This page is part of the official website of the CFAA, the Canadian Fire Alarm Association
If you arrived at this page directly from an internet  search engine or bookmark, you may wish to 
visit and bookmark our home page and site map
The handy keyword-searchable CFAA website contains a wealth of information relating to the fire safety indusrty.

Canadian Fire Alarm Association (CFAA)

IMPORTANT: By accessing any part of this website, you are agreeing to be bound by specific Terms of Use and Disclaimer under which you may use this website. If you have not already done so, please read the TERMS OF USE DISCLAIMER. If you do not accept the stated Terms of Use and Disclaimer, then do not use this website.

And the Children Slept 

An Independent Investigative study concerning reports of non-arousal by Children to Activated Smoke Detection Alarms and Related Research


BY:      Thomas M. Cunningham
U.S. Naval Academy Fire Department
        Annapolis, Maryland

  The text of this publication may not be reproduced, nor may talks or lectures based on material contained within the document given, without the consent of the author.

An applied research project

December 2002


            For years the fire service in the United States and various support organizations have attempted to spread the message through various public fire education programs that residential smoke detectors that are installed, have working batteries, and are tested monthly increase the possibility of individual(s) lives being saved by at least 50%. These claims are not unfounded, but rather are supported by statistical data collected through the NFIRS (National Fire Incident Reporting System) data collection system. Reports and statistics are put together and distributed by organizations such as the United States Fire Administration (USFA) and the National Fire Protections Association (NFPA).

  The message communicated to the public in varying degrees throughout the years during fire prevention week and public fire education programs has been that “Smoke detectors save lives.” This has been proven through statistical data showing a 50% increases in survivability if a smoke detector is installed and maintained. But there is also another side of reality, in that a simple and inexpensive smoke alarm may not save all the lives that were previously thought. Recent findings from various fire departments nationwide have found that children are sleeping through the alarms generated by an activated residential smoke detector.

  Studies in Wisconsin, Indiana, Texas, and California have all came upon the same results. There is visual evidence from the use of video to validate these findings that all four drills produced. The drills were originally conducted to show the effectiveness of family escape planning among school-aged children from 5 to 14 years of age.

  What the fire departments prevention staff and the news media witnessed was that family escape plans are as useless as the smoke detectors themselves if they do not audibly arouse a sleeping individual. A recent undisclosed study sponsored by the Allstate Foundation has validated these findings from recent testing. Fire prevention educators throughout the nation and internationally have all taken notice of these recent findings.

  While investigating these stories medical pediatric doctors were told of the findings and the medical experts were not surprised by these findings. The doctors questioned for these news reports all stated that children do experience deeper sleep architecture patterns than adults do. The research for this report has uncovered various studies that have been directly or indirectly related to smoke detector audibility and arousal in sleeping children.

  The following research questions were posed: 1) Examine the background, facts and exercises conducted by the North Shores, Fishers, Arlington, and Los Angeles fire departments that first revealed children failed to be awakened by audible alarms of residential smoke detectors. 2) Research previous or related studies conducted on this condition of non-arousal 3) Recommend short-term corrective measures until further research and studies can be conducted.

  The recommendations from this research report include:

  1)      Additions and modification to public fire education to include the “First up, last out” © criteria

  2)      That funding and grants be provided to examine the effectiveness of smoke alarms on all age groupings

  3)      That the public be made aware of nonarousal or arousal dysfunction of children during sleep.  

  4)      That parents should not panic over this condition, but rather make the proper corrections in case of a nighttime fire emergency

  5)      Reinforce that installation of smoke detectors increases your survivability by 50%, but that sleeping areas with individual detectors further increases reaction and survivability rates.

List of Contents

 - Abstract
 - Table of Contents
 - Background
 - North Shores F.D
 - Fishers F.D
 - Arlington TX, F.D
 - Los Angeles F.D
 - Stow F.D
 - NFPA Report
 - Research
 - Recommendations
 - Reference List 


 BRK electronics in 1969 designed the first battery operated smoke alarm to receive the Underwriters Laboratories (UL) listing. BRK Electronics would later launch the First Alert brand of residential smoke detectors as a branch of the Pittway Corporation. During the 1970’s the fire alarm division flourished as the fire service and related industries pushed the general public to install these alarms within their homes. This was seen as an effective and affordable way of reducing death from a fire. Over the past 23 years the research data suggest that all these claims have been right, that “smoke detectors do save lives.” As these alarms became more affordable, battery and AC powered, the quality of the product rose as well. By 1995 it was estimated that 93 percent of all homes in America were equipped with smoke detectors. But statistics also show that 43% of these installed detectors are nonoperational due to exhausted batteries, or have no batteries. In the mid 1980’s we saw the development of laws that required the installation of smoke detectors in new and existing dwellings. The fire service through public fire education efforts delivered the message to the masses concerning the benefits of installing smoke detectors in their residences. The public is made aware frequently that smoke detectors that are installed and maintained properly provide the best early warning system for the money to protect property and lives, providing people with the chance to escape a potentially dangerous fire and effects. 

            The effectiveness of residential smoke detectors can be seen through facts and data collected and analyzed by the United States Fire Administration or the National Fire Protection Agency. But data still shows that two age groups are still the largest loss of life and injury from the effects of fire (FEMA 1997). The two age groups in question that have the highest death rates are children and seniors. The facts concerning children are that they still die at averages higher than that of the general population. Annually 20% of all deaths by fire can be contributed to children having perished in a fire. Children under 1 y ear of age accounted for 2%, ages 1-4 account for approximately 12%, and children 5-9 account for approximately 6%. Estimates state that 2,800 children 14 years old and younger are injured each year by fire and at least 850 are killed in residential structure fires. Of these 70% is under the age of 10 and 40% under the age of five.


Figure 1. Average Percent of Child Fire Injuries and Deaths by Month. Source: NIFRS

There has been a decrease in the number of children’s deaths over the years. Children age 0-9 years of age have seen decreased data numbers In 1994 there were 940 deaths accounting for 23%, in 1995-96 these total deaths went down to 734 and 660 and accounted for 20% and 18 % of the population. Declines like this in fire deaths can be contributed to the increase in public fire education and the widespread use of smoke detector alarms in residential dwellings. The rate of death in preschool children is double what is the norm for the relative population combined. Studies statistics also show that young children may also have difficulty escaping burning structures even if a smoke alarm has been activated (Bruck 1998, Bruck Horasan 1995). 

Gender and ethnicity also varies of the rates of fatalities. Studies have proven that children of African American decent are more susceptible to death by fire than are white children (Duncan 1999). Winter months provide for the greatest number of deaths for children (figure #1). This is consistent with the fact that December and January are the two months that produce the greatest number of deaths for the general population. Having a working detector in your home at any time during the year doubles your chances for survival as shown in fire related data. 


North Shore Fire Dept.

The North Shore Fire Department, under the direction of Lt. Jonathon Cohn, conducted testing on two volunteer families. Test group one consisted of three males ages six, eight and 10. Cameras were installed in the bedroom area so that observations and reaction times could be recorded for accuracy, as well as the child’s reaction to the nocturnal emergency. Having fallen asleep, fire department personnel would then come into the home and with the use of non-toxic smoke generation machines, they would activate the residential fire alarms. The six-year-old male was awakened first, within a minute he was at his bedroom window preparing to exit the residence. The 10-year-old male was also awakened but returned back to sleep. The 10-year-old male was finally awakened after hearing the six-year-old call for him repeatedly. The eight-year-old was never alerted until both males verbally called for him. All three males got out in 4.5 minutes. This could have been disastrous if it were a real fire emergency. Fire research and modeling have concluded that within 4 to 4 ½ minutes a home can become a raging inferno making escape impossible. Due to this research safety organizations and officials have concluded that a home evacuation should take no more than two and a half minutes to execute and complete. The nearest smoke alarm for North Shore test was located in the hallway outside the boy’s bedrooms. A second test was conducted utilizing the same three children. In the second test, the smoke alarm sounded for 10 minutes before the first child woke up and alerted the others. The reaction times showed an increased deficiency during the second test.

In test group two, this study utilized five children. During the test, the 14-year-old was the first alerted. But before finally awakening to the activated alarm, the 14-year-old fell back to sleep twice. The 14-year-old evacuated the residence after five minutes had elapsed. The four other children were never awakened by the alarm. During the test the alarm could be heard from outside of the dwelling by fire department evaluators. All four children were finally awakened after a parent was sent into the room and awakened them. From the start of the test to completion total test time was 10 minutes. Interviews conducted afterward by fire department officials with the children uncovered two
startling facts: (1) that they never heard the activated alarms, and (2) when they did hear them they did not recognize the sound or what it meant.

A second test was conducted, and the results showed that within two minutes of activation two of the children were alerted. This demonstrated a marked improvement in the reaction time over the results in test session number one. With multiple alarms sounding, a third child was also alerted. Two of the children were never awakened after 10 minutes. An activated smoke alarm was then placed within one foot from the children’s heads. This action resulted in another other child being alerted. The youngest child who is nine years of age was never alerted to the danger. This is disturbing because the smoke alarms’ decibel range was measured at 85 dB; this is equal to that of a lawnmower.

Fishers Fire Dept.  

On 11/4/02 results from a second evaluation with similar results surfaced. The Fishers Fire Department of Indiana conducted a similar exercise under the direction of Deputy Fire Marshall Ron Lipps and his staff. The test and the results were shown on WRTV channel 6 in Indianapolis. Statistics show that annually 21 children lose their lives to fire in the state of Indiana and annually in the United States approximately 600 to 800 die. The Fishers experiment used two test subjects, both 8-year-old females. Both test subjects were shown what actions were to be taken if awakened by an activated smoke detector alarm. These actions included:  

·        Getting out of bed and to the floor to avoid the smoke layer over their heads.

·        Crawling on hands and knees to the door.

·        Feeling door with back of the hand to determine if the door was hot to touch.

·        If the door was cool, they were to open slowly and determine if escape through the house is a possibility.

·        If the child determined that escape through the house was not possible, they would then go to the bedroom window and attempt an escape through it.

·        Once outside they were to go to the designated meeting place with the other family members at the predetermined meeting site. 

Both test subjects were aware that this would be a test and that they would need to complete all the objectives to successfully pass the test. Test subject No.1 was asleep when Chief Lipps began his test. The smoke machine was started and the smoke detectors activated. The main problem noticed by evaluators immediately was that the girl was failing to awaken. The fact is that she did not even stir in her sleep. After a period of time the evaluators sent her mother into the room to awaken her. When the mother finally succeeded in awakening the child, she got out of bed and had forgotten everything that the evaluators had reviewed with her prior to this exercise. She did not crawl, instead she walked to the door and opened it. She then proceeded downstairs still groggy from having been awakened abruptly. She then walked into the garage area. This garage area is where all the gasoline and flammables were stored. She then stood there dazed and confused instead of proceeding to the designated meeting area.  If this had been a real fire, test subject No.1 would be another death statistic. Deputy Marshall Lipps decided that test subject No.1 would be tested again within the month. 

Test subject No. 2 passed the exercise on the first attempt. She completed all the objectives required. Participant No.2 was immediately alerted by the smoke detector and was awakened, she then performed the all task required in the allotted time period. The results showed Chief Lipps that test subject No.1 could possibly be the norm rather than the exception. Test subject No.2 had practiced the fire escape plan and showed good retention skills. But the second test on subject No.1 again provided findings considered to be disturbing to for Captain Lipps. 

Test subject No. 1 was revisited again. The smoke alarms were activated and the girl still was still not awakened by the alarm. The mother was sent in and on the third time of calling her name she was awakened. The mother had to verbally shout her name. This time the girl got up and completed the escape plan by performing all the functions required of her. This test subject failed on both test by failing to wake up without assistance.  During a real nighttime activation or nocturnal emergency the girl would now be dead. 

Arlington Fire Dept. 

The Arlington Fire department in Texas conducted a similar test to the North Shore and Fishers Fire Department. Arlington’s test aired on the NBC evening news with Tom Brokaw. Arlington had also discovered what all the fire industry has never known up until now, and that is an activated smoke detector alarm will not awaken most children to the danger of fire in the home. 

Arlington Fire Lieutenant Randy Ingram and staff supervised this drill with the help of volunteer families. The drill and its results were recorded by NBC5 KXAS in Fort Worth, Texas (11/5/02). Before the evaluation was to begin evaluators questioned the children concerning how they would perform when awakened by a smoke detector? The children in this test told evaluators that they would know what to do in event of a fire in their house. Using smoke generating machines with non-toxic smoke evaluators began to fill the residence. Once the drill had begun and the alarms sounded. None of the children were even awakened or alerted by what was occurring because they were still asleep. One parent was very disturbed about how their child could sleep through the sound of an activated smoke alarm. 

In every residence that was tested the detectors were activated for a period of six minutes and the children continued to sleep. Lt. Ingram stated that he never realized that the children would not awaken to the sound of smoke alarms, instead evaluators thought that it may take two to three minutes for the children to be awakened. 

This test also clearly demonstrated that these children did not respond to audible stimuli by the sound that was being generated by the alarms. They also failed to associate it with imposing danger. One test subject was awakened when a parent called their name. The child sits up on the bed, the room has the haze of smoke present and says “what?” In the case with two of the families in the test, neither activated smoke alarms or shouting of the children’s names had an effect on their arousal. One parent entered the sleeping area and attempted to awaken the children by prodding and telling them that there was a fire that they have to go. The parent then picks up the other child from the bed and this child continued to remain asleep. The child’s body resembled that of a bag of potatoes. This parent finally succeeded in alerting the children but did so after experiencing a certain degree of difficulty. 

The children had never heard the alarm but another point of interest was that they had never practiced their escape exercise with the alarms going off. Practicing an escape plan with the smoke detectors activated may be just one part of uncovering this puzzling dilemma that faces the fire industry. Humans are creatures of habit meaning that we become accustomed to doing things a certain way and it soon becomes habit. 

The families then practiced fire escape planning for two weeks before another evaluation would take place. The children’s rehearsals were done with the detectors activated. This was done in order to accustom the children with the sound of an activated smoke alarm. Lt. Ingram then conducted a second test. Smoke detectors were activated and the children continued to sleep. 

L.A. County Fire Dept. 

            LA County Fire Inspector Mike Brown has now conducted another test (11/20/02) with the assistance of TV station NBC4 in Los Angeles. The first test conducted by Insp. Brown used three children, one female, and two males. All three children were questioned prior to the exercise and the children felt confident enough that they would know exactly what too do if awakened by a smoke alarm. Smoke generators were used to activate the alarms and upon activation the children continued to sleep. The children were awakened only after the parents moved in to get an arousal response. The next test conducted in LA county used two males ages 12 and 10. Both boys practiced smoke alarm drills several times a year and felt confident that they would be alerted by an alarm. During this drill the 12 year old woke immediately and proceeded to the ten-year-olds room and alerted him that there was a need to evacuate. 

The results seen by these varying fire departments test all show similar results. That the audible alarms produced by early warning smoke detectors are not awakening children 5 years of age and older. 

Stow Fire Dept. 

The Stow Fire Department originally conducted a study of approximately 600 children before 1997 to determine if children would sleep through an activated smoke detector alarm. The results of their initial testing showed that 70% of the children tested were not awakened. The testing was done in-home by parents who activated smoke detectors while the children slept and reported their findings. The Allstate Foundation then awarded the Stow Fire Department with a grant of $2,000.00 in 1997 to conduct a more thorough study. The study required the purchasing of 10 detectors and 10 strobe-horn units. The study was performed on 76 children ages 14 and under. The testing required the installation of:  

In all six tests were performed. Utilizing all the different signals available greater than 55% of the children failed to be awakened. The best results came from the use of the horn type signal when placed inside the children’s sleeping area. The Stow fire department was hoping to achieve a 95% auditory awakening response rate among the studies participants. The test showed that two children under the age of five were responsive to the alarms, while 12 children between the ages of 10 and 14 were alerted during the test. 

Of the participants 11 of 76 were alerted when the detector was located in the hallway. The numbers increased significantly to 35 children awakened when detectors are located inside of the sleeping area.     

NFPA Report Statistics  

            According to the NFPA report on patterns of fire causalities in home fires by age and sex (Hall, 2001) we can view actual statistics that relate to children. Between 1994-98, 73% of all fatalities resulting in burns and smoke inhalation occurred in the home. Death by smoke inhalation accounted for 20% and death burns resulted in 5%. The national mortality database that maintains death certificates shows that 73% of fire deaths are due to exposure to byproducts of combustion. Statistics like this are indicative of the importance of why smoke detectors should be installed and maintained within residential dwellings. Detectors offer an early warning system to the occupants that significantly increase chances of survival in a fire by 50%.  

Nature of injury

Age 6-9

Age 10-19

Burns/smoke inhalation

132 (79.0%)

152 (71.6%)

Smoke inhalation Only

32 (19.1%)

46 (21.7%)




Ref 1. Source: National estimates on NFIRS and NFPA survey           

Ref. 1 shows that only three children ages 6-9, and 15 children 10-19 have died from some other cause other than burns and smoke inhalation. These age groups have very high numbers when compared to other age groupings that are old enough to save their own lives. The activity that the victims were involved in has also has an impacted this study. According to the NFPA figures, this studies statistics show that 41% of all victims were sleeping at the time of the fire. It also shows that 27.2% were attempting to escape from the emergency. Combine these figures and 68% of all victims were either asleep or attempting to escape. 

This raises two questions: 

 Unable to act accounted for 13.6% of the total. Children ages 10-19 that were attempting to escape accounted for 29.6% of these averages. For general knowledge concerning death by fire, children under the age of five are considered to be unable to act or attempt to escape either due to not having the physical or mental capacity to do so. 

Activity at the time of the fatality


Age 6-9

Age 10-19


94 (56.4%)

103 (48.3%)

Attempting to escape

47 (28.2%)

63 (29.6%)




Ref 2. Source: National estimates on NFIRS and NFPA survey 

Ref 2 shows a very disturbing figure. Mainly that 56% of deaths by children 6-9 years old occurred while they were asleep. 48% of 10-19 years of age also show a high number of deaths occurring when sleeping. We know by use of statistical data that most of these deaths occur in homes were there are no smoke detection devices or devices that are not operating. What we do not know is how many of these children slept through alarm activations because statistical data does not take this into account, thus showing clearly that a problem as arousal dysfunction exist. 

Approximately 40% of all fire deaths occur in homes between the hours of midnight to 6am. This is at hours when most people are asleep, only 29% occurred between 9pm and midnight and between 6am to 9am. Ref 3 Shows that the averages between the two age groups. In the 6 to 9 age group, 42 out of 167 death occurred between midnight and 6am, while 86 deaths occurred between 6am and midnight. In the 10-19 grouping we see a difference. 123 children died between midnight and 6am, and between 6am and midnight 86 perished. Here we also see another higher death rate while in the sleep hours.   

Time of Fire

Time of Fire

Ages 6-9

Ages 10-19

12:01am to 3am

40 (24.2%)

63 (29.5%)

3:01am to 6am

42 (25%)

60 (28.4%)




Ref 3.  National estimates on NFIRS and NFPA survey 

Even though 94% of American homes are equipped with some sort of early warning smoke detection device (1996) only 42% of fires occurred in homes without an early warning device. It has been estimated that fire deaths are 40-50% lower in homes equipped with smoke detectors. In the homes that have smoke detectors approximately one-third are not operating due to dead or missing batteries. Statistics clearly show that children are more likely to be sleeping at the time a fire occurs than are older individuals. A 1994-98 study by the NFPA shows that out of 577 children ages 6-9, that 261 or 45.2% perished due to being asleep at the time. 775 out of 2,264 equaling 34.2% of children 10-19 died while sleeping.  


After having conducted preliminary research into this problem concerning children sleeping through activated fire and smoke alarms. The facts have uncovered interesting and provocative data to confirm what these various departments have discovered in every one of their videotaped test sessions. 

A motivated response can be solicited from a known stimulus during sleep. Grace (1997) looked at the effect of motivational responses to smoke detector alarms and their waking effectiveness. This study determined that an individual is able to distinguish between sounds recognizable and unrecognizable to the participants in the study while they are in the sleep state. Could signals or sounds produce an emotional response from the participants? Grace in this study determined that people could distinguish between sounds and signals while in the sleep state. The use of a persons name as the stimulus was found to have had the best response averages overall. Individuals who are sleeping can be prepared to respond to the sound of an activated smoke alarm. The effectiveness of the stimuli depends upon the emotional content. This is important when testing concluded that subjects could be awakened from their sleeping state. Motivational response depends upon the incentive. Grace (1997) concludes that self-programming can be accomplished by the individuals to awaken to certain types of stimulus to a certain extent. 

A report (Busby & Pivik, 1983) on “Failure of high intensity auditory stimuli to affect behavioral arousal in children during the first sleep cycle.” This study examined children 8-12 years of age. This research attempted to determine children’s arousal to auditory stimuli occurring during the first cycle of sleep to further understand both physiological and behavior indices. The test sessions were conducted to study the effects on hyperactive children using non-hyperactive controls. During this testing researchers used 123 dB sounds (pressure level of 0.0002 dynes/cm2) and obtained no auditory response or sustained awakenings. Researchers conducting the study consider this to be 90 dB above normal waking threshold values. 50% of stage 2 and 25% of stage 4 sleep produced partial response or momentary physiologic arousal in the children. Any arousal that occurred was short-lived in duration and the participants immediately returned to the Sleep state. With continuing or increased auditory stimulus intensity there was still no response. 

A second study (Busby & Pivik, 1985) “Auditory arousal thresholds during sleep in hyperkinectic children” also produced results similar to those divulged in Busby & Piviks first report. Hyperkinetic is defined as a “general restlessness” or “excessive movement” such as those characterizing children with attention deficit disorder (ADD) or hyperkinesis.  

Arousal thresholds have usually been determined through the use of auditory stimuli with various criteria to determine arousal state. The criteria (Bonnet MH, Johnson L.C 1978) for subawakening indications including those considered being physiological and confirmed by specific behavioral responses (Price LJ, Kremen I. 1980). Studies  (Zimmerman WB 1970) have indicated that intensity of stimulus, sleep stage, and time of night produce variables that are important determinates in the arousal process (Keefe FB, Johnson L.C., Hunter EJ 1971). Busby & Pivik attempted to determine the auditory response in “hyperkinetic” children to auditory stimuli during various sleep cycles. Researchers were attempting to find evidence of sleep related arousal dysfunctions. Researchers attempted to prove that children suffering from ADD (attention deficit disorder) have associated symptoms that include impulsiveness, deficits in attention, motor control, and arousal dysfunction. When conducting this study, researchers used three control groups that included medicated hyperkinetic children (MHK), non-medicated hyperkinetic (NHK) children, and control children. The control group were evaluated and cleared as not experiencing hyperkinetic symptoms and who were not on any type of medication. 

            The study also provides startling results. Researchers found a behavioral response inability in all three-study groups to be aroused by intense audible signals up to 123 dB. They were also astonished that the group that had the most responses was the NHK group. Twenty-four preadolescent (8-12 years old) male subjects participated in this study. The participants would be subjected to auditory stimuli for durations of 3 seconds on, 3 seconds off. The auditory stimulus was delivered through the use of earphone inserts. To produce a response from the participants, signal delivery would be intensified by increments of 2-5 dB. These signals would be delivered starting within the waking threshold values determined for children. The study showed children experienced arousal dysfunction at intensities up to 123 dB. This is 90-100 dB above the normal waking threshold levels. 

A non-response by the participants occurred in 227 out of 432 attempts, this equals a 52.5% failure rate. 13% of the participants had non-sustained physiological arousals, these were short in duration and in all cases the participants returned to their sleep state even with continued and increasing auditory stimulus. Responses were produced from the participants either vocally or by the participant pressing a button to acknowledge that they were aroused. These arousals occurred in 148 out of 432 attempts, equaling a 65.7% failure ratio. The study did show that arousals from REM sleep exceeded those of stage 2 sleep. 

The control (normal) children while in SWS (slow wave sleep) had no response to 23 attempts over 100 dB. In stage 2 sleep 28 out of 89 (31.5%) responded at 111.1 dB. In REM sleep there was 26 out of 47 (55.3%) at 107.5 dB. The total arousal/attempts for the control children resulted in 54/159 or a 34% response rate. 

The results of this study suggest that the intensity to elicit arousals among children are much smaller and higher than comparable values found in research studies involving adults. Researchers believe that this is due to the absence of cortical or behavioral arousal associated with parasomnias such as sleepwalking, sleep talking, or enuresis (bedwetting). What was also noted in this study is that it was virtually impossible to gain a positive response during the first cycle of sleep. There was also a noticeable diminution in overall arousals demonstrated throughout the night by the participants.

Researchers have concluded that a dysfunction related to “waking underarousal,” which will assuage the increase of sensitivity to auditory stimulation normally associated with sleep. 

Another report, (Duncan.1999) from the University of Canterbury in Christ Church New Zealand gives interesting insight to the current phenomenon that has been uncovered concerning underarousal in children. “The Effectiveness of the Domestic Smoke Alarm Signal” was released in March of 1999. This study was conducted to determine the response times to alarm activations and to determine what age groups were at the greatest risk from the effects of fire. The focus of the study was to examine whether inexpensive smoke detector alarms could perform their function, and to what degree of effectiveness the early warning signal alerted the occupants in residential dwellings. The study also provided recommendations on minimizing risk to occupants. 

New Zealand fire fatality statistics show that the highest number of fatalities occur to 20-29 year olds, and “children” under the age of nine. The study revealed that children under the age of nine experience the deepest sleep of all the age groups that were studied. That the time actually spent in deep sleep (stage 3 sleep) shows to be the most detrimental to a child being alerted by an activated smoke detector alarm. 

Twenty juniors from six to 17 years of age participated. The results showed that 85% of these participants slept through one or both of the alarm activations. The study also showed data that adults exposed to the same stimulus were awakened by the alarm signal 100% of the time. To give validity to the study, smoke detectors were activated between the times of 2am-6am. This was determined by utilizing fire data that showed that these were the hours when most fire related fatalities occur within a residential structure. 

The study performed 229 individual exposures to activated smoke alarm event. 26 participants (74%) reported not being alerted by the smoke alarm, and 18 (69%) cases were children under the age of ten. Studies show that children ten years of age and under may need a stronger stimulus to be alerted when sleeping. Eight of the studies participants were under the age of ten (from 1 to 9 years of age) and of equal sexes (4 males and 4 females). All participants were exposed to at least one alarm activation. The decibels in the children’s rooms were measured as being 75dBA when doors were left open. With the doors closed the audible sound was measured at 67dBA. These dBA ranges were not significantly different from what was measured in the parent’s bedrooms. 

This decibel range was enough to cause arousal in the adults. This test resulted in only one child (age 9) wakened by the alarm activation. The result of this study shows that 87.5% of the junior participants failed to be aroused when exposed to two smoke detector activations. The study shows conclusive evidence that children cannot be reliably awakened by the signal generated by a domestic smoke alarm within a residential structure. If an activated smoke alarms sounds at night this study suggest that children will not be alerted and will not respond accordingly. 

A case study showed that a 3 year old male child exposed to 5 activations by an alarm measured at 65dBA (detector in hallway) and 91dBA (detector inside of bedroom) failed to be alerted by the smoke alarm activation. The location of the detectors failed to give a positive result in this case study. 

The study concluded that: 

·        Test alarms were generally terminated in less than thirty seconds.

The Fire Code Reform Research Program (March, 1998) released a study on sleep arousal in children and adults from smoke detectors. Another study was conducted at Victoria University in Melbourne Australia were Professor Dorothy Bruck is on staff at Victoria’s Department of Psychology. The study gives insight into the ineffectiveness of inexpensive residential smoke alarms when attempting to arouse children from sleep. 

The focus of this study was on whether children could be awakened to smoke detector activations at 60 dBA when the detector is located in a hallway and the bedroom door is left open. At the time of this study this was the standard location placement of smoke detectors that was being advocated by fire organizations. Another study (Nober, et al 1981) used adult participants. The Nober study showed that adults could respond within 21 seconds to a smoke detection alarm activation during sleep. 

Participants in this study were between 18 to 24 years of age. They failed to be awakened 30% of the time after exposure to a 60dBA smoke detector alarm (Bruck, D., Horasan, M.1995). This is consistent with studies that show children and young adults experience longer deep sleep (slow wave sleep) periods than adults and seniors. Prior to this there had been no studies performed concerning children’s arousal performance to smoke detector activations.

What is also noted in this study is the fact that in newer homes we also see a significant change in design. The change occurred in the distance between the sleeping areas of residential structures. Distances between children’s and adults sleeping areas have grown distant when compared to that found in older structures. In event of a fire this may add to a minor delay in alerting or rescuing children if they fail to awaken from their sleep state independently. 

For this study 20 junior participants were recruited and used. The ages of the participants ranged from 6-17 years. There were 9 males and 11 females for an average mean of 11.6 years. 

All test subjects were screened for problem factors such as hearing, sleep patterns, illness, etc. All were found to be healthy and within the parameters predetermined as acceptable by the evaluators. Participants were also required to undergo other testing to determine suitability for participation. All alarms activations would occur between the time of 1:00AM and 4:30AM. 

Testing was conducted and the results showed that there existed a noticeable problem among the younger participants. In the junior age group (6-17 years) 69.3% or 2/3 of the group slept through the three-minute activation period. This average was determined over a two-night period. 

Three of the twenty children in the test were awakened by the alarms on both nights leading to results that showed the unreliability of smoke detectors to awaken children. Six children were awakened during one night of testing, but all other eleven participants slept through the smoke detector activations on both nights. The studies final results showed that 85% of the junior participants failed the test when audible alarms at 60 dB were used. The 16 and 17-year-old participants were alerted on both nights of the test. 

Another published study (Bruck, 1999) examined non-awakening in children in response to a smoke detector alarm. This study was published in the Fire Safety Journal no.32. The study was conducted in response to the need for further research on children’s non-arousal to activated smoke detectors as loud as 60 dB. The testing would occur at approximately midpoint during the households sleep period. The time coincided with the documented times when the greatest number of fire fatalities occur within residential dwellings. The time chosen was between midnight and 6 am. The study utilized 20 child participants between the ages of 6-17 years of age. In this group there were 9 males and 11 females with a mean average age of 11.6 years. 

The study concluded with the findings that children would not reliably awaken to a smoke detector that was located in a hallway. 85% of the participants failed to awaken on either night of the test. This shows a definite need for detectors to be located within the sleeping area as well. 

These results were received during the middle third of the children’s sleep period. This period was selected because the greatest degree when non-responsiveness would occur during the first third of a child’s sleep period. Measurements were taken prior to testing to ensure that all participants received a 60 dB signal at pillow level + or – 3 dB. This study gives creditability to the fact that severe implications to children can arise from a nocturnal fire emergency. 

Researchers noted that due to the amount of deep sleep that children experience, a smoke detector alarm cannot and should not be relied upon to awaken children, even if the detector is rated at 85 dB. A reaction could possibly be solicited if a detector is placed close to the child’s head, but this is practically unacceptable.  



Ahrens, M. Experience with smoke alarms, Quincy MA. NFPA. Jan 2000. 

Berry, C. H. (1978) Will Your Smoke Detector Wake You?, Fire Journal, July, 105-108. 

Bonnet MH, Johnson L.C., Relationship of arousal threshold to sleep stage distribution and subjective estimates of depth and quality of sleep. Sleep 1978,1:161-8. 

Bruck, D. (1998) Non-awakening in children in response to a smoke detector alarm,  Fire Safety Journal, 32.  

Bruck, D., Horasan, M. (1995) Non-arousal and Non-action of Normal Sleepers in Response to a Smoke Detector Alarm, Fire Safety Journal, 25, 125-139. 

Busby K, Pivik RT. Failure of high intensity auditory stimuli to affect behavioral arousal in children during the first sleep cycle. Pediatric Res. 1983;17:802-5 

Duncan, C. The Effectiveness of the Domestic Smoke Alarm Signal. Fire Engineering Research. 1999. 

Feinberg L. Changes in sleep pattern with age. Psychiatr Res+ 1974; 10:283-306 

Federal Emergency Management Agency, Children and Fire in the United States: 1994-1997. United States Fire Administration, 1998. 

Grace, T. (1997) Improving the Waking Effectiveness of Fire Alarms in Residential Areas, Fire Engineering Research Report: 97/3, Canterbury University. 

Hall J.R. Jr. Burns, toxic gases, and other hazards associated with fires, Quincy, MA: Fire analysis & research division, March 2001 

Hall J.R. Jr. Patterns of Fire Causalities in Home Fires by Age and Sex. Quincy, MA: National Fire Protection Association, Fire Analysis and Research Division 2001. 

Hall J.R. Jr. Children Playing with Fire. Quincy, MA: National Fire Protection Association, Fire Analysis and Research Division; 1999 

Hume B. Sitting of domestic smoke alarms. Joint committee on fire research, UK. 1997; research report number 71. ISBN 1-85893-761-2 

Hutt SJ, Hutt C, Lenard HG, Von Bernuth H, Muntjewerff WJ. Auditory responsivity in the human neonste; nature 1968; 218:888-90 

Keefe FB, Johnson L.C., Hunter EJ, EEG and autonomic response patterns during waking and sleep stages. Psychophysiology, 1971; 18:198-212 

Nober, E. H., Peirce, H., Well, A. (1981) Waking Effectiveness of Household Smoke And Fire Detection Devices, Fire Journal, July, 75 #4, 86-130. 

Price LJ, Kremen I. Variations in behavioral response threshold within the REM period of human sleep. Psychophysiology, 1980; 17:133-40 

Rechtschaffen A, Kales A (eds). A manual of standardized terminology, techniques and scoring systems for sleep stages of human subjects. Los Angeles, California. Brain Information Service/Brain Research Institute, University of California at Los Angeles.1968 

Robinson MD, Seward PN. Hazardous Chemical Exposure in Children. Pediatr Emerg Care. 1987; 3: 179-183 

Schmidt K, Bims B. The behavioral arousal threshold in infant sleep as a function of time and sleep state. Child Dev. 1971;42:269-77 

Zimmerman WB, Sleep mentatron and auditory awakening thresholds. Psychophysiology, 1970; 6:54-69   



Permission is granted to create WWW pointers to this document. All other rights reserved.

This document is maintained by the CFAA WEBMASTER. Please report broken links.

Copyright © 1999 CFAA