Accommodating Individual Differences through an Adaptive User Interface by David Benyon Presented by Alison Nichols 10/21/94 >OVERHEAD # 1 Computer systems vary in: structure functionality purpose size representation of internal workings Humans differ in: background sex education personality cognitive skills and preferences (such as interpreting computer displays, command names and icons) motivation goals mood The social context in HCI involves: organizational and political factors available user support pressures of a situation How can we improve the usability of computers systems to make them easier to learn and use? How do we make the interaction more effective and more satisfying? We can continue to improve the human-computer interaction design so that the average user can better use it. This approach assumes that there is a single design which will suit everyone. But, if we provide a variety of help for various user levels, then the system becomes too complex. Many functions of the system will not be needed by many users, and if they do find that they need the help at a future time, they may very well have forgotten how to access the data. >OVERHEAD #2 A potential solution to this problem is unique to computer systems. If the system is supplied with a suitable theory of interaction and how interaction can be improved, the computer system can: * monitor the interaction and * change its functioning, structure, or the representations provided by the interface to better match preferences of the users > The computer can adapt itself to individuals or groups of users Mechanisms of adaptivity are programmable. Theory needs attending to. Mechanisms of adaptivity need to infer likely events. May be based on user representation, applications or previous interactions. Theory of adaptivity should help us to decide when an adaptive system solution to usability problems is appropriate. There is no point in adapting to some user characteristic if that characteristic cannot be reliably inferred from the interaction. Individual differences People differ from one another Different personalities...what is personality? Different cognitive skills and preferences....what are they & how do they vary? For our purposes, what differences are useful and which one will have an impact on the interaction? Early studies of individual differences and attempts to isolate individual characteristics hare generally unacceptable now. Since human activity is inevitably social, human characteristics interact with the social, environmental conditions in complex ways. We must consider both the characteristics of the individual and the characteristics of the task and understand how these interact. Psychological tests can be used for a quick test of characteristics, performance skills, assessment of abilities. Important to indentify a stable and reliable correlation between the task and the tests. Computer systems must be used by a wide variety of people in a mostly cognitive activity. HCI is concerned with the acquisition, manipulation and expression of abstract symbols which signify something else. HCI involves information processing. The user of the system must evaluate information based on the cognitive processing of external symbols systems. Therefore, the individual differences in cognitive abilities, preferences, methods and techniques become important in HCI. >OVERHEAD # 3 Individual Differences in Cognition Cognitive abilities, or cognitive skills seek to describe methods by which humans process information. Cognitive abilities are relatively stable and change very slowly. Underlying distinctions in a general model of cognition: how data from the world is perceived its input to the brain how data is stored (memory) how data is processed (inductive/deductive reasoning, problem- solving abilities and strategies, mental imagery) how solutions and decisions are expressed (language ability) Cooper and Mumaw's review of research on spatial ability concludes that the spatial aptitude literature is quite clear in showing that a broadly defined spatial factor exists independently of verbal and quantitative factors. Evidence shows holistic and serialistic learning styles, though an individual may utilize both. Differences in cognitive abilities can be measured between individuals, but it is hard to judge exactly what the difference is. A promising approach seems to be to identify the components of complex tasks and the cognitive abilities which are required to perform them. >OVERHEAD # 4 Individual Differences in Personality Personality ..."those individual, relatively enduring patterns of reacting and interacting with others and with the world..." Personality concerns characteristics which remain stable over time and across situations. In this paper, he separates personality from cognition. Personality having to do with one's approach to the world rather than spatial or verbal ability. Personality has more to do with strategy than ability, and tends to dominate preferred approaches to specific tasks. Such as: sober vs. enthusiastic moralistic vs. expedient tough vs. sensitive practical Vs. imaginative conservative vs. radical Individual Differences and Human-Computer Interaction Van der Veer classifies some cognitive and personality factors along a "resistance to change" dimension. Personality factors are most difficult to change. It is not practical to expect people to change to adapt to a computer system. The computer can change the way it appears or operates in order to suit the needs and preferences of the user. D. E. Egan summarizes work on individual differences and HCI into three main areas of large differences: text editing information retrieval programming He concludes that spatial ability is an important determinant of performance in evaluating detailed spatial patterns (e.g. locating particular strings of characters) and that reasoning ability underlies the development of appropriate searching and problem-solving strategies and in producing accurate symbolic expressions (e.g. forming query statements or editing commands). Verbal ability did not seem to predict performance, nor did personality. Dealing with Individual Differences Egan's summary indicates that we should look at the extent of individual differences when designing computer systems and see how they may be accommodated. Redesigning he interface will be an effective way to deal with these differences in some cases. Sometimes just adding additional commands will increase the performance of those with poor spatial ability. However, sometimes these options won't be enough and an adaptive solution will be the only viable answer. This was the focus of Benyon's research. > OVERHEAD # 5 Egan and Gomez' three stage approach to dealing with individual differences: Assess the extent of the differences Identify and isolate individual differences which have an impact on HCI Accommodate these features > OVERHEAD # 6 Benyon's database retrieval system each subject tested on five personality/cognitive variables: spatial ability verbal ability field dependence short term memory thinking/feeling personality Significant differences observed between the high and low spatial ability and verbal ability groups on the menu interface, but the most significant was found betweent he performance of the high and low spatial ability groups on the command interface. Questions: How will users feel about having to take an IQ test prior to using a system? What about privacy issues? Are there other ways to ascertain cognitive abilities? If personality differences could be isolated, how would those factors translate into interface design?