In the 1950's and 1960's the dominant theory of attention was the bottleneck theory, stating humans act as a single communication channel of limited capacity at some part in an information processing sequence. Dichotic listening tasks were devised to illustrate this theory; early studies showed that people are very good at processing only one of two physically distinct concurrent sources of information. The resource theory however assumes that attention can be regarded as a single reservoir of information processing resources. (Reason, 1990)
The idea of attention is critical when examining actions and intentions. Normans and Shallices attention to action theory argues there are two control structures; horizontal and vertical threads. Horizontal threads comprise of processing structures called schemas and vertical threads interact with the horizontal threads to provide the means by which attentional and habitual factors activate schemas. Horizontal threads govern habitual activity without the need for attentional control. (Reason, 1990)
Later on, the role in which schemas play in action slips, an error which occurs when a person does an action that is not intended (Norman, DA, 1981) will be examined as will the role of slips and mistakes with expert users of computer programs.
Slips can be categorised into; errors in the formation of an intention, faulty activation of schemas and faulty triggering (Norman, DA, 1981). To interpret the categorisation of action slips a theory known as Activation-Trigger-Schema system (ATS) must be applied. This theory assumes that action sequences are controlled by schemas (a sensori-motor-knowledge structure). According to the ATS system information passes from high order schemas to low order schemas. High order schemas or parent schemas are initiated by lower order schemas or child schemas. The ATS theory assumes that skilled action, such as that by computer users need only be specified at the highest level. Each schema has a specific condition needed for it to be triggered (Norman, DA, 1981).
Slips resulting from error in the formation of the intention are due to errors in classifying the situation and from an ambiguous or incompletely specified intention. A type of this slip is mode errors in which the resulting action may be appropriate for the analysis of the situation but not for the actual situation. Another type of this slip is description errors, where by the relevant information is not available to form an appropriate intention this leads to performance slips such as placing the sugar lid on the coffee pot. (Norman, DA, 1981)
Slips resulting from faulty ...
This is a preview of the whole essay
Slips resulting from error in the formation of the intention are due to errors in classifying the situation and from an ambiguous or incompletely specified intention. A type of this slip is mode errors in which the resulting action may be appropriate for the analysis of the situation but not for the actual situation. Another type of this slip is description errors, where by the relevant information is not available to form an appropriate intention this leads to performance slips such as placing the sugar lid on the coffee pot. (Norman, DA, 1981)
Slips resulting from faulty activation of schemas are faulty in two ways; a schema being unintentionally activated leading to action intrusion and a schema loosing activation before its time to control behaviour, which results in omission of a component of an action sequence. When a schema is unintentionally activated this is called a capture slip, a familiar habit substituting itself for the intended action sequence. If a habit is particularly strong, even partial matches are enough to trigger parent schemas. Loss of activation of a schema because of interference or decay of primary memory causes the desired intention to become lost, the behaviour however continues until it reaches a logical stop point. (Norman, DA, 1981)
Slips resulting from faulty triggering occur when a schema is properly selected but triggered improperly, at wrong time or not at all. (Norman, DA, 1981)
Much of the data gathered on slips was gathered naturalistically. This is the only way to gather data of people under extreme stress. However the disadvantages in this methodology include; the difficulties in recording exactly what has occurred, observers may not always be around or not ready to make the detailed observations required, retrospective records are unreliable, observers are selective in what they record and interpret as an error. Evidence of action slips has high external validity but lack internal validity and control, lab experiments could be used to validate findings on action slips (Norman, DA, 1981).
Strength of Normans ATS theory is that it allows for multiple sources of attention (Norman, DA, 1981).
Heckhausen and Beckman (1990) argue actions are guided by mentally represented intentions. Intentions are made up of goal intentions and instrumental (contingent) intentions. Should there be problems in the initiation, implementation or termination of an intended act instrumental intentions will be generated to overcome these. There are two modes for conscious processing of goal intentions; wide goal span which relieves action control from having to focus attention on current activity and permits conscious processing of overlapping cognitive activities, which may be related or unrelated and narrow goal span whereby all attentional capacity is absorbed by the ongoing activity.
Action slips are most likely to occur whenever a course of action has become automated to such an extent that it no longer requires conscious control. There are three major categories of action slips; initiation slips, implementation and termination slips.(Heckhausen and Beckman, 1990)
Initiation slips include missing an opportunity; an interpolated action cannot be initiated because it is embedded in an automated segment of the mind. Overspecified opportunities occur because of a failure to recognise an opportunity because it was too narrowly or broadly delineated during the formation of initiation intent. Underspecified opportunities are the pouncing on an inappropriate opportunity because of its underspecification. Falsely specified opportunities occur when the initiating act is inappropriate, in order to subvert conscious control there must be a functional or structural similarity between the erroneously specified act and the current act, this could occur with expert computer users who are often switching between similar applications. (Heckhausen and Beckman, 1990)
Termination slips can be caused by the premature termination of an act, an overlapping of an automatic action by a related parallel action demanding conscious control resulting in the parallel action terminating the main action. The premature deactivation of the initiation intent, retarded deactivation of intentions, and continuation of a consummatory act and repetition of the consummatory act can all result in termination slips. (Heckhausen and Beckman, 1990)
The last category of slips is implementation slips. Triggering slips can lead to skipping because after a triggering slip an act may be omitted or repeated. Triggering slips leading to reputation occur because the completion of an act has not been triggered e.g. when a computer expert has finished typing text, if this completion of the act has not been triggered the action may be repeated and the text typed again. If two fairly similar objects are handled in an automated fashion at nearly the same time they can be mixed up resulting in the performance of an inappropriate act, this is known as substitution act co-ordination. Other types of implementation slips are falsely specified implement acts and actions going out of sequence because of a relapse during learning. (Heckhausen and Beckman, 1990)
Freuds 'rebellion theory' argues that slips are caused by inappropriate, suppressed or displaced intentions striving to gain access to behaviour to express themselves. However this doesn't explain why an intention is displaced or why an intention gains access to behaviour at any given time. Freud's theory shares three components with the Heckhausen and Beckman model; the assumption of volitional forces in attention, the overlap of related process and the mutual sharing of processes of a portion of the capacity for conscious control. (Heckhausen and Beckman, 1990)
One strength that the Heckausen and Beckman model has over the Norman (1981) model is that the view of automated activity reducing the load on conscious control, which plays no part in the Norman model. However the research by Heckhausen and Beckman (1981) never defines what they interpret as an action slip. There is also no definition of an implementation slip nor do they describe what a falsely specified act slip is.
Kitajama and Polson (1983) developed a human-computer interaction model to account for the rate of slips made by expert computer users. Skilled computer users have a schematic representation of a task made up of a hierarchal structure of task and device goals. A task goal is associated with one or more device goals which specify device states to be satisfied in order to achieve an associated task goal. Model generates representation of display, which is then elaborated by retrieved information, models evaluation of display is in a context that is defined by task and device goals. Candidate object selection is performed on the basis of evaluation.
Card, Moran and Newell (1983); as cited in Kitajama and Polson (1983) studied skilled computer users and found errors were made on 37% of command sequences describing editing. Over half of these errors were detected and corrected. 21% of the commands resulted in an incorrect edit and required additional edits to correct, showing that a slip has only been detected after an action sequence has been performed therefore no conscious monitoring of behaviour was occurring. However the majority of errors made by expert computer users were detected and corrected before a slip generated an incorrect result. Card at al (1983), experts accept high error rates because error recover can be done easily and rapidly. Experts trade speed for accuracy, resulting in slips.
Strength of Kitjama and Polson's (1983) model is that it generates correct actions without assuming a special set of mechanisms for erroneous action.
The theory by Norman (1981) and Heckhausen and Beckman (1990) both categorise slips into types, some of which overlap. It is difficult to interpret slips made by expert computer users in light of these theories, Kitjama and Polson used a different model to explain slips. To further understand and interpret errors made by expert computer users further research applying these theories to these errors would develop knowledge in this field.