluation. 
A critical factor in any form of evaluation is the type of information the 
evaluation is supposed to provide: quantitative or qualitative. Quantitative 
data is numeric and is analysed using a range of statistical and mathematical 
methods. This, in some regards, makes it easier to process large amounts of 
data in order to provide statistical evidence. However, unlike qualitative 
data, it does not provide the detailed subjective data or opinions that give 
an insight into how users actually perform tasks using the interface. Qualita-
tive evaluation does yield this information although the resulting data is 
more difficult to analyse in that it consists of a wide variety of diverse in-
formation which is often expressed in very different ways by users. 
However, in the case of this study, we are interested in examining the 
way users work with the interface in a significant amount of detail. As we 
discussed in Chapter 4, usability is determined not only by how quickly or 
efficiently users work with an interface but also by users’ opinions, satisfac-
tion and attitudes to the interface. In addition to quantitative data such as 
how quickly users perform tasks, we are interested in finding out how easy 
users perceive the use of the interface to be. In addition, the decision as to 
180
sen s heuristics (Nielsen & Molich 1990). Comparative experiments, on the 
’

---

Approaches to Empirical Evaluation
whether to gather quantitative or qualitative data has significant implica-
tions for the design of the experiment, the methods used and the number of 
subjects involved in the experiment. 
In the following sections, we will examine the various methods for con-
ducting usability evaluations with users in order to gather a combination of 
both qualitative and quantitative data.