From the beginning of human existence, humans have found various ways to record and share information with others. From cave drawings to the Dead Sea Scrolls, we have seen how data was recorded in our past. Today’s world is very different, with the emergence of computer technology to share information. CleanKeys is a company that is developing a sealed, touch-input keyboard that can be easily cleaned to prevent the spread of bacteria. The company needed to know how individuals type today, and this study served to help them answer that question. Previously, there were no studies that analyzed the typing styles of individuals. Typists were described as “touch typists” and “hunt-and-peck typists”. Based on the data collected from observation of individuals typing and analysis of the excerpts they were asked to type, we have identified three groups of typists rather than the previously described two groups. This research showed how what we had originally thought about various age groups and typing speeds, accuracy, and typing styles did not reflect reality. Occupational therapists can utilize this data to set realistic goals for their patients in a clinical setting.
Keywords: Typing Styles, Health Risks, Keyboard Technology
From the early cave paintings of the Paleolithic period, humans have recorded significant life events for future access (The Archive, ND). Through the medieval era, the act of recording was largely the province of those with special skills or training, commonly called “scribes” (Tillotson, 2006). With the spread of education, more people were becoming trained to read (consume) and write (store) information, and the process of reading and writing soon became central to social interaction.
Since more people are creating and communicating information between one another, we need to understand how that process is done today. From the early 1900s to the 1980s, information was predominantly stored on paper, in files, and encoded using pencils, pens, and typewriters. With the digital revolution, we began storing information digitally, and accessing it electronically (Kunde, 1986). “These technologies provide the ability to gather, sort, process, transfer, and display information about highly complex events that occur in wide geographic areas” (Nye & Owens, 1996). Rather than going to a library, file room, or document storage facility, workers could call up information at their desktop using personal computers and Internet connections. This capability, developed for business, also became common in everyday life. This change made the ability to effectively use computer keyboards central to the daily life of workers, students, and everyone else in pursuing their personal interests.
Computers are used in nearly every workplace from offices to warehouses and even the home. Many corporations make use of the Internet to communicate and coordinate operations between widely separated members of their organization. Employees whose offices are located in outlying facilities, without information technology, may not be able to attend daily meetings at the "home office" nor participate in the free-flowing exchange of information that generates new ideas. Electronic mail is distinctive in allowing management to provide employees with information regarding company events, strategies and goals (Kraut, 1996). The availability of affordable computers has allowed smaller organizations to capture and analyze complex information as well. Both national and local grocery stores are able to use the information from scanning checkouts to monitor the flow of products from their shelves (NA, 2010). In addition to work-related computer applications, keyboarding has become central to social participation. Instant messaging, Facebook and Twitter have made keyboarding an almost constant activity throughout the day among their aficionados.
The act of keyboarding is a central part of the modern American lifestyle, but little is known about how individuals actually type. We do know that keyboards can be found in a variety of settings, including schools, homes and offices. We know that young students in school are taught keyboarding techniques, but we do not know how well those lessons are followed in everyday life. We do not know the keyboarding techniques used by adults who were not formally taught to type or who were taught to type on mechanical typewriters where the physical demands are quite different from those of computer keyboards. We do not know, for typical typists, what speed and accuracy might be reasonably expected. We do not know typical productivity levels. We, as occupational therapists working with clients with disabilities, do not know where to set goals that are reasonable for our clients. Since occupational therapy is based on the precept that participation in meaningful activity supports health, and since computer-based activities have become primary meaningful activities for the majority of people in our country, occupational therapists must know how to engage their clients in computer-based activities, both as a therapeutic medium and to enable participation in daily life.
As the value of keyboarding, and the hours spent using keyboards each day increased, so did awareness that the keyboard is not an unmixed blessing. Prolonged keyboard use on a daily basis appears to present risks to the health of the user. The most widely recognized of these risks is a condition once referred to as Carpal Tunnel Syndrome, and now more commonly referred to as repetitive stress injury (RSI).
“RSI is a soft-tissue disorder that results from the repetitive use of some part of the body” (Heintzelman & Pfeiffer, 1997). Among computer users, RSI is associated with excessive use of the mouse and keyboard. Untreated, keyboard related RSI can lead to chronic numbness, pain, weakness, and even paralysis of the hands, fingers, and wrist. While not all keyboard users will develop RSI, any individual who uses a computer more than 2 hours per day is at increased risk (Heintzelman & Pfeiffer, 1997). This level of exposure is common among computer programmers, hotel and airport clerks, and young adult’s communication via social media as their daily activities.
While the association of RSI with extended use and poorly designed workstations is generally recognized (Rempel, Serina, & Klinenberg, 1997), the magnitude of the problem is often overlooked. It has been estimated that, at the turn of the century, the total cost to business of RSI was $20 billion annually (Okamoto, 1998). Many individuals working with computers are aware of this common injury risk yet do not take preventative measures. “Varying the activity and position are the simplest and most effective ways to prevent RSI. Micro-breaks - brief rest-periods taken at frequent intervals throughout the day help to reduce a user’s risk of injury” (Heintzelman & Pfeiffer, 1997).
In addition to the ergonomic risks of keyboard use, epidemiologists have recently become concerned about a second type of keyboard risk: disease transmission.
Those who use keyboards daily may notice that, over time, the keys become discolored and dirty. Because keyboard manipulation depends on the same parts of the body (the hands) used to manipulate other feature of the environment, environmental soils are transferred to the keyboard. This “cruft,” as it is called by computer hobbyists, is composed of many things, but includes bacteria that the user comes into contact with on a daily basis (see Figure 1). For personal computers, these bacteria pose little risk, since they simply return to the hands that delivered them to the keyboard in the first place. In shared settings, however, the keyboard can distribute infections between users by spreading bacteria of each user via the keyboard surface.
One setting where this is of particular concern is the hospital. The bacterial ecology of hospitals represent increased hazard for two reasons: increased virulence and increased susceptibility to infection. While most of the bacteria in hospitals are the same as those of the surrounding area, the concentration of disease-causing organisms (as opposed to the environmentally innocuous majority) is higher in hospitals (Petroudi, 2009; Weinstein, 1998). This increased concentration occurs because individuals who have succumbed to bacterial infection are admitted to hospitals for treatment, and the disease-causing bacteria are “shed” into the environment. Because bacteria share genetic material between strains, hospitals provide the opportunity for the development of new, more virulent varieties, such as Methicillin-resistant Staphyloccus aureus (MRSA) (Mayo Clinic Staff, 2010; Tletjen, Bossemeyer, & Mcintosh, 2003).
In addition to the potentially increased virulence of hospital-bred bacteria, hospital patients are at greater risk of infection than the general population. In addition, the “deconditioning” brought about by enforced bed-rest, increased stress, and disrupted schedule of a hospital stay increases susceptibility to infection (Aizen, Ljubuncic, Ljubuncic, Aizen, & Potasman, 2006). In some cases, hospital patients may be receiving treatments (radiation, chemotherapy, anti-transplant rejection drugs) that further suppress or diminish immune response, making the patients even more at risk of infection.
These combined influences have resulted in a high level of cross-contamination in hospital settings (Tietjen, Bossemeyer, & McIntosh, 2003). Diseases acquired in the hospital (nosocomial infections) can be controlled using proper infection control techniques, especially washing of hands and contaminated surfaces.
Shared keyboards of hospital workstations have been implicated as transmission agents for disease (Bures, Fishbain, Uyehara, Parker, & Berg, 2000; Dumke, 2005; Kramer, Schwebke, & Kampf, 2006). First, they are in contact with the hands, which are the most likely part of the body to be carrying pathogens picked up from the environment. Second, the interior of the conventional keyboard is “open” to the environment, so that it is easy for pathogens to move into the spaces between keys, from which they are difficult to extract. Since the keyboards that are used in hospitals and many other work settings are not waterproof it is hard to disinfect them. They are so much a part of the environment; they are often not even seen as a vector of infection. While shared keyboards in the hospital are a major concern, shared computer keyboards in other locations also have potential risks. Public access computers in libraries and schools, shared computers in offices and warehouses, and reservation system at airports all represent potential vectors for disease transmission.
The primary defense against environmental contamination is to clean the contaminated surfaces. While cleaning keyboard keys using antiseptic wipes can reduce the direct contact with pathogens, this approach is limited on the conventional keyboard because only the exterior surface can be cleaned. Conventional keyboards use individual micro-switches under each key to detect keypresses. Operating these switches requires a degree of vertical movement of the key relative to the keyboard. In most keyboard designs this entails gaps between keys and between the keys and the body of the keyboard which allow contaminants to enter the space below the keys.
Insert Figures 2 & 3 about here
One means of reducing the risks of keyboards is to seal the surface of the keyboard beneath a membrane. Existing keyboards can be covered with a flexible “skin” which is placed over the keys (see Figure 2). Alternatively, a sealed keyboard can be provided. Keyboards such as the iKey DT-5K-MEM-TP have been developed for use in environments such as machine shops where the conventional keyboard would soon be disabled by contamination (see Figure 3). In both cases, the intent of the design is to protect the keyboard from the user rather than the user from the keyboard, and no assurance of cleanability is provided.
Currently, Cleankeys, Inc. , of Edmonton Alberta is developing a cleanable keyboard that is specifically intended to limit transmission of pathogens (see Figure 4 and Figure 5). Rather than using individual switches for each key, the Cleankeys keyboard uses a multi-touch surface to identify finger locations, and “accelerometers” (sensors that can detect extremely small movements including the effect of a finger coming into contact with a firm surface) to triangulate the location of a key touch.
The process of developing this keyboard has shown that an understanding of the mechanics of typing are important in assuring that the keyboard can accurately determine where on the solid surface of the keyboard a finger was pressed (Marsden, 2009). Because a touch typist uses all four fingers of the hand, some keys are struck with the strong index or middle fingers, while others are struck with the weaker ring or little fingers. In contrast, a hunt-and-peck typist would use the stronger index and middle fingers primarily while typing, thus putting more equal pressure throughout all the keys instead of varying level of force with the weaker ring and pinky finger.
In spite of the ubiquity of typing in modern society, and the concerns about the physical and biological risks of typing, little is known about how typical people interact with the keyboard. Most discussions of typing methods focus on two extreme cases: hunt-and-peck or touch typing.
In hunt-and-peck typing, the typist has not committed the layout of the keyboard to motor memory. To type each key, the typist must first locate the key visually (the hunt phase), then press it with the one or two fingers used to type in this method (the peck phase). Typically, in hunt-and-peck typing, the typist will memorize or compose a few words or sentences prior to typing, then type those words from memory before referring once more to the source. The typists’ attention is constantly cycling between the source, the keyboard (to locate the keys) and the screen (to check for accuracy of typing).
In contrast, a touch typist has learned the layout of the keyboard, and developed motor-patterns for typing each key, and often for many words and phrases. When the hands are positioned on the home-row (aided by the tactile nubs on the “F” and “J” keys), these motor plans allow typing without visual surveillance of the keyboard. The attention of the typist can remain focused on the source material, with occasional references to the screen to assure accuracy. As the typist becomes more fluent, they cease to think about the location of the keys and the position of their fingers, and are able to focus on the words to be typed.
These two recognized approaches to typing differ along two vectors. Hunt-and-peck typing, as generally described, requires continuous visual attention (while typing), and uses just one or two fingers. Touch typing, by contrast, demands little or no visual attention, and uses all available fingers (9, usually. The left thumb is seldom used.). It is likely, however, that these conditions represent the end-points of two spectrums, and that most typists use methods that are at the endpoints of neither spectrum. For example, with experience, even a two-fingered typist will learn on which side of the keyboard a key resides, and thus limit the visual search. The degree of visual attention may diminish as more and more keys are learned. Similarly, over time, a typist may integrate more fingers into the typing process. Conceivably, any portion of the available fingers might be integrated into the typing process.
Recognizing the importance of keyboard use to modern life and the poorly characterized typing styles identified by the various methods by which individuals actually type, the proposed study will explore typing styles of the typical keyboard user in the community. We hope to answer the following questions:
What are the typing strategies used by typical typists. In addition to the usually discussed “hunt-and-peck” and “touch-typing” styles, are there identifiable strategies that are shared between typists? In addition to visual dependence and finger counts, are there additional variables that determine speed and accuracy of typing?
The population of interest for this study was adults who use computer keyboards in their daily life. The sample consisted of 75 community-dwelling adults, stratified by age. The sample included subjects in each of five age ranges: 18-29, 30-39, 40-49, 50-59, and 60+. All subjects in the study were able to see well enough to read 12 point, Times Roman print, hear well enough to participate in a conversation with or without hearing aids, could sit unsupported in a conventional chair without armrests, and were able to use both hands to access the keyboard (though two-handed typing was not required).
Subjects in this study typed for a measured 10 minutes from a provided source text. The number of words typed was divided by 10 to provide an average typing speed over the duration of the sample. Burst typing speeds were not recorded.
Typing errors can occur in a wide variety of ways. Within an individual word, letters may be dropped, extra letters may be inserted, or the order of letters may be inaccurate. More than one of these errors can occur within a single word. At the document level, words or lines may be skipped or repeated, or incorrect words substituted for the intended one. These varying types of errors make the process of determining accuracy difficult.
For purposes of this study, the subject's typed document was compared to the source document using Microsoft Word’s “Compare Documents” feature. Any identified “block” of text that was different between the source text and the subject’s typing was considered a single error, regardless of how many individual differences might occur within the block.
Based on the letter frequency of the English language and the arrangement of the keys on the standard keyboard, the expected frequency of finger usage was just over 20% for each index finger, and just under 2% for the little finger of the right hand. It is possible, therefore, that within a ten-minute trial, a finger that would be used in prolonged typing might not have the opportunity to be utilized in the short time frame. Because of this, we used two indicators of finger use.
The first indicator was observed finger use. If the researcher observed a finger being used in typing, this will be recorded on the Observation Form (see Appendix). The second indicator is the resting position of the fingers during typing. During typing, the fingers that are available for use may be held in a semi-extended, relaxed position over the keyboard. Those fingers that are not in use may be flexed into the palm to minimize the risk of accidental keystrokes. At the beginning and during the typing trial, the researcher observed the finger positions of each hand, and recorded this information on the observation form.
When typing from a source document, the visual attention of the typist may move between three targets: the source document, the computer screen, and the computer keyboard. Because the computer monitor and document holder were both in the vertical plane, and the keyboard was on the table top in front of the typist, the researcher was able to observe the number of times the subject directed visual attention to the keyboard. This was documented on the observation form. A second indicator of visual dependence was moving the hands away from the keyboard to allow visual inspection of the keyboard labels. These movements were also recorded on the observation form.
The computers used for this study were three laptops running Windows XP. These computers had a minimum Pentium processor operating at 2 GHz, with at least 512 MB of RAM.
To control for differences in keyboard size and feel, all keyboard input was via a Microsoft USB Media keyboard 3000 connected to the laptop via an available USB port.
The computers used in this study had screen sizes varying from 12 to 17 inches. To control for variations in screen size, user input was via an external 17-inch Acer monitor, with a resolution of 1024 x 768, set to mirror the primary screen of the computer.
The user did not have any mouse interaction in the course of this study.
Input timing was via a Polder digital countdown timer.
The computers were set up prior to the arrival of the subjects. The laptop was positioned so that the subject did not have a view of the laptop screen. The 17-inch monitor was connected to the laptop’s video output port, and the resolution set to 1024x768, with screen mirroring activated. (Mirroring shows the same view on the external monitor as the laptop screen, rather than extending the desktop.) This monitor was positioned approximately 20 inches from the leading edge of the table initially, though the subjects were free to move the monitor closer or farther if desired.
The USB keyboard was connected to the laptop and positioned in front to the subject’s screen. Initially, the keyboard was centered in front of the screen, but the subject was free to adjust its position as needed for comfort. A copy stand was placed to the left of the subject’s monitor, with the source text clipped to the stand, but not initially visible to the subject.
Following completion of the informed consent process, the subjects were seated at the computer workstation, facing the 17-inch monitor and keyboard. The subjects were allowed to adjust the position of the chair, the monitor, and the keyboard as needed to provide a comfortable position. When this adjustment had been completed, the researcher instructed, “When I say ‘GO!,’ I’d like you to type this text (revealing the source document) as quickly and as accurately as you can for 10 minutes. After 10 minutes, the timer will signal the end, and I’d like you to stop typing, even if you are in the middle of a word. Do you understand? Are you ready? Go!”
At the word “Go,” the researcher started the countdown timer, and place it so that the display did not face the subject (to avoid distraction).
During the subject’s typing, the researcher recorded the position of the subject’s fingers (flexed or extended) on the data collection form, as well as indicating each finger that was observed to be used during typing. In addition, the researcher recorded a tick each time the subject's hands were removed from the keyboard or eye-gaze directed at the keyboard.
When the timer signaled that 10 minutes had passed, the researcher said “Stop!”, and indicated that the subject was to stop typing. The subject was then thanked for their participation and politely dismissed. The subject’s typing sample was saved to an external USB drive, with a file name including the subject’s ID number (e.g. Subject45.doc).
Each typing sample was evaluated for typing speed and accuracy. The “Word Count” feature of Microsoft Word 2003 was used to provide the number of words typed during the 10-minute trial. While the Word Count feature has some inaccuracy in long documents, it was sufficiently consistent for short documents such as those we obtained. A small number of subjects fully completed the typing sample in less than the 10 minutes allotted. In these cases, the time required to complete the sample was recorded and used in the data analysis.
To evaluate typing accuracy, the subject’s typing was compared with the source document using the Compare Documents feature of Microsoft Word 2003. This feature highlights any blocks of difference between two texts, including misspelled words, missing words or phrases, and extraneous words or phrases. The number of blocks of text difference (excluding the large block of untyped text at the end) was counted to provide a repeatable count of “errors.”
The number of fingers used in typing and degree of visual dependency was evaluated by counting the ticks on the observation sheet for each subject.
The purpose of this study was to discover how well and how people in the community actually type. To discover this, we examined the typing samples in a number of ways, both qualitatively and quantitatively.
In order to identify typing styles, the raw data was analyzed for emergent patterns. The number of fingers observed to be used in typing was counted for each subject and recorded. The frequency of each finger-use pattern was plotted in a histogram (see Figure 6). From this graph, it appeared that true hunt-and-peck typing (using one to two fingers, looking for each letter before typing) was very rare. Only one two-fingered typist was included among our 75 subjects. Of the remaining typists, there appear to be two groups: those using from three to seven fingers to type, and those using from eight to ten fingers, with the frequency of seven fingers forming a division between the two groups. The bimodal distribution of the finger-use data makes it clear that, in addition to the rare hunt-and-peck typist, there are two additional groups of typists based on finger-use.
Next, the data was evaluated for the number of times typists looked at the keyboard. Because it was conceivable that each typist looked at the keyboard a different number of times, we grouped the raw data in blocks of 10. If the subject looked at the keyboard between 0 and 9 times, they were assigned one visual group, from 10 to 19, the next group, and so on.
Again, the data showed that the subjects fell into three groups: those who were highly visually-dependent, looking at the keyboard from 120 to more than 200 times during the 10 minute trial, those who looked from 50 to 99 times, and those who were relatively independent of vision, looking at the keyboard less than 30 times. As with fingers used, there appeared to be natural divisions between these groups, with only three subjects in the range from 30 to 49 glances, and only two in the range from 100 to 120 glances. The membership in these groups was very similar to the groupings for fingers used, though there was some overlap.
Our assumptions about finger position being an indicator of typing style were not borne out by observation. Our subjects tended to hold their fingers in extension whether or not they were used in typing. Similarly, very few of the subjects removed their fingers from the keyboard, and those that did, did not do so in patterns that supported grouping.
Based on these observations, we categorized typists into three fairly distinct groups (see Figure 6 and Figure 7). The first group, true hunt-and-peck typists were distinguished more by their visual dependency than the number of fingers used in typing. We had only a single typist who used just two fingers to type, but 18 subjects (24%) who looked at the keyboard from 120 to over 200 times. From this observation, we concluded that hunt-and-peck typing was more characterized, in modern typists, by the "hunt" phase, where the individual locates the target character, than by the "peck" phase, as our subjects typed with whichever strong finger was nearest the target letter.
The second group, which had also already been characterized, was the touch-typists. These typists were relatively vision independent for typing, looking at the keyboard fewer than 3 times per minute, and used eight or more fingers to type. We observed that a significant portion of touch typists do not use their little fingers when typing. Since the little finger is generally much weaker than the other fingers, this typing behavior was common among those who learned to type on mechanical typewriters, but it was also observed in younger typists, who were unlikely to have ever used mechanical typewriters.
The third group of typists identified in this study was "transitional typists." This group used from three to six fingers when typing, and looked at the keyboard frequently, but not consistently. In the 10-minute typing trial, this group looked at the keyboard from 50 to 100 times. We hypothesize that they have learned the location of a significant portion of the keyboard, and only look to verify their finger positions and the location of lower frequency characters.
While it might be supposed that these groups would reflect typing exposure and age, this was not the case. Neither the number of fingers used to type nor the degree of visual dependency was significantly correlated with age. There were touch typists in all age groups and transitional typists in all age groups. The sole single-digit typist in our sample was near the middle of the age spread of the study.
Age was, however, a predictor of typing speed. When the mean typing speed of each age group is plotted, there is a clear decrease in average typing speed with increasing age. This chart, however, shows an anomaly in this relationship. From the age 60+ group, with an average typing speed of 15 words per minute to the youngest typists (aged 18 to 30 years), there is a steady increase in typing speed of about 10 words per minute per decade. However, this pattern interrupted by the typists aged 30-39, who typed at almost the same speed as the 40-49 year old cohort. One possible explanation for this pattern is that the 30+ typists may be essentially self-taught (typing training was available to this group, but was neither mandatory nor frequently taken), while the 18-30 year old typists had keyboarding training in the classroom (see Figure 8). The average speed of 32 words per minute may represent a plateau of typing speed for self-taught typists, with the continued jump to 47 words per minute representing the effect of classroom instruction. This can be tested in the next decade by determining whether a new plateau of keyboard speed appears, or if input speeds continue to climb. (This assumes that the keyboard will remain the primary input method for the next decade, which is far from certain.)
As was noted earlier, typing styles were not strongly correlated with age. However, typing speed change with typing styles, though not in an expected pattern. Whether grouped by number of fingers used or visual dependency, the transitional typists, on average, typed somewhat more slowly than the hunt-and-peck typists. It was only the true touch-typists who showed increased speed. This may relate to the increased cognitive overhead of partially memorizing the keyboard. However, the typing speed of those who had memorized the keyboard and learned to use more fingers was double that of the transitional and hunt-and-peck typists. We hypothesize that typing speeds increase once it is possible for the typists to develop motor plans for complete words rather than for single letters. Since this is only possible once the full keyboard is memorized, and enough fingers are in use, it is only the touch-typists who can make this leap.
We had originally considered two possible relationships between typing speed and accuracy. One possibility was that, as the typist focuses on getting words on the page, more errors would be made, and there would be an increase in errors with typing speed. An alternative relationship would be that poorer typists (those who type slower), might also be less accurate, so that errors would decrease with typing speed. In fact, neither of these patterns emerged from the data collection and analysis.
In our study, typing speed decreased with age, but the number of errors that occurred in each age group was approximately the same. In each age (and speed) range, we found between 1 and 1.5 errors per minute of typing, regardless of the number of words actually typed. From the available data, we do not have a working hypothesis as to why this should be true.
Due to the limited time and funding for this research, some adaptations to our original plan for data collection were adopted. Our original intent was to have a larger participant size of 100. We also initially planned to utilize two camcorders for observation of fingers used during typing while also recording the looks at keyboard by the participant during the timed session. However, we were unable to obtain the camcorders and supporting equipment. The camcorders would have made analysis of typing style more robust, and perhaps allowed identification of more subtle typing variations. It would have also increased the inter-rater reliability of this study. While we practice the data collection techniques prior to using actual subjects, it was not possible, without recordings, to check the actual recorded data. Also, because our study was the first to assess the community’s typing style, speed, and accuracy we did not have any other studies with which to compare our methodologies and findings.
Prior to this study, only two styles of typing were commonly characterized: hunt-and-peck and touch-typing. Our data showed an additional style of typing, which we are labeling as “transitional typing,” which involves less visual dependence, more fingers, but no increase in typing speed over touch-typing. This group should be studied more fully to determine more fully how they type.
The data from this study indicated that average typing speeds decrease with increasing age at 8 to 10 words per decade. This should be taken into account when setting goals for keyboard-related therapy activities. Because this study was a single-probe cross-sectional design, it is impossible to determine if people’s typing slows as they age or if this is a result of exposure to keyboards and training. Longitudinal studies of typing behavior, examining typing speed through the life-span could show if a given individual maintains early typing speeds or if speeds slow over time.
This study had two distinct purposes that were independent of one another but would both benefit from the research collected during our study. The first was to assist the developers of the CleanKeys keyboard in their sensing algorithms. The second was to assist therapists in understanding how non-disabled individuals type, to assist in formulating reasonable goals for keyboard-related therapies.
We identified three distinct styles of typing, rather than the two that had been previously described. The smallest of these groups was the hunt-and-peck typists, who use relatively few fingers (1 - 3) to type, and who look at the keyboard very frequently when typing. These fingers were the index and middle fingers, which are the strongest of the fingers of the hand. The second group was transitional typists, who are less visually dependent, and use more fingers when typing (4 - 6). In this group, the ring fingers are added those used by touch-typists. These are also relatively strong fingers. The largest group was the touch typists, who are relatively visually independent when typing, and who use most of their fingers when typing. However, we also observed that very few typists used their little fingers when typing, independent of the typing style. Most typing is done with the index, middle, and ring fingers, each of which is relatively strong.
This study showed a correlation between typing speed and age. Where young typists typically type around 50 words per minute, this speed decreases by 8 to 10 words per minute with each decade of aging, regardless of typing style. Even more unexpected, we found that accuracy was strongly related to neither typing style nor typing speed. Instead, we found a fairly consistent rate of errors at 1 to 1.5 errors per minute of typing across all age groups and typing styles.
We also found an unexpected anomaly of typing speed among typists of 30 to 39 years of age. Rather than the predicted typing speed of around 40 words per minute, this group typed at about the same speed, 33 words per minute, as their next older cohort. It is not clear what caused this divergence from the observed relationship between age and typing speed.
This data aids in understanding that errors are not dependent on the typing styles and speeds but are approximately constant rate among different types of typists.
As the modern world becomes more and more dependent on computer- and keyboard-based activities, it becomes increasingly important to understand how, and how well, people use keyboards. These data can help CleanKeys design their multi-touch keyboard surface to adapt to these common typing styles to most most accurately determine the user’s input.
As discussed, the observed typing styles differ both in visual dependence (which may be important in labeling the keys) and the number and strength of the fingers used. While touch typists use more fingers than the other approaches, a minority of typists use the little fingers in typing. This information should prove useful in developing keyboard algorithms for maximum keyboard accuracy.
Results of this study indicate that typing speeds vary with age, and that, when planning how long a keyboard-based activity should take, age is an important factor. This information can be used in setting therapy goals for keyboard-based activities. It is not clear whether the changes in typing speed with age stem from reduced exposure to keyboards in early life or biological changes of aging. Further exploration of this variation is an avenue for future research.
We found no prior research on typical typing accuracy. Our findings did not support either expected model for typing accuracy. We did not find that less-skilled (slower) typists made more errors nor than faster (less attentive) typists made more errors. Instead, we found an approximately constant rate of errors that was independent of typing speed and method. While this provides useful information on how error-tolerant data-entry processes should be, the mechanism of errors and this constancy of rate are not well understood, and should be further investigated.
The assumptions about typing prior to this study included the idea that, while touch typists use all fingers to type, hunt-and-peck typists use just one or two. This study has shown that a large portion of typists uses several fingers to type, and that even touch-typists often do not use their little fingers in typing. This information has particular implication for the design of multi-touch keyboards where location of a keypress may be refined by the relative force of the impact. The results of this study should assist CleanKeys in designing optimal target-identification algorithms.
Aizen, E., Ljubuncic, Z., Ljubuncic, P., Aizen, I., & Potasman, I. (2006). Risk factors for methicillin-resistant staphylococcus aureus colonization in a geriatric rehabilitation hospital. Journals of Gerontology: Series A, 62, 1152-1156.
Bures, S., Fishbain, J. T., Uyehara, C. F. T., Parker, J. M., & Berg, B. W. (2000). Computer keyboards and faucet handles as reservoirs of nosocomial pathogens in the intensive care unit. American Journal of Infection Control, 28(6), 465-471.
Dumke, N. M. (2005). Hospital computer keyboards implicated in transmissions of nosocomial infections. Journal of the National Medical Association, 97(8), 1204.
Heintzelman, M., & Pfeiffer, P. (1997). Machines, statues, and people: strategies for promoting RSI awareness in computing curricula. ACM SIGCSE Bulletin, 29(1), 296 - 300.
Kaufman, H., Wheeler, T., Dotray, P., & Keeling, W. (2001). Remote sensing tools can add precision to your farming operation. In T. A. M. University (Ed.). Lubbock, TX: Texas A&M.
Kramer, A., Schwebke, I., & Kampf, G. (2006). How long do nosocomial pathogens persist on inanimate surfaces? A systematic review. BMC Infectious Diseases, 6(130).
Kraut, R. E. (1996). Media use in a global corporation: Electronic mail and organizational knowledge. In S. Kiesler (Ed.), Research milestones on the information highway. Mahwah, NJ: Erlbaum.
Kunde, B. (1986). A brief history of word processing (Through 1986). Retrieved from http://www.stanford.edu/~bkunde/fb-press/articles/wdprhist.html
Marsden, R. (2009). A solid-surfaced infection control computer keyboard. Edmonton, Alberta: Cleankeys, LTD.
Mayo Clinic Staff. (2010). MRSA infection. Retrieved from http://www.mayoclinic.com/health/mrsa/DS00735
NA. (2010). Point of sale case study. Retrieved from http://www.waspbarcode.com/point_of_sale/case_sunglasses.asp
Nye, J. S., & Owens, W. A. (1996). America's information edge. Foreign Affairs, March/April: 23 (20-36).Okamoto, J. (1998). Repetitive Strain Injury: Causes, Treatment, and Prevention. Proceedings of the 12th Systems Administration Conference (LISA '98). Retrieved from http://static.usenix.org/event/lisa98/invited_talks/okamoto_html/tsld001.htm.
Petroudi, D. (2009). Nosocomial infections and staff hygiene. [Technical Note]. Journal of Infection in Developing Countries, 3(2), 152-156.
Rempel, D., Serina, E., & Klinenberg, E. (1997). The effect of keyboard keyswitch make force on applied force and finger flexor muscle activity. Ergonomics, 40, 800-808.
The Artchive. (ND). Cave paintings (40,000 - 10,000 BC). Retrieved Nov. 3, 2010, 2010, from http://www.artchive.com/artchive/C/cave.html
Tietjen, L., Bossemeyer, D., & McIntosh, N. (2003). Preventing nosocomial infections. Infection Prevention Guidelines for Healthcare Facilities with Limited Resources. Baltimore, MD: Johns Hopkins Foundation. Retrieved from http://www.reproline.jhu.edu/english/4morerh/4ip/IP_manual/ipmanual.htm.
Tillotson, D. (2006). The Practice of writing in the middle ages. Retrieved from http://medievalwriting.50megs.com/literacy/writing.htm
Weinstein, R. A. (1998). Nosocomial infection update. Emerging Infectious Diseases, 4(3), 416-420.