HUMAN

INFORMATION
PROCESSING
MODEL
IE-311 Ergonomics 2
 If the organism carries a ‘small scale model’ of external reality
and of its possible actions within its head it is able to try out
various alternatives, conclude which is the best of them, react to
future situations before they arise, utilize the knowledge of past
events in dealing with present and the future, and in every way
to react in a much fuller, safer and more competent manner to
the emergencies which face it.

CRAIK, K. 1943
 Human–machine interaction depends on a two-way exchange
of information between the operator and the system. Designers
usually have detailed, explicit models of machines and machine
behavior that can be used to improve human–machine
interaction. According to Preece (1993), a general model of the
user’s cognitive processes and cognitive behavior is also needed
to:
 Provide knowledge about what can and cannot be expected of
users
 Identify and explain the nature and causes of problems
 Supply modelling tools to help build more compatible interfaces

THE NEED OF A MODEL

 According to Reason (1990), in broad terms, the human
information processing system (HIP) can be thought of as
 A general-purpose pattern recognizer
 With limited information processing capacity
 Using heuristics (rules of thumb) to simplify the information processing
load
 And acting as a ‘satisficer’ rather than an optimizer

A GENERAL INFORMATION
PROCESSING MODEL OF THE USER
 Sensation occurs when environmental stimuli impinging on an
organism give rise to neural events. These events may or may not
be processed further. Perception is the construction of a model
of the event that gave rise to the initial sensation. The percept is
not a replica of the external world but a representation of it that
depends both on the neural ‘hardware’ and on previous
experience.

FROM SENSATION TO PERCEPTION

 Physical level: This is the letter ‘A’ written in italics.
 Phonetic level: This word sounds like ‘cheese’.
 Semantic level: This flashing red light means danger.
 Other levels: This is a nice photograph of my grandmother

FURTHER PROCESSING OF THE

INCOMING INFORMATION
 Banners are distinctive perceptual objects located on web
pages to help users orientate themselves during search (Figure
12.2). According to Benway (1998), these banners are often
missed by web users, particularly when they are placed on top of
the webpage, away from the links at the bottom. It seems that
people don’t just look at web pages, they look for things to click
on and if banners are distant from links, or cannot be clicked on,
their presence is not registered. Banner blindness illustrates how it
is possible to look at something, but not see it (that is, processing
never gets beyond the lower stages of the hierarchy in Table
12.1).

BANNER BLINDNESS
 Only a fraction of the information impinging on the senses
reaches conscious awareness. As soon as environmental energy
impinges on the senses, it is transduced, distorted and coded by
the very processes that transmit it to the higher centers of the
brain. The concept of coding is central to the information
processing approach. To use an analogy with computing, data
are often coded in some way before being entered into the
machine (names might be abbreviated, for example) according
to the requirements of the software. Having been entered and
stored in some way, the coded data can then be operated on
by programs. Thus, what is operated on is a coded version, or
representation, of the external (real world) data.

CODING
 Since cognitive processes act on percepts, not on stimuli, the
way the system codes a stimulus determines what can or will
follow. The stimulus configuration and the coding mechanisms of
the perceptual system determine what will be perceived.
Memory also plays an important role. This has major implications
in ergonomics – particularly in display design. The information
displayed to an operator must be designed to be perceived in
an appropriate way for the task. This requires that the task
requirements be analyzed early in the design stage and that the
skills and knowledge of the operator be taken into account.

CODING AND COGNITION

 Short-term memory (STM) can be likened to a temporary store (or
buffer) in which small amounts of information are briefly retained
while a particular mental or physical operation is carried out.
 Long-term memory (LTM) contains symbolic structures built up
through learning in which new data can be embedded. General
knowledge of the world and of life events is stored in LTM. We
may sometimes be unable to retrieve this information but can be
reminded of it by others, evidence that LTM storage is more or
less permanent. LTM is associative in nature: new data can be
represented in the context of past behavior, but this takes time.

TWO KINDS OF MEMORY

 The probability of correct recall was found to diminish as the
retention interval increased – from about 80% after a delay of 3
seconds to 10% after a delay of 18 seconds. (Peterson and
Peterson 1959)
 Mental operations reactivate items in STS that otherwise decay
rapidly.
 The drop in recall probability over time was assumed to be due to
the ‘decay’ of the memory trace of the consonants in the short-
term store (STS).

STORING MEMORY
 Miller (1956) concluded that the ‘magical number 7, plus or minus 2’ was
the limit on the capacity of STS.
 STM limitations have important ergonomic implications. People who have
to use short-term memory in their jobs are easily distracted by interruptions
of even a few seconds – particularly by conversation – or by having to
carry out verbal secondary tasks (such as answering the phone while
doing mental arithmetic). Tasks that require STM storage from multiple
sources can be said to be of high mental workload.
 A multichannel theory of attention was proposed by Allport et al.
(1972). Humans are multimodal in the sense of being able to
process and represent information in many modalities (e.g. visual,
auditory, semantic) and many purposeful activities, such as
walking and standing can be carried out without being
consciously attended to. Attention is seen more as a problem of
allocating processing resources to tasks. The multichannel model
of attention is presented in Figure 12.5.

HOW MANY TASKS CAN WE DO

SIMULTANEOUSLY?
 A limited-capacity processor handles incoming data but specialized
subroutines requiring no processing capacity can be used to free
the processor for additional tasks. The main features of the
multichannel view are summarized in Table 12.3.
 The model is intuitively attractive in that it can account for how
we can walk and talk at the same time but might have difficulty
in walking and juggling or in talking and reading simultaneously.
Activities that occupy different modalities are fed by different
channels to dedicated subroutines and can be carried out
simultaneously. Activities that occupy a single modality can only
be carried out simultaneously if channel capacity is not
exceeded.

RATIONALE
 Allport et al. (1972) and Schaffer (1975): A multi-channel view would
say that the shadowing task and the memory for pictures task utilize
different limited-capacity channels that can operate independently
of each other. Copy typing from visually presented text does not
involve the processing of verbal information (skilled typists do not
read text when typing, they just type in the letters), so the copy
typists had spare verbal capacity to process the auditory message.
However, audio typing does require verbal processing capacity, as
does shadowing a visual message, and insufficient capacity was
available to do both tasks together.
 Vidulich (1988) has argued that the traditional all manual human–
machine interface leaves a pool of human resources untapped.
New ways of enabling humans to interact with machines promise to
make these resources available and lessen the conflicting demands
on operators in situations where more than one task has to be
carried out at a time.

RELATED STUDIES
 In terms of the multichannel theory, the experienced drivers had a
subroutine for handling gear shifting, leaving attentional resources
free for other tasks. Novice drivers, on the other hand, performed
significantly less well on the signal detection task when driving the
car with manual transmission. According to the researchers, up to 2
years’ driving experience is needed to fully automate the skill of
gear-shifting.
 Lamble et al. (1999) found that the time taken for drivers to detect
that the car ahead was decelerating was reduced by 0.5 second
when they carried out a memory and addition task.
 The view of the ROSPA (the Royal Society for the Prevention of
Accidents) in the UK is that hand-held and hands-free cellphone use
while driving should be banned, because drivers need to allocate
100% of their attention to driving 100% of the time

SHOULD PEOPLE BE ALLOWED TO USE

CELLPHONES WHILE DRIVING?
 Response selection and execution may occur automatically, as
the output of a specific subroutine, or they may be initiated by
the executive processor. This corresponds to the automatic or
‘reflex’ response of the skilled performer or the rapid avoidance
behavior that overrides all other outputs to avoid impending
danger. When initiated by the executive processor, response
selection and execution are more akin to a measured or
calculated decision to act.

RESPONSE SELECTION AND

EXECUTION
 The requirements of tasks can be analyzed to determine whether
the performance of one of them will interfere with the performance
of the other. Tasks that are unlikely to interfere can be said to be
compatible and tasks with compatible components can be said to
be of low mental workload if the capacity of either the subroutine of
the processor is not overloaded
 A skilled operator has subroutines to handle the low-level aspects of
task performance that spare the executive processes. The
subroutines that handle motor actions (sometimes referred to as
motor programs) are of particular importance in integrating different
muscular actions into a unified whole.
 The unskilled operator is said to be tied to the present through lack of
automatic processes – the mechanics of the task require all his
attention, leaving none for other activities.

SKILLED VERSUS UNSKILLED

OPERATORS
 One of the interesting and important implications of this finding is that perhaps less
attention should be paid to minimizing the number of errors in the operation of a
system and more should be paid to teaching users what types of error can occur and
effective error-handling skills. In other words, design effort should not be spent only in
minimizing errors, since even experts are error-prone, but on error recovery and on
minimizing the consequences of an error. This seems to be one of the key requirements
in the operation of interactive systems.
 When evaluating existing systems, data are needed not only on the number of errors,
accidents or unexpected events but also on their effects on system operation and the
resources required to restore the system to its previous level of functioning.
 A first step in the design of any human–machine system, then, is to characterize the
expected skill level of the user population since the interface design requirements may
be completely different for different groups. Of particular importance are the
constraints on display complexity and layout and the requirements for on-line help,
training manuals and operating instructions, warning signs and labels and feedback.

CONCLUSION