Broadbent’s (1954) filter model suggests that humans selectively attend to only some of the cues and tune out much of the rest. According to Broadbent, the bottleneck occurs very early in processing and is based on the major physical properties of the incoming stimuli. This model is supported by evidence from dichotic listening task and split-span procedure. However, it fails to take into account the cocktail-party situation, as well as several shadowing experiments in which participants’ recall included meaningful information from the non-shadowed message.
In 1960, Gray and Wedderburn also proved that attention could be switched easily from one channel to another. In their shadowing technique, they simultaneously presented the sequence ‘ OB - 2 - TIVE’ in the left of ear of participants, and presented the sequence ‘6 JEC 9’ in the right ear. Results indicated that participants acted intelligently and paid attention to meaning, and repeated ‘OBJECTIVE’, contradicting Broabent’s theory of the physical nature of stimuli on attention.
Treisman’s (1964) attenuator model solved Broadbent’s predicament. Although it is also an early selection model based on physical characteristics of the information, Treisman’s filter model suggests that unattended information is not eliminated but is attenuated, or weakened. Therefore, if necessary, information is still available for higher level processing, where it reaches the threshold level of intensity. Hence, explaining the cocktail-party situation. Critics of Treisman’s attenuator model believe that it is too simple to account for attention, and does not explain how the process of attenuation occurs.
In 1963, Deutch and Deutch proposed the Pertinence model, where they claim that all information, both attended or unattended, is analyzed for meaning in order to select an input for full awareness. The same evidence supporting Treisman’s model supports this theory. However, the Deutsch & Deutsch model explains the process of focused attention more simply. More support is derived from Von Wright et al (1975), who paired electric shocks with a word to condition a galvanic skin response when the word was spoken. A response was produced even when the word was presented to the unattended ear and the participants were unaware of it. However, critics of this model believe that there is uneconomical analysis if every message is to be analyzed before processed. Also, it is found to be rigid and inflexible.
Many researchers question whether any filter theory, which assumes a single, general purpose, limited capacity processor can account for the complexities of selective attention. Early shadowing investigations overlooked the effects of factors such as practice, (i.e. when individuals develop strategies, there is less demand for central resources) similarity of tasks, (i.e. modality of input or output) and also, difficulty. Most evidence arises from divided attention experiments, which utilizes the dual-task technique, as it actually measures processing capacity.
Schaffer (1975) investigated on how a skilled audio typist would perform if in one ear, he/she were to be presented an audio typing, while a shadowing or reading aloud from text was presented in the other. Due to similarity in task, causing interference, poor performance at both tasks was observed. Allport et al (1972) investigated if piano playing were to be affected if pianist were required to both read music and listen to a prose. Results indicated that these different tasks had no impact on piano playing. This means that when using different modules, (in this case, sight and hearing) attention can be easily divided. Also, the more practice involved, the less attention one has to allocate to a task. It may become automatic if the task carried out need no thinking, as it is fast and unavoidable.
Schiffren and Schneider (1977) indicated the contrast between automatic and controlled processing. This distinction is supported by evidence that simple visual features such as shape or color can be detected automatically, but combined features such as shape and color need focused attention. With practice, these controlled processes can become automatic and difficult to change. Automatic processes can interfere with tasks requiring controlled processing. An example of this is the Stroop effect; participants were required to report the color of ink that the word was written on, regardless of the actual word itself (i.e. the word ‘blue’ written in red ink must be reported as red). Participants took longer to report ink color than to report the actual word. This indicates that the meaning is, in general, more important than the color thus explains automatic process of information.
Kahneman stated that attention can be focused on more than one subject at a time; however, some tasks may have more awareness than others, depending on how much mental effort is provided for each task. The concept of divided attention led Kahneman (1973) to suggest that a limited amount of attention is allocated to tasks by a central processor involved in resource allocation. Kahneman indicated that when an individual is alert and aroused, the attention resources of that individual is higher than that of when that individual is tired. However, the capacity theory fails to explain exactly how the allocation decisions are made.
Allport (1980) proposed that a number of limited-capacity processing modules exist. This concept is able to explain how an individual can easily divide their attention between dissimilar tasks (i.e. using different modules) but not between similar tasks (i.e. competing for resources from the same module). However, it does not explain how modules are coordinated together.
Great uncertainties arise from filter theories of focused attention, as it cannot account for the complexities of selective attention; verification, therefore, mainly surfaces from divided attention experiments, as it evaluates processing capacity. However, to this point, it is still doubtful that any of the theories or models is accurate, as they all seem too contain weaknesses and ambiguity.
