Reading Time: 15 minutes.

Safety and Production – A Comprehensive 1920’s Study

This article summarizes a comprehensive study of industry in the United States1 to determine whether increasing production has any relationship to reduced accident rates. Some study results remain important for industry today.

The report defines the problems to be addressed by the study and introduced the source of the problem that led to the study. The report summary addresses insights of the study committee, the methods for the study, the volume of participants, and the overall results, conclusions and recommendations.

The majority of the 414 pages of the study report cover detailed statistics for 17 industry groups:

  • Cement Industry
  • Chemical Industry
  • Coal-Mining Industry
  • Coke Industry
  • Electric Light and Power Industry
  • Electric Railway Transportation
  • Gas Industry
  • Iron and Steel Industry
  • Machine-Building and Metal-Working Industry ! Mineral-Mining Industry
  • Paper, Pulp, and Paper Products Industry
  • Quarry Industry
  • Steam Railway Transportation
  • Telephone and Telegraph Industry
  • Textile Industry
  • Wood-Working Industry
  • Miscellaneous Industries

Origin of the Study and Definition of the Problem

In the very early part of the 20th century, there developed a significant “safety movement.” A number of major companies initiated strong emphasis on taking care of employees and reducing injuries, illnesses and deaths at work. Several other efforts emerged.

One example is the American Museum of Safety established in 1907 in New York City. It developed displays and presentations to teach companies, workers and the public about safety in their domains of work, transportation and daily life. Insurance companies also led the way already before 1900 to help educate people about safety through such publications as Insurance Engineering.

Another example is the steel industry that established the first safety congress in Milwaukee in 1912. Before 1900 organizations involved in the first safety movement were the American Society of Civil Engineers, the American Society of Mechanical Engineers and the National Fire Protection Association.

New organizations devoted to safety and health in the early 1900s included the National Safety Council and the United Association of Casualty Inspectors that became the American Society of Safety Engineers. A lot of the efforts were aimed at immigrants from various countries, a labor force that was new to industrialization and the use of machines in manufacturing. In addition, there was a rapid growth among states passing workmen’s compensation laws. A common campaign was “Safety First.” These factors formed some of the context for the first safety movement.

During WWI and afterward, major industries found ways to significantly increase production. Much of that first came as a result of filling production needs in support of the war. At the same time there was a reduction in job- related injuries and illness, but an increase in severity among cases that occurred. In addition, the insurance industry that dealt with the increased production and the new workmen’s compensation climate, saw an increase in costs and experienced financial losses.

Some called this era the “second safety movement.” It had brought some useful results for production and lowered injury rates, but also created the increases in severity and the unsatisfactory financial conditions for insurers.

The issues leading to the study of the relationships between production and safety appeared in the study report published in 1928. The report stated the following foundation for the problem:

The problem rests basically upon the following conditions:

a. An insurance experience since the war of increasing industrial accidents and, in particular, increasing accident costs.
b. A substantial increase during the past decade in the productivity of the American industrial worker.
c. A generally accepted, but unproved thesis that “the safe factory in general is the efficient factory, and the efficient factory in turn is the safe factory.”
d. A further development of this thesis (also, hitherto unproved) that safety in the factory and efficiency of production are both aspects of good executive control, and that a right organization in industry will, with proper managerial effort, bring about both efficiency and safe production, and, further, that neither can be satisfactorily achieved by itself alone.

Figures 1 and 2 illustrate the trends at the time for work places.

Accident Mortality per 100,000 of Population in the Registration Area of the United States

Figure 1. Accident Mortality per 100,000 of Population in the Registration Area of the United States, as Reported to the United States Bureau of the Census.

Relative Industrial Accident Mortality Rates

Figure 2. Relative Industrial Accident Mortality Rates in Terms of Earned Payroll Among Workmen’s Compensation Insurance Cases in 33 States, as Reported to the National Council on Compensation Insurance.

Who Conducted the Study?

The National Bureau of Casualty and Surety Underwriters requested the study and provided financial assistance. The Committee on Safety and Production of the American Engineering Council planned and conducted the study. Eight people served on that Committee and devoted countless hours to the project.

The American Engineering Council, formed in 1922, was composed of eight national and twenty state and local engineering societies. Some of the national organizations included:

  • American Institute of Chemical Engineers
  • American Institute of Electrical Engineers
  • American Institute of Mining and Metallurgical Engineers
  • American Society of Mechanical Engineers
  • Society of Industrial Engineers

Each year the American Engineering Council conducted a study on some timely industrial or economic problem. The study problem for 1925-1926 was the relationships between accidents and production. The Committee established the general scope and character of the survey, the industries to be surveyed, the character of data to be collected, and the procedural plan for the study.

The Research Method

The Committee solicited help with the study by establishing investigation centers led by field engineers and local committees to help with data collection. There were investigation centers in 15 industrial cities, with assigned territories for each. In addition, there were local subcommittees for the study in 42 cities across the United States. The workers in these sub-groups solicited data from 5,306 companies. The effort produced a response rate of 21.6% with data provided by 1,145 companies. Field engineers who participated in the data collection visited companies to help ensure effective participation.

The collected data included the following:

  • Contact information for the company and its representative.
  • Time period for data.
  • Employee data included total number, gender distribution, American or foreign born, and race.
  • Production and accident data included:
    • Year.
    • Average number of employees.
    • Total number of man-hours worked.
    • Total production and production units of measure.
    • Number of lost-time accidents and their distribution by death, permanent disability, temporary disability, and total cases.
    • Total number of days lost and their distribution by death, permanent disability, temporary disability and total cases.

There were instructions for recording all of these data.

The Participation

The study compiled a huge amount of data. After completing follow up efforts and after evaluating the data supplied, the data represented 13,898 companies. Data from 428 companies were unusable. For example, some companies did not have accident data. In a few cases, the data did not extend beyond one year, a necessary requirement to follow trends in production and accidents. Overall, the data represented 122,028 company-years (8.8 years per company), 2,464,413 employees, 13,143,569 man-years and 54,430,707,000 man-hours. Researchers organized companies into at least 59 industrial groups. When addressing various research questions, not all data were applicable to some of the individual research questions.

The Fundamental Questions and Results

The study focused on four research issues.

First Research Issue

1. Have industrial accidents under the conditions of intense production of the past few years increased in (a) frequency and/or (b) severity?
2. Are these increases in the frequency and severity of industrial accidents proportionately greater or less than the increase in production?

To answer these questions, the study compared production and accident rates in 1922 to those in 1925 for 359 companies in 34 of the industrial groups. The overall combined changes in experience among all of these companies were:

Productivity increased 14.4%
Accident frequency decreased 10.4%
Accident severity increased 2.5%

Overall, about 80 percent of the companies showed an increase in the rate of production and a decrease in both accident frequency and severity. The improvement in production for the group increased at a higher rate than did the reduction in accident frequency. Because the rate of increase in accident severity was significantly lower than the rate of increase in production and because accident severity links to higher costs of such accidents, the concern by the insurance industry about the increased cost of claims is not conclusive.

Second Research Issue

3. Can industrial accidents be controlled under modern conditions of production?
4. If so, how?
5. Does safety interfere with production?

The study found evidence that accidents can be controlled along with production increases. They point to the fact that improved production results from good management of the production processes and that accident potential under production changes can also result from good management.

One form of supporting evidence came from tabulating the no-accident periods for many of the companies and their facilities. The study created a sizable list of no-accident periods extending from 69 days to more than 4 years at plants operating under branches of the United States Steel Corporation.

Evidence for extended periods of no-accident production were compiled from 14 units of E. I. DuPont De Nemours and Company. The no-accident periods ranged from 170 days to 1,439 days (nearly 4 years).

In addition, the study compiled data from 19 industrial groups covering 513,083 employees that showed both an increase in production for at least 3 and up to 10 years while experiencing a decrease in both accident frequency and severity during the same time periods.

For another group of companies representing 332,548 employees, there was an increase in productivity while experiencing a decrease in accident severity, but also an increase in accident frequency at a rate much lower that the rate of increase in production.

The study also charted performance for two other much smaller groups of companies that experienced increased productivity, but less consistent changes in accident rates. One of these two groups saw an increase in accident frequency and a decrease in accident severity. The other group saw the reverse, an increase in accident severity and a decrease in accident frequency.

The study concluded that in most industries accidents were controlled at the same time that production increased and that “safety does not interfere with the orderly processes of manufacture.”

Third Research Issue

6. Does a positive correlation exist between the safety performance and the efficiency of production?
7. Is the safe factory the efficient factory, and the efficient factory the safe factory?

The authors first acknowledged that there are many factors that impact production as well as accident rates. Those factors are not the same across industry groups or between similar companies. In addition, they recognized that changes in industry can and were currently changing rapidly. Furthermore, they conceded that a correlation coefficient is an estimate of a relationship between variables and is not a basis for generalizing to all participating organizations. All of these make it difficult to produce precise correlation statistics that apply in all cases. At best, the resulting statistic is an indicator.

After considering data pooled from four different groups identified in the study, the committee computed a coefficient of correlation of 0.835 between an increase in production and a decrease in accident rates. They stated that the relationship is not a causation; increasing production does not by itself infer decreased accident rates.

They also looked at four groups when reviewing study results, as follows:

Increasing production and decreasing accidents 845,631 employees
Increasing production and increasing accidents 72,200 employees
Decreasing production and decreasing accidents 7,754 employees
Decreasing production and increasing accidents 7,377 employees

They applied the term “safe” factory to a decreasing accident rate and “unsafe” factory to an increasing accident rate.


Based on the overall and partitioned results, they reached the following main conclusion:

A “safe” factory is eleven times more likely to be productive than is an “unsafe” factory.

Their overall conclusion was:

“There exists a high degree of correlation between industrial safety and industrial productivity and that the combination of low accident rates and high production rates is possible for any of the industrial groups.”

They also concluded that:

”Plants which appear to have made the greatest progress in accident prevention have also made the greatest progress in productivity.”

In general, they found that:

“The safe factory is the efficient factory.”

Fourth Research Issue

8. Is safety an executive responsibility? Is the key
to the new safety movement the executive?
9. What degree of improvement can be made in safety performance? What reduction can be expected in the costs of industrial accidents?

They found that excellent management, an increase in production, and a decrease in accidents go together. They considered performance data and found companies with a strong reputation for excellent management confirmed their conclusion. They stated that good production and reduced accident rates stem from the same source: “intelligent and progressive management” that takes every opportunity to improve production and to reduce accidents and their costs. The two results come from the application of managerial skills. And “initiating safety improvement rests squarely and unescapably upon the shoulders of the leading executives.”

The First and Second Safety Movements

The report referred to a “new” or “second” safety movement. The first safety movement occurred just prior to and following the turn of the century. That movement simply emphasized the importance of safety and the need to protect employees. It had a strong moral flavor and a change in thinking. The change progressed from risks being assigned to workers and moved toward an employer sharing or accepting the risks at work. One element of that shift was manifested with the shift in legal theories and the introduction of a no-fault, workmen’s compensation system. Much of the first safety movement fell to the efforts of people who had titles of factory inspector, safety man, and safety engineer. The effort to identify risks and reduce them was led by insurance companies and trained specialists such as engineers and others who pushed for increased job safety. The first safety movement continued for many employers well into the 20th century.

The second safety movement resulted from the rapid increase in production that began with the United States participation in WWI. To meet military demands, many companies had to find ways to increase production. The focus on increased production continued after WWI. However, as noted earlier, the premise for this study resulted from the increased cost of claims for employer insurance companies.
The finding of this study established that increased production paralleled reductions in accident rates. One of the most valuable results was recognizing that increasing production requires a lot of management attention and leadership, while recognizing that reducing accidents requires just as much management attention and leadership. With each step forward to advance production further, additional management and executive leadership was required. Similarly, continuing to reduce accidents, whether in frequency or severity, requires that same kind of management and leadership ingenuity and attention.

The authors stated:

“Action in line with the study recommendations will bring in a new safety movement which will far eclipse the old. There are hidden sources of strength in executive control which hitherto have not been applied to the accident situation. If American industrialists will adopt the same executive policy toward safety which they have fully developed toward production, we may confidently expect a decreasing number of deaths, permanent disabilities, and temporary disabilities, with their attendant costs.”
“A responsibility which cannot be evaded rests upon the managers and executives of industry to make safety a major interest and a continued care.”

The Detailed Findings

The following is a summary of findings for this huge study:

  1. Industrial accidents can be controlled under modern conditions of highly efficient productivity.
  2. The rate of change in production per man- hour for the industrial groups studied is greater than the rate of change in accident frequency per man-hour or the rate of change in accident severity per man-hour. Quantitatively, the production rate was 14.4% higher in 1925 than in 1922, the accident frequency rate 10.4% lower in 1925 than in 1922.
  3. The experience of a large group of companies shows that material reductions in accident rates can be obtained simultaneously with an increase in production rates.
  4. Major industrial executives have as much responsibility to initiate accident prevention as to initiate improvement in productivity.
  5. Efforts to improve safety performance do not interfere with production.
  6. Maximum productivity is ordinarily secured only when the accident performance tends toward the irreducible minimum.
  7. The production and accident performances of the best plants in each industry studied clearly show that tremendous improvements can be achieved by the majority of plants in each industry.
  8. The incidental or accompanying cost of industrial accidents is a loss in industrial operation which should not be neglected.
  9. Organized safety work is being carried on in a relatively small percentage of industrial plants.
  10. A large number of industrial establishments keep no accident records and make no attempt to analyze their experience as the first step in decreasing accidents.


The Committee on Safety and Production that conducted this study recommended the following:

  1. That the same executive direction and control be given to decreasing industrial accidents as is given to increasing productivity.
  2. That those agencies which collect and disseminate accident statistics adopt uniform terminology and standardized records so that all data will be compiled on a nationally comparable basis.
  3. That the executives of those plants having high accident frequency and accident severity rates initiate, direct, and control ways and means of lowering such rates to at least the low rates obtained by other plants in their industry.
  4. That industrial trade associations, engineering societies, and other agencies concerned with the improvement of industrial operation bring to the attention of their members the necessity of improvement in safety performance as a vital step in the strengthening of their industrial position.
  5. That industrial trade associations secure, compile, and analyze accident statistics for the purpose of determining the lowest accident rate possible of attainment for their respective industries.
  6. That industrial trade associations endeavor to secure such action on the part of executives of their industries as will result in each plant having the lowest accident rates obtainable.

Author’s Comments

One of the things in this study that captured the author’s attention was the strong focus on the importance of close participation by senior managers and executives to achieve improvements in safety performance. The study authors emphasize that such company and plant leaders must be engaged in safety enhancements just as much as they are engaged in production enhancements and other aspects of a successful business. What is interesting is that this strong emphasis and associated recommendations occurred nearly a century ago. The report cites the work of Lewis Amory De Blois found in his 1926 book 2. He was a member of the study team.

For those who are also students of management methods, there is a philosophy incorporated in this report that suggests that production is no better than the methods applied to the production process and in a similar sense, safety performance is no better than the methods applied to preventing accidents in connection with production improvements.

The concept is very close to the management principles of Deming3 and Juran, which were promoted a few decades following this study. These management leaders demonstrated to all kinds of work groups that one cannot perform better or safer by trying harder, working harder or learning the tasks better. Work group performance can be no better than the work methods and safety methods established by the management system and its leaders. The principles presented in the study report are akin to the extensions of Deming’s management philosophy known as Total Quality Management4 and later Six Sigma5 when applied to safety. Today it even extends to the formal approaches called for in Safety Management Systems (SMS). A formalized example of a safety management system appears in the aviation industry by the Federal Aviation Authority6. The concepts also appear in the newest safety management standards, such as ANSI Z107, ISO 14001:20188 and ISO 45001:20189.

Footnotes and References
  • 1 Safety and production; an engineering and statistical study of the relationship between industrial safety and production; a report. American Engineering Council. Published by Harper & Brothers Publishers, New York, 1928.
  • 2 Lewis Amory De Blois, Industrial Safety Organization for Executive and Engineer, McGraw-Hill Book Co., New York, 1926.
  • 3 Mary Walton, The Deming Management Method, Perigee Books, New York, 1986.
  • 4 Joseph & Susan Berk, Total Quality Management, Implementing Continuous Improvement, Sterling Publishing Company, Inc., New York, 1993. and

    Peter Capezio & Debra Morehouse, Taking the Mystery Out of TQM – A Practical Guide to Total Quality Management, Career Press, Franklin Lakes, NJ, 1995.

  • 5 Greg Brue, Six Sigma for Managers, McGraw- Hill, New York, 2002. and

    Subir Chowdhury, Design for Six Sigma – The Revolutionary Process for Achieving Extraordinary Profits, Dearborn Trade Publishing, Chicago, 2002.

  • 6 Safety Management System; https://www.faa.gov/about/initiatives/sms/explained/
  • 7 Occupational Health and Safety Management Systemshttps://webstore.ansi.org/standards/asse/ansiaihaassez102012 
  • 8 Occupational heath and safety; https://www.iso.org/iso-45001-occupational-health-and-safety.html 
  • 9 Occupational health and safety management systems; https://www.iso.org/standard/63787.htm

Pin It on Pinterest