Human Behaviour in Smoke

People often underestimate the risks posed by smoke. They do not evacuate rooms or locations filled with heavy smoke or do not do so with the required speed, or they even consciously choose to go through the smoke. This article explains why many people behave in this manner.

Human Behaviour in Smoke: People often underestimate the danger and do not immediately leave a building or a room despite heavy smoke.
Fig. 1: People often underestimate the danger and do not immediately leave a building or a room despite heavy smoke. (Source: FeuerTrutz)

By Dr. Laura Künzer, Dr. Gesine Hofinger. Around 400 people die every year in German in fires. 70% of the victims are taken by surprise by a fire at night while they are asleep. According to experts, it is the smoke, however, that represents the greatest danger and not the fire: As few as three to five breaths can be deadly [1]. 80 to 95% of fire victims are not killed by the flames but rather by the smoke gases.

Even when thick smoke has developed, people are not usually awoken by the smell of smoke or because they have difficulty breathing. Smoke thus poses an acute danger to life. This seems obvious to members of the fire services but people who are not used to dealing with smoke nevertheless underestimate its dangers and effects [2]. Findings from practical experience and research demonstrate that people by no means always avoid smoke – contrary to many expectations. However, there is hardly any data on which behaviour is exhibited, how often it is exhibited and under which circumstances.

The effects of smoke need to be considered from a systematic perspective because various different factors influence one another. Alongside the local conditions, physical fitness and health also play a role. These factors appear to improve human resistance to toxic smoke, while smoking cigarettes reduces it. The inhalation of smoke is also dependent on increased respiration. Pregnant and physically active persons (e.g. those fleeing a dangerous situation in a stressed state) have a higher rate of oxygen consumption and breathe more quickly than calm people. As a result, the inhale higher volumes of smoke. These and other physiological limitations in different groups of people, e.g. due to old age or restricted mobility, are not systematically investigated in relation to behaviour when faced with smoke in the relevant literature. Yet these people, on the one hand, move more slowly and are thus exposed to any possible toxic effects for longer, while on the other hand, they are more vulnerable.

Understanding the impacts of smoke and vulnerability does not mean that we understand how humans behave. Therefore, the findings from research and practical experience about the behaviour of people in fires with smoke will be discussed below. By way of illustration, we will use the behaviour of passengers exposed to the fire at the underground railway station ‚Deutsche Oper‘ in Berlin in 2000.

The article was published in FeuerTrutz International, issue 1.2018 (January 2018).
More information about eMagazine FeuerTrutz International

The behaviour of people faced with smoke

In the scientific literature and in media reports, there are numerous events described in which people have remained in smoke for a long time or moved through smoke ([3],[4],[5]). The evaluations are based on fire events spanning more than 40 years and thus give a good impression of human behaviour in smoke. Despite evidence to the contrary from practical experience and research, the ‚myth‘ that people will not choose to move through smoke still perpetuates even amongst the emergency services themselves [6].

People will choose to remain in smoke or move through it when they personally have a ‚good‘ reason to do so. They will even do this when there is a clear deterioration in the conditions ([7],[5]).

An evacuation is often necessary in the event of a fire accompanied by the development of heavy smoke. When analysing the behaviour of people, two phases of the evacuation need to be considered: The pre-movement time or response time describes the timespan from when the event is first noticed, e.g. by smelling the smoke or hearing a fire alarm ([8],[9]). This is followed by the movement time. In both phases, various factors and motives play a role in behaviour.


During the Love Parade in Berlin in 2000, a fire broke out at the underground railway station ‚Deutsche Oper‘. The last carriage of a subway train caught fire near to the (at the time) only exit from the underground railway station. Some of the around 350 people present did not take the danger posed by the smoke seriously despite the development of heavy smoke on the platform, even though the other passengers become agitated [10]. They remained on the platform, responded to the development of smoke with applause or even went back into the smoke-filled subway carriages to take photos. A summary of the fire brigade’s incident report can be read in [10].

Behaviour during the response phase:

People need information

People need information in an exceptional situation in order to decide what they want to do [11]. The decision to leave a smoky area can be viewed as a risk assessment based on personal knowledge, experience and the processing of information about the situation.

Human Behaviour in Smoke: An evacuation is often necessary in the event of a fire accompanied by the development of heavy smoke
Fig. 2: An evacuation is often necessary in the event of a fire accompanied by the development of heavy smoke. When analysing the behaviour of people, response phase and movement phase have to be considered. (Source: Künzer, Hofinger)

People thus have to initially perceive the danger and then understand that it concerns themselves. In the example of the ‚Deutsche Oper‘, the smoke itself evidently was not a sufficient indicator of the danger posed to health and the passengers remained in the station.

In many hazardous situations, people lack information to make a sensible, correct decision. This is especially the case when they didn‘t listen to alarms and warning messages, considered them as irrelevant, or didn‘t understand them. They then start to actively look for further information ([9],[11]) because they want to know what is going on. This information-seeking requires time and can delay leaving a smoke-filled area. Communication and information about the dangers and the design of warning signs and alarms can thus play a decisive role for influencing behaviour ([12],[13]).

Motivation determines behaviour

Individual motivation is also relevant for the decision to evacuate. In the example of the ‚Deutsche Oper‘, it appears that the motivations of amusement and curiosity were predominant [14], which were supported by the party atmosphere and the craving for sensation amongst the visitors to the Love Parade at the time. The smoke was seen as part of the entertainment programme rather than as a danger.

People sometimes decide against evacuation despite the presence of smoke because they do not want to interrupt an of action [3]. In the incident at the ‚Deutsche Oper‘, some passengers continued to wait for the next underground train until it became absolutely clear that services had been suspended.

On the other hand, people avoid uncertainty: Most passengers are not aware that tunnel systems between underground train stations, as well as emergency exits in the tunnels, are escape routes. This uncertainty also encouraged people in the aforementioned example to decide against evacuation through the tunnel and remain in the smoke.

Social influence on decision making

Indications of danger can come from the environment (smoke, fire). Yet other people can also be a source of information. An example is the so-called herding behaviour which means that people tend to do what others do [15]. If most of those present stay on the platform despite the development of smoke, a person can (incorrectly) interprete this as ‚those people are staying so it can‘t be too dangerous for me‘ and underestimate the danger. Social pressure to conform can also work against an evacuation: ‚If everyone else is staying, I’m not going to be the one to run away!‘ Social influences can mean that people will even choose to move through the smoke [16].

Behaviour during movement phase: Returning to/moving through smoke

The reasons given in studies for remaining in a smoke-filled area included e.g. gathering information and firefighting or to warn and help other people. In this context, there are differences between the genders [2]: Men tended to gather information, while women were quicker to alert the fire services and rescue relatives.

People even often go back into a building through the smoke to help others [17]. The person‘s own ‚social role‘ is relevant here [8], e.g. parents go back through smoke to rescue their children. Returning to smoke-filled areas can be explained by emotions. In case analyses, people also returned to fetch objects such as their jacket [7].

The exceptional circumstances and the poor visibility caused by the smoke can induce fear [18] and lead to stress [20]. People under stress often find it difficult to reflect on alternative courses of action or to seek new options. A smoky but familiar route thus appears attractive to many people as an escape route. In addition, indirect routes are aversive [20]: A short route through the smoke is therefore considered the better option than a longer (and unknown) route e.g. the escape route through an underground railway tunnel.

In a scientific study, people went down a smoky corridor when asked to do so even though they felt uneasy about it [18]. Social pressure and ‚incorrect‘ leadership can thus encourage people to move through the smoke. In the example of the ‚Deutsche Oper‘, leadership had a positive effect: The passengers present began to remove themselves from the smoke when the fire services (with respiratory apparatus) assumed leadership and initiated the evacuation through the tunnel.


People choose to remain in smoke or move through smoke when they have ‚good‘ reasons to do so. These reasons are of a psychological nature: a lack of knowledge about the dangers of smoke, social motivations, etc. Although research focuses above all on explanations for this behaviour, the findings can be utilised to ensure people are evacuated from smoke-filled areas more quickly and to prevent them moving through the smoke:

  • Providing knowledge about the dangers of smoke (ahead of a fire)
  • Providing information about the danger while smoke is developing
  • Understanding and using social mechanisms
  • Designing alarms and warnings clearly and with a sufficient level of urgency
  • Giving leadership (via personnel) and thus also helping to reduce uncertainty.


Dr. Laura Künzer and Dr. Gesine Hofinger: Qualified psychologist, partner in the HF Team– Human Factors Research Consulting Training (Ludwigsburg, Germany); research associates at the Friedrich Schiller University Jena. Areas of activity include humans and safety, crisis management, evacuations, designing warnings and announcements, visitor safety for large events. They publish, give lectures and teach at various universities and teaching institutions. ( )


[1] Blomeyer, R. (2010). Fünf Atemzüge reichen zum Sterben . BRANDaktuell (27/10). Zugriff am 26.04.2015.

[2] Hofinger, G. (im Druck). „f (risk)“ – Wann und warum gehen Menschen durch Rauch? Projektinterner Bericht im Forschungsprojekt ORPHEUS [bei den Verfasserinnen erhältlich]

[3] Bryan, J. L. (2002). Behavioral Response to Fire and Smoke. In P. J. DiNenno (Hrsg.), SFPE Handbook of Fire Protection Engineering (3rd ed., 3-315-3-340). Quincy, Mass.: National Fire Protection Association; Society of Fire Protection Engineers

[4] Fahy, R., Proulx, G. (2002). A Comparison of the 1993 and 2001 evacuations of the World Trade Center . Proceedings – Fire Risk and Hazard Assessment Symposium. Zugriff am 12.02.2013.

[5] Wood, P. (1980). A survey of behaviour in fires. In D. Canter (Hrsg.), Fires and human behaviour (83–95). Brisbane: John Wiley and Sons

[6] Künzer, L. (2015). Mythen der Räumung und Evakuierung. FeuerTRUTZ (4), 44–47

[7] Fahy, R., Proulx, G. (2009). ‘Panic’ and human behaviour in fire . Zugriff am 05.12.2012.

[8] Kuligowski, E. D. (2016). Human Behavior in Fire. In M. J. Hurley, D. T. Gottuk, J. R. Hall Jr., K. Harada, E. D. Kuligowski, M. Puchovsky et al. (Hrsg.), SFPE Handbook of Fire Protection Engineering (2070–2114). New York, NY: Springer New York; Imprint; Springer

[9] Tubbs, J. S., Meacham, B. J. (2007). Egress design solutions. A guide to evacuation and crowd management planning. Hoboken, NJ: John Wiley & Sons

[10] Kircher, F. (2000). Feuer im U-Bahnhof Deutsche Oper. Brandschutz (9), 883–893

[11] Hofinger, G., Künzer, L., Zinke, R. (2013). „Nichts wie raus hier?!“. Entscheiden in Räumungs- und Evakuierungssituationen. In S. Strohschneider & R. Heimann (Hrsg.), Entscheiden in kritischen Situationen. Umgang mit Unbestimmtheit (249–263). Frankfurt/M.: Verlag für Polizeiwissenschaft

[12] Fitzpatrick, C., Mileti, D. S. (1994). Public risk communication. In R. R. Dynes, K. Tierney (Hrsg.), Disasters, collective behaviour, and social organisation (71–84). Newark: University of Delaware press

[13] Wogalter, M. S. (Hrsg.). (2006). Handbook of warnings (Human Factors and Ergonomics). Mahwah, NJ: Lawrence Erlbaum Associates

[14] Künzer, L. (2009). Dynamische Fluchtwegslenkung – Anforderungen aus psychologischer Sicht, Unveröffentlichte Diplomarbeit. Universität Regensburg

[15] Helbing, D., Farkas, I., Vicsek, T. (2000). Simulating dynamical features of escape panic. Nature, 407 (6803), 487–490

[16] Latané, B., Darley, J. M. (1968). Group Inhibition of Bystanders Intervention in Emergencies. Journal of Personality and Social Psychology, 10 (3), 215–221

[17] Fahy, R. F., Proulx, G (1997). Human Behaviour in The World Trade Center Evacuation. In Y. Hasemi (Hrsg.), Fire Safety Science – Proceedings of the Fifth International Symposium (713–724). Boston, MA: International Association for Fire Safety Science

[18] Jin, T. (1997). Studies on Human Behavior and Tenability in Fire Smoke. Fire Safety Science – Proceedings, 5th International Symposium (3–21).

[19] Proulx, G. (1993). A stress model for people facing a fire. Journal of Environmental Psychology, 13 (2), 137–147

[20] Helbing, D., Molnár, P., Farkas, I. J., Bolay, K. (2001). Self-organizing pedestrian movement. Environment and Planning B: Planning and Design, 28 (3), 361–383

The article was published in FeuerTrutz International, issue 1.2018 (January 2018).
More information about eMagazine FeuerTrutz International

  • FeuerTrutz International

    Our free newsletter will regularly provide you with information on new issues of "FeuerTrutz International" and important news from the industry, the specialist media offers from Rudolf Müller Mediengruppe and products offered by our advertising partners. By providing my e-mail address and clicking on the button "send", I consent to you providing me with this information by e-mail. This consent is valid until it is withdrawn, which I am free to do at any time. I can deregister from the newsletter at any time via the unsubscribe link displayed in the newsletter.

    Fields marked with * are required

    Please take note of our Privacy Policy.