Correspondence analysis used complementary to loglinear analysis

Loglinear analysis and correspondence analysis provide us with two different methods for the decomposition of contingency tables. In this paper we will show that there are cases in which these two techniques can be used complementary to each other. More specifically, we will show that often correspondence analysis can be viewed as providing a decomposition of the difference between two matrices, each following a specific loglinear model. Therefore, in these cases the correspondence analysis solution can be interpreted in terms of the difference between these loglinear models. A generalization of correspondence analysis, recently proposed by Escofier, will also be discussed. With this decomposition, which includes classical correspondence analysis as a special case, it is possible to use correspondence analysis complementary to loglinear analysis in more instances than those described for classical correspondence analysis. In this context correspondence analysis is used for the decomposition of the residuals of specific restricted loglinear models.     

Some evidence for correlated separate activation in a simple letter-detection task

This study investigates the separate activation versus coactivation issue for redundant targets in a simple letter-detection paradigm with latency as the dependent variable. The results of a one-response visual-search task are reported. Since, on single-target trials, only the target was presented and no accompanying noise element, no “distraction decrement” caused by irrelevant noise elements (Grice et al., 1984) was to be expected. The data obtained showed a clear redundantsignal effect. Subsequent detailed analysis of the latency data using Miller’s (1982) procedure indicated that the results were consistent with a separate activation model and failed to provide convincing evidence in favor of coactivation models. A further analysis of the data indicated that, in the present study, the separate channels were negatively correlated for a range of fast RTs and positively correlated for intermediate and larger RTs. No evidence in favor of Grice et al.’s (1984) distraction-decrement hypothesis was found. The conclusions of this study are that (1) a separate activation model summarizes the essential features of information processing in this simple visual search task, and (2) no convincing evidence in favor of coactivation in visual search tasks has been reported in the literature up to now.     

Two stages in postcategorical filtering and selection

Memory &; Cognition - In a modified Stroop paradigm, Kahneman and Henik (1981) varied the spatial location of incompatible color words in relation to the position of the ink color to be named. A...     

The role of position in object selection in vision

Summary Within contemporary visual-information-processing psychology, two classes of selective-attention theories can be distinguished: position-not-special theories and position-special theories. The position-not-special theories postulate that attentional selection by colour, by form, and by position are equivalent selective operations. The position-special theories assume that selection by position is more basic or direct than selection by colour or by form. Examples of both types of theory are briefly described, and irrelevant and relevant evidence is critically discussed. It is concluded that the relevant evidence is directly compatible with the position-special views and that the position-not-special theories require additional extraneous assumptions. The position-special model presented in Van der Heijden (1992) is elaborated in further detail. It is shown that this model is compatible with two important and often substantiated assumptions of the position-not-special theories: the assumption that pre-attentive analysers organize the visual scene in objects against a background, and the assumption that visual-selective attention can be directed at objects isolated in this way. This position-special theory is a parsimonious theory because it can identify the mentalistic conceptselective attention with the materialistic conceptspatial position.     

Time courses in visual-information processing: Some theoretical considerations

Summary In the literature two different views with regard to the time course of activation of recognition units —hypothetical entities that represent identity information —are proposed. Both views are derived from a restricted set of phenomena in visual perception. The first view gets its intuitions from phenomena such as those summarized by Bloch's law; these phenomena show integration or summation of activation over time. This grow-and-grow view assumes that recognition units accumulate activation gradually over times The second view finds its inspiration in phenomena such as Mach-bands and simultaneous contrast; these phenomena show inhibition or steady-state suppression over time. This grow-and-shrink view assumes that the activation levels increase fast initially and then start to decrease.The perception literature strongly suggests that it may be light intensity that determines whether mainly integration (and phenomena such as Bloch's law) or inhibition (and phenomena such as Mach-bands) is obtained. So it is also likely that both information-processing views apply only to a limited range of intensity conditions; the grow-and-grow view to conditions of low-flight intensities, and the grow-and-shrink view to conditions of intermediate-and high-light intensities.It is argued that, to arrive at a complete view, a visual phenomenon that applies to the complete range of light intensities should be taken as a background for generating hypotheses about time courses in visual-information processing. The Broca-Sulzer effect is such a phenomenon. Starting from this phenomenon a simple, synthesizing, information-processing model is derived that is compatible with both the grow-and-grow view and the grow-andshrink view. It is shown that, besides the relevant temporal properties, this model also produces the appropriate spatial properties.     

Time courses in visual-information processing: Some empirical evidence for inhibition

Summary In the literature two different views on the time course of activation of recognition units — hypothetical units that represent identity information — are proposed. The first view assumes that the recognition units accumulate activation gradually over time. The second assumes that the activation levels increase fast initially and then start to decrease. Hagenzieker and Van der Heijden (1990) have proposed a more general information-processing model, which is compatible with both views. In this model a treshold determines what type of time course will be obtained. For below-threshold-activation values a gradual accumulation of activation will be found; for above-threshold values activation levels will first increase and then, because of inhibition, decrease. This model served as a starting-point for the present paper. Elaboration of this model led to two testable hypotheses with regard to identification performance. The first prediction is that evidence for inhibition should be observed only under relatively bright luminance conditions. The second is that, given inhibition, the amount of the effect of this inhibition should increase over time. A partial-report bar-probe task was used to test the predictions. The experiment involved two luminance conditions and the information was sampled at two different moments in time. To control for possible localization artefacts a digit-naming task was also used. Evidence for inhibition was found in both luminance conditions. In accordance with the model the amount of the effect of inhibition increased over time.     

Location errors in partial-report bar-probe experiments: In search of the origin of cue-alignment problems

In studies using Averbach and Coriell’s (1961) partial-report bar-probe paradigm with linear arrays, most errors involve the naming of an item that was in the display but in a position other than the cued one. Up to now, there is no general agreement on the origin of these location errors. Point of departure in this paper is that part of the location errors arises from inappropriate application of the cue. It is tested whether this originates from problems to perceive the position of the cue (“cue-displacement hypothesis”) or from confusion about the order of the items in the array (“item-order hypothesis”). The results of two bar-probe experiments are reported. A novel, crucial, finding in both experiments is that, among the location errors, there was a preponderance of response letters that came from the central side of the cued item. In the second experiment, this was observed not only in the usual postcue conditions but also when the cue preceded the array. These results positively corroborate the cue-displacement hypothesis and do not support the item-order hypothesis: The cue tends to be perceived more toward the center of the visual field than it actually is exposed—that is, there is a central drift of the cue.     

Feature-specific interference in letter identification

It is a well documented finding that letter identification may be impaired by the presence of neighboring characters. There is much less agreement about the exact nature of such lateral interference. Evidence has been obtained for feature-specific interference, and even feature-specific facilitation. Santee and Egeth (1980) suggested that feature-specific interference will mainly be observed in data-limited situations where stimuli are presented for short durations and positional uncertainty is involved. They further suggested that interference would be revealed in accuracy rather than latency measures.In our first experiment, similar exposure conditions were used to those suggested by Santee and Egeth. In a second experiment, however, a longer exposure duration was used, and latency measurements were also collected. In a third experiment, positional uncertainty was removed by foveal presentation, and with this high quality stimulus presentation only latency measurements could be obtained.In all these experiments clear evidence for feature-specific interference was found, both in accuracy and latency data, and these results also held up under the detectability/criterion analysis proposed by Estes (1982). From these findings it can be concluded that feature-specific interference occurs under a wide range of conditions, including those representative of normal reading.     

WAIS-III factor index score patterns after traumatic brain injury

Profile subtypes, based on the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III) factor index scores, were examined in a sample of 166 patients with traumatic brain injury (TBI) by means of a two-stage clustering procedure. Three reliable subtypes were found that were differentiated primarily by level of performance across all factor index scores, although each of them demonstrated a relative weakness on the Processing Speed index. These subtypes were then validated on the basis of demographic variables, injury parameters, and additional psychometric measures that had not been included in the clustering procedures. The results indicated that performance on the WAIS-III after TBI was affected by both injury severity and level of education. It is concluded that there is no unique "signature" profile on the WAIS-III after TBI, except that a relative strength on the Processing Speed index is uncommon with this condition.