Perception and recollection of fire hazards in dwelling fires

FSEG have published a new journal paper concerning the perception and recollection of fire hazards in dwelling fires. The paper is important for two reasons:

• It demonstrates that the recollection of fire hazard size can be improved through the use of layperson-friendly descriptors, both text and image based. This has implications for fire investigation, in particular when witnesses are interviewed concerning their experiences. 
• It suggests that while individuals are less willing to attempt to extinguish a fire or enter a smoky room as the size of the fire and the amount of smoke increases, a worryingly high proportion of people are prepared to tackle the largest fires or enter smoke filled rooms. This observation could explain why so many people are injured in dwelling fires and suggests that we must rethink our public education messaging concerning the dangers of fire.

You can download the paper for free using the following link until December 29 2019. https://authors.elsevier.com/c/1a1f03IVV9glkC

 Our current understanding of dwelling fire injury outcomes is impacted by data limitations, confounds, and failures to adequately examine occupant behaviour. For instance, research rarely considers: occupant perception of fire hazard properties (e.g. size of flames/smoke when first encountered); resultant engagement (enter smoky room, tackle flames); whether hazard size percepts are accurate when recollected for investigators; and what the best recollection method is.

Prof Ed Galea and Dr Lynn Hulse making final preparations in test kitchen

Two experiments (N=141, 132) presented short videos of kitchen fires where hazard size was either Small, Mid or Large.

The fire compartment, a residental kitchen

Immediately after seeing this (Experiment 1), or after a delay (Experiment 2), participants’ performance at recollecting hazard size and their willingness to (hypothetically) engage with the hazards was tested. 

(a and b) small, (c and d) mid, (e and f) large fire and smoke conditions

Recollection performance was compared across three methods,

•  Free Recall
• With the aid of lay-person-friendly text descriptors
• With the aid of lay-person-friendly picture descriptors. 

Text and image descriptors

The results of the recollection experiments can be summarised as follows:

• Free recall resulted in the poorest recollection performance (35% for flames and 19% for smoke).
• Recollection performance improved by 2–3 times when using the two types of descriptors that allowed hazard size to be referenced to other scene elements (text: Flames 61%, Smoke 49%, pictures: Flames 66%, smoke 55%).
• Picture descriptors had a slight advantage over text descriptors.
• Larger hazards were recollected less accurately than small ones, albeit still somewhat meaningfully; the exception was mid-sized smoke possibly due to narrowing effects.

Frequency of recollocection performance for instant recall

Concerning the willingness to engage with the fire:

• An important observation was that while increased hazard size reduced willingness to engage, a concerning percentage of participants nevertheless considered engaging with the largest hazards.
• Concerning the flames: for the immediate recall group, 27% suggested that they were prepared to attempt to extinguish the largest flames, while 32% said that they were not sure. So 59% were either not sure or would attempt to tackle the large flames.
• Concerning the smoke: for the immediate recall group, 18% suggested that they would enter the room with the greatest amount of smoke, while 32% said that they were not sure. So 50% were either not sure or would attempt to enter the smoke filled room.
• The delayed recall groups had similar overall results with a higher proportion suggesting that they would attempt to extinguish the flames or enter the smoke filled room with a smaller proportion being uncertain.
• Such risky behaviours may explain injury outcomes in fires.

View during experiment

• Prior fire experience and gender affected recollection and willingness, often interacting with hazard size. Delayed recollection and individual differences did not.

Willingness to engage with hazards

These findings suggest occupant behaviour, characteristics and hazard size data need capturing to help assess fire injury risks. These two experiments demonstrate that it may be possible for people experiencing dwelling fires to recollect, even after a delay, hazard properties such as the size of flames and smoke when first encountered and their behavioural responses to that.

 However, an important finding from this work is that the recollection of hazard size is likely to be more forthcoming and meaningful when people are provided with tools to help them, and the investigators, visualise the hazards in relation to other elements of the scene, i.e. the occupant’s body and the built environment.

 Larger hazards may result in less accurate recollections, in terms of precision due to attentional narrowing potentially, but may still be relatively meaningful (if describing flames at least). 

 Larger hazards may also deter behaviours that would bring the occupant into closer proximity with them and thus more danger. So risks posed by hazard size do appear to be perceived, to some extent by a proportion of the public. However, it is also important to be aware that, based on the findings here, the number of occupants deterred by larger hazards may be relatively small and the results on this matter raise concern.

 It would appear that certain factors – linked to occupant characteristics such as gender and prior personal experience of a fire but not individual differences in risk-taking and locus of control – will lead some people to nevertheless attempt to tackle flames of a height beyond the control of an untrained individual or enter a room containing sufficient smoke to impair one’s functioning and health. Such behaviours in turn may explain, or at least contribute to, the high incidence of dwelling fire injuries.

Thus, if non-fatal injury outcomes are to be better understood and the number of such injuries reduced, the links between hazard size, occupant behaviour, occupant characteristics, and the likelihood of being injured need to be researched further in domestic settings.

 All comments welcome. You can download the paper for free using the following link until December 29, 2019: https://authors.elsevier.com/c/1a1f03IVV9glkC

The full citation for the paper is:

Lynn M. Hulse, Edwin R. Galea, Owain F. Thompson, David Wales; Perception and recollection of fire hazards in dwelling fires, Safety Science, Volume 122, Feb 2020, 104518. https://doi.org/10.1016/j.ssci.2019.104518

 Acknowledgements

This work was supported by Innovate UK and the EPSRC as part of a Knowledge Transfer Partnership between Kent Fire & Rescue Service and the University of Greenwich (KTP number 9153) and the LIFEBID (Lessons In Fire & Evacuation Behaviour In Dwellings) project.