» home

» overview

» products

» studies

» testimonials

» the company

» FAQ

The Results of a Representative Group of DataHand Productivity Studies

Showing Average Thirty-Day
Proficiency Improvement of DataHand Workers
Compared to Their Prior Flat Keyboard Performance

Abstract

Thirty-three sets of worker data gathered at five different companies demonstrate an average DataHand performance proficiency 116% (plus or minus 9%) of prior flat keyboard speeds (in conformance with a 95% confidence interval). The minimum average proficiency expected within thirty days of DataHand work is 107% (seven percent above prior flat keyboard speed). The maximum projected thirty-day average speed is 125% (twenty-five percent above prior flat keyboard speed). The thirty-day time period allowed workers the time needed to master the DataHand touch and operating method as well as to achieve the proficiencies recorded.

Introduction

Studies designed to record and analyze the speed and efficiency of DataHand keyboard users began as soon as the first preliminary models of the DataHand keyboard were manufactured. Improvements to enhance DataHand performance have since been made, especially to the most recent Professional II DataHand model, but the efficiency results have been impressive since the beginning of the DataHand experience.

Learning time of approximately thirty days is needed before typical workers begin to demonstrate the level of achievement they can expect. Because the DataHand keyboard is greatly different from any other keyboard, time is required to learn the lighter touch and the less demanding, more gentle movement pattern.

The first four workers who were tested on the first model of the DataHand keyboard worked part-time, an hour or two a day, and in some cases only a couple of days a week for up to approximately 105 hours. This group of average entry-level keyboard workers showed performance improvement ranging from 151.9% to 177.7% using the DataHand keyboard. Although the study was designed to learn the number of hours required by average workers to learn the DataHand system of keyboard work (equal or begin to exceed their flat keyboard performance,) it allowed learning and data tracking to proceed long enough to collect preliminary data on proficiency improvement.

Since then all but one of the DataHand productivity studies have been conducted by companies evaluating the capabilities of the DataHand ergonomic keyboard prior to placing an order. The one exception was conducted by a consulting firm involved in the creation of DataHand training materials. The results of this study are cited in the brief summary on the lead page of the Studies Section, but the data are not included in the summary results below. This study, called the Evers-Koon Group Study, gathered data from workers using a DataHand ten-key model in typical one-handed, banking-type operations. This summary report includes only data from workers using the full two-handed DataHand keyboard, so the Evers-Koons data is excluded.

Some of the most significant productivity results have been logged by DataHand users at U.S. Postal facilities, but most of these data also come from workers doing their keyboard work with one hand. Data from workers at one mail handling center in Phoenix, Arizona have been analyzed by Roberto Fernandez of Stanford University. The Fernandez/Stanford study compares the performance of workers using a standard ten-key device with the performance of workers using a single-handed DataHand ten-key model specially engineered to operate the U.S.P.S. mail sorting equipment. The Fernandez report is posted under the heading of Speed and Fatigue Study.

(Click here to view the Fernandez/Stanford Speed and Fatigue Study).

A two-handed U.S.P.S. study conducted with two workers at a facility in Pittsburgh is included in the summary. The postal service has many workers doing two-handed mail sorting in addition to those working with a one-handed device.

The other studies included in the summary are:

• Card Catalogue Company study involving four workers entering words, sentences, and numbers on typical library cards.
• Wachovia Bank study involving nine workers performing three different banking tasks. Separate performance data was recorded for each of the three tasks.
• Sun Trust Bank study designed to be similar to the Wachovia study, but involving two different types of bank accounting work both performed by five workers. As with the Wachovia study, separate data was recorded for all workers performing both tasks.
  

Methodology

Each company where the productivity studies were performed gathered their own data according to their standard methods of worker performance evaluation. Some companies count keystrokes per second—as is typical in the banking industry; others count units of work per hour. The Postal Service counts the number of letters handled in an hour. Some others count words per minute as well as completed units of work per hour.

In all cases the same method of measurement was used for both the flat keyboard work and the DataHand work. The companies furnished their raw data to DataHand Systems, where the analysis was performed using standard statistical methods. Questions about the analysis can be addressed to Dennis Monroe, DataHand Systems's Vice President for Engineering and Manufacturing.

Combined Results of the Five Studies

Five study sites and thirty-three sets of worker data are included in this analysis. Twenty-five of the thirty-three subject data sets (75%) achieved results higher than their flat keyboard speed in the thirty day (160 hour) learning period. Fifteen of the thirty-three (45%) achieved results higher than 116% of their prior flat keyboard speed, and eight of the group (25%) achieved results less than 100% in the allotted thirty-day time period. With 95% confidence, the average proficiency of the entire group is 116% plus or minus 9%. The minimum average proficiency was 107% or seven percent greater than prior flat keyboard speed. The maximum average proficiency was 125% or twenty-five percent above the flat keyboard baseline.

These data enable a projection of the likely achievement of DataHand workers within thirty days of work. Approximately 25% of a typical worker population would not be expected to achieve their flat keyboard speed within thirty days. This portion of a typical worker population would likely need additional time to achieve their flat keyboard speed. Each person has a likely learning speed, but other factors affecting the learning process are the speed the worker is able to work on the flat keyboard and the worker's level of manual dexterity.

If the five highest performers and the five lowest performers are excluded from the analysis, the average productivity is reduced by one percentage point from 116% to 115%. After this step is taken, conformance with 95% confidence is retained by extending the range of the average by plus or minus 6%. This exercise in cautious affirmation of the average raises the minimum average to 109% but also lowers the maximum anticipatable average to 121% from 125%.

Discussion

Given more time, the performance of the slower learners is typically improved, but no systematic record of the data for a longer time period and a larger number of workers has yet been recorded. Daily or monthly data for six months of work would be valuable when more extended studies become possible. The time required by different workers to learn the DataHand skill conforms to a bell curve pattern of distribution with the peak of the curve falling within the days just prior to the end of the thirty day test period. The fastest learners have been found to need as little as a few hours to master DataHand operation while the slowest learners may need as much as sixty days to reach a relative peak of performance. This information defines the outer edges of the bell curve, but more data is needed to complete the picture and affirm the preliminary analysis.

DataHand workers have been found to reach a relative plateau from which they continue to consolidate their skill over a period of additional weeks or months. During this extended time, error rates decline as the DataHand touch becomes increasingly automatic or tactily habitual. Observation of DataHand learning shows differentiated tactile reinforcement stemming from the uniquely different feel of all the DataHand keys. This feel coupled with the five different movement directions needed to activate the DataHand keys tactily reinforces both DataHand learning and DataHand operation.

Another factor affecting speed of learning and the achievement of productivity targets in the case of injured workers is the amount of pain residual from previous flat keyboard work. In many cases learning cannot advance quickly and DataHand productivity cannot accelerate until pain begins to ameliorate. Sometimes, several weeks or even months of DataHand work are required before the effects of the reduced stress experienced by DataHand workers begins to lower residual levels of pain. These factors were not evaluated in the summary of the five productivity studies. In general, individual experiences vary widely, just as do the combinations of injury symtomology.

The summary compares the performance numbers without attempting deeper examination of the specific details of each worker's motivation, normal pattern of learning, learning speed, state of prior health, or subsequent DataHand work experience. Data was accepted from the companies where the tests were done without the provision of any medical history or other personal information. The assumption is made that all data fits within broad averages, but the basis for making this assumption was not documented.

The analysis fails to adequately explore one important observation about worker productivity on the DataHand keyboard: the greatest margin of performance improvement seems to be captured by average workers, who are not particularly fast on the flat keyboard. If this hypothesis proves correct upon more detailed analysis, the fastest flat keyboard workers may not be able to improve their speed by using the DataHand keyboard as much as slower workers can. Accordingly, the margins of improvement at the high end of the flat keyboard performance scale may be less than they are at the low end of the scale.

Further study of high speed workers is needed, but if margins of improvement are higher for slower flat keyboard workers, this suggests the economic value of converting slower workers to the DataHand system first. If productivity were the only consideration, slower flat keyboard workers would enable the fastest payoff on the DataHand investment. On the other hand, faster flat keyboard workers may experience higher risks of musculoskeletal injury. If this is so, the injury risk may equalize any productivity-based tendency to favor slower workers in prioritizing workers for DataHand training.

The margin of performance difference is not the only reason to prefer working on the DataHand keyboard. Some extremely fast typists prefer to work on the DataHand keyboard just because it is less stressful, less fatiguing, and more comfortable. They may not be able to exceed their flat keyboard performance in the short bursts typically used for typing speed tests, but over long hours of work, they feel they are able to accomplish more with less stress, more comfort, and less risk of injury. This finding shows the DataHand keyboard to be more important as a marathon typing device which shows its value over long hours of work. Intensely athletic high speed typing for short bursts of time is much less valuable than high performance with less fatigue and less risk over full workdays and many continuous months of work.

The DataHand keyboard is clearly most valuable for workers who typically do long hours of keyboard work each day. It could be less important for people who work in short bursts and do not mind the intensely stressful, athletic requirement associated with flat keyboard work. If the work routine of this latter group allows frequent breaks in the work flow and a low amount of total keyboard work each day, the workers in the group may be able continue with the old keyboard paradigm without becoming injured or suffering unmanageable discomfort. If productivity and comfort are not particularly important to the workers or their managers, working with the method already known may be preferred. Analysis is needed to weigh long-term economic benefit against short-term cost. Neither workers nor company management readily engage the required analytical process to reach the optimal decision.

The key design features contributing to the DataHand productivity advantage can be summarized as follows:

• Greatly reduced finger travel distance among the keys lessens total work and improves efficiency of finger movement on the DataHand keyboard.
• Elimination of hand movement increases speed, reduces stress, and lowers fatigue.
• Lower keyforces reduce both work requirement and fatigue.
• Hand support reduces fatigue and increases the sustainability of worker speed.
• Convenient, efficient location of the Backspace/Delete key enables immediate correction of errors without time consuming hand movement needed to reach the Delete key. Similar easy access to Return, Tab, Control (Command), Alt (Option) also speed the use of these keys without dependency on hand movements and long finger reaches.
• Easy access to numbers and symbols without long reaches and hand movement enhances speed and lowers fatigue.
• Performance of mouse work without hand movement saves time while also reducing musculoskeletal stress and fatigue.
• Easy visibility of the key location display enables infrequently used keys to be identified and utilized without removing the fingers from their normal position. Unlike the system of key identification used on the flat keyboard, the fingers do not cover the key display on the DataHand system.
• Custom engineering analysis and programming to match particular data entry requirements. Customization to meet specialized needs often results in improved productivity.
• DataHand adjustability enables the keys to be located in the optimal, most convenient place for each user. When the tool is made to fit the user, the user is not contorted and stressed to fit the tool. This is a primary design intention of the DataHand ergonomic keyboard.

Conclusion

Because of the lower fatigue factor observed in the case of DataHand keyboard use, the efficiency of workers does not deteriorate during the day as it does with flat keyboard workers. Fatigue reduction is a major factor in accelerating the rate of DataHand investment pay-off but it is not the only factor. Proficiency can result from both the reduction of fatigue and from the utilization of the other proficiency enhancing improvements incumbent in the DataHand design.

The DataHand productivity advantage results from design which addresses the limitations long identified with the traditional flat keyboard, but nothing proves the case so well as personal experience. The DataHand productivity benefits need more than just a few minutes to become persuasive. The DataHand ergonomic keyboard is sold with a thirty-day trial period to enable DataHand users an extended trial. To fully prove the DataHand benefits within the thirty-day period, purchasers should begin DataHand work immediately upon receipt of their new keyboard. The more time spent operating the DataHand keyboard the more clear the DataHand difference and the productivity advantage becomes. This result is documented in the comments of many DataHand workers.

DataHand Systems encourages all purchasers to make good use of the available trial time, so they can make sure their own work needs and style of keyboard work is well served by the DataHand concept. The company recognizes the first few hours of DataHand work can be daunting, even humbling, but as learning proceeds, the benefits become increasingly clear. In the beginning, errors are frequent because the lighter DataHand touch is so radically different from the touch workers know from previous keyboard experience. The DataHand benefit has been demonstrated by a wide range of worker experience. Reports of these findings are available elsewhere on this website.

Acknowledgments

DataHand Systems, Inc. thanks each of the companies which provided the data used to perform the combined analysis. Corporate performance data were provided by the United States Postal Service in Pennsylvania, Wachovia Bank in North Carolina, Sun Trust Service Corporation in Georgia (Sun Trust's DataHand keyboard operations have subsequently moved to Florida), and The Card Catalogue Company in Kansas. The fifth study cited in the summary data was performed at DataHand Systems, Inc. Also appreciated is work performed by Stanford University and the Evers-Koon Group.

DataHand Ergonomic Keyboards: Feel the Difference

©2005 DataHand^® Systems, Inc.
3044 North 33rd Avenue
Phoenix, Arizona 85017-5247
(602) 233-6000; (800) 875-7171; FAX: (602) 233-3434
datahand@datahand.com