During the experimental design stage, the consortium identified a large number of variables of interest, and it soon became evident that it would be difficult to limit the number of independent variables and insufficient test evacuations were available to obtain adequate replications of each test condition. It was finally agreed that the evacuation trials would not be conducted as an experiment; instead, the evacuation trials would take the form of a series of evacuation demonstrations, which could then be used to explore possibilities for future research.
The final programme that was agreed was to use the trials to explore passenger movement in three types of situation: where passengers had a free choice between available exits on both decks, where passengers on the lower deck were required to move to the upper deck, and where passengers on the upper deck were required to move to the lower deck. The trials were implemented by the Human Factors Group, within the Cranfield University Large Cabin Evacuation Simulator, the facility was modified to provide key physical features of a generic wide-bodied cabin with two decks. The planned test programme was completed and no evacuations were halted. In total, 336 individuals participated in the evacuation demonstrations. It is believed that the trials produced passenger behaviour representative of non-competitive evacuations and the crew behaved in a manner that might be expected under a set of simulated operational conditions in which no additional training concerning the use of stairs for evacuation was provided.
Although a number of pilot trials had been conducted, the trials did not proceed in the controlled manner that was originally planned. Due to the small number of data points provided by these trials, there is insufficient data upon which to claim statistical significance for any of the observations documented within the report. In the event, the cabin crew behaved in a number of ways that differentiated from that which had been expected. It was apparent that in all the trials, crew played some role in managing the passenger flow on the stairs. It must be remembered that all crew (except those located at UR1) were line cabin crew who were trained in specific operator emergency procedures with the aim of reducing the overall evacuation time of the aircraft. Ethically it could be argued that if the cabin crew were trained in behaviours that conflicted with their normal procedures, this could be potentially detrimental to their later performance in a genuine emergency situation. Although cabin crew knowledge and experience is crucial to our understanding of aircraft emergency evacuation, the trials have demonstrated that in exploratory research where specific crew commands and behaviours are fundamental to the experimental design, in particular where these are not identical to those implemented by the operator, the use of researchers trained as cabin crew should be carefully considered.
Unfortunately, the Cranfield University analysis was limited to descriptive analysis only on the passenger evacuation times, as inferential analyses of the evacuation data could not be conducted, as insufficient data was available to conduct comparisons across conditions. However within the free choice evacuations, there did appear to be differences in evacuation rates between the two demonstrations, with lower mean evacuation times, faster evacuation rates, and lower overall exit evacuation times evident on the last trial of the programme. However, this may simply be a function of the cabin crew, who by this time would have gained significant additional experience in passenger management and evacuation situations. Within the conditions involving ascending the stairs, there did appear to be marked differences in evacuation rates between UR1 and UL1. The UR1 exit involved passengers evacuating down a slide whereas UL1 was out onto a platform. This difference in time through UR1 is most likely a function of the caution exercised (due to safety) by Cranfield University cabin crew at the UR1 exit. Finally, within the evacuations involving descent of the stairs, the mean evacuation times, evacuation rates and overall exit evacuation times do appear to be broadly similar across the evacuation trials conducted. The trials have identified a number of areas where future research needs to be conducted to generate essential data to improve our understanding of passenger performance, cabin crew performance and passenger-crew interaction within VLTA configurations.
The next step should be to form clearly identifiable research objectives and to develop detailed research programmes combining partial experimental evacuation testing including statistically reliable results, evacuation computer modelling and qualitative analysis, in an attempt to address the complex issues relating to the safe evacuation of VLTA.