This paper describes and evaluates a computational model of anomalous data integration. This model makes use of three factors: entrenchment of the current theory (the amount of data explained), the relative probability of the contradictory explanations (based on conditional probabilities as part of the domain-knowledge), and the availability of alternative explanations based on learning. In an experimental study we found that the enu-enchment of a theory and the availability and likelihood of an alternative explanation influenced solution speed and the correctness of inferred causal explanations. However, in detail, the single levels of both factors were not cleariy distinguishable and did not follow the predictions. These findings suggest that entrenchment itself is not a major factor in determining the difficulty of a task. Instead, we hypothesize that task difficulty is dominated by a person's ability to construct an alternative explanation of a given situation, a factor that is only indirectly related to entrenchment.

People look at emptied spatial locations where information has been presented during encoding. There is evidence that this socalled ‘looking at nothing’ behaviour plays a functional role in memory retrieval of visuospatial and verbal information. However, it is unclear whether this effect is caused by the oculomotor movement of the eyes per se or if covertly shifting attention is sufficient to cause the observed differences in retrieval performance. In an experimental study (N = 26), participants were manipulated in being able to shift either their eyes or their focus of attention to a blank spatial location whilst retrieving verbal information that was associated with the location during a preceding encoding phase. Results indicate that it is not the oculomotor movement of the eyes that causes the facilitation while retrieving verbal materials, but rather covert shifts of attention are sufficient to promote differences in retrieval performance.

When finding a best explanation for observed symptoms a multitude of information has to be integrated and matched against explanations stored in memory. Although assumptions about ongoing memory processes can be derived from the process models, little process data exists that would allow to sufficiently test these assumptions. In order to explore memory processes in diagnostic reasoning, 29 participants were asked to solve a visual reasoning task (the Black Box paradigm) where critical information had to be retrieved from memory. This study focused on differentiating between processes that take place during the encoding and the evaluation of symptom information by comparing eye movement measures (the number of fixation and fixation duration per dwell). Results will be discussed in light of existing theories on sequential diagnostic reasoning. Further, it will be discussed to which extent eye movements can be informative about memory processes underlying sequential diagnostic reasoning.

Based on the established framework of Cognitive Load Theory, the presented research focuses on the inspection of cognitive load factors in an interrupted learning task. The task itself is inspired from basic cognitive research and demands participants to learn abstract symbol combinations of varying complexity. In addition, they have to deal with interruptions while performing the task. Experimental results indicate the influence of task complexity on how interruptions effect learning performance. However, questions on underlying learner cognition persist, rising the need for a more in-depth way of examination. For this purpose, a cognitive model within the cognitive architecture ACT-R is developed to clarify cognitive processes and mechanisms within different conditions of the task. Preliminary results from a first model for the easy task condition already indicate some fit between human and model data. Modeling work continues with adjusting the current model and implementing a model for the difficult task condition.

Using a process model of skill acquisition allo- wed us to examine the microstructure of subjects' performance of a scheduling task. The model, im- plemented in the Soar-architecture, fits many qua- litative (e.g., learning rate) and quantitative (e.g., solution time) effects found in previously collec- ted data. T h e model's predictions were tested with data from a new study where the identical task was given to the model and to 14 subjects. Again a general fitof the model was found with the restrictions that the task is easier for the m o - del than for subjects and its performance impro- ves more quickly. T h e episodic memory chunks it learns while scheduling tasks show h o w acquisition of general rules can be performed without resort to explicit declarative rule generation. T h e model also provides an explanation of the noise typically found when fittinga set of data to a power law — it is the result of chunking over actual knowledge rather than "average" knowledge. Only when the data are averaged (over subjects here) does the smooth power law appear.