Researchers Anna Pereira, David L. Lee, Harini Sadeeshkumar, Charles Laroche, Dan Odell, and David Rempel published the results of a study they conducted to investigate the effect of key spacing on typing speed, error, usability, forearm muscle activity and wrist posture. They identify this as Part 1, with Part 2 to be published at a later date. The study focused on conventional mechanical keyboard design, not software generated on-screen keypads such as those seen on mobile computing devices. This article is a summary of their research and findings.

Background

Pereira et al indicate that conventional keyboard key spacing may be influenced more by industry practice and earlier research from the 1970's than it is by ergonomics issues like typing speed, biomechanics, error rates and usability. They note that International Organization for Standardization (ISO) and American National Standards Institute and Human Factors and Ergonomics Society (ANSI/HFES) each recommend that the horizontal and vertical center-to-center distances (when viewing a keyboard from above) each be 19 mm +/- 1 mm, though not all keyboard designs abide by these standards.

For example, as laptop/notebook computers have become smaller, so have some keyboard designs, which create some advantages, for some people and organizations, including:

• lighter and more portable designs;
• reduced manufacturer cost;
• better usability for smaller hand sizes;
• better usability for users with narrower shoulder widths; and
• reduced reach distances to a mouse situated nearby the keyboard.

For the most part, though, most keyboard key spacing, even for laptops/notebooks, abides by the conventional 19 mm standard.

Pereira et al also review the findings of the relatively few other studies that have looked at performance related to key spacing, including these previous findings:

• Japanese research: for Japanese with small fingers, there were no performance decreases when comparing a range of spacing differences from 15 mm to 19.7 mm; however, performance did decrease for those with large fingers for key spacings of 16 mm or less (the researchers caution against applying these same results to USA or other country populations because of differences in hand anthropometry, and also note study design issues that could further confound the results);
• a different study found greater input time and error rates with numerical keypads when spacing increased from 19 mm to 21 mm, though this study also had design problems;
• a 1972 literature review that looked at keyboard design parameters of that time concluded that 'it is due to design conventions rather than empirical data . . . that the typical spacing between key centers on these keyboards is 18.1 mm';
• a 1987 study found that when typing speed, error rate and user preference were considered together, a 19 mm spacing was best (when compared to 14.3 mm, 16.6 mm and 21.4 mm spacings), although that study, too, had experimental design problems that bring validity into question.

None of the above studies considered key spacing effects on biomechanical or physiological measures.

Recognizing that people with smaller fingers would likely adapt better to smaller spacings, Pereira et al focused on people with larger fingers, and in this Part 1 article they primarily examine horizontal key spacing (vertical spacing will be reported in their future Part 2 article).

Methods

The following discussion of study methods is significantly abbreviated, and interested readers are directed to the complete article, cited below.

• this is a laboratory study, not a field use study;
• 37 males between 18 and 65 yrs old, with the ability to touch type at least 30 words per minute, participated in the study;
• participant population had a mean right middle finger length of 8.74 cm ± 0.30 cm (ranging from 7.65 to 9.47 cm; 8.61 %tile to 97.5 %tile); right mean middle finger width of 2.22 cm ± 0.15 cm (ranging from 1.91 to 2.51 cm; 1.25 %tile to 94.3 %tile); and a mean right hand length of 11.5 cm ± 0.53 cm (range 10.6 to 12.7 cm)
• an Ergodex DX1 customizable keyboard system was used to create QWERTY layouts for the key spacing test conditions;
• 5 keyboard configurations differing only in horizontal and vertical key spacing were tested;
• the size of the tops of all keys were controlled at 14.7 mm horizontal and 13.7 mm vertical;
• key activation forces were controlled to be between 63 and 77 grams;
• participants were provided with warm-up periods and rest breaks prior to beginning the experiment;
• typing tasks were presented on screen, consisting of 8th-9th grade reading level;
• passages included capitalization requiring a shift key, but did not include numbers of punctuation;
• using an adjustable workstation (chair and worksurface), the researchers initially set the system for each participant to support feet comfortably on the floor, worksurface height at seated elbow height, and keyboard placed 18 cm from the worksurface edge in front of the participant;
• participants were familiarized with the system adjustments and encouraged to adjust the workstation to the most comfortable postion during the practice session, but were not allowed to make adjustments once the experiment began;
• the 5 keyboard conditions and 15 typing passages were presented randomly to each participant in a manner creating 3 trials per keyboard condition;
• software calculated gross typing speed and percentage error averaged over the 3 trials;
• the ISO keyboard questionnaire was administered after each keyboard configuration was tested, collecting measures for usability and fatigue, including "force required to activate the keys, keying rhythm, fatigue in hands or wrists, fatigue in arms, fatigue in shoulders, posture required for keying, and overall use";
• forearm electromyography (EMG) was used to collect wrist ulnar and radial deviation muscle activity;
• wrist extension/flexion and ulnar/radial deviation posture was measured using electrogoniometers;

Results

Again, interested readers are directed to Pereira et al's complete article for full details, but their results include:

• gross typing speed was significantly slower for the 16 mm horizontal by 19 mm vertical spacing keyboard configuration when compared to all other spacing configurations;
• percentage error was significantly higher than all others for that same 16 mm by 19 mm configuration;
• there were no significant differences in speed, error rates or subjective usability ratings between the 17 mm, 18 mm or 19 mm horizontal key spacings;
• no learning effect or accumulated fatigue or usability effects were detected;
• Pereira et al attribute decreased performance with the 16 mm horizontal spacing to biomechanical factors, "such as fingertip size or inadequate motor control precision, are interfering with productivity measurements," and further review the data to suggest that "key spacing may be limited by finger width, not finger length";
• the effect of key spacing on wrist posture and forearm muscle activity "were minimal," although "There was a trend for muscle activity to increase in the left and decrease in the right forearm with decreasing horizontal key spacing, but the differences were small, only 2% to 5%";
• however, when taken in light of other ergonomics research, from keyboarding activities to industrial meatpacking work, Pereira et al state that "... although the observed differences are small, these small differences may be important for tasks performed for many hours a day"
• as key spacing decreased, there were significant differences in left wrist posture, with ulnar deviation decreasing and extension increasing with decreasing key spacing; A similar pattern for the right wrist, however, was not statistically significant.

What This Might Mean to Ergonomists

This is the kind of research that is very important at the consumer product design level, but it may seem esoteric, or it's value may not seem readily apparent in the everyday practice of ergonomics. On the design level, it's hard for me to believe, but not at all surprising, that for the past 30 or more years something as important as key spacing on a keyboard has been driven by manufacturer convenience and unfounded conventional wisdom. I applaud Pereira et al for stepping up to finally apply a human centered research and design approach -- ergonomics -- to something as ubiquitous as the keyboard, and look forward to their Part 2 publication. This kind of research will be valuable in the formulation of future keyboard design standards, and those standards will finally be based on human/user performance requirements instead of technical, manufacturing and market convenience.

On an everyday application level, however, you may have heard the phrase "I fat-fingered it," meaning the person missed the button, target, or in this case the proper keyboard key and made an error, placing the blame on user error rather than poor design. This study provides scientific evidence for the "fat-finger" effect.

This study should also cause us to reflect on the "one-size-fits-all" keyboard design (or any other one-size-fits-all design). Such designs rarely, if ever, fit all well, and often, in fact, fit few well. As Pereira et al point out in their review of their and others' research, people with smaller fingers may benefit from less key spacing, while those with larger fingers may benefit from greater spacing. Will we see a trend toward different keyboard designs for different sized people? Is this something that ergonomists should begin advocating, or is the on-size-fits-all approach "good enough" when it comes to keyboard design?

In the mean time, for "standard" keyboard design, Pereira et al offer this advice:

On the basis of these findings, keyboard designers are encouraged to consider designing keyboards with horizontal and vertical key spacing of 17 mm or 18 mm to gain the benefits of smaller keyboards (e.g., smaller and lighter laptops, reduced cost to manufacture, better usability for smaller users, and reduced reach to the computer mouse) while still accommodating the needs of typists with large fingers.

Reference

Anna Pereira, David L. Lee, Harini Sadeeshkumar, Charles Laroche, Dan Odell, and David Rempel, (2012). The Effect of Keyboard Key Spacing on Typing Speed, Error, Usability, and Biomechanics, Part 1. Human Factors, published online before print, December 6, 2012, doi:10.1177/0018720812465005.

At the time of this writing, the complete article was available to Human Factors or Sage Publications subscribers at: http://hfs.sagepub.com/content/early/2012/12/04/0018720812465005.full

Originally published on ergoweb.com November 5, 2013.