This article reviews selected literature related to the credibility of information, including (1) the general markers of credibility, and how different source, message and receiver characteristics affect people's perceptions of information; (2) the impact of information medium on the assessment of credibility; and (3) the assessment of credibility in the context of information presented on the Internet. The objective of the literature review is to synthesize the current state of knowledge in this area, develop new ways to think about how people interact with information presented via the Internet, and suggest next steps for research and practical applications. The review examines empirical evidence, key reviews, and descriptive material related to credibility in general, and in terms of on-line media. A general discussion of credibility and persuasion and a description of recent work on the credibility and persuasiveness of computer-based applications is presented. Finally, the article synthesizes what we have learned from various fields, and proposes a model as a framework for much-needed future research in this area.

When a letter and a tone are presented in rapid succession, there is substantial interference if both require an overt response (the dual-task paradigm), but there is very little interference if the tone is a signal to inhibit the response to the letter (the slopping paradigm). These effects were replicated under common stimulus conditions in two experiments, and they were compared with effects in a change paradigm, in which the tone required an overt response, like the dual-task paradigm, but was also a signal to inhibit the response to the letter, like the stopping paradigm. In the change paradigm, the overt response to the tone suffered interference only if subjects failed to inhibit the concurrent response to the letter; inhibiting the response to the letter virtually eliminated the interference. The results suggest that differences in response competition are primarily responsible for differences in interference between the dual-task and stopping paradigms. More generally, the results support response-competition interpretations of dual-task interference.

Speeded choice responses (reading or naming) to a relevant stimulus under conditions of spatial uncertainty are delayed by the simultaneous occurrence of other events. This "filtering cost" occurs despite high discriminability of target and distractors, which allows parallel detection of the target in search through the same displays. Reading is also delayed when the removal of irrelevant objects from the field coincides with the onset of the target. Filtering costs are caused by the processing of events rather than by the mere presence of irrelevant items. They are eliminated by advance information about the location of the target or by advance presentation of maintained distractors.

The latency of reading a single word is increased by 20 to 40 msec if another object is present in the display. The delay is affected by the spatial organization of the display: a colored frame causes less delay when it surrounds the word than when it is shown on the opposite side of fixation. A small gap in the frame is also more efficiently detected as a secondary task when the frame is around the word than when the two are spatially separate. The advantage of integrated over separate presentation suggests that a "filtering cost" is incurred when two distinct perceptual objects compete for attention. Attention in filtering tasks operates on perceptually distinct objects rather than on nodes in a semantic network.The act of reading is usually assumed to be automatic, in the sense that it occurs both without voluntary control and without requiring attentional resources, if a word or letter is sufficiently clear and close to the fovea. Yet a consistent delay in naming a single letter is produced by the addition of irrelevant objects to the display, even when these objects are highly discriminable from letters (Eriksen & Hoffman, 1972; Eriksen & Schultz, 1978, p. 18). Distractors such as black disks or color patches add about 30 msec to the latency of letter naming. Eriksen and Schultz labeled this effect cognitive masking. We have found a similar delay in reading a single word when an irrelevant but highly discriminable object is added to the display. We further found that the delay in reading increases as more objects are added, although probably at a decreasing rate; it can be eliminated by precuing the location of the word on each trial, and it is reduced or eliminated, with the same displays, when the subject is asked to press a key whenever a word is shown, instead of to read it (Kahneman, Treisman, & Burkell, in press).The fact that it disappears with precuing links the delay to attention rather than to peripheral interference. The fact that it is reduced in search or detection suggests that the delay arises not in finding the word but in allocating attention to it and/ or in filtering out the irrelevant objects. Focused attention to the word is not required when the response is determined directly by the detection of its presence. We suggest attention must be narrowed down to the relevant stimulus, however, when the choice of a response demands further processing. We therefore interpret the observed delay of read- ing as a filtering cost. The cost of filtering should be distinguished from interference caused by confusion between the relevant and irrelevant items and from interference that originates in response conflict: Filtering costs occur even with highly discriminable stimuli, or with irrelevant stimuli that cannot be read and are not easy to name. (Of course, both these factors could be sources of further interference if they were introduced.)It has been suggested that selective attention to objects is more effective than attention to properties (Kahneman, 1973;Treisman, 1969), and that the allocatio...

Aims and objectives : Access to health information is important for the wellbeing of people living in the community after spinal cord injury (SCI). In order to design appropriate information interventions, it is critical first to understand the information sources typically used. The goal of this study therefore is to identify the information-seeking practices of this group. Sample and methods : A sample of 207 individuals living in the community following traumatic spinal cord injury were surveyed regarding their ongoing information needs and practices for seeking information. Results : The results reveal that respondents have unmet information needs, despite the fact that they typically access information through a variety of channels. SCI specialists are the most commonly used source, although they are viewed as relatively inaccessible. By contrast, the Internet (used by a relatively high proportion of respondents) is viewed as comparatively accessible, although there are some concerns about the quality of information found there. Conclusions : These survey results point to the need for an information source that is accessible and delivers high quality information. Although respondents use a variety of information sources, none meets this ideal profile. Information professionals should consider this gap in the design of information interventions.