Development of a Fire Safety Training Tool for Staff in Retail Stores

DEVELOPMENT OF A FIRE SAFETY TRAINING TOOL FOR STAFF IN RETAIL STORES

Dmitry A Samochine, Jim Shields 2 , Karen Boyce 2

1 Academy of State Fire Service of Russia, Building 4, B. Galushkina Str, Moscow, 129366, Russia

2 FireSERT, University of Ulster at Jordanstown, Shore Road, Newtownabbey, Co Antrim, N. Ireland

ABSTRACT

A novel approach to staff fire training for Marks and Spencer staff is presented and discussed in this paper. The contextual backcloth is the current training techniques employed and the requirements generally of Health and Safety Legislation in terms of duties imposed on staff, particularly with respect to the safety of others. The training approach articulated in this paper is rooted in the notions of occupancy where the responses, actions and behaviours of individuals in time and space are a function of the environment in which they find themselves.

The concept of ‘occupancy’ was used to analyse video taped recordings of five unannounced evacuations of Marks and Spencer stores. From these five different settings which imposed different constraints and opportunities for staff were identified. A computer video presentation for each of the settings for each of three cues, ie alarm, smoke and flame were presented to 172 Marks and Spencer staff to whom a follow up questionnaire was administered. The results of the work are presented in this paper. These results indicated that the percentage of staff who actually followed their current training instruction was surprisingly low. The computer video presentation highlights the inadequacies of the current training methods, exposes staff to real situations for a range of cues and offers the possibility of realistic, useful staff training assessments.

INTRODUCTION

The past two decades have witnessed a global transition in developed countries from prescriptive fire safety legislation to functional fire safety regulation based on actual performance of building elements and people in fire, with fire safety engineering being accepted as a creditable design profession.

Building fire evacuation studies have revealed that apart from the architectural design solutions and engineering systems, the behaviours of staff in the event of an emergency can make a significant contribution to the timely evacuation in the event of a fire emergency, and therefore the life safety potential of occupants. The influence of staff on the evacuation of customers, both in terms of initiating response and giving direction was particularly evident in five unannounced evacuations of retail stores1. Further analysis of these evacuations2 indicated that almost 80% of staff had a direct beneficial influence on customer behaviour. In contrast, analysis of fatal fires in retail stores3,4 confirms that inappropriate staff behaviours can contribute to the loss of life.

Detailed analysis of the five evacuations of retail stores also indicated that the response of staff to the alarm varied significantly with setting2. This is fully consistent with the theory of ‘occupancy’ defined as “the constraints on, conditions and possibilities of knowledge and action afforded by the social, organisational and physical locations occupied by people over time” developed at FireSERT and articulated in 5. In this context, the building is considered as an information system, as a social organisational system of communications. Occupants differ, not in the sense of the physical environment, but because of the knowledge afforded by different locations occupied depending on their activities. This manifests itself such that building characterisation as ‘occupancy type’ e.g. hospital, retail store, office building is insufficient and a further level of specification such as settings is needed. Settings in this sense are a cumulative indicator of the micro activity of a person, and give us a context for the search for greater understanding of human behaviour in fire.

However, the behaviour of staff to a greater or lesser extent is a manifestation of the organisational training environment which in turn is influenced by legislative requirements.

Current legislation with regard to Health and Safety at Work6,7 establishes general duties and responsibilities for both the employer and employees. In essence the health and safety requirements are that staff work in health and safe environments, that they do not put themselves of work colleagues at risk and that they are trained in the performance of their duties with respect to health and safety at work. Health and Safety legislation however does not expressly require staff to put out fires or to manage the evacuation of the premises in the event of fire. Consequently, fire safety design solutions, incorporating fire safety management as an essential component, require much more than mere compliance with health and safety legislation. Another contextual consideration, is the shift in continuing control mechanisms from an annual fire certification process for premises administered by UK fire brigades to a risk assessment process undertaken by owner/occupiers. This also has serious consequential training issues.

The changing environment which influences fire safety management is the contextual setting for this paper, which presents the findings of a programme of research which set out to use the concept of occupancy to understand the behaviour of staff in fire evacuation of Marks and Spencer stores. The development and application of the Computer Video Interview Technique (CVIT) which was used to determine the response behaviours of staff in different settings to various fire cues is discussed,. The development of this technique as a training tool which can meet the demands of the changing environment influencing fire safety management in the UK today is proposed.

METHODOLOGY

It is recognized that retail stores can be complex environments and that in understanding behaviour, the setting, the person and the fire must be considered together, ie, not as separate elements. Traditional human behaviour research methods such as unannounced evacuations or interviewing survivors of real fires were therefore not considered sufficient in that one or more of the essential factors, eg fire, details of setting were not readily available, it was considered necessary therefore to develop a simulation method that could capture the entirety of the process.

It was also considered prudent to determine staff behaviors3 in response to different fire cues and settings. For the purposes of this study, three fire cues were chosen ie fire alarm, flame and smoke. Video analysis of previous evacuations of Marks and Spencer stores suggested that staff behaviours in response to the alarm vary between settings, eg:

• changing rooms; staff members have clear responsibilities for customers in the changing rooms and in this respect they and customers may be constrained by the setting,
• coffee shop; staff are serving customers and may be constrained by the setting. Previous fire investigations have revealed the occurrence of longer customer pre-movement times in this setting resulting from different staff behaviour3,4,
• customer service desk where staff are serving customers and may be constrained by the setting and often have additional fire duties,
• shop floor; staff have less well defined responsibilities, i.e. they have freedom to move and are not in direct contact with customers, and
• till banks; cashiers have responsibility for money and may be reluctant to cease customer servicing; in this respect they may be constrained by setting.

Figure 1. Video clip contents

Video presentations were therefore developed for each setting (changing room, coffee shop, customer service desk, shop floor and tills) with each cue (alarm, fire and smoke).

The generic content of the presentations were video clips showing the retail store environment with the fire cue introduced at a certain time intervals. Staff were tested individually and were asked to watch the video presentation, and to imagine that they were the member of staff in the workplace.

In an attempt to introduce the environment to the participant and to make the participants imagine themselves as staff within the particular settings, several film clips were produced and compiled together. First, the staff member was presented with an outside view of the store, then a general inside view followed by a clip showing a staff member taking up position in one of the above settings, and finally the staff member working in the setting, Fig 1A. These video clips were recorded in Marks and Spencer, Belfast store on an ordinary weekday, with customers and staff present. In all presentations a second clip was introduced to the screen showing an area which would be visible to the staff member from their workplace, Fig. 1B. After an initial period of time, the fire or smoke cues emerged in this clip, Fig. 1C. The fire and smoke clips comprised real video footage of flames and smoke emerging from behind sales goods which were filmed in the Belfast store at night when no customers or staff were present. In the case of the fire alarm cue, the fire alarm was simply superimposed on the original sequence of clips, Fig. 1D.

The required video components were recorded and processed (MPEG 1 standard). Two computer applications MVSetup and MultiVid were developed within Borland Delphi 5 environment. MVSetup was developed to establish the precise locations and time instances of the clips appearing and disappearing. The objective of MultiVid program is to play the clips with determined characteristics by the MVSetup on the computer’s screen and record some experimental parameters i.e. decision making time. In total, 15 presentations were created (three fire cues for five settings). The duration of each was approximately 1.5 - 2 min. In order to ensure that participants responded appropriately within the experimental settings a familiarisation film was also developed in which fire cues were substituted by non fire related cues to which the staff member would be expected to respond.

As noted previously, the generic content of the presentations were video clips showing the retail store environment with the fire cues introduced at a specific time. Each participant was asked to imagine themselves in the role of the member of staff in the film and press a red bar on the keyboard, if they felt they had a decision to make about anything unusual that they saw or heard on the screen.

Having indicated that they had a decision to make, the staff member was then asked what he/she would do in response to the cue. Their subsequent actions until they decided to evacuate the building were also elicited. Staff actions were categorized into generic behaviours distilled from previous research.8.9,10,11,12,13 namely, continue with activity, wait for information, alert, investigate, fight the fire, guard the fire, prepare to evacuate, request help, evacuate customers, check the area, direct/assist staff, evacuate self. Profiles of staff members in terms of age, gender, job title, time employed in the store, nature and details of the training received, particular fire duties as well as their views on the experience were also elicited. Other pertinent factors indicating their awareness of fire were also elicited, eg, did they have fire alarms in their houses, had they installed them. Had they ever experienced a real fire emergency, etc.

SAMPLE

One hundred and seventy two staff from eleven Marks and Spencer stores across Ireland and Wales participated in the study. The sample represented on average 20% of the sales floor staff of each store, and in terms of gender (80%female) and time employed in the Marks and Spencer (59% employed between 1 and 10 years) they are representative of the Marks and Spencer general staff profile10.

A breakdown of staff exposed to each fire cue and setting is given in Table 1.

Table 1 Breakdown of participants across fire cues and settings

Settings / Experimental cue / Total
Alarm / Flame / Smoke
Changing room / Frequency
% from settings
% from fire cue / 13
27.7%
30.2% / 18
38.3%
27.3% / 16
34.0%
25.4% / 47
100.0%
27.3%
Coffee shop / Frequency
% from settings
% from fire cue / 3
23.1%
7.0% / 5
38.5%
7.6% / 5
38.5%
7.9% / 13
100.0%
7.6%
Customer service / Frequency
% from settings
% from fire cue / 4
26.7%
9.3% / 6
40.0%
9.1% / 5
33.3%
7.9% / 15
100.0%
8.7%
Shop floor / Frequency
% from settings
% from fire cue / 11
23.9%
25.6% / 17
37.0%
25.8% / 18
39.1%
28.6% / 46
100.0%
26.7%
Till / Frequency
% from settings
% from fire cue / 12
23.5%
27.9% / 20
39.2%
30.3% / 19
37.3%
30.2% / 51
100.0%
29.7%
Total / Frequency
% from settings
% from fire cue / 43
25.0%
100.0% / 66
38.4%
100.0% / 63
36.6%
100.0% / 172
100.0%
100.0%

RESULTS

A comparison of the response behaviours of staff across the three fire cues, regardless of their sequence and setting is given in Table 2

Table 2 Summary of staff responses for different fire cues

Action / % of staff
Alarm / Flame / Smoke
Alert
Evacuate customers
Investigate situation
Fight fire
Wait
Request help
I don’t know / 9.3
95.3
48.8
0
7.0
0
0 / 95.2
84.9
13.6
28.8
7.6
4.5
1.5 / 97.0
84.1
27.0
4.8
7.9
3.2
3.2

From Table 2 it can be see that the response of staff will vary with cue. For example, it would appear that a lesser percentage of staff faced with real fire manifestation (flame or smoke) would engage in customers’ evacuation in comparison with staff exposed to the fire alarm. It is also evident and perhaps surprising, that a large percentage (48.8%) of staff, on hearing the fire alarm, would investigate the situation; this is rather more that those who responded to other fire cues i.e. 27.0% for smoke and 13.6% for flame. A possible explanation is that the alarm is actually an ambiguous fire cue for staff, because they consider that it might be a false alarm etc. In contrast, flame and smoke are obvious manifestations of fire, with the greater percentage of staff investigating the situation with smoke than with flame being explained by the fact that smoke is a perhaps a more ambiguous cue than flame. Also of interest from Table 2 is that the percentage of staff who would fight the fire in case of flame is five times greater than for smoke. This again may imply that the staff do not consider smoke as a fire manifestation to the extent that they do the flame.

Detailed analysis of the responses of individuals to each of the fire cues by settings is presented in 2.This analysis indicated that the majority of staff would perform four or five actions, while others (mostly managers) would perform as many as eight. Due to the complex nature of the data, it is not possible to present the detailed results in this paper

However, the analysis of the data indicates clear differences in the response of staff in different settings to different cues. The analysis clearly shows that the staff in the changing room and at the tills reacted similarly and in contrast to shop floor staff for each fire cue. Shop floor staff were more likely to take active actions (i.e. evacuate customers, in the case of the fire alarm, or actively investigate the situation and alert about it in case of flame and smoke cues). Shop floor staff response was not necessarily the most appropriate, but certainly the most active. It indicates that the absence of customer servicing duties and absence of bounded space induces different response. It was also found that staff across all settings did not respond as appropriately to smoke as they did to the flame cue, e.g. more staff across all settings investigated the situation in the smoke conditions in contrast to flame (in general twice as many staff), also more staff requested help, didn’t know what to do, and less staff phoned the fire brigade.

IMPLICATIONS FOR TRAINING

The analysis of the data obtained in this study clearly indicated wide variability in staff response to the various fire cues. It was evident that some staff would not perform the actions required of them (ie evacuate customers in event of alarm, or activate fire alarm, evacuate customers in event of fire cue), that some staff would perform the required actions but in the wrong order and that some staff would perform additional, unnecessary actions such as waiting for information, investigate, double alerting.

A comparison of the actions of individuals in response to the various fire cues with the actions required of them with respect to their current training was made. The main criteria for action evaluation was performance of the required actions in an appropriate order. The percentage of staff across all settings and cues who followed the training requirements was surprisingly low ie only 37.2%. A breakdown of the percentage of staff who followed training by fire cue is given in Table 3.

Table 3. Percentage of staff who followed training procedures by fire cue

Cue / Followed Training
(%)
Alarm / 44.2
Flame / 39.4
Smoke / 30.2

From Table. 3 it can be seen, that the percentage of staff who followed training varied with cue. The most appropriate responses were by staff who were exposed to the fire alarm with a less appropriate response to the flame, and the least appropriate to smoke. This may be explained by the fact that staff are trained to respond to the fire alarm through drills, and rarely trained to respond to flame and smoke. The more inferior response to smoke than for flame is perhaps again related to the fact that smoke is more ambiguous than flame.