/pdf/feeling-and-thinking-preferences-need-no-inferences-q4mw2hj8rx.pdf

Feeling and thinking : Preferences need no inferences

Publisher Summary This chapter presents a study of expert–novice differences between computer users. The results of expert–novice difference studies of computer users are consistent with corresponding research in other problem solving domains. The knowledge-based approach to assessing expert–novice differences offers a fruitful theoretical framework. The list of four types of knowledge offers a starting point, but future theoretical work is needed to specify the nature of what experts and novices know. The logical goal of this theoretical work is the development of precise models corresponding to expert and novice computer users. Research on expert–novice differences in computer users also suggests some tentative practical implications, all of which require additional study. First, to the degree that experts and novices differ with respect to the knowledge they possess, there is a role for training and education. To become an expert requires the acquisition of a body of knowledge, much of which can be taught. Second, this line of research offers some suggestions concerning how programming should be taught.

From Novice to expert

Stimulus structures and mental representations in expert comprehension of computer programs

People have severe problems wayfinding in large virtual worlds. However, current implementations of virtual worlds provide little support for effective wayfinding. We assert that knowledge about human wayfinding in the physical world can be applied to construct aids for wayfinding in virtual worlds. An experiment was conducted to determine whether people use physical world wayfinding strategies in large virtual worlds. The study measures subject performance on a complex searching task in a number of virtual worlds with differing environmental cues. The results show that subjects in the treatment without any additional cues were often disoriented and had extreme difficulty completing the task. In general, subjects' way finding strategies and behaviors were strongly influenced by the environmental cues in ways suggested by the underlying design principles.

Wayfinding strategies and behaviors in large virtual worlds

As important as navigation is to human performance in virtual worlds, it is an often overlooked problem in the design process. This article reports an experiment intended to show that real‐world wayfinding and environmental design principles are effective in designing virtual worlds that support skilled wayfinding behavior. The study measures participant performance on a complex searching task in a number of virtual worlds with differing environmental cues. The worlds are augmented with either a radial grid, a map, or both a grid and a map. The control condition provided no additional navigational cues. The results showed that navigational performance was superior under both map treatments as compared to the control and grid conditions. The grid was, however, shown to provide superior directional information as compared to the other conditions. The control condition provided the worst performance, with participants often becoming disoriented and experiencing extreme difficulty completing the tasks.

Navigating large virtual spaces

Subjects compared pairs of animal, country, and car names, which had been ranked along concrete (perceptual) and abstract (nonperceptual) dimensions by an independent group (e.g., animals were rated for size and ferocity). For judgments based on abstract and concrete dimensions, comparison time decreased as the difference between the rated scale values of the items increased. Furthermore, comparison time was better predicted by the interval than by the ordinal distance between test pairs. This finding suggests that judgments are based on some kind of long-term memory representation which has analog characteristics, in that it preserves more than rank-order information. The question of whether one common memory code or two separate codes are involved in comparisons based on abstract and concrete dimensions is discussed. Three alternative models are proposed and the role of visual imagery in memorial comparisons is considered.

/pdf/mental-comparisons-for-ordered-information-on-abstract-and-1wz7j75oyk.pdf

Mental comparisons for ordered information on abstract and concrete dimensions.

Subjective estimates of the distance between sites on a familiar/unfamiliar map were obtained under various conditions (prior knowledge, after brief study, and while in view). Nonmetric multidimensional scaling was used to develop a cognitive map for each condition. Differences were noted in the relative compression of the two map axes and in the degree of site clustering that occurred. The systematic distortion observed when observers depended on their prior spatial knowledge was of particular interest. The scaled map for this group did not display the rectangular properties of the actual map, but was “squared up” in the observer's long-term store. In addition, all of the mental maps revealed a tendency for individual locations to cluster into three groups. This tendency was greater in the memory conditions than in the perceptual conditions.

Memory and Perception of Cartographic Information for Familiar and Unfamiliar Environments

Intersite distances were obtained using both subjective pair-distance judgments and map sketching, either perceptually with the map in view or from memory following a study period. The perceptual distance estimates were more accurate than those made from memory. In addition, when responses were made from memory, intersite distances on the sketched map were more accurate than the same distances estimated in the pair judgment task.

/pdf/judgment-accuracy-in-pair-distance-estimation-and-map-1uqkemel83.pdf

Judgment accuracy in pair-distance estimation and map sketching

Human memory organization has been shown to be important in the processing of natural language. Evidence is provided in this investigation which suggests that human memory organization is also important in processing programming languages. Subjects were divided into experimental groups which studied programs with or without documentation, and with or without hierarchically indented statements. Subjects studied and recalled five Fortran programs. The pattern of recalled statements at logic segment boundaries was compared to the recalled pattern within segments. In addition, the recalled boundary patterns of the experimental groups were compared to each other. The results indicate that algorithmic logic segments form a cognitive organizational structure in human memory for computer programs. Statement indentation and internal program documentation did not apparently enhance the organizational process or the recall of programming statements.

Human Memory Organization for Computer Programs.

Subjects made repeated magnitude estimates of the lengths of lines or the areas of irregular polygons either from memory or while viewing the stimuli. As had been found previously for cartographic distance and area (Kerst & Howard, 1978), memory exponents approximated the square of perceptual exponents. Memory exponents differed for length and area, which suggests that more than ordinal information is preserved in memory. These results suggest important similarities between perceptual and memorial processes.

https://link.springer.com/content/pdf/10.3758/BF03333895.pdf

Stephen M. Kerst

Papers

Mental comparisons for ordered information on abstract and concrete dimensions.

Memory and Perception of Cartographic Information for Familiar and Unfamiliar Environments

Judgment accuracy in pair-distance estimation and map sketching

Human Memory Organization for Computer Programs.

Magnitude estimates of perceived and remembered length and area