Psychologists have invested a significant amount of effort in researching handedness and brain lateralization. Unfortunately, this effort has failed to result in many robust conclusions. Bold statements on the topic of handedness are quickly disputed or refuted. Yet, the work persists, presumably because of an abiding hope that eventually some hypotheses will prove vigorous under the pressure of repeated experimentation. Though it seems that few such hypotheses exist at this point in the field of handedness research, there are nonetheless hypotheses that are interesting and backed by at least some empirical data.
The work on handedness revolves around several central questions. One of the most important of these questions is on the origin of handedness, and of sidedness as a whole. Though the different hypotheses presented by researchers in this area certainly have implications for all the other areas of handedness research, I do not intend to focus on these issues. Rather, I will examine another very important part of handedness research-- the effect of handedness on individual differences. A vast amount of research, some of which has been summarized in the popular press, has grappled with the question of what it means for a human to favor one side of the body, particularly one hand. What makes a "righty" different from a "lefty"? Psychologists have attempted to find connections between handedness and differences in cognitive ability, cognitive style, personality, health, etc. As noted above, so far the results are far from creating a coherent picture of how individual differences relate to handedness, but with further refinements in methodology, the studies may yet bear enduring fruit.
As with the study of handedness, the study of human-computer interaction sometimes finds itself trying to answer questions of individual differences. Work on individual differences in HCI has garnered some significant results. Certain differences in cognitive abilities have been shown to have an impact on user performance at the interface, and there is hope that research into cognitive style and personality may eventually result in robust predictive theories on performance (Dillon and Watson, 1996). Yet the work on individual differences in HCI is not mature. I believe that it is in these questions of individual differences that research on handedness and research on HCI may have a mutually beneficial cooperation.
I contend that there are at least three ways in which these fields of study may aid one another. The first is in interface design and training procedures. Some tentative, and contested, conclusions have been reached about the cognitive differences associated with different handedness patterns, and further careful research may lead to more stable conclusions. Knowledge of cognitive differences between left-handed and right-handed individuals may enable designers to customize an interface to make it more amenable to a user of a certain hand preference. A possibly more important use of the knowledge of cognitive differences is that it may help those responsible for training people on computer applications to understand better how to deal with the particular problems encountered by a left-handed or right-handed student. Furthermore, there has been some research into certain occupational and other groups in which left-handed people are over- represented compared to their proportion in the general population. An awareness of this research may lead those designing, or purchasing, hardware and software for these groups to consider more ambidextrous computer designs . Also, the strengths and aptitudes of left-handed individuals, if such do exist, may come to constitute an argument for including them in computer design teams.
As I have already noted, the field of handedness and individual differences is in need of more robust theories. HCI could provide another avenue in which answers may be sought to some of the persistent questions on handedness. That is, computer applications could function as research tools aiding in the study of individual differences and handedness. Along with looking at the SAT scores of left- and right-handers, researchers could test subjects on various computer tasks.
Moreover, there is a possibility that better conclusions on the connection between handedness and cognition could further knowledge of user cognition at the interface. If it is established that individuals of a certain handedness excel in specific cognitive tasks and also excel at interacting with a particular computer application, researchers may be better able to theorize on the form of cognition at that interface.
I intend to survey the current research on handedness and individual differences and see if I can detect any general trends in the chaos of empirical studies in this field. Also, I will propose some possible ways handedness research might be incorporated into HCI, and suggest some studies that may reap benefits for specialists both in handedness research and in HCI.
Nearly every handedness study may find its foil somewhere else in the body of literature. Such issues as handedness and body size, handedness and sexual orientation, handedness and alcoholism, handedness and allergies, and handedness and longevity (Coren (1992) argues that left-handed individuals have an average life-span nine years shorter than right- handed individuals) are contested in the literature. When the focus narrows to handedness and cognition, the state of affairs does not get much better. The differences between left- and right- handed individuals in performance on verbal and spatial tasks, in personality, in level of musical talent, in geographical knowledge, etc., all remain disputed. Yet I will examine the recent research with an eye to making some tentative suggestions about general trends.
A number of studies have looked at the question of whether left-handed individuals perform better or worse than right-handed individuals on either spatial or verbal tests. Miller (1971) cites and summarizes both Levy (1969) and Annett (1964 and 1967). The last two researchers proposed different theories for lateralization of functions in the human brain, theories that made contrary predictions on the relative strengths of left- and right-handed individuals in spatial and verbal tests. According to Miller's synopsis, Levy's model proposed that while right- handers have verbal functions mediated only by the left hemisphere, left-handers spread verbal processes across both hemispheres. The result, according to Levy, is that the left-handers' right- hemisphere verbal functions in a sense take up space that would normally be dedicated to spatial functions, which thus suffer in left-handers. Levy administered a WAIS test to differently handed postgraduate students and discovered that the left-handers did indeed score below the right-handers in spatial intelligence. Lewis and Harris (1990), in prefacing their own study of spatial ability and handedness, relate that researchers have termed Levy's position the "cognitive crowding" hypothesis. Lewis and Harris also discuss the reanalysis of her earlier data done by Levy (1974). A more careful categorization of the sub-tests of the WAIS led her to conclude that left-handers scored lower than right-handers on sub-tests dealing with right-hemisphere tasks and higher than them on ones for left-hemisphere tasks.
Miller's own study examined a group of first-year psychology undergraduates. He compared right-handers to mixed-handers on both verbal and spatial tests. He found that the mixed-handers, while having a mean score on the verbal test virtually identical to that of the right- handers fell significantly behind in the spatial test. He thus reports that his research supports Levy's hypothesis.
Annett (1964), as described by Miller, argues rather that there are very few true left- handers in the world. Most individuals regarded as left-handed are in fact mixed-handed. The true right-handed individuals in this schema have verbal functions localized in the left hemisphere, while the group of mixed-handed individuals have bihemispherical verbal functioning. Moreover, Annett (1967) contends that in brain-damaged children spatial functions, when in competition with verbal ones within a hemisphere, will develop more strongly, to the detriment of verbal functions. Miller combines these two studies in order to show that Annett's research does not support either his or Levy's research and leads instead to conclusions that left-handers will be disadvantaged verbally rather than spatially. In fact, Annett's later work has taken her further. In Annett and Kilshaw (1982), the authors argue that enhanced mathematical ability is related to a decreased dextral bias. They tested a group of subjects for handedness, handedness of subjects' relatives, eye and foot preferences, and peg moving with each hand. Those who excelled in mathematics, both students and teachers, were more biased to the left side than the controls. Annett and Kilshaw conclude that these results arise because many right-handers are too biased to left-hemispheric language. This over-specialization makes it more difficult for many right-handers to function well with both hemispheres. And mathematics is a skill that requires an individual to put spatial constructs into linguistic terms. Thus the activities involved in mathematics require the cooperation of left-hemisphere language functions and right-hemisphere spatial functions, and this sort of cooperation is harder for extremely right-biased individuals to achieve. On the other hand, mixed- handers, who Annett believes make up a large portion of apparent left-handers, are better able to coordinate the hemispheres because of their lack of bias to one side.
Another very influential theory of handedness and its effects is that held by Geschwind and his various colleagues. Bakan (1990) summarizes this body of research. The Geschwind theory holds that a natural bias of humans toward left-hemispheric dominance may be mitigated by excessive levels of prenatal testosterone. The testosterone deters the development of the left hemisphere and leads to changes in the patterns of brain lateralization and a shift to the left in handedness. This stunting of the left hemisphere, according to the theory, can also lead to various learning disabilities and mental retardation. Furthermore, the testosterone also supposedly has an adverse influence on the development of the thymus gland, an integral part of the immune system. The Geschwind hypothesis thus posits an increase of immunological problems among left-handed individuals. On the other hand, Geschwind and Galaburda (1985) also contend that the cerebral areas outside those hampered by high prenatal testosterone may develop more as slow growth in some regions is compensated for by faster growth in others. Thus certain talents associated with these latter areas may be developed more fully in individuals whose left-hemispheric dominance has been compromised, i.e. left-handed individuals.
Much of the most recent work on handedness has sought to provide evidence for one of the hypotheses on the nature of handedness already described. Natsopoulos and Xeromeritou (1989) attempted, like Miller (1971), to resolve the contention between the work of Levy and that of Annett on verbal abilities and handedness. They looked specifically at the verbal capabilities of children. Their study ends up supporting neither the position of Levy nor the one of Annett, as Natsopoulos and Xeromeritou were unable to find differences in verbal ability between left-handed and right-handed individuals. Natsopoulos, Kiosseoglou, and Xeromeritou (1992) then examined spatial ability and its relationship to handedness. They found that the left-handed individuals in their study had enhanced spatial, mathematical, and general intellectual abilities. They thus agree with the conclusion of Annett that extreme dextral bias has deleterious effects for certain aspects of intellectual performance.
Likewise, Benbow (1986) found that left-handers and mixed-handers were over- represented among highly gifted adolescents who excelled in either mathematical or verbal reasoning ability. This study is essentially an attempt to verify empirically some of the conclusions of the Geschwind theory, so she examined immune disorders and myopia along with handedness. Her data did show elevated levels of asthma, allergies, and myopia among the gifted youths. On the subject of handedness, she concludes that individuals with cognitive functioning less localized to the left hemisphere are more likely to excel in tasks that involve reasoning, whether the reasoning is verbal or mathematical. The results for verbal reasoning stem from the fact that complex verbal reasoning, such as was tested by her means of assessment, the verbal SAT, is more likely to have right hemisphere components than more simplistic verbal tasks. The individuals with these less localized patterns of cognition are ones who are left-handed, mixed- handed, or even right-handed with left-handed relatives. In her reanalysis of her 1986 data (O'Boyle and Benbow, 1990), Benbow states that her data support the theory of Annett that a reduced dextral bias enhances some cognitive functioning. For all of Benbow's subgroups, except the mathematically talented females, there was a shift away from consistent right-handedness.
Unfortunately for anyone trying to find a consistent pattern in the research on handedness and cognitive abilities, Wiley and Goldstein (1991) specifically refute the findings of Benbow on the basis of their own analysis of gifted youths. They report no significant connections between left-handedness and extreme giftedness. Flannery and Liederman (1995) tested the Geschwind thesis and found that there was an increased incidence of non-right handedness among individuals with generalized neurological problems such as mental retardation but not among individuals manifesting problems regarded as resulting from left-hemisphere dysfunction. Moreover, they could not find evidence supporting any other postulate of the Geschwind thesis, including that a shift away from right-handedness is accompanied by an increased occurrence of certain special talents.
O'Boyle and Benbow (1990) survey a good cross-section of the work arguing for enhanced abilities of left-handers. They mention that Springer and Deutch (1989) discuss the incidence of left-handedness among great artists such as Leonardo da Vinci and Michelangelo Buonarotti. Several other studies have found that left-handers figure prominently in contemporary surveys of architects. Coren (1995) cites the work of Cranberg and Albert (1988), who report a high incidence of left-handed chess masters. These studies seem to reinforce the position that left- handers do indeed have heightened spatial ability considering the importance of spatial competence to artists, designers, and chess-players. Stanley Coren, in his Left-Hander Syndrome (1992), a book filled with statements both strong and controversial, states that some of the research on handedness tends to place the ability of left-handers at the extremes. In other words, left-handers appear to be over-represented both among the extraordinary intelligent and among the mentally disabled. Yet in saying this he is focusing on only one strain, albeit a very important one, of the research. A great number of the questions on handedness and different cognitive abilities have yet to be conclusively answered.
Many of the studies on handedness have considered factors beyond just a basic measure of right- or left-handedness. These other factors create a good deal of confusion in the accounts of the effects of handedness, but they may also be the key to finally unlocking some of the mysteries surrounding cerebral lateralization. For instance, many researchers see the importance of also taking sex differences into account and have found that different handedness patterns do not seem to have the same effects in males and females.
Lewis and Harris (1990) looked at a group of male and female high-reasoners, as determined by ACT scores. They administered four different spatial tests to each subject. There were two tests designed to measure spatial visualization, one for spatial perception, and one for mental rotation. The left-handed females outscored right-handed females on all four tests, while right-handed males outscored left-handed males on two of the four tests, with the differences on the other two insignificant. The results indicate that perhaps the cerebral organization of spatial functions differs between male and female individuals as well as between right- and left-handed individuals.
Snyder and Harris (1996) administered a somewhat modified version of the "Gallup Geography Test" to a group of college students. They were examining the effects of both handedness and sex differences on performance. They found that men outscored women, with little difference between the right-handed and left-handed men. Among the women, however, the left-handers performed significantly better on the test than did the right-handers, perhaps making a case for enhanced spatial understanding among left-handed relative to right-handed women. Ceiling effects caused by a good number of the men answering all the questions correctly precluded any conclusions about differential performance by differently handed male subjects. One interesting result of this study was that left-handed women were the only group within which there was not a significant relationship between individuals' scores on the geography test and scores on a generally administered spatial reasoning test. This finding suggests that perhaps left- handed women process geographical information in a way less related to their spatial reasoning than do the members of the other three subgroups.
While such studies exist supporting the hypothesis that left-handed women have spatial advantages over right-handed women and left-handed men are inferior to or nearly the same as right-handed men in spatial measures, there are yet again voices to the contrary. Sanders, Wilson and Vandenberg (1982) did a large-scale study in which subjects were divided by sex, ethnicity (all were of Chinese, Japanese, or European ancestry), and handedness (left, right, or mixed) and tested for spatial ability. For all three ethnic groups, strongly left-handed males outperformed strongly right-handed males. For the Asian groups, the mixed-handed males fell below each of the other two groups, though for Caucasians the mixed-handers performed between the other two groups. For females, the findings were essentially the opposite. For all three ethnic groups, left- handed females turned in poorer performances than right-handed females. Also, O'Boyle and Benbow (1990) cite a study by Klintenberg, Levander, and Schalling (1987) that showed that left- handed females scored lower on a maze task than any other group in the study. Some of the discrepancies between these studies and Lewis and Harris (1990) may be the result of Lewis and Harris using only high-reasoners. Yet the various conclusions still remain difficult to reconcile.
Martino and Winner (1995) examined spatial and verbal ability as manifested in individuals of different sex, handedness, and college major. They discovered that men were more likely than women to have higher spatial reasoning and more verbal problems. Left-handed and right-handed men did not perform differently. Yet within the group of male spatial majors, the left- handers had the highest spatial ability. They also had more verbal deficits than any other subgroup in the study. Though the exact classification of subjects within this study is somewhat different from that used by Lewis and Harris (1990), the two studies may still be roughly compared. That left-handed males in spatial majors outperformed all others in spatial ability would not be predicted from Lewis and Harris' results. For male as well as female subjects, the handedness by sex picture is unclear.
Therefore, controlling for sex does not make the relationship between handedness and individual differences come into clear focus. Many researchers have also seen degree of preference for a certain hand and family history of handedness as means to settle some of the disputes in the field. I have already alluded to the work of Annett, who sees an important break between the pure right-handed individual, who uses his right hand consistently and has a pronounced bias to left- hemispheric language, and the mixed-hander, who, regardless of which hand he writes with, has a more diffuse pattern of cognitive organization. Thus consideration of the strength of hand preference is important within her research, as is family sinistrality (i.e. the presence of left-handed relatives).
Several studies cited by O'Boyle and Benbow (1990) concluded that left-handed individuals with a positive history of family sinistrality (FS+) score lower than either other left- handers or right-handers on various cognitive tests, especially ones testing for some sort of spatial ability. She also discusses the findings of Yeo and Cohen (1983), who found that FS+ handicapped spatial ability in right-handers, and Burnett, et al. (1982), who found instead that a FS+ led to increased spatial ability for right-handers. Thus while it is rather unclear how family sinistrality affects right-handers, there may be some hope that the connection between positive family sinistrality and deficits in spatial performance in left-handers is a robust one. Yet Natsopoulos and Xeromeritou (1989) and Natsopoulos, Kiosseoglou, and Xeromeritou (1992) failed to find differential performance between FS+ and FS- children.
In an effort to better tease out the strands of the question, Searleman, Herrmann, and Coventry (1984) looked at the performance of left-handers on the SAT and found that strongly left- handed FS+ individuals had lower scores than moderately left-handed individuals, with or without family sinistrality, or strongly left-handed FS- individuals. This last subgroup had the highest scores among the four, a finding the researchers report as a surprise. Searleman, et al., conclude that strongly left-handed FS- individuals lacking family sinistrality may in fact have certain cognitive advantages relative to the general population. Snyder and Harris (1993) partially support the findings of Searleman, et al. The former discovered that consistently left-handed FS+ individuals perform significantly worse than right-handers on a mental rotation test while inconsistently left-handed FS+ individuals did not. Snyder and Harris thus also tag the consistently left-handed FS+ individual as the cause of the performance deficits sometimes shown for left-handers.
If one were to extrapolate from these findings, one could hypothesize, as many of the researchers do, that the inconsistencies in the research into handedness and individual differences may be the result of failures to account for strength of handedness and family sinistrality. Perhaps studies in which left-handers show deficits in ability are ones in which strongly left-handed FS+ individuals are over-represented, while ones that demonstrate enhanced abilities in left-handers are ones in which there is a preponderance of strongly left-handed FS- individuals.
The studies that consider multiple factors, by proposing a reason for the disarray of handedness and individual differences research, have taken a step toward making sense of a knotty issue. Yet the field still seems far from settling on some generally accepted conclusions. So what is the relevance of this chaotic field to HCI?
Researchers have proposed various cognitive abilities that lead to differential performance with computer applications. For example, Allen (1994) examined the ability of subjects to identify references to articles that dealt with the same topic as an article read by the subjects. It was discovered that individuals scoring higher on a Diagramming Relationships Test, a test of ability to identify the most accurate graphical representation of the relationship between three concepts, were able to better identify relevant article references than those with lower scores on the relationships test when the references were presented in a standard reverse chronological order. When the references were presented in ranked order of relevance, however, performance discrepancies largely disappeared. Thus Allen has demonstrated a relationship between general cognitive abilities and performance on a computer system, and at the same time found a way to minimize performance variations.
Allen points out at the end of his article that much work remains to be done on the cognitive characteristics of users as they relate to performance. As more research on individual differences and human-computer interaction is done, a clearer picture of what cognitive factors influence performance at the computer interface should emerge. With this knowledge, the field of human-computer interaction may be able to further the research into handedness, cerebral lateralization, and cognition. If, as Allen posits, accuracy in information retrieval tasks is related to cognitive ability, differential performance among handedness groups in such tasks may yield solid evidence on individual differences and handedness. Computer applications thus might become another means beside more traditional sorts of tests of cognitive abilities for trying to discern differences between groups.
One possible area of research is hypermedia environments. Baecker et al. (1995) point out that Conklin (1987) contends that one of the major problems with hypermedia is the tendency of users to become disoriented and lose track of structure and connections. It might be valuable to look at whether certain handedness groups have this problem more than others. It may answer questions either about cognition in hypermedia or about cognition and handedness, depending on what one used for a starting point.
For instance, a hypermedia study structured similarly to the geography study carried out by Snyder and Harris (1996) may yield worthwhile results. Groups of users, controlled for sex and handedness, could be introduced to an intentionally somewhat confusing hypermedia environment. After a timed and possibly guided tour of the environment, subjects could be questioned on its structure and navigation. Perhaps a second test relating more to the information content of the hypertext could then be administered. It is possible that the subgroup performance of subjects on a test of hypermedia structure would mirror the results that Snyder and Harris obtained for knowledge of real-world geography. Such a finding would provide evidence, though not proof, that gaining an understanding of a hypermedia "landscapes" is cognitively similar to gaining an understanding of real-world ones. If, however, the patterns of performance diverge from the Snyder and Harris results, there would be evidence that a rather different cognitive skill is used in navigating hypermedia. Performance of different handedness and sex groups with respect to structural knowledge of the hypertext could then be compared to their performance on the content test. It may be found that the differential performance patterns are not the same and thus that comprehension of structure and comprehension of content are very different tasks. This proposed experiment is one that would take advantage of extant handedness research to try to gain a better understanding of user cognition in a hypermedia environment.
There is of course a converse (and perhaps more prudent, considering the inconsistency of handedness research) way of looking at research into handedness and hypermedia. Instead of taking as a starting point a study such as that of Snyder and Harris that draws conclusions about handedness and cognition, one could instead attempt to ascertain the nature of user cognition in hypermedia environments using spatial batteries, geography tests, and tests of performance within the hypermedia space, without controlling for handedness or sex. Established conclusions about this user cognition might then be used to gain a knowledge of differences in handedness groups. For instance, if solid evidence suggests that hypermedia navigation is essentially a spatial task, testing groups controlled for handedness and sex (and possibly the other factors handedness researchers argue should be considered) on their ability to navigate in these environments could provide further evidence on which subgroups exhibit spatial advantages or deficits. Thus, hypermedia environments could become a tool for handedness researchers. For example, the results of a study in which both navigational knowledge and content knowledge of a hypertext are tested for and then summed may indicate that non-right-handers perform better. Such results may constitute evidence in favor of the thesis of Annett, since they would demonstrate that dextrally biased individuals perform worse than others in tasks involving both spatial and verbal components. This argument is essentially the one she made for why non- right-handers are over-represented among those skilled in mathematics. At any rate, both handedness and hypermedia have a number of questions surrounding them, and maybe some of the answers can be found by bringing these questions together.
I hope that these suggestions for future research demonstrate that there is the possibility of a mutually beneficial blending of handedness research and HCI research. If one is willing to jump to some tentative conclusions about the trends in the research on handedness and cognitive abilities, there are other factors relevant to HCI that appear.
If the evidence presented by Searleman, et al. (1984) can be relied upon, there are certain subgroups of left-handers (strongly left-handed FS+) who experience cognitive deficits and others (strongly left-handed FS-) who tend to be cognitively gifted. The evidence of Allen suggests that low-reasoners need more affordances to perform at the same level as high-reasoners. A rather Draconian plan of action based upon the association of these findings might be one in which it is proposed that the aforementioned FS- group be selected for in employment decisions and the FS+ group be selected against. After all, as Dillon and Watson (1996) argue from their survey of several meta-analyses of ability tests, cognitive ability accounts for about 25% of variance in performance in a work setting. A more modest appraisal may lead managers to a better awareness of individuals who may need extra assistance or training and designers to attempts to create affordances for poorer performers. Providing extra assistance to certain workers is especially important when dealing with computer systems, since computerization has vastly increased the variations in performance across individuals, widening the gulf that exists between the best and the worst actors (Egan, 1988). If handedness by family sinistrality and strength of hand preference is indeed a robust indicator of cognitive ability, computer support resources may be focused on handedness groups most likely to need them. Furthermore, the handedness research may provide clues not only on who to help but also on how to help them. For example, the Searleman, et al. study posits that strongly left-handed FS+ males lag significantly in verbal ability. An interface customization or training protocol that helps this user subgroup understand the computer application in more spatial terms may serve to nullify the performance deficit. Unfortunately, Snyder and Harris (1993), along with other researchers, found this group deficient in a mental rotation task, so the question of how to help this group must remain open for the time- being.
Another consideration is that, as ever younger children are introduced to computer technology, instructors working with this user group need to be aware of which children may potentially have the greatest problems learning the necessary skills. The study of Tierney et al. (1984) argues that children with mixed or not yet established handedness exhibit cognitive deficits relative to established right-handers, so this might be a subgroup to pay close attention to when dealing with a group of children containing both individuals with established handedness patterns and ones without. Research on handedness may eventually enable instructors to quickly and easily determine which young learners require greater attention
As with handedness and differences in cognitive ability, work has been done on handedness and differences in personality. Lester (1987) found that there was a correlation between handedness and extraversion for females. He posits from a study of nearly two hundred undergraduate age students that left-handed females are less extraverted than right-handed females. Coren (1992) cites the work of Hicks and Pellegrini (1978), who looked at a group of 266 university students and discovered that left-handers score higher on measures of anxiety. The study of Weinrich, Wells, and McManus (1982) explicitly takes exception with the Hicks and Pellegrini study. The former group argues instead that anxiety seems to be greater among those of extreme handedness, whether left or right, and among women relative to men. More research cited by Coren, that of Schuenenmann, Pickleman, and Freeark (1985), backs up the conclusions of Hicks and Pellegrini. Schuenenmann, et al., examined surgical residents at Loyola University and discovered that left-handers were more prone both to react strongly to stress and to act cautiously. Coren's own research, which employed the Interpersonal Adjective Scale to test for a variety of personality characteristics found that left-handers were more likely to be introverted, distant, harsh, and somewhat belligerent than right-handers.
The implications of this work for HCI are somewhat unclear, especially considering that, as Dillon and Watson (1996) point out, there is a dearth of good empirical evidence on the interaction between personality traits and work. A possible area of future research, in light of what has already been done on handedness and anxiety, is whether anxiety about technology, specifically computer technology, is related to handedness. Since some research has also demonstrated that women are more anxious than men, it would behoove anyone studying computer anxiety and handedness to control for sex as well. If researchers are able to gather solid evidence that left-handedness positively correlates with computer anxiety, an imperative for systems implementation may arise. It may very well be that left-handed individuals will need extra assistance in becoming comfortable with technological changes. And if the traits derived by Coren are truly characteristic of left-handers, it would seem that providing this assistance without it being solicited would be wise. For, if Coren's research is correct, it would appear that left-handers may be more likely than others to quietly and resentfully resist a technology change rather than ask for help.
Pipe (1990) surveys a large number of studies on handedness in mentally disabled groups. This area is one of the few areas in handedness research in which reasonably consistent results have been attained. There is little doubt that mentally handicapped individuals are significantly more likely to be left-handed than the members of the general population. The very recent study of Flannery and Liederman (1995) confirms the work of the earlier researchers. There are a number of theories on why this is the case. One essentially entails an idea that generalized cerebral damage destroys natural hand bias, which would most likely have been to the right, and creates left- or mixed- handers who were "naturally" predisposed to right-handedness. Another argument is that lesions in one hemisphere that are severe enough to cause mental disabilities can switch dominance to the other hemisphere and change hand preference. This switch is more likely to create "pathological," non-natural left-handers than "pathological" right-handers because more individuals are right-handed prior to the damage (Pipe 1990; Flannery and Liederman 1995).
Regardless of the origin of this elevated incidence of left-handedness among the mentally disabled, it is a fact that has to be dealt with by those who work with mentally disabled individuals. Lipson (1984) writes on how teachers of learning disabled students might better customize classroom layouts and teaching styles to deal with the unusually high number of left- handed students among this population. As computers become more mundane tools, their usage by individuals who are mildly retarded or disabled will presumably increase. With this occurrence, computer instructors who work with this growing user population should keep some of Lipson's recommendations in mind. For instance, ideally at least some computer work stations in a training setting for disabled individuals should be set up so that left-handed individuals can manipulate input devices, such as mice, more easily. Lipson suggests that certain non-traditional teaching methods may be more effective with left-handed learning disabled children, and this may be true for computer instruction also.
One of the few studies that has taken account of handedness differences at the interface was conducted by Marquez and Lehman (1992). They looked at placement of icon buttons in a hypermedia application, placing them either to the top and right or to the bottom and left of the screen. The study employed only five left-handed subjects. (Only one of them "moused" with the left hand, probably because most of them were used to right-handed work stations). Yet despite this small sample size, there were some interesting conclusions. It was discovered that subjects inexperienced with the use of a mouse (only one left-hander fell into this group) had a difficult time manipulating the mouse to click on the icon buttons in a task performance test. Also a survey of subjective preference in interface design revealed that all the right-handed users who selected "Right Handedness" as one of the main reasons they liked one design over the other preferred the right-top design. The lone individual who moused with his left hand preferred the left-bottom design and gave left-handedness as the main reason why. Marquez and Lehman suggest further research into the benefits to interface design of including movable icon buttons and considering individual differences among users. I would argue that the implications of this study are profound for computer use among the disabled. If novice mouse users of normal intelligence have difficulty with this input device, there is a good chance that disabled individuals will have even greater problems acclimating themselves to the mouse. And if a computer setting has exclusively right- side mice that are difficult to move to left side because of various factors in the layout of the whole work station (e.g. the hard drive or the edge of the table could be to the left), the learning process may become that much more frustrating. Furthermore, customization of icon placement toward the bottom and left for disabled left-handers might also make the first steps into computing easier.
There is also research that indicates that certain other user groups have higher than expected proportions of left-handers. As mentioned above, architects seem to be overly sinistral in their biases. Schacter and Ransil (1996) looked at handedness in nine occupational groups. They found that architects followed by lawyers and accountants had the highest incidence of left- handedness among the nine groups, with architects showing a much greater tendency to left- handedness than the population at large. Their work was preceded by that of Peterson and Lansky (1974 and 1977) who found that left-handers are more successful in architecture and make up a remarkably high (29%) proportion of architecture school faculty members. Aggleton and Wood (1991) dispute these findings, but the preponderance of evidence seems to be that at least among architects there is indeed an increased presence of left-handers.
The conclusion on architects suggests on the one hand that student computer labs within architecture, and for that matter graphic arts, departments in universities should have work stations in which rearrangement for the left-handed is relatively easy. While it does not take much coordination to simply click an icon while mousing with the non-preferred hand, the fine muscle control involved in computer-aided art and design is for most individuals confined to one hand. For many architecture students, that hand is the left. The discovery that accountancy has many left-handed members militates in favor of designing some keyboards with a number keypad on the left rather than the right side. Researchers in HCI might also be interested in the question of whether designers for a computer environment show a similar tendency toward left-handedness as the one shown by architects. And if the proportion of left-handers is significantly different, it may point to some fundamental and intriguing differences in the nature of architectural design and software or interface design.
Coren (1995) completed a large and impressive study looking at handedness and divergent thinking. He factored in sex and degree of handedness, dividing individuals into consistent left, mixed left, consistent right, and mixed right groups. In one of three tests for divergent thinking, an alternate uses test, no significant differences were found between handedness groups. On two additional divergent thinking tests, an originality test and a intellectual flexibility test, analysis of the scores for females still showed no significant differences across handedness groups. For males, however, the divergent thinking scores rose steadily as a function of increased sinistrality,with consistently left-handed individuals scoring the best. A test for convergent thinking as a control in order to eliminate the possibility of simply measuring general cognitive ability showed that right-handers as a whole scored slightly higher. Coren seems to argue from his results that it might be enhanced divergent thinking rather than enhanced spatial ability that accounts for elevated numbers of left-handers among those talented in art, mathematics, or chess. Further research into this question might be done with the aid of computers, considering the ability of computer designers to create games or other applications in which divergent thinking abilities are tapped. And if further research supports Coren, HCI researchers might want, as I have already suggested, to take a look at those very designers and see how many of them are left- handed themselves, since divergent thinking is so important to software design. Moreover, if it is discovered that there are very few left-handers among software designers, robust research on the connection between left-handedness and divergent thinking may indicate that there should be more.
The work of Hassler and Gupta (1993) found that left-handed individuals scored higher on a test for musical talent than did right-handers. They cite the earlier studies of Byrne (1974) and Deutch (1980) as having similar results. Hassler and Gupta also discovered that musicians exhibit less hemispheric specialization for linguistic processing than do non-musicians. O' Boyle and Benbow (1990) state that Witelson (1980) reached a similar conclusion.
This body of research has interesting implications as the field of HCI theorizes on computer applications achieving greater exploitation of the resources of the auditory channel. Baecker et al. (1995) discuss the underutilization of sound in computer systems and make reference to one study (Brewster, Wright, and Edwards, 1993) that demonstrated some benefit to users from the use of melodic sound. One may hypothesize about future computer interfaces that attempt to use more complicated sounds, including melodic ones, to interact with the user. And it may very well be that there will differential performance among handedness groups at this interface if left-handers are truly better able to process complex auditory stimuli. Also, the effects might be influenced by the placement of sound output devices if the differences between left-handers and right-handers stem from the degree and site of the localization of speech functions. This area is one that hopefully will receive more attention as computer applications become more multimedia in nature.
Gabbard, Hart, and Gentry (1995) looked at motor skills in young (approximately seven years) children and discovered that left- and mixed-handed children are inferior to right- handed children in motor performance. On the other hand, in a study of older subjects, Kilshaw and Annett (1983) found that those with extreme dextral biases were rather uncoordinated with their left hands and were deficient to left-handers in overall coordination. Left-handed males were even quicker with their left hands than right-handers with their right hands.
These two studies taken together have implications for HCI. The former, if confirmed by further research, reinforces the need to allow left-handed children to have work stations set up with them in mind. Their coordination problems will be exaggerated in a setting prejudiced against them. The latter study suggests the need for further research on differential performance between left-handed and right-handed adults in computer contexts in which performance depends more on motor skill than reasoning ability. The Kilshaw and Annett study would imply that left-handers would perform better, especially if both hands are used a great deal. Coren (1992) does in fact point out that a great number of champion typists have been left-handed. He attributes it largely to the QWERTY keyboard's design, which favors the left hand. Yet it may instead stem, at least partly, from the ambidextrous nature of typing.
I have discussed here some of the research that has been done on handedness and individual differences. Even with so many questions on handedness still undecided, implications for HCI are present in handedness research. These two fields should begin to look more at one another in order to establish a mutually beneficial relationship. .
Allen, B. (1994) Cognitive abilities and information system usability. Information Processing and Management, 30, 177-191.
Annett, M. and Kilshaw, D. (1982). Mathematical ability and lateral asymmetry. Cortex, 18, 546-568.
Baecker, R. et al. (1995). Hypertext and Multimedia. In Baecker, R. et al (Ed.), Readings in human-computer interaction: Toward the year 2000 (pp.833-842). San Francisco: Morgan Kaufmann Publishers, Inc.
Baecker, R. et al. (1995). Speech, Language, and Audition. In Baecker, R. et al (Ed.), Readings in human-computer interaction: Toward the year 2000 (pp.525-537). San Francisco: Morgan Kaufmann Publishers, Inc.
Bakan, P. (1990). Nonright-handedness and the continuum of reproductive casualty. In S. Coren (Ed.), Left-handedness: Behavioral implications and anomalies (pp. 33-74). New York: Elsevier Science Publishers B.V. (North-Holland).
Benbow, C.P. (1986). Physiological correlates of extreme intellectual precocity. Neuropsychologia , 24, 719-725.
Coren, S. and Porac, C. (1982). Lateral preference and cognitive skills: An indirect test. Perceptual and Motor Skills , 54, 787-792.
Coren, S. (Ed.). (1990). Left-handedness: Behavioral implications and anomalies . New York: Elsevier Science Publishers B.V. (North-Holland).
Coren, S. (1992). The Left-hander syndrome . London: John Murray.
Coren, S. (1995). Differences in divergent thinking as a function of handedness and sex. American Journal of Psychology , 108, 311-325.
Dillon, A. and Watson, C. (1996). Individual differences and the analysis ofusers in HCI [Online]. Available: http://www-slis.lib.indiana.edu/PrePrints/adillon-inddiff.html [1996, December 8].
Egan, D. (1988) Individual differences in human-computer interaction. In M. Helander (ed.) Handbook of Human-computer Interaction . North-Holland: Elsevier. 543-568.
Flannery, K.A. and Liederman, J. (1995). Is there really a syndrome involving the cooccurrence of neurodevelopmental disorder, talent, non-right handedness and immune disorder among children. Cortex , 31, 503-515.
Gabbard, C., Hart, S., and Gentry, V. (1995). General motor proficiency and handedness in children. Journal of Genetic Psychology , 156, 411-416.
Geschwind, N. and Galaburda, A.M. (1985). Cerebral lateralization: Biological mechanisms, associations, and pathology: II. A hypothersis and a program for research. Archives of Neurology , 42, 521-552.
Hassler, M. and Gupta, D. (1993). Brain organization, handedness, and immune vulnerability in relation to musical talent. Neuropsychologia , 31, 655-660.
James, W.H. (1995). Handedness and size. Perceptual and Motor Skills , 80, 1250- 1250.
Kilshaw, D. and Annett, M. (1983). Right- and left-hand skill: effects of age, sex, and hand preference. British Journal of Psychology , 74, 253-281.
Lester, D. (1987). The relationship between handedness and personality traits: Extraversion and neuroticism. Personality and Individual Differences , 8, 437.
Lewis, R.S. and Harris, L.J. (1990). Handedness, sex, and spatial ability. In S. Coren (Ed.), Left-handedness: Behavioral implications and anomalies (pp. 319-341). New York: Elsevier Science Publishers B.V. (North-Holland).
Lipson, A. (1984). Left-handed connections. Academic Therapy , 20, 179-187.
Marquez, M.E. and Lehman, J.D. (1992). Hypermedia user-interface design: The role of individual differences in placement of icon buttons. Journal of Educational Multimedia and Hypermedia , 1, 417-429.
Martino, G. and Winner, E. (1995). Talents and disorders-- Relationships among handedness, sex, and college major. Brain and Cognition , 29, 66-84.
Miller, E. (1971). Handedness and the pattern of human ability. British Journal of Psychology , 62, 111-112
Natsopoulos, D. and Xeromeritou, A. (1989). Verbal abilities of left-handed and right-handed children. Journal of Psychology , 123, 121-132.
Natsopoulos, D., Kiosseoglou, G., and Xeromeritou, A. (1992). Handedness and spatial ability in children: Further support for Geschwind's hypothesis of "pathology of superiority" and for Annett's theory of intelligence. Genetic, Social, and General Psychology Monographs , 118, 103-126.
O'Boyle, M.W. and Benbow, C.P. (1990). Handedness and its relationship to ability and talent. In S. Coren (Ed.), Left-handedness: Behavioral implications and anomalies (pp. 343-372). New York: Elsevier Science Publishers B.V. (North-Holland).
Pipe, M. (1990). Mental retardation and left-handedness: Evidence and theories. In S. Coren (Ed.), Left-handedness: Behavioral implications and anomalies (pp. 293-318). New York: Elsevier Science Publishers B.V. (North-Holland).
Sanders, B., Wilson, J.R., and Vandenberg, S.G. (1982). Handedness and spatial ability. Cortex , 18, 79-89.
Schacter, S.C. and Ransil, B.J. (1996). Handedness distributions in 9 professional groups. Perceptual and Motor Skills , 82, 51-63.
Searleman, A.S., Herrmann, D.J., and Coventry, A.K. (1984). Cognitive abilities and left- handedness: An interaction between familial sinistrality and strength of handedness. Intelligence , 8, 295-304.
Snyder, P.J. and Harris, L.J. (1993). Handedness, sex, and familial sinistrality effects on spatial tasks. Cortex , 29, 115-134.
Snyder, P.J. and Harris, L.J. (1996). Where in the world am I-- Sex and handedness differences in knowledge of geography. Perceptual and Motor Skills , 82, 1379-1385.
Tierney, I. et al. (1984). The McCarthy's scales of children's abilities: sex and handedness effects in 128 Scottish five-year-olds. Bristish Journal of Educational Psychology , 54, 101-105.
Wienrich, A.M., Wells, P.A., and McManus, C. (1982). Handedness and sex differences. British Journal of Psychology , 73, 69-72.
Wiley, J. and Goldstein, D. (1991). Sex, handedness, and allergy: Are they related to academic giftedness? Journal for the Education of the Gifted , 14, 412-422.
Wood, C.J. and Aggleton, J.P. (1991). Occupation and handedness: An explanation of architects and mail survey biases. Canadian Journal of Psychology , 45, 395-404.