manage safety and health matters. The majority of them were mining engineers or graduates in some other engineering discipline. Interviews with them about their backgrounds in safety and health indicated that they had very limited education, and no direct experience, in these matters. In light of the importance of health and safety to productivity and profits in mining ventures, it appears that the training of most mining industry professionals in these matters is insufficient. As a result, safety officers often have poorly defined duties and, as previously indicated, are selected with little regard for qualifications. Supervisors under such managers generally do not seek to acquire special skills in the health and safety area on their own initiative.

If the top manager has properly executed the planning, organizing, and staffing functions related to safety and health, the direction and control functions are made much easier. This is especially true if the organizational arrangement makes all supervisors fully accountable for safety and health in their departments. This arrangement was a prominent feature of the mining operations that were observed to have the best safety and health conditions.

In the discussion above, the phrase "good safety and health conditions" is used several times without defining what "good" means in this context. Specific evaluative considerations are discussed in the "Field Evaluation Plan" section. It is sufficient here to note several points that are basic to the selection of a measurement for determining the "goodness" of health and safety conditions:

A measurement designed to record personal injury experience is not necessarily a sound way to measure safety performance.

The occurrence of accidents is a probabilistic phenomenon. Management actions that reduce the probability of accident occurrence cannot reduce the probability to zero.

Accidents that produce personal injuries do not necessarily result in a larger total cost to a mining enterprise than those that do not. In most mining operations, the noninjury accidents are far more numerous and their total cost is larger.

As much or more can be learned about accident reduction from accidents that "almost occurred"- that is, "near misses" - as can be learned from accidents that actually occurred (those that are required to be reported to the government).

Contrary to popular belief, it is not true that "unsafe acts" are the cause of 80 to 90 pct of all accidents in mining. The reasons for accident occurrence are much more complex.

The literature review and discussions with experts brought forth a great many ideas that might be candidates for inclusion in a model health and safety program. Nearly every safety practicioner and author had some favorite technique to which he gave credit for whatever accident reduction had been achieved. Often data were presented to show that a reduction in the "MSHA1 incidence rate" had been experienced during a specified period. Aside from questions about the statistical accuracy and significance of the data, the precise relationship of the data to the variable under consideration could not be determined. It is a rare case when only one element of a company's safety program is altered and the program then remains unchanged for even as long as a year. Nearly every program element

4 Mine Safety and Health Administration.

alteration or innovation is accompanied by a great many other changes, both obvious and subtle, during its implementation. Even when the variable can be "isolated," there may still be questions about the phenomenon that produced the result. For example, was it the physical improvement of a work area that reduced accidents or was it the socalled "Hawthorne effect"? That is, did the specific physical improvements induce the recorded improvements in work behavior, or did the work behavior results occur because of the workers' response to management's increased attention to them and their workplace conditions? That is, might the results have been the same with entirely different physical improvements?

For almost every safety improvement method advocated by some practicioner, expert, or author, there could be found someone else with apparently equal credentials to impugn the method. For example, there are many advocates of safety contests. Crapnell (7) declares that "safety contests and promotions can make safety fun- no small accomplishment in itself." He also quotes a remark concerning "safety bingo" made by a manager of industrial relations: "(It) creates tremendous peer pressure for workers to perform their jobs safely." These statements are representative of many made by persons who use and endorse safety contests. Others raise some questions that the contest advocates usually do not address. Gilmore (8) observes that the performance of a plant or work group "cannot really be compared" to that of another plant or work group. Instead, the plant or work group "can show improvement only as it improves its own experience." Of course, contests can be designed so that winning requires the greatest improvement in the group's own experience record. Accurate measurement is necessary, but difficult to achieve. Gilmore points out that "meaningful measurement depends on the sincere desire to learn the truth and is hindered by the desire to appear safer than someone else in a contest situation." Hampton (9) observes that, although contests often achieve the intended result, they also produce many unintended results. After citing several actual example, he states, "Neglect, conflict and dishonesty are among the principal side effects of contests. Through defects in design and implementation, contests can implicitly reward activities inconsistent with organizational goals."

The use of safety committees is another practice that some safety experts recommend and others discourage. Mims (10) writes, "I have found worker safety committees the most practical way to foster genuine grassroots concerns with safety-the 'they is us' attitude... it spreads this feeling of responsibility and safety awareness to every worker who is serving on it or has served on it in the past." However, DeReamer (10) opposes the use of safety committees, arguing, "Their basic principle-dispersing responsibility - simply does not work where safety and health is concerned. Employee participation in safety is essential, but there are better ways than committees to achieve it."

In the early part of the background research it seemed appropriate to attempt to select for the model program those practices that were favored by the preponderance of available experience, information, evidence, and opinion or supported by especially impressive data. However, soon it became evident that this was not appropriate because

However successful a given safety improvement is for one or several mining enterprises, it will not necessarily be a sound method in all.

Field observations indicated that very different sets of methods produced similar results at different mines, if both sets were carefully managed.

Interviews with experts and field observations led to the conclusion that no single improvement method was essential in a model program to attain and maintain good safety and health conditions.

What is essential in a model health and safety program for mines is for management to create and sustain a few fundamental conditions.

This section describes the criteria established for a model health and safety program and then defines the model program in terms of five fundamental conditions.

CRITERIA

The background research work suggested that the model program should satisfy the following criteria in order to be effective and acceptable to mining industry managers:

1. The model must be readily adaptable to all types and sizes of mines in the coal, metal, and nonmetal mining industries.

2. The model should identify all of the essential features for illness and injury reduction effectiveness. All of the essential features are required so that the scope of the model is well known at the beginning of implementation and so that the evaluation does not show an unfavorable result because of program incompleteness. Nonessential features should not be included in the model; they can make the model less acceptable to mine management and the evaluation confusing.

3. The model program must be manageable in ways that are compatible with the management concepts that are customary in the United States. This criterion addresses especially the avoidance of "intervention mechanisms" that require the services of specialists or alter the accountability structure of an organization. The reason for this criterion is primarily to improve the probability of obtaining the cooperation of mining companies in evaluating and, later, adopting the model.

4. The loss reductions attainable through the model program implementation must exceed the cost of the implementation. In 1981-82, the state of the national economy and of the mining economy in particular was such that few mining companies would have been likely to consider any new program unless there was an expectation of a reasonable return on any investment required to implement the program. Safety and health matters are more likely to be dealt with effectively if increased profit is the expected result.

5. The model program should be acceptable to management and labor on its merits alone. No inducement to adopt the program should be necessary, other than the expectation of reduced losses due to accidents and illness. Model program acceptance should be entirely voluntary.

FUNDAMENTAL CONDITIONS

There are many ways in which a model program might be described or defined. The way chosen was to define the model program in terms of a small number of fundamental conditions. All of the conditions can be created and maintained by competent managers in the performance of the normal functions of management.

In framing the model, the research team considered rather carefully what "normal functions of management"

means. There are hundreds of management and business administration textbooks and articles that list and explain management functions. One of the most useful lists is presented by Dale (11). Dale identifies and discusses seven functions of management:

Planning
Organizing
Staffing

Direction

Control Innovation Representation

Dale's list is especially appropriate to mining management because it includes the functions of innovation and representation-functions not included in many other management function listings. Innovation is especially important in mining. Dale notes that management is "a creative rather than adaptive task"-a concept entirely consistent with the model program concepts.

Some methods of defining the model that were considered created problems in satisfying one criterion or another. Satisfying the first and third criterion required the rejection of several ideas for model program features because the ideas were too specific to be acceptable (or useful) to small mines. For examples, many of the experts interviewed advocated a particular organizational pattern for the mine safety office, including a full-time safety official. A typical small mine management would not feel that it could afford the arrangement advocated or that the arrangement was needed. In other words, the small mine management could not foresee a satisfactory return on the investment. The managements of some medium-size mines (say, 50 to 100 miners), and some larger ones as well, might also oppose the arrangement on the grounds that a same, or better, return would be available through a smaller investment.

Another example relates to a proposed model program feature that called for regularly scheduled safety meetings at all levels. It was rejected because it was inconsistent with the first fundamental condition (explained below) and because of a particular finding during the field observations. A mine shift foreman expressed his strong opposition to safety meetings, saying that they were largely a waste of time that they were "rituals" which "gave safety a bad name," and that they dealt publicly with matters which, in his opinion, could be dealt with effectively only in private, one-on-one, supervisory actions. Later it was observed that this foreman had superior safety performance on his shift, as well as an excellent production record. He never conducted formal or informal safety meetings; and some of his subordinates, with a mixture of amusement and admiration, said they had never heard him use the word "safety." Safety meetings were not essential in this operation. There are many ways to deal effectively with safety training and safe work performance.

No Separation of Production and Health and Safety

There must be no separation of production and safety and health in the management of the mining operations. The best "safety program" is none at all-that is, there is no management activity that is separate and distinct from production management. "Built-in" safety and health should be evident in every work aspect at every level.

Probably the best way to assure that safety and health are integral to all production tasks is to prepare complete written work descriptions, or "safe job procedures," that define the right way to do each task. These are developed through job safety analyses, methods engineering, worker's suggestions, various industrial standards, and equipment manufacturer's recommendations. They are designed to guide the training and supervision of workers. If they are to be effective, these descriptions cannot be mere "read-andsign" documents designed primarily to "prove" that the worker was trained in case of an accident.

The most important prerequisite for attaining this fundamental condition is that each employee must be provided with precise and comprehensive instructions concerning the proper performance of his or her job, with the instructiions taking into account safety as well as other issues. Assuming that employees accept and follow these instructions, and that management is properly performing its other normal functions (i.e., planning, organizing, and directing), the achievement of this fundamental conditions should require no additional investments.

Honest Commitment to Health and Safety

There must be an honest commitment by the top managers to constantly improve the operation's health and safety performance. This commitment must not only exist, it must be regularly demonstrated. Lippert (12) states, “. to the extent that the chief executive perceives that 'a safe operation' is important in fulfilling his role, to that extent will safety have importance in the perception of the lower levels of management and supervision." Frequently the safety literature cites the need for the top manager to support the safety department, but such support is not an impressive manifestation of the manager's commitment in the eyes of the workers. The safety department should be devoted to supporting the top manager, who should be leading the way by precept, encouragement and example. With regard to leading by example, Shaw (13) writes, "If, on an underground inspection, a manager or supervisor passes an unsafe condition with no comment and then raises cain over production, no amount of later talking about safety is going to convince the workers that he or she is really committed to safety. Therefore, step one in getting safety is the commitment from management."

Several authors have pointed out that failure of upper level managers to effectively demonstrate their commitment to safety, as perceived by the first-level supervisor, will result in the failure of a safety program. Walters (14) writes:

The firstline supervisor tries to do what is requested by his supervisor, but if the secondline supervision and other supervision at higher levels do not act in a manner that supports what they say, then the firstline supervisor soon notes this and governs himself accordingly. As a result, the time and expense

devoted to any program is often wasted. In
many cases it is the second level of supervision
on up that causes the failure of the otherwise
successful and well designed policy.

Every manager and supervisor must accept that he or she is perforce the principal safety official for those activities under his or her supervision and that he or she must constantly act in a manner that reflects dedication to proper performance of that responsibility. Representation, direction, and control are the key management functions pertinent to this fundamental condition.

Management Health and Safety Training

Managers and supervisors at all levels must receive basic training in safety and health (more broadly, loss control) management and additional periodic courses to update and upgrade their training in more advanced management techniques. All managers and supervisors must thoroughly understand the costs of occupational illnesses and injuries, and accidents that result in equipment and facility damage. They must be able to correctly relate these costs to productivity and profit losses in the operations for which they have responsibility. A working knowledge of countermeasures to reduce the probability of these losses is necessary, including mastery of simple analyses of cost-to-loss reduction relationship. They must learn to view accidents and job-related illnesses and injuries as failures of management to adequately perform one or more of the management functions. In particular, they must recognize that properly training their subordinates is always a major part of the management functions of direction and control. Many task training needs (identified as part of establishing the first fundamental condition) require careful supervisory planning and innovation. These essential training activities are usually done properly only by managers who are themselves properly trained. However, training, no matter how intensive and thorough, cannot by itself reduce the probability of accidents to acceptable levels. Counter-measure training of managers must include demonstrations and practice in the reduction of workplace safety and health hazards-that is, the technology as well as analyses of alternatives in cost-versusbenefit terms. Management training also must show the manager that the Federal 30 CFR standards and State standards should not consitute a set of management goals, but rather, in general, only a minimum foundation. Finally, the manager must understand that there are some (although few) attractive safety practices or countermeasures that are simply too expensive to implement. There are sometimes advantages in accepting risks-but this will never be true if the degree of risk being accepted is not accurately known.

Management Emphasis on Health and Safety in

Organizing and Staffing Functions

Managers must include special emphasis on health and safety in their performance of the organizing and staffing functions. Job-selection processes should include thorough evaluation of employment experience, accident and injury history, physical condition, and learning disabilities and any other handicap. Safety and health risks to handicapped persons and to their employer should be controlled by identifying handicaps accurately and defining training, job assignments, and appropriate performance restrictions accordingly. Periodic performance evaluation (a control activity within the staffing function) must include a loss control

component commensurate with loss control elements that should be included in work descriptions such as those mentioned in the discussion of the first fundamental condition. Employees unable to meet sensible performance standards after corrective retraining, closer supervision, and counseling must be reassigned or discharged. Periodic organizational analysis should be conducted to evaluate the organizational structure and the jobs that comprise it in terms of the current and near-future goals and standards of the mining enterprise. Organization should be dynamic. Like the people in it, it needs correction and redirection from time to time.

Reliable Feedback Mechanisms

There are several reliable feedback mechanisms that have reliable and readily detectable safety and health components. One of the most important of these is the investigation of all accidents and "near misses" using a method designed to identify all factors that may have contributed to the event. ("Accident" is defined as any event that produces an unplanned cost.) Information from such investigations provides management with knowledge about the effectiveness of training, hazard removal, and any other actions that may have been taken to improve safety. It also identifies actions that need to be taken in the future. The "near miss" is an equal, or better, information source than an accident. Palisano (15) reports that Exxon Corp. has very successfully

used a near-miss investigation program to reduce accidents. The near-miss information is published in newsletters and used to reenact some of the near-misses for video taping instruction material.

Another very valuable feedback or control mechanism is the safety and health inspection. These inspections may be done in many ways. Jones (16) reports on a safety audit technique used by Allied Chemical Corp. It measured employee compliance with safety rules; unsafe acts, conditions, and equipment; and the health and safety environment of the plant. The top manager, superintendent, supervisor, and safety officer took part in a twice-monthly audit of the 3,000-employer plant. The audit technique, like the near-miss investigations, provides information about the status of earlier safety and health improvements and the need for new improvement actions.

There are many other feedback methods that produce good results in some mines (job-performance sampling, which is discussed later, is an example), but the two methods discusssed above can be effective in any mining operation- especially if they include the participation of appropriately selected members of the work force. Whatever the methods, they must incorporate accurate records, maintained with care and available for examination by anyone in the workforce. This last fundamental condition is essentially an element of the control function, but it also involves the planning, organizing, staffing, and representation functions.

Gilmore (8) notes that ". . . loss control is the art of attaining the optimum balance of loss potential, loss probability and profit." The fundamental conditions discussed in the previous section describe, collectively, the management environment that makes practice of the art of loss control most effective. It would have been desirable, if it had been possible, to evaluate the model health and safety program in terms of objective measurements of loss potential, loss probability, and profit. However, for a variety of reasons, such direct measurements were not possible.

It was decided that the evaluation plan should include considerations such as

Ease of implementation and administration - A measure of the simplicity, or lack of it, in setting up the program and running it on a continuing basis.

Cost effectiveness-A measure of both the initial installation costs and ongoing costs as compared to the program's effectiveness in reducing accidents and injuries.

Acceptability - A measure of the willingness of labor and management to use the new program.

Compatibility-A measure of how well the program fits into the existing overall mine situation.

Some comments about these considerations will help the reader relate them to the evaluation methods planned for actual use in the field.

It was concluded that there would be no way to measure acceptability except in a subjective way- that is, by using anecdotal data from the mining companies that agreed to allow field implementation and evaluation in their operations. The fact that they agreed indicates acceptability by management, and an expectation that the work force would find the model program acceptable too. Although some mining companies declined the invitation to participate in this

study, none of them gave as reasons anything that suggested management found the model unacceptable.

The compatibility measure, it was decided, would also have to be made from anecdotal data-or perhaps from the lack of such data. Since, as several mining executives noted, the model program dealt with "getting managers to do what they should have been doing all along," it was believed that compatibility would not be a problem except with respect to those managers who might simply resist change-any change.

The ease of implementation and administration measurement could be made satisfactorily by compiling a list of recommended actions to implement the model program, discussing them with the top manager and recording his or her decision concerning each action, and observing and rating the progress made over a reasonable period of time. This technique is explained in more detail in the section on program implementation and appendixes A and B.

The cost effectiveness measure was more difficult, especially the determination of accident and injury reduction effectiveness. During the field observations, it appeared that only about 5 pct of the mining operations visited recorded and reported accidents and injuries exactly as required by 30 CFR 50. Based on the field observations, it was judged that the mining companies that agreed to participate in the evaluation might not have good records of accidents and injuries (as those terms are defined in 30 CFR 50) and would not have reported as required by 30 CFR 50. In other words, it was anticipated that there would be no suitable baseline data available. (There was a high correlation between good record keeping and correct reporting and the quality of safety and health management. The companies with good records and correct reports probably would not be candidates for participation in the model program, since

they would not see a need for improvement.) It is a part of the last fundamental condition that accurate records be kept of all injuries and all accidents, not just those defined in 30 CFR 50, and of near misses as well. If this condition was introduced during program implementation and maintained during evaluation, it was considered very likely that more accidents would be recorded and reported than in the previous year, even if fewer actually occurred. In addition, the probabilistic nature of accidents makes it possible that there could be more accidents in one short period of time than in a second similar period, even though the first period has a lower mean probability of accidents. However, a sustained reduction of the mean probability of accidents (and illnesses) will reduce the long-term accident risk.

One very useful way to judge whether the mean probability of accidents is being reduced is to use a job performance sampling technique. The sampling procedure is a form of inspection. If conducted properly, it not only provides feed back, but also an opportunity for a management person to demonstrate concern for safety and health in a very favorable setting: one-on-one communication at the employee's usual workplace. A job performance sampling procedure was designed expressly for the model program evaluation. Through observation of a randomly selected sample of jobs, the procedure provided a basis for estimating the proportion of the work force that was working with

Individual job performance deficiencies that could be corrected by the employee alone (PA), or

Individual job performance deficiencies that could be corrected only through employer action beyond the control of the individual employee (PB), or

Deficiencies that were common to several jobs of the same kind, or various kinds, and correctable only through substantial changes by the employer in the mining plan, operating policies, or job structuring (Pc).

The sampling process is explained in more detail in the section on implementation and in appendix C.

A second measure related to cost effectiveness is the change in the slope of the cumulative cost of accidents, illnesses, and injuries curve, from the beginning of the program implementation through the evaluation period.

Decreases in the slope indicate that losses are costing less. Unfortunately, it was not expected that all loss data would be available at the beginning. It was estimated that several months of training and other management actions would be necessary before all losses could be estimated with reasonable accuracy. However, there is one element of loss data that seemed likely to be available for all the mines from the payroll records: records of work days lost due to work injuries or illnesses. This statistic is certainly not an acceptable surrogate for the total cost of accidents, injuries, and illnesses or even for the total cost of work injuries and illnesses, but it is likely to be a dependable indicator of the trend in those statistics.

A third measure related to cost effectiveness is the "expectation calculation for various health and safety management improvements. It can be used to evaluate the "worth," that is, expected return on investment, of hazard countermeasures, production practice changes, and injury or severity reduction methods. The "expectation" is a representation of the fact that if two or more possible, mutually exclusive events or outcomes can be expressed as numerical values, and each can be assigned a probability of occurrence, the "expected value" of a trial is the sum of the value of each event multiplied by its probability. In short, the "expectation" of the outcome of a chance process is the weighted average of all values of the variable, that is, the arithmetic mean. The "expectation" can be expressed mathematically as follows:

Consideration of possible sites for model program implementation and evaluation began during the field research observations (task 2). The following conditions for evaluation mines were established:

Two mine sites should be evaluated, an underground coal mine and mine of a different type in a different part of the country.

The minimum employment at each mine should be approximately 100 persons.

The mines should have significant, but not atypical, safety problems.

There was another condition that took precedence over all of the others: The managements of the candidate mines would have to be willing to participate in the model program evaluation.

Fifteen mines were identified as candidate evaluation sites. These were either mines that were known to members

of the research team through work on other projects or mines that were recommended by MSHA mine inspectors who knew about the model program and whose recommendations had been sought by the researchers. From the 15 candidate mines, the researchers chose 8 to be invited and encouraged to become participants in the field evaluation processes. Each of these eight mines received a detailed explanation of the model program and the implementation and evaluation plans. Each also received a formal letter that defined the cooperative arrangement proposed and requested a favorable response. Two mines never responded at all. Two others responded with polite, detailed explanations of their reasons for declining to participate. The reasons had to do mostly with programs of their own, or of their parent companies, which had been recently introduced to improve their health and safety performance. The researchers withdrew the invitation to one mine when they learned of some management problems that appeared to make an agreeable arrangement unlikely. After 2 months,