Safety Research: You and Your Lawn Mower

A. M. RameySmith, National Bureau of Standards
V. J. Pezoldt, National Bureau of Standards
J. J. Persensky, National Bureau of Standards
ABSTRACT - The Human Factors Section at the National Bureau of Standards was recently involved in the evaluation of a Lawn Mower Safety Standard for the Consumer Product Safety Commission. This report describes the results of three studies performed in conjunction with this effort as related to blade contact injuries and the implementation of a dead-man control requirement.
[ to cite ]:
A. M. RameySmith, V. J. Pezoldt, and J. J. Persensky (1978) ,"Safety Research: You and Your Lawn Mower", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 508-514.

Advances in Consumer Research Volume 5, 1978      Pages 508-514


A. M. RameySmith, National Bureau of Standards

V. J. Pezoldt, National Bureau of Standards

J. J. Persensky, National Bureau of Standards



The Human Factors Section at the National Bureau of Standards was recently involved in the evaluation of a Lawn Mower Safety Standard for the Consumer Product Safety Commission. This report describes the results of three studies performed in conjunction with this effort as related to blade contact injuries and the implementation of a dead-man control requirement.


The Human Factors Section at the National Bureau of Standards (NBS) has provided support to the Consumer Product Safety Commission (CPSC) during the development and evaluation of a power lawn mower safety standard. Part of this support took the form of providing CPSC with data concerning how lawn mower users interact with power mowers. Three empirical studies were performed to assist CPSC in resolving specific questions concerning requirements in the proposed safety regulation based on a report from Consumers Union (Consumers Union, 1975).

In October 1974, CPSC accepted the offer of Consumers Union of United States, Inc. (CU) to develop a standard under Section 7 of the Consumer Product Safety Act [15 U.S.C. 2056(b)]. The CPSC received a proposed standard from Consumers Union in July 1975 (CU, 1975). In May 1977, CPSC published in the Federal Register (CPSC, 1977) a proposed safety standard based on the CU proposal with modifications and additions made by the CPSC staff.

CPSC estimated that there were 178 000 power lawn mower-related injuries in 1975. Costs associated with these injuries are estimated to be in excess of 73 million dollars. According to the National Electronic Injury Surveillance System, operated by CPSC, over half of the injuries resulted from body contact with the lawn mower blade (CPSC, 1976). In an attempt to reduce the number of blade contact injuries, the CPSC included two requirements in the proposed safety regulation which directly relate to blade contact.

The first requirement, intended to protect feet from blade contact, describes a region under the lawn mower housing through which the blade cannot pass. This is accomplished by using an anthropometric foot probe in a performance test. The second requirement is for a "dead-man" control which must be continuously activated by the mower user in order for the blade to operate. The "dead-man" control is primarily intended to protect hands and fingers from blade contact. The three studies reported here, all performed during the evaluation phase of the standard development process, deal with these two requirements.

The first experiment addresses the problem of describing the region under a mower housing through which the blade should not pass. This was accomplished by a comparison of three available foot simulator probes with a sample of potential lawn mower operators' feet, especially with regard to dynamic characteristics.

The second experiment provided data to answer the question "How long after the dead-man control is released should the blade be allowed to rotate?" Alternatively, one might ask "How long does it take an operator to move from the handle location to an area of potential contact with a moving blade?"

The third experiment is related to the prospect that consumers might intentionally defeat the dead-man control. To meet the proposed standard the blade may be stopped either by means of a blade clutch/brake system with the engine continuing to run or by engine shut-down. If this latter method is employed to stop the blade, lawn mower users would likely be required to restart their mowers repeatedly throughout a period of use. The CPSC has suggested that this inconvenience may encourage consumers to defeat the dead-man control unless the mower is easy to restart. This experiment, therefore, was designed to provide CPSC with data which can be used to objectively define the subjective judgment of easy-to-restart for pull-started lawn mowers.

Complete technical details and data from the three studies can be found in the NBS reports on the individual studies (Ramey and Persensky, 1977; Pezoldt and Persensky, 1977; Persensky and Ramey, 1977).

Study I. Evaluation of Anthropometric Foot Probes

The foot probe test developed by CU for inclusion in the proposed mandatory lawn mower standard makes two major assumptions: first, that the probe or foot simulator is representative of the feet of the population of lawn mower users; and second, that the probe test determines potential for foot contact with the blade.

NBS evaluated the adequacy of available generic foot probes, emphasizing a comparison between the probes and the dynamic characteristics of the feet of potential lawn mower users. The probes used in the study were: the Consumers Union (CU) proposed probe, adapted from an Underwriters' Laboratory probe; the United Kingdom (UK) probe, designed by the British Standards Institute; and the American National Standards Institute (ANSI) probe.

The apparatus used in the study (Figure 1) consisted of two parts. First a camera and strobe were mounted on a platform focused on a gray-on-black measured-graph backdrop. The film used produced positives for immediate examination and negatives for future analysis and was capable of multiple exposure. The second part of the apparatus was a frame with a movable metal partition 15 cm wide which could be adjusted to allow vertical openings of 6 to 10 cm in 1 cm increments to simulate different lawn mower housing heights.

The subjects in the study included 127 males and 74 females ranging in age from 9 to 66 years. Foot length data for these subjects were compared to those collected by other researchers for specific populations. This comparison demonstrated that the present sample was representative of the range of potential operators' foot sizes. No determination however, could be made of the representativeness of the distribution of foot sizes.



After putting a white nylon stocking (to increase image contrast) over their right foot, subjects were instructed to:

Insert their right foot under the metal panel as far in and to the left as comfortable, keeping the foot flat on the base. Then,

Rock back on the heel of the foot, lifting the front of the foot as far off the base as possible, curling toes upward, yet keeping the foot as far in as possible.

Keeping the front of the foot elevated and toes curled up, pull the foot out from under the partition until the base of the toes contacted the metal panel.

At each of the three positions described above a picture was taken, creating a triple exposure print and negative as graphically depicted in Figure 2.



Triple exposure photographs, one at each of the three housing heights (i.e., 6, 8 and 10 cm), were taken both with shoes on and shoes off. This procedure resulted in six triple exposure photographs, one exposure at each of the three positions described above, for each participant.

The negative photographs were used for data reduction. Measures were taken for the point representing the height of the longest point of each of the three positions and the point corresponding to the length of the highest point at each position. (Position 1 was foot flat on floor; Position 2 was foot raised and toes curled; and Position 3 was base of toes contacting the partition.) There were, therefore, two sets of coordinates for each position, or a total of six per negative (housing height). The same procedure was repeated for each participant at each housing height both with shoe on and shoe off--a total of 36 sets of coordinates per participant. The same three positions were measured for each of the three generic foot probes and for a shoe probe made by affixing a shoe on the end of a length of wood. The shoe probe was constructed from a 95th percentile size female's shoe.

Horizontal and vertical insertion distances for the generic probes, the shoe probe, and the foot data were used to develop safety envelopes for each housing height. Figure 2 shows the measurement points used to develop the safety envelopes. Figure 3 is a composite envelope at a housing height of 10 cm for each of the three generic probes (CU, UK, ANSI), the shoe probe, and the 95th percentile foot data.



Of particular interest is the fact that, of the three generic probes, only the UK probe passes through the space above the plane of the housing opening. However, none of the probes approaches the vertical insertion distance achieved by the 95th percentile foot data. This region, above the housing, is of primary importance in providing foot protection since mower blades are located above the plane of the bottom of the housing. Although the CU probe exceeds the other generic probes and the foot data in terms of horizontal insertion distance, it does not provide any additional foot protection over that provided by the UK probe since the mower blade rotates above the lower edge of the housing.

In addition to comparing safety envelopes derived from the aggregate foot data with the generic probe data, a comparison of each individual's data was made with each generic probe to determine the percentage of participants who would be completely protected by each generic probe and by the shoe probe. These data indicate that at least one point of each individual's foot movement data falls outside the safe area defined by each of the generic probe envelopes. Therefore, a lawn mower which satisfies the criteria of any of the generic foot probes would not completely protect any of the participants in study. The shoe probe performed somewhat better than the generic probes. However, as shown in Table I, this probe would not provide the magnitude of protection which would be obtained from a probe designed from the 95th percentile foot data.



Clearly, there is a serious problem with all of the existing generic probes, as shown by comparison of the individual data with the generic probe data. It seems clear that further research should be performed toward the development of a more adequate probe.

Study II. Time-to-Blade-Access

Several recommendations have been made for the maximum allowable blade stopping time, that is, the time from release of the dead-man control until the blade comes to a complete stop. The power mower industry, through the Outdoor Power Equipment Institute (OPEI), suggested a stopping time of seven seconds (OPEI, 1975). Consumers Union recommended a three or four second stopping time based on a study employing college students (CPSC, 1977). The National Bureau of Standards had previously recommended a blade stopping time of one second, based on data reported by the University of Iowa (Porter, Jones, and Persensky, 1974). In view of the inconsistency among these recommendations, CPSC requested NBS to perform a basic empirical study of operator movement time to provide an ergonomically sound recommendation for blade stopping time for walk-behind power mowers.

Of critical interest is the time it takes a lawn mower user to move from the operating position at the handle to a position of potential blade contact. Participants in this study were tested using a reaction time device which was designed to measure mower users' approach times to various points representing areas of potential blade contact. The test apparatus (Figure 4) permitted measurement of reaction time--or time to release a simulated dead-man control at the onset of a cue light and, more importantly, movement time--or the time from the release of the dead-man control to activation of one of five switches as surrogates for blade access areas.



The subjects' task was to move from the handle to the appropriate switch designated by a cue light, then depress the switch by hand. The distances of the switches from the handle were at intervals of 25.0 cm from 50.0 cm to 150.0 cm. The shortest and longest distances were based on the distance of the closest rear wheel and furthest front wheel, respectively, as measured on a sample of existing lawn mowers. The 100 subjects tested (64 males and 36 females) ranged in age from 16 to 62 years. Each received five trials at each distance on a randomized schedule. Subjects were instructed not to rush to strike the switch plates, but rather to walk at a normal rate.

Blade-area access-time data are summarized in Table II. Analyses of these data reveal that, as expected, there are no statistically significant differences in reaction time as a function of movement distance; however, movement time increases with movement distance. Average movement times observed in this study ranged from 0.6 sec. to 3.3 sec. The median movement times at the shortest and longest distances were 1.4 and 2.2 sec., respectively. Clearly, however, a "worst case" strategy seems more fitting when recommending an appropriate blade stopping time. That is, the selected time should protect significantly more than half of all users. Table II also presents percentile information for movement time to each of the five simulated blade contact areas. Inspection of the table suggests that a blade-stopping time of 0.7 second would protect all but five percent of the sample from potential contact with the rotating blade at the shortest distance tested.



The time data generated in this study reflect only direct movement time from the handle to areas of potential blade contact. In the absence of any data to the contrary, however, it must be assumed that in the worst case lawn mower operators do, in fact, move directly from the operating position to an area of potential blade contact. Unless other movement scenarios can be established, a more valid movement time criterion cannot be determined.

Study III. Ease of Pull

This study was designed to provide CPSC with data based on subjective judgments, which can be used to objectively define "easy-to-restart" for pull-started power lawn mowers. Many factors are involved in establishing criteria, among the more important of which are:

- the force of pull required to start the engine;

- the distance through which a pull must be made;

- the average number of pulls required per engine restart;

- the number of times the engine must be restarted during the mowing period; and

- the time interval between restarts.

In addition to these factors, which are all external to the individual starting the mower, a number of human characteristics are of at least equal importance. To the obvious factors of age, sex and physical condition, must be added the kinesthetic and proprioceptive feedback cues experienced when pulling the starting cord and the connotations placed upon the terms "easy" and "hard."

This study did not attempt to explore all of these factors. The data generated by this effort, therefore, cannot be construed as providing a definitive answer to the question "What is easy to restart?" Rather, the present study attempts to define easy to pull under experimentally controlled conditions. In psychological terms, the problem becomes one of determining the relationship between physical stimuli and the psychological responses to such stimuli. In this case the stimuli are the forces which are exerted on a simulated pull-start mechanism and the responses are subjective judgments about the ease or difficulty involved in applying these forces.

Two simulators were designed and built for use in this study (Figures 5 and 6). The simulators were operationally identical with the exception of the locus of pull. Pulls on the "housing" simulator were executed from a height typical of several current lawn mower designs, i.e., 38 cm (15 in). The "handle" simulator was designed to be pulled from a position on the handle, 83 cm (33 in) above the ground. Although most current lawn mowers are more similar to the housing simulator, lawn mowers incorporating a dead-man control may have the pull handle located in an area similar to that on the handle simulator.





The simulators permitted pulls similar to those experienced when pull-starting a power lawn mower. Pulls could be made through four distances and the difficulty (resistance) of pulls could be manipulated by the experimenter. Difficulty of pull could only be controlled in a very gross sense by manipulating an adjustable cylinder on the simulator engine; the actual force applied was determined by the participant. Measurement of the forces applied when pulling the start cord was accomplished by incorporating a load cell in the pull handles. These load cells were interfaced with a digital readout, peak load indicator. Peak force (i.e., the maximum force exerted at any point in the pull) was measured to the nearest pound (4.45N). Forces required to actually start the engine could not be determined since neither gasoline nor spark plugs were present in the simulators.

The subjects' task consisted of making 36 pulls at one of the simulators on each of four consecutive days. Each day the subjects were provided with a different length cord which determined the distance through which they were to pull. The distances through which they pulled were approximately 46, 61, 76, and 91 cm (18, 24, 30, and 36 in). The order in which subjects were assigned to the various distance categories was randomly determined. Subjects were instructed to pull the "starting" cord in the same manner they would to start a lawn mower. Immediately following each pull, subjects were requested to judge whether that pull was easy or hard. A forced choice paradigm was employed, requiring subjects to judge each pull as either "easy" or "hard." No intermediate responses were allowed. Participants were instructed to consider the simulators as "perfect lawn mowers," that is, they would start on every pull. In fact, of course, the engine never started. Subjects made six pulls in a five minute test period, followed by a rest period of approximately 10 minutes. Six test periods were completed each day by each subject.

A modified staircase method was employed to determine compression ratio adjustments to the simulator after each pull. Generally, if a subject judged a pull "easy", the compression ratio of the simulator engine was increased one unit for the next pull. If a pull was judged "hard," the compression ratio was decreased. Some variation in this procedure occurred in an attempt to assure that subjects pulled over a broad range of forces. After each pull, the judgment about the pull, either "easy" or "hard," and the peak force exerted during the pull were recorded. Peak force, rather than any measure of force over time, was employed for several reasons. First, peak force proved simple to measure. Second, pilot tests showed the measure to be capable of discriminating between pulls judged easy and pulls judged hard. Finally, peak force was felt to be the most easily adaptable to test method development.

A total of 5,372 pulls were made from the handle position and 5,009 pulls from the housing. For most analyses, the peak force measures of primary interest were the maximum forces judged easy (Easymax) and the minimum forces judged hard (Hardmin) for each subject. For purposes of analysis, peak forces were grouped in 22N (5 lbf) categories.

With one exception, there were no statistically significant differences for either Hardmin or Easymax as a function of the distance of pull. The peak forces, therefore were combined for all four pull distances. Tables III and IV show the cumulative distributions (number and percent) of Hardmin and Easymax for males and females using the handle and housing simulators respectively. These data are shown graphically in Figure 7 for female subjects using the handle simulator.

Figure 7 provides the basis for defining "easy to pull" under the conditions of this study. The cumulative percentages of Hardmin and Easymax displayed in this figure define the lower limit of "hard to pull" and the upper limit of "easy to pull" respectively for females pulling from the handle position. Similar figures, not presented here, have been constructed for males using the handle simulator and for males and females using the housing simulator. For any given percentage of the sample, the Hardmin line defines the minimum peak force which was judged hard. Similarly, the Easymax line defines the maximum peak force which was judged easy for any specified proportion of the sample. These two functions may be viewed as providing conservative (Hardmin) and liberal (Easymax) definitions of the peak forces considered to be easy to pull. Inspection of Figure 7 reveals the very large difference between the "liberal" and "conservative" estimates of easy to pull. This difference reflects the large variability in judgments of easy and hard both between individuals and within a single individual. This variability was evident for males and females using both the handle and housing simulators.

It was recommended that a "conservative" stand with regard to determining a force value for easy to pull be adopted by CPSC. This was suggested so that the greatest percentage of the population could "easily" restart lawn mowers and so that the incentive for defeating the dead-man control would be reduced. In practice, this would mean determining the peak force by using the Hardmin distribution for females and selecting the sample percentage at 80 percent or greater. That is, the value for pulls from the handle position should be approximately 140N (31 lbf).

This study does not, nor was it intended to, answer the question of why one judgment is "easy" and another "hard." Neither does it provide a definitive answer to the question of what is easy to restart. The data generated in the study do, however, provide a practical basis upon which a policy decision regarding the upper force limit for "easy to restart" can be based.








The three studies reported here provide empirically based answers to specific questions which have arisen during the development of a proposed safety standard for power lawn mowers. As a result of these studies, the following recommendations have been made to CPSC:

1. In lieu of development of a new foot probe, the United Kingdom probe should be used for testing compliance with the standard since, of available probes, the UK probe most closely approximates movement of human feet under a lawn mower housing.

2. Based solely on the elapsed time from release of the dead-man control to finger access to areas of potential blade contact, a maximum of 0.7 seconds should be allowed for blade stopping.

3. The peak force necessary to restart a lawn mower with a handle mounted pull cord should not exceed approximately 140N if restarting is to be considered easy by a substantial portion of lawn mower users.

All of these recommendations have been made with certain qualifications. However, the three studies reported here do provide data which should be considered when CPSC makes policy decisions in these matters.


Consumers Union of United States, Inc. Proposed Safety Standard for Power Lawn Mowers, Submitted to the U.S. Consumer Product Safety Commission, July 1975.

Outdoor Power Equipment Institute, Technical Comments and Recommendations on Proposed Lawn Mower Safety Standard, Submitted to the U.S. Consumer Product Safety Commission, July 1975.

J. J. Persensky and A.M. Ramey Power Lawn Mowers: Evaluation of Anthropometric Foot Probes. National Bureau of Standards NBSIR 77-1294, May, 1977.

V. J. Pezoldt and J. J. Persensky Power Lawn Mowers: Ease of Pull. National Bureau of Standards NBSIR 77-1298, June, 1977.

L. G. Porter, C. E. Jones, and J. J. Persensky Some Aspects of Lawn Mower Safety, Status Report to U.S. Consumer Product Safety Commission, December, 1974.

A.M. Ramey and J. J. Persensky Power Lawn Mowers: Time-to-Blade-Access. National Bureau of Standards NBSIR 77-1299, May, 1977.

U.S. Consumer Product Safety Commission. Product Profile: Power Mowers, October, 1976.

U.S. Consumer Product Safety Commission. Power Lawn Mowers, Proposed Safety Standard and Extension of Time. Federal Register, Vol. 42, No. 87 - Thursday, May 5, 1977, 23052-23072.