This reading list is intended to help you become acquainted with the body of knowledge on mistake-proofing as quickly as possible. It will allow you to bypass many of the references that may not be readily available.
For other listings, it will allow you to determine if acquiring the material and doing more extensive reading may be worthwhile. If you want to buy these books, click on the book cover. This page is part of Amazon.com’s Associates Program. When you click on a book on this page, it will take you to an amazon.com page. Putting the book in your “shopping cart” from that page will provide me with revenues to support the cost of www.mistakeproofing.com and give you the secure transactions and delivery reliability of Amazon.com.
The literature on mistake-proofing is continually expanding, making a comprehensive literature review a moving target. Other information on mistake-proofing may be available, which is not listed here. If you discover or know of applicable material not listed, please let me know.
- Medical literature
- Engineering literature
- Mistake-proofing and poka-yoke literature
- Psychological literature
- Software & human-computer interaction literature
Poka yokay systems to ensure zero defect quality manufacturing
Bandyopadhyay, J. K. 1993. International Journal of Management 10 (1): 29-33.
Bandyopadhyay provides a very basic treatment of poka-yoke. Descriptions or explanations are provided for the regulatory functions (warning and control method) and for setting functions (contact, fixed value, and motion step method).
In-Process Quality Control for Manufacturing
Barkman, W.E. 1989.New York: Marcel Dekker. (Preface and chapter 3 are of particular interest.)
This book discusses the use of high-tech electronic approaches to in-process control. Typically, one thinks of mistake-proofing as a more low-tech approach; however, the basic theory and thinking are closely related. Barkman uses the term “deterministic manufacturing,” where determinism means “a doctrine that acts of will, occurrences of nature, or social or psychological phenomena are determined by antecedent causes… Belief in a ’cause-and-effect’ environment leads to the idea that one means of attaining control of a manufacturing process is to monitor and/or control the key parameters that are most significant in establishing the quality of the finished product.” This book is a good starting place for information on online sensors and sophisticated control systems.
Using Poka Yoke (Mistake Proofing Devices) to Ensure Quality
Bayer, P.C. 1994. IEEE 9th Applied Power Electronics Conference Proceedings 1:201-204.
This paper gives a basic description of mistake-proofing and discusses specific implementation at AT&T Power Systems. Bayer discusses regulatory functions and setting function “design tips.” Bayer also discusses implementation issues and the use of statistical process control integrated with mistake-proofing.
The Basics of Mistake-proofing
Beauregard, M.R., R.J. Mikulak and R.E. McDermott. 1997. New York: Quality Resources.
This book is what its title says it is: basic. As evidence of how basic it is, the first three chapters appear on just five pages. The authors do make slight additions to the practice of mistake-proofing by identifying the sensory alert regulatory function and by adding “making it easy to do it right” to the setting functions described by Shingo.
Beauregard et al. identify four mistake-proofing methods in order of preference: control, shutdown, warning, and sensory alert. Control methods eliminate the opportunity for errors to occur. Shutdown methods stop the process when an error occurs. Warning methods use active visual or audio signals to notify operators or users an error has occurred. Sensory alert methods make errors obvious but depend on the operator or user to actually perform the check. Their addition to setting functions, making it easy to do it right, encourages the use of color-coding, shapes, symbols etc., to help the operator or user sense when an error is occurring. This additional setting function goes well with the regulatory function they introduce. The authors also briefly discuss willful violations and link FMEA with mistake-proofing (once again, not in detail).
The trend: 100 percent quality verification
Bodine, W.E. 1993. Production (June): 54-55.
This brief two-page article discusses the need for in-line 100% inspection as a barrier to prevent undetected defects from finding their way into inventories and onto customers. Bodine discusses the growing use of programmable logic controllers (PLCs) to perform these inspections without slowing the process.
Despite fuzzy logic and neural networks, operator control is still a must
Bose, Rana. 1995. CMA 69: 7.
Bose accurately points out that after a decade of trying to engage the worker in thinking about and improving processes, we should not replace worker intelligence with artificial intelligence. Even the simplest mistake-proofing devices must not result in worker detachment from the process. Rather, workers need to be involved in designing and enhancing the process and its control mechanisms, one of which is mistake-proofing.
Set Phasers on Stun and Other True Tales of Design, Technology, and Human Error
Casey, Steven. 1993. Santa Barbara, CA: Aegean.T
This book provides a series of case studies where design, technology, and human error coincide to create catastrophes. An example is the case of a Salomon Brothers trader who single-handedly caused a 15-point drop in the Dow Jones Industrial Average in the last 5 minutes of trading due to misunderstanding the units of a “sell” order. He mistakenly sold 11 million shares ($500,000,000) of stock instead of 11 million dollars of stock. The Wall Street computers did the rest.
Make your service fail-safe
Chase, R. B., and D. M. Stewart. 1994. Sloan Management Review (Spring): 35-44.
This is the first article on mistake-proofing services. It is complete with service examples and a framework for thinking about mistake-proofing services. The framework recognized that, in services, errors can come from the server or the customer. They create task, treatment, and tangible server poka-yokes. Customer poka-yokes include preparation, encounter, and resolution. Chase and Stewart then discuss the steps of mistake-proofing, which they superimpose on a process map. They conclude with an automotive service example. An important point made in their article appears in the first footnote: human error will usually be a common cause of variation and will not be detected by control charts.
Mistake-proofing: Designing errors out
Chase, R.B., and D. M. Stewart. 1995. Portland, Oregon: Productivity Press.
Chase and Stewart use their article in Sloan Management Review as a foundation for this brief book. They include all of the concepts from the article in the book. They also include information on the psychology of human error. They lay out Rasmussen’s skill-, rule-, and knowledge-based (SRK) framework of cognitive control. They also broaden the regulatory function categories by using the six cues of Norman and Lewis: warning, gagging, nonresponse, self-correct, let’s talk about it, teach me. Chase and Stewart also broaden Shingo’s setting functions. They classified setting functions as physical, sequencing, grouping, and counting, which corresponds roughly to Shingo’s contact, motion-step, and fixed-value methods. Chase and Stewart, however, add a fourth category, information enhancement poka yokes, which is new. These poka yokes are designed to ensure that information required in the process is available at the correct time and place and that it stands out against a noisy background.
It is both more readable and broader in application than Shingo (but does not have a catalog of examples like Shingo). Regrettably, the Chase and Stewart book was only printed for two years. Used copies may be available.
Or download a free PDF (shared with permission from the authors)
An Economic Analysis of Inspection Costs for the Mistake-Proofing of Binomial Attributes
Downs, B.T. and Grout, J.R. Forthcoming in the Journal of Quality Technology.
In this research, processes that have attributes as primary quality characteristic are considered. Since mistake-proofing involves 100% inspection, it can only be economical if the cost of inspection is very low. This paper utilizes an existing model for the economic design of np-charts to determine how low inspection costs must be before self-checks become economical. An existing model for checking proper operating conditions will be used to find how low the cost of source inspection must be in order for it to be economical.
Poka-Yoke Designs Make Assemblies Mistakeproof
Dvorak, Paul. 1998. Machine Design, (March 10): 181-184.
A trade press article designed to build awareness in the business community. A readable article that could function as an executive summary in introducing mistake-proofing to design engineers and managers.
Guidelines for Preventing Human Error in Process Safety
Embrey, D. American Institute of Chemical Engineers. New York, NY.
Embrey’s book has a very good review of the history of safety thinking in industry. He argues persuasively that more training is not an effective response to incidents caused by experienced workers. He also provides a comprehensive set of hazard analysis tools.
Impact of a Poka-Yoke Device on Job Performance of individuals with Cognitive Impairments
Erlandson, R.F., Noblett, M.J. and Phelps, J.A. 1998. IEEE Transactions on Rehabilitation Engineering 6(3): 269-276
This article reports on the use of a poka-yoke device for individuals with cognitive disabilities. The device allowed more severely impaired workers to be able to perform tasks correctly that they were previously unable to perform. Productivity increased by 80%. The error rate dropped from 52% to less than approximately 1%. Morale and worker attitudes also improved. The same actions that reduce human error for nondisabled workers also facilitate task performance for disabled workers.
Human Error: Designing for Error in Medical Information Systems
Felciano, Ramon M. 1995. Stanford University Web Page.
Felciano defines human error as “an inappropriate action, or intent to act, given a goal and the context in which one is trying to reach that goal.” He then describes the Rasmussen’s SRK framework. He then discusses “slips” and “mistakes.” Slips are actions where the intent is correct, but the execution goes wrong. Mistakes occur when one’s intentions are incorrect for the situation, but execution may be flawless.
Felciano provides a human error case study involving the Therac-25, where a patient died from receiving 125 times the usual radiation exposure called for in the treatment of a tumor. The case demonstrates a “latent” error in the system. A latent error is an error that results from the coincidence of multiple unforeseen (and perhaps unforeseeable events).
Attitudes toward error in medicine and aviation are compared in a similar manner to Leape (see below). Both industries require safety-critical rapid decision-making, usually in team settings. The approach of the two industries is quite different. Aviation has done extensive work on human error and how to reduce it. They tend to analyze how to keep the problem from recurring. In medicine, doctors are ultimately responsible for patient care. Errors are resolved by assigning blame and malpractice litigation.
Human Error in Medicine Bibliography
Felciano, Ramon M. 1995. Stanford University Web Page.
Felciano identifies 26 sources of human error and provides a short annotation of most sources. Some of his sources have not yet been incorporated into this bibliography.
Fail-safing and Measurement Control Charts
Grout, J.R. and Downs, B.T. 1998. Quality Management Journal 5(2): 67-75.
This paper analyzes the use of fail-safing devices (or poka-yoke devices) for controlling processes that are managed using measurement control charts. Grout & Downs show that inspecting production for defects as work is completed (Shingo’s self-checks) does not result in lower process variance. This outcome results from using specification limits to control the process instead of using control limits. If go/no-go inspections using control limits are used instead of specification limits for self-checks, information is lost as measurements are converted into binomial data. The ability to estimate the variance is reduced. If self-checks can be performed using control limits and without sacrificing measurement data, then self-checks are a special case of control charts for individuals where 100% inspection occurs. Using Source Inspection to inspect conditions for high-quality production is shown to involve a binary decision, therefore, source inspections using binomial data do not result in lost information. Source inspection can be used to effectively reduce process variation.
Mistake-Proofing Production
Grout, J.R. 1997. Production and Inventory Management Journal 38(3):33-37.
The basics of mistake-proofing are reviewed. Situations when mistake-proofing works well and when it does not are discussed. Several examples are provided.
Where Mistake-Proofing Works Well:
- where manual operations where worker vigilance is needed
- where mispositioning can occur
- where adjustment is required
- where teams need common-sense tools and not another buzzword
- where SPC is difficult to apply or apparently ineffective
- where attributes, not measurements, are important
- where training costs and employee turnover are high
- where mixed model production occurs
- where customers make mistakes and blame the service provider
- where special causes can reoccur
- where external failure costs dramatically exceed internal failure costs
Where Mistake-Proofing does NOT work well:
- where destructive tests are used
- where the production rate is very fast
- where shifts occur more rapidly than they can be responded to
- where control charts are used effectively (for successive and self-checks only)
Mistake-Proofing: Process Improvement through Innovative Inspection Techniques
Grout, J.R. 1998. The Quality Yearbook, 1998 Edition: 405-414.
This paper, is one article in a very large book. The logic of the five inspection techniques is presented as flow charts. Error opportunity, real-time completion checks, and self-correcting process inspections are added to the traditional source inspection and self-check.
Make No Mistake
Henricks, Mark, 1996. Entrepreneur (October): 86-89.
A popular press article designed to build awareness in the business community. A very short, readable article that could function as an executive summary in introducing mistake-proofing to management. The last quote by Grout should read “…average net savings of around $2500 apiece…” not average cost.
The role of variation, mistakes, and complexity in producing nonconformities
Hinckley, C.M. and Barkan, P. 1995. Journal of Quality Technology 27(3):242-249.
In this article, Martin Hinckley and the late Philip Barkan delineate some very clear and novel thinking in quality control. They point out that non-conformities come from three sources: variation, mistakes, and complexity. They identify the tools that correspond to each of these sources. Variation is managed with SPC and statistics. Mistakes are addressed using poka-yoke. Complexity is controlled during design. A major contribution of the paper is their linking of design for manufacturability with complexity and complexity with conformance quality. They provide quality design tools that can be used very early in the design process to predict the defects per unit of final designs.
Make No Mistake!: An Outcome-Based Approach to Mistake-Proofing
C. Martin Hinckley, 2001. Portland, Oregon: Productivity Press.
Listen to a podcast with the author and Mark Graban:
Martin Hinckley expands on his article with Barkan above in this substantial book. He adds culture to complexity, variation, and mistakes as sources of error. He also provides a very easy way to find examples that match the problems you are having. Some of the examples are new others come out of Poka-yoke: Improving product quality by preventing defects byNikkan Kogyo Shimbun/Factory (listed below).If you want to buy only one book on the topic, this book is the one I recommend.
Plant Design for Safety: A User-Friendly Approach
Kletz, Trevor. 1991. New York: Hemisphere Publishing Corp.
This book focuses on designing safe chemical manufacturing plants. “It is the theme of this book that, instead of designing plants, identifying hazards, and adding on equipment to control the hazards or expecting operators to control them, we should make more effort to choose basic designs and design details that are user friendly.” Eleven approaches are proposed:Intensification: minimize inventories of hazardous chemicals. “What you don’t have can’t leak.”Substitution: substitute safer chemicals for hazardous ones.Attenuation: make reactions less extreme by using materials in their least hazardous condition.Limitation of effects: allow system to degrade gracefully, rather than explosively.Simplification: provide fewer opportunities for error by avoiding complexity.Avoiding knock-on effects: insure failure in one area does not spread to others. Making incorrect assembly impossible: design the process so it can only be done the correct way.Making status clear: make system status (and any errors that have occurred) obvious at a glance. Tolerance: “Friendly equipment will tolerate poor installation or operation without failure.”Ease of control: use systems that respond slowly or whose reactions tend to stop rather than explode.Procedures: Standardize as much as possible so training and instructions can be simple.
Your Job: Managing Error in Out of Control
Kelly, Kevin, 1994. New York: Addison-Wesley.
Kevin Kelly discusses human error in software development. Kelly persuasively argues for need to mistake-proof the creation of software code. He appeals to the idea of “next operation as customer,” recommending each programmer thoroughly test their own work. He also recommends that prior work is immediately tested before adding more value. Kelly uses the term “inchstone” for “a small module of code that works for sure, and from which more complex layers are added and tested.” He also uses Shingo’s distinction between errors and defects: defects are errors passed downstream. Defects should be detected and corrected early and often. Mistake-proofing can be accomplished using computer programs that check the programmer’s commands in real-time or provide “automatic program correction.” Object oriented programming is also a means of avoiding errors.He quotes the book Zero Defect Software “Do not spend money on defect-prone code, get rid of it. Coding cost is nearly irrelevant compared to the cost of repairing error-prone modules. If a software unit exceeds an error threshold, throw it out, and have a different developer do the recoding. Discard work in progress that shows a tendency toward errors because early errors predict late errors.”Kelly also notes that “as software programs mushroom in complexity, it becomes impossible to exhaustively test them at the end. Because they are discontinuous systems, they will always harbor odd corners or a fatal response triggered by a one-in-a-million combination of input that eluded detection of both systematic and sample-based testing. And while statistical sampling can tell if there are likely to be faults left, it can’t locate them.”Kelly is also realistic that you probably can’t eliminate the possibility of bugs, but he argues your are better off with the smaller number of less complicated bugs that are likely to result from a zero defect approach to coding.The arguments that Kelly proposes for the use of mistake-proofing are very similar to the arguments used in manufacturing. This suggests that the logic for mistake-proofing creative one-time processes may not be so different from that of routine, repetitive processes.
Cpk of 2 not good enough for you?
Lafferty, J.P. 1992. Manufacturing Engineering (October): 10.
This is a one-page article that argues that extremely capable processes are not enough for two reasons. 1) if the world really can be described using bell shaped curves then the customer will always receive defects. 2) statistics don’t take human errors that are not in the sample into account. Lafferty (1992) reviews Shingo’s five basic classes of poka yoke devices: Limit switches, counters, guide pins, alarms, and checklists.
The best quote from the article is “Shingo brought his Poka-Yoke devices to America in the mid-’80s…Unfortunately, the reception to Shingo’s methods in this country is similar to our response to Dr. W. Edwards Deming in the 50’s. It took us 30 years to become convinced Deming was right about statistical control. Must we wait 30 more to believe Shingo?”
Error in Medicine
Leap, Lucian L. 1994. Journal of the American Medical Association 272(23): 1851-1857.
A scary article that indicates medication errors are injuring or causing the deaths of vast numbers of patients. Leap then discusses why the error rate is so high and lays out the psychology of human error in terms of slips and mistakes and latent errors. He also discusses the differing approach to error in medicine and aviation. These are discussed in Ramon Felciano above. Leap is one of the sources used by Felciano in preparing his materials.
Company Cuts the Risk of Defects during Assembly and Maintenance
Marchwinski, Chet (ed.) 1996. MfgNet: The Manufacturer’s Internet Newsletter Productivity, Inc. Norwalk, CT.
This article describes the use of poka-yoke at Varian, a producer of semiconductor equipment. Managers at Varian have changed their attitude toward errors. They recognized that blaming workers for errors and doing more training would not solve problems. Rather, they now view errors as the result of poor engineering design. Errors occur because the design allows them to occur. Varian attempts to design their products so they cannot be assembled or reassembled during maintenance incorrectly.
Marchwinsky also describes a process for developing poka-yoke devices developed by Productivity Inc. It involves identifying errors and assessing error rates, determining when errors are made and detected, and identifying deviations from the current standards for the process. Additionally poka-yoke developers should try to identify the root causes of error by identify conditions in the process that commonly lead to error. Alternative approaches to preventing the error and corresponding device designs should be considered. After selecting one of the alternatives, the device should be fabricated and tested.
Mistake-proofing
Marchwinski, Chet (ed.) 1995. Productivity 17(3): 1-6.
This article describes the implementation of mistake-proofing at United Electric, AT&T Power Systems, TRW, and Honeywell. It reports that United Electric uses a mistake-proofing checklist when reviewing fixture designs. They are looking for conditions that may cause errors like “tooling changes, multiple steps, lack of standards, rapid repetition, high volume, and poor housekeeping.” AT&T Power Systems reports substantial quality improvement and savings due to mistake-proofing. They also used a catalog of devices to promote and expedite the duplication of devices. Poka-yoke developers are paid a royalty when others use their device. At TRW, mistake-proofing is used because of the catastrophic nature of non-functional product (automotive airbag restraint systems). Implementation and lessons learned are discussed. Honeywell reports defects being reduces by 52%.
SPC vs. ZQC
Marchwinski, Chet (ed.) 1997. Productivity 18(1): 1-4.
This article discusses the work of Grout and Downs on when to use SPC and ZQC. (ZQC is the acronym for mistake-proofing.) These ideas are also discussed in the Grout and Downs papers above.
Microsoft, HP Use Poka-Yoke to Squash Software Bugs
Marchwinski, Chet (ed.) 1997. Productivity 18(5): 1-4.
An article summarizing the work of HP’s Harry Robinson and Microsoft’s James Tierney. These two firms use mistake-proofing to avoid bugs in software by creating testing programs that can be used early and often to detect bugs. Minimizing the delay between action and feedback are critical components of mistake-proofing.
Poka-Yoke Prevents Accidents As Well as Defects
Marchwinski, Chet (ed.) 1997. Productivity 18(5): 4-7.
Mistake-proofing focuses on human error in a broad sense. This article provides examples of how mistake-proofing devices can be used to avoid human errors that result in safety accidents as well as quality defects.
Poka-Yoke and the Art of Motorcycle Maintenance
McClelland, S. 1989. Sensor Review 9(2): 63.
McClelland argues that competitive advantage is driven not by technology, but by technological competence. This distinction means that engineers have to seek out the simplest technology to solve problems, not the more complex “latest technology.” He then reviews Shingo’s book (see below). He states “The whole treatise [of Shingo’s book] is a testimony to what can be done with a little money and a lot of thought. Ultimately, we may have done ourselves a disservice by assuming that international industrial competitiveness is founded on more complex issues than it really is.”
Toyota Production System
Monden, Y. 1983. Norcross, GA: Industrial Engineering and Management Press. (pp.10, 137-154)
Monden is also a Toyota industrial engineer, like Shingo. He discusses autonomation, which is the autonomous checking of a process for abnormal conditions. Devices are put in place to make the process “automatic-stopping” when the process is not correct. Monden includes poka-yoke as one example of autonomation. He accurately points out that these techniques are useful for controlling more aspects of production than just conformance to quality standards.
Boka/yoke-ing your way to success
Myers, Marc 1995. Network World. September 11, 1995: 39.
Baka/yoke is an early term, which was later replaced by poka-yoke. Baka/yoke is translated as fool-proofing. Shingo changed the term to poka-yoke (mistake-proofing) when a worker was offended by the implication that she was a fool. Myers interviewed AT&T Power Systems’ quality manager. He reports on three levels of mistake-proofing:
- catching errors before they create defects,
- catching errors during the process of creating defects,
- catching errors that have created defects and keeping the defects from going further in the process.
Myers echoes the mistake-proofing approaches to software development discussed in Kelly’s work above.
A Case History Development of a Foolproofing Interface Documentation System
Nakajo, T., Azuma, I. and Tada, M. 1993. IEEE Transactions on Software Engineering 19(8): 765-773.
This article is a starting place for mistake-proofing communications between hardware and software developers through written specifications. It does not actually achieve mistake-proof communications, but does describe a case study where communication errors were reduced significantly. The principles of mistake-proofing the authors use are 1. adjust the operations and objects to the operators’ abilities, 2. distinguish differences/changes in the operations/objects, and 3. decrease the number of differences/changes in the operations/objects.
The principles of foolproofing and their application in manufacturing
Nakajo, T. and Kume, H. 1985. Reports of Statistical Application Research, Union of Japanese Scientists and Engineers 32(2): 10-29.
Nakajo et al. study 1014 mistake-proofing devices on assembly lines. They divide mistake proofing into prevention of occurrence and minimization of effects. Each is further subdivided. Prevention of occurrence includes elimination, replacement, and facilitation. Minimization of effects includes detection, and mitigation. They manipulate and probe the data in a series of lengthy detailed tables.
Poka-yoke: Improving product quality by preventing defects
Nikkan Kogyo Shimbun/Factory Magazine, (Ed.). 1988. Portland, Oregon: Productivity Press.
This book is known by many as just “the big red book.” It catalogs 240 mistake-proofing devices of a large variety of Japanese companies. It provides ideas so firms can avoid reinventing the wheel. Shingo wrote the Preface. Hiroyuki Hirano provides a worthwhile overview of poka-yoke, which includes lists of different kinds of errors and devices, and principles of basic improvement. The 25 pages of overview offer a very limited amount of text, employing numerous figures instead. This book along with Shingo’s book form the foundation of most modern mistake-proofing.
The Design of Everyday Things
Norman, D.A. 1989. New York: Doubleday.
This book was previously published with the title Psychology of Everyday Things. Norman provides an entertaining discussion of how the design of an object and human psychology interact to create error. Here’s an excerpt from the preface:”We are surrounded by large numbers of manufactured items, most intended to make our lives easier and more pleasant. In the office we have computers, copying machines, telephone systems, voice mail, and fax machines. In the home we have television sets, VCRs, automated kitchen appliances, answering machines, and home computers.””All these wonderful devices are supposed to help us save time and produce faster, superior results. But wait a minute–if these new devices are so wonderful, why do we need special dedicated staff member to make them work–‘power users’ or ‘key operators’? Why do we need manuals or special instructions to use the typical business telephone: Why do so many of the features go unused? And why do these devices add to the stresses of life rather than reduce them?”Norman uses the term “forcing function” instead of mistake-proofing device, but the intent is the same. He also discusses how to “put knowledge in the world.” An approach to design that makes the correct action easier to determine and makes errors more obvious. Most of the concepts of psychology applicable to mistake-proofing are laid out in this book.
Mistake-Proofing for Operators: The ZQC System
The Productivity Press Development Team. 1997. Portland, Oregon: Productivity Press.
This book is a distillation of Shingo’s basic ideas for workers at all levels of a firm. While little new material is revealed, the basics of mistake-proofing are presented in a straightforward, methodical way. The book discusses the integration of the do and check phases of the traditional plan-do-check cycle for quality improvement. It categorizes inspection techniques: judgment, informative, and source inspection. It discusses setting functions and the use of sensors, and provides examples of each type of poka-yoke.
New Technology and Human Error
Rasmussen, J. et. al. (Eds.) 1987. New York: John Wiley and Sons.
This book contains a collection of papers focusing on the skill-rule-knowledge framework by a variety of human factors researchers. Both Reason and Rasmussen make multiple contributions along with many others; thirty-two papers in all. The papers are presented in sections titled:
- Definitions and taxonomies of human error
- Cognitive error mechanisms and error control
- Errors of judgment and the social context
- Human error and safety at work,
- Industrial studies of human error I and II
- Error sources, reasons or causes
- Review of discussions
- Epilogue
Human Error
Reason, James. 1990. New York: Cambridge University Press.
Reason’s book presents a comprehensive description of the body of knowledge in the area of cognitive psychology and human factors. Reason provides a meticulous discussion of the theory of human error. He lays out the generic error-modeling system (GEMS) which extends Rasmussen’s skill-rule-knowledge classification system. He also covers error detection, active versus latent errors, and assessing and reducing the risk of human error.
The limited role of statistical quality control in a zero defects environment
Robinson, A.G. and D. M. Schroeder. 1990. Production and Inventory Management Journal 31(3): 60-65.
Robinson and Schroeder discuss the attractions and limitations of statistical quality control (SQC). The attractions for SQC are large amounts of information from small samples, structured information collection, and the understanding how to use data to control the process. The limitations are an over reliance on statistics, which does not in itself improve the process, the “false sense of security” provided by being in-control. SQC can also be costly, and can disenfranchise workers and management.
In contrast, poka-yoke focuses on process improvements that improve the effectiveness of quality control performance at minimal cost. The basics of source inspection, self-checks, successive checks, and autonomation are presented. Robinson and Schroeder point out that the simplicity and common sense involved in mistake-proofing may be more easily adopted by teams than the statistical concepts and formulas of SQC.
Modern Approaches to Manufacturing Improvement: The Shingo System
Robinson, A.G. (Ed.)1991. Portland, Oregon: Productivity Press.Robinson combines the best from Shingo’s books on manufacturing into one inexpensive, concise volume. It includes poka-yoke, single minute exchange of die, etc.
How To Achieve Error-Proof Manufacturing: Poka-Yoke and Beyond: A Technical Video Tutorial
SAE International.
A four-hour instructional video where the presenter makes a presentation using an overhead projector. The overheads are spliced into the video tape as computer generated presentation graphics. SAE Promotional material state that the user will be able to “find cause-and-effect relationships between human error, and quality and safety problems; zero in on the root cause of human error; manage the concept of Poka-Yoke (error-proofing); define and discuss the five implementation steps of error-proofing.” Numerous other topics are presented as well.
Zero quality control: source inspection and the poka-yoke system
Shingo, Shigeo. 1986. trans. A.P. Dillion. Portland, Oregon: Productivity Press.
This is the seminal book on mistake-proofing. Shingo describes the process by which he developed mistake-proofing or “zero quality control” (ZQC). Shingo places inspection techniques in three categories: source, informative, and judgment. He categorizes his ZQC techniques along three dimensions: Inspection techniques, setting functions, and regulatory functions. The inspection method determines when the inspection occurs and what is inspected, the product or the process. A setting function checks a process parameter or a product attribute. The setting function is linked to a regulatory function, which is a warning, notification, or cue to the worker that a process parameter or a product attribute is incorrect. Shingo identified three setting functions: contact, fixed-value, and motion-step. The contact method checks to insure the physical attributes of the product or process are correct and error-free. The motion-step method checks the precedence relationship of the process to insure steps are conducted in the correct order. The fixed value method facilitates checking that matched sets of resources are available when needed or that the correct number of repetitions has occurred. A Regulatory function is any cue or attention-getting signal about the status of the product or process. Regulatory functions that control the process and keep it from proceeding (control methods) are more powerful than those that simply provide a warning (warning methods) are.Shingo’s book also provides a catalog of sensors and 112 examples showing before and after situations and categorizing devices on the three dimensions of ZQC. Each example indicates the cost of the device.
Using Poka-Yoke Concepts to Improve a Military Retail Supply System
Snell, Todd and Atwater, J. Brian 1996. Production and Inventory Management Journal 37(4): 44-49.
Snell and Atwater present a case study application of mistake-proofing to a service operation. They apply the concepts of Chase and Stewart’s Sloan Management Review article to a military auto parts supply outlet at Fort Carson, Colorado. Customer and service errors are described along with devices for their solution. The results reported are impressive: part location accuracy in the warehouse climbed from 65% to 98%. Incorrect “receipt closures” to suppliers plummeted from 90% to 0%. A specific class of errors consistently made by customers dropped from 22% of orders to 0%. Request process also declined from 12.5 days to 1.6 days.
Eradicating mistakes in your software through poka yoke
Tierney, James.
Alternate title: “From Deming to Shingo: using Japanese Quality Assurance Principles to Eradicate Problems from Your Software Development Process.” A video tape of a presentation at a local ASQC meeting, MBC Video, Inc 14415 N.E. 64th St., Redmond, WA 98052. (206) 885-7934.
Promotional materials state, “many people have tried unsuccessfully to use highly successful Japanese QA principles on software development: James describes the Poka-Yoke method, which is highly applicable to software QA, and is widely used at Microsoft and acclaimed for its strong positive effect. James Tierney is Test Training Manager at Microsoft and had previously been Director of Test for Microsoft at Work.” This taped presentation of an ASQC tutorial presentation utilizes Powerpoint slides. It will be of most interest to software engineers concerned with software testing and the use of debuggers. It discusses some of the testing procedures mentioned in Marchwinsky’s “Microsoft, HP Use Poka-Yoke to Squash Software Bugs” and Kelly’s Out-of-Control.
Implications of fool proofing in the manufacturing process
Tsuda, Yoshikazu. 1993. In Quality through Engineering Design, edited by W. Kuo. New York: Elsevier.
Tsuda categorizes mistake-proofing 4 ways, two of which broaden Shingo’s initial approach. Mistake detection is equivalent to Shingo’s self-checks and successive checks for discovering when defects have occurred. Mistake prevention is Shingo’s source inspection, which eliminates the possibility of making a mistake. Tsuda adds the concept of mistake prevention in the work environment which means avoiding ambiguity by simplifying processes, making gages more intuitive, practicing good housekeeping, and insuring adequate quality and safety instruction to employees. For example, a book of work instructions that has two products’ instructions on facing pages can be improved by formatting the pages so that a worker can only see one product’s instructions at a time. Tsuda’s also adds preventing the influence of mistakes. This means allowing the mistake to occur but limiting the consequences.
On training for mistake-proofing
Vasilash, G.S. 1995. Production (March): 42-44.
Vasilash describes the training for and use of mistake-proofing at TRW’s Inflatable Restraints (airbags) group in very general terms. Mistake-proofing is used because airbag testing is destructive and because it was determined to be a methodology that “would allow unit volume increases without any decrease in quality or productivity.” Production volume increased from 2 million to 13 million between 1990 and 1995. Mistake-proofing allowed the introduction of new workers and machinery without deterioration of quality.
Sustained Attention in Human Performance
Warm, J.S. 1984. New York: John Wiley and Sons.
This book contains a collection of papers on human attention and vigilance. Including “An introduction to vigilance” by Warm and “Vigilance and Inspection” by Weiner. These papers discuss the difficulties faced by human beings when they are expected to maintain high levels of attention while performing repetitive or monotonous tasks.
Improvement in transfusion safety using a new blood unit and patient identification system as part of safe transfusion practice
Wenz, B. and Burns, E.R. 1991. Transfusion 31(5): 401-403.
Wenz and Burns describe a device called “Blood-Loc” which is a disposable plastic bag with a plastic dial combination lock. The letters needed to unlock the bag can only be obtained from the patient’s wrist band. This insures that the nurse reads the wrist band and properly identifies the patient. The nurse will not be able to incorrectly administer a unit of blood that was intended for another patient. The product was tested in use and did prevent incorrect transfusions. “Blood-Loc” is one example of an information poka-yoke as described by Chase and Stewart.