Amberlight's Hierarchy HCI Briefings Library Amberlight Whitepapers

Attention

What is attention?

“It is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought…… it requires withdrawal from some things in order to deal effectively with others."

- William James, 1890

Attention is what enables us to process information about the outside world. We are only aware of things when we attend to them. Attention is often anecdotally referred to as the spotlight we shine on various objects in our environment to make them stand out. Once we have focused on an object in this way we are then aware of it and begin to process and interpret it. Attention can be internally driven by specific goals. This type of attention is voluntary and intentional, such as when we are learning, being trained or searching for something specific. We can use voluntary attention to filter out irrelevant information by focussing our visual scanning on a particular shape, colour or word.

Our motivations, expectations or intentions all play a part in what we attend to. Alternatively, a person’s attention can also be steered externally by environmental cues or objects within our sensory field. This is exemplified when we turn our head after we hear a loud noise or when someone concentrates on a warning light flashing in the corner of a display. This type of attention is spontaneous, stimulus driven and mediated by salience.

Attention has a huge impact on performance and this applies to interaction with computers and other interactive technologies. Interfaces must focus the user’s attention on what they need to be looking at or listening to for any given stage of a task, guiding their attention to the relevant information.

Types of attention

There are 3 distinct types of attention:

1. Selective attention. This is often referred to as cognitive tunnelling, and occurs when a person attends to a single stimulus or task so fervently that they are completely unaware of other stimulus or objects in their environment. A person focussed on a computer task, or on the road ahead whilst driving, may fail to perceive a question asked of them by a second person.
2. Focused attention. Focused attention could be considered the beneficial form of selective attention in the sense that people purposefully ‘tune out’ interfering stimulus or cues in order to complete a task. Blocking out certain screen elements on a ‘busy’ display to locate an important cue, and a receptionist in an office blocking out the chatter of co-workers in order to carry on a conversation on the phone are both examples of focused attention. Pre-attentive processes help us decide what to pay attention to and what to filter out and ignore.
3. Divided Attention. Sometimes it is not possible to focus attention on one task when other stimuli start interfering. For example, if we can hear someone talking about us while we are having a separate conversation, it is difficult to focus exclusively on our own conversation. Sometimes the nature of a computer task forces users to focus on several different things.

Relevance to Human Computer Interaction (HCI)

Understanding how users attend to stimuli, and how things block their attention, is integral to interface design. The information users extract from their interaction with an interface depends on:

• Their motivation
• Their goal and task
• What they already know
• The content of the message
• The salience of the message

Interface designers must understand the impact of these factors on task performance - how they might combine or interfere with each other, and the underlying principles behind them - in order to design effective interactive systems.

Salience

An interface should facilitate the focussing of users’ attention to salient information. The location, size, colour and distinctness of a word, image or screen element can make it particularly salient to the user. Techniques can be used to effectively draw users’ attention to links and features that are critical to their task, or to the service provider’s business goals.

It is important to know which elements should be made salient, at which point and in what way, in order to avoid situations where elements interfere with each other, the screen becomes over-cluttered and important information is dismissed. A common mistake in web design is to place critical task or business information in a large, colourful box, only for this to be ignored by users. Users have learnt to ignore banner adverts on websites, and selectively attend to elements and features that match their expectations of meaningful content in terms of visual design. This is known as “banner blindness”.

Overuse of salient elements on a particular page can reduce the effectiveness of each at grabbing the user’s attention, because they end up competing against each other ('too many voices'). Users can also habituate to a salient stimulus if it is presented too often without any variation. This is important for the design of warnings in safety or task critical systems.

User expectations

Violating existing design conventions for particular features can cause them to be overlooked by users. In terms of navigation it is important to make links (and particularly related and relevant links) salient. It is also important to make buttons associated with task-critical actions stand out clearly. It is vital to understand users’ expectations of such features, their spontaneous stimulus-driven reactions to various sensory stimuli and their acquired tendency to filter cues that are perceived to be irrelevant, in order to ensure the design of an interface facilitates the tasks and user behaviours that are important or desired at that point (i.e. users may overlook the navigation options if these options are unconventional and lack salience).

Facilitating focused attention

The completion of some tasks requires a user’s focused attention. At these points in a task flow, competing and attention grabbing elements need to be removed from the interface. Flashing adverts can be very distracting if a user is attempting to write an important email.

Similarly, irrelevant options may inhibit user performance and make a task more difficult to learn or less efficient to perform. In software design the use of secondary windows (e.g. the Print dialogue window) helps to focus user attention onto a limited number of currently relevant options (and filter out the remaining functionality). The user is probably unaware of how their attention is manipulated and shaped by such design elements, and this can also be a goal of user-centred design.

Focused attention can be engineered to help draw attention to security issues, errors or other important system messages. For example, when trying to save a document in Word under a name that would overwrite a pre-existing document, the warning on the interface is designed in such a way as to prevent users from attending to anything else on the screen aside from the warning message. External cues and screen elements (such as the boxing around the warning) focus users’ attention on a single screen element in order to stop them from making a potentially major error.

Query by attention

This is a term used to describe how users can find specific content within interfaces dense with information. The user uses their attention to distinguish and isolate features of interest. An interface can be designed to facilitate this by associating different and distinct properties with different nodes of information (colour, shape, position etc.). The user manipulates their attention and not the interface; this can be very effective because the user can quickly adjust their query parameters and obtain immediate feedback.

Designing for divided attention

There are also instances where users need to attend to two or more things at the same time. A person’s ability to divide their attention explains why people can multi-task – i.e. perform two or more tasks at the same time. Watching the news, while talking on the phone with a friend is an example of divided attention. However attention is a limited resource, which can only be stretched so far. The amount of different stimuli a person must attend to in order to complete a task is often referred to as their mental workload. The greater the mental workload the greater the chance a user will fail at one or all of the tasks they are trying to perform.

When designing an interface it is essential to keep mental workload at a constant and comfortable level and this usually involves cutting tasks into chunks or steps. If a task requires the monitoring of two or more stimuli or the simultaneous processing of different strands of information then the task steps need to be smaller (i.e. by removing everything else from the user’s attentional field). Divided attention can become a critical factor in vigilance tasks (those where the user must monitor an interface for change over a long period of time) such as air traffic control, security scanning, or nuclear power plant process control. It is also a factor in iDTV where various streams of information across several modalities are superimposed onto a single screen.

Divided attention and interference

Interface designers should be aware that two stimuli can interfere with each other, disrupting our ability to control attention. This interference may be triggered by an over-demand on the user’s attentional bandwidth (task difficulty), but it can also occur for a number of other reasons. If two stimuli are similar in modality (i.e. two audio streams) they are more likely to interfere with each other, as are two stimuli similar in content or meaning (e.g. the Stroop effect).

Human short term memory is mediated by a cognitive structure known as the articulatory loop which involves repeating an auditory stimulus about every 2 seconds (even if this is done mentally). Any task involving speech disrupts the user’s ability to memorise textual or verbal information appearing on the screen. When two stimuli, actions or screen elements compete for a similar cognitive resource they interfere (or combine streams in some cases) often disrupting one or both tasks. It is important to design interfaces in a way that allows users to focus their attention on important information without interference from other sources, particularly if divided attention is required. An interface in a noisy environment should not rely on auditory feedback alone (and this is not simply an issue of audibility), or it should select tones, pitches and durations which are distinct.

Vigilance and sustained attention

Many users rely on computer technology to help them perform effectively at work. This reliance on computers is often greatest in time or safety critical systems, particularly where the user must monitor feedback from a computer display over long periods of time. Examples of this include share and stock trading, air traffic control or safety systems on oilrigs, power stations or factories. As the operator becomes fatigued over time it becomes increasingly more difficult to focus attention and vital cues may be missed and errors become more likely. It is important for the designers of such systems to understand the precise nature of the task and the limits of human attention.

Designers should ensure that critical new information can be attended to quickly and effectively and that the system minimises the attentional demand on the user to an absolute minimum.

Central and peripheral attention

An important concept to consider when discussing selective, focused and divided attention, and vigilance tasks is that of where the users visual attention is directed at any given moment. Essentially this is related to the area of the interface which falls directly onto the fovea (the central area of the retina where the concentration of photoreceptors is greatest) or the point of focus within the user’s entire visual field. This is often referred to as the UFOV (useful field of vision). When the user employs selective or focused visual attention they are using their UFOV to scan an interface or focus on an area of the interface. The UFOV is typically between 1 and 4 degree of the visual angle and this is important to consider in monitoring tasks where information outside of this area may be overlooked.

Designers must understand the principles of peripheral vision and the visual cues that can be employed in this region to “grab” the user’s attention away from the UFOV. The peripheral field is most sensitive to cues, which involve movement, flashing and sharp changes in contrast. Designers must understand the attentional focus involved in certain tasks in order to employ the correct techniques required to shift the users attention when required.

Summary

An intelligently designed interface can improve task performance by subtly manipulating the user’s focus of attention and by ensuring that the demands of the task at each stage are within the limits of human attention.

Further Reading

Basic introduction to human attention and its limitations:
Chapter 5 in: Eysenck, M. W. & Keane, M. T. (2000) Cognitive Psychology (4th Ed.) Hove: Psychology Press

Attention and visual displays:
Chapter 3 in: Wickens, D. W. & Hollands, J. G. (2000) Human Engineering and Human Performance (3rd Ed.) London: Prentice Hall

Core texts on attention:
Paschler, H.E. (1997) The Psychology of Attention. MIT Press
Styles, E. A. (1997) The Psychology of Attention. Bucks: Taylor and Francis Psychology Press