3

Review of the Civil Aerospace Medical Institute Project

The previous chapter summarizes the stated objectives, experimental setup and methods, data analysis results, and conclusions reached in the 2019–2020 Civil Aerospace Medical Institute (CAMI) research project as documented in the report Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress.¹ The study committee’s review of this work is provided in this chapter, organized in a manner similar to Chapter 2’s description of the work and giving attention to many of the project’s details. The review is intended to fulfill this study’s charge, which calls for an assessment of whether the CAMI project’s objectives and research methods, design, and procedures are appropriate and sound. The charge also calls for the committee to consider how the results of the experimental research may be used, and in this regard the committee also critiqued the work, particularly when considering the policy and legislative interests that prompted CAMI to commission the work in the first place.

The chapter begins by considering the stated objectives of the CAMI research project and reviewing them for clarity and coherence and their relationship to the hypothesis that the project was designed to test. That hypothesis is that seat dimensions should not affect the time it takes to evacuate an airplane because crowding at the exit door and queuing in the aisle are far more impactful on evacuation performance than a passenger’s

___________________

¹ Weed, D. B., et al. (2021). Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress. https://www.faa.gov/sites/faa.gov/files/2022-04/Effects_of_Airplane_Cabin_Interiors_on_Egress_I.pdf.

speed in exiting the seat. Consideration is then given to the single-study test group that was formed for the project’s data collection experiments, each of which is then reviewed in turn. The chapter ends with the committee’s thoughts on how well the project’s design satisfies the interests of policymakers who worry that seat space on airplanes is becoming too small as Americans are becoming larger to the potential detriment of expeditious and safe emergency evacuations.

CLARITY AND COHERENCE OF STATED OBJECTIVES

The study committee wanted to be sure it understood the purpose of the CAMI research, but this required parsing of the language in the project’s report where the research objectives are stated. As noted in Chapter 2, the project’s two main research objectives are worded differently in two places in the report. When initially cited on page 1, the first research objective reads as follows:

• to determine the percentage of the American population for whom ergonomic minimums are being violated at the lowest average seat spacing, and the percentage affected if that spacing was further reduced.

Later, on page 7, this first objective reads as follows:

• to determine what percentage of the American population, based on anthropometric measurements, would not be able to sit in transport airplane passenger seats at the currently narrowest and even narrower seat pitch.

In the first instance, the statement uses the term “ergonomic minimum,” whereas the second statement is more direct by referring to a person’s ability to sit in a passenger seat. The report cites previous Federal Aviation Administration (FAA) research² as the source of the term “ergonomic minimum” but does not offer a definition. On page 7, the report explains that the project’s interest in investigating a person’s ergonomic minimum was to establish when a person “would be unable to sit, and thus be unable to fly, should the occupiable seat space … be further reduced.” Also, with

___________________

² McLean, G. A., et al. (2002). Access-to-Egress I: Interactive Effects of Factors That Control the Emergency Evacuation of Naïve Passengers Through the Transport Airplane Type-III Overwing Exit. DOT/FAA/AM-02/16. Washington, DC: U.S. Department of Transportation. https://www.faa.gov/sites/faa.gov/files/data_research/research/med_humanfacs/oamtechreports/0216.pdf.

regard to these two instances, the alternate references to “seat spacing” and “seat pitch” caused some confusion given the project’s interest in seat width as well as seat pitch.³

Without a definition of “ergonomic minimum” within the context of the CAMI study, it is difficult to precisely understand the implications of its usage throughout CAMI’s report, such as in describing the study objectives and interpreting study findings. The committee suggests that the term “ergonomic minimum” refers, in fact, to seat dimensions and not passenger dimensions. The way to think about this is that for each seat dimension, there is a maximum passenger body size that can fit into the seat. Given the variation in body size within the population, there will then be corresponding variation in ergonomic minimum seat dimensions. Thus, within the context of this research, it is appropriate to identify the minimum seat dimensions (ergonomic minimum) that a significant proportion of the American public can fit in. Furthermore, it would be useful to identify the proportion of the American public that can fit into the current ergonomic minimum seat configuration (e.g., 16-inch seat width and 28-inch seat pitch).

As the committee’s understanding of the CAMI project evolved, the purpose of this first objective became clearer, inferred from the project’s interest in forming a study test group that was generally representative of the flying public and the project’s pre-trial experiments intended to assess the ability of study participants to sit in airplane seats configured at the lowest seat pitch now in airline service and at a pitch that is even lower but not currently in service. The first objective sought to identify the seat pitches and widths that should be the subject of the follow-on evacuation trials because they are the lowest currently in use or have the practical potential to be introduced by airlines if enough passengers would be able to sit in them. Observing the ability of study participants to sit in seats would allow CAMI to make such determinations, as long as its study test group resembled the airline customer base to a sufficient degree. From what the committee could gather, these determinations were the central purpose of the pre-trial seat experiments, and it would have behooved CAMI to have stated its first objective more directly.

The research project’s second objective, as stated on page 1, reads as follows:

• to determine what, if any, effect various seat pitch and width configurations have on the speed of a simulated airplane evacuation.

___________________

³ See Research Question 1 as listed in Appendix B.

Later, on page 7, the objective reads:

• to determine the effect of seat pitch and seat width on individual egress time.

Here the terms “egress” and “evacuation” are used interchangeably, as they are throughout the CAMI report. For reasons explained in Chapter 1, the committee believes “evacuation” is the more appropriate term. There are other differences in how the objectives are stated that can cause misunderstanding, such as substituting “speed” for “time” and referring to only “individual” egress time in the second instance, when group times were also of interest. Here again, this objective should have been stated more clearly and consistently. Significantly, and as will be discussed later in this chapter, the objective’s interest in determining the effect of seat pitch and seat width configurations on the time required to evacuate a simulated airplane says little about how germane that simulated airplane’s passenger composition should be to the mix of passenger compositions that will undoubtedly occur across thousands of flights in daily service. Here it is important to note that there are two distinct population profiles that can be characterized and that may have potential relevance to the study: the profile of the overall U.S. flying public and the profile of passengers on individual U.S. flights (i.e., the flight profile). While the flight profile will be made up of individuals from the flying public, the demographics of the population on each flight will not necessarily satisfy the flying public demographic or anthropometric composition. Is the population used in the simulated airplane evacuation supposed to represent a flight in which the demographic and anthropometric composition of the passengers align closely to that of the flying public in general or the flight profile?

Finally, even though the CAMI report frequently cites two objectives for the research, a third objective is also given. That objective is to collect “an amount and type of anthropometric data from each participant that has not been collected in previous evacuation research performed at CAMI, with the intent to use these data for ergonomic analyses of current and future seat designs and use in various computer-modeling efforts.”⁴ In other words, CAMI viewed the research project as contributing to an anthropometric database that can inform and provide a basis for potential future research.

In the sections that follow, these three study objectives—the two main ones and the ancillary one—are considered when assessing the research

___________________

⁴ See p. 7 in Weed, D. B., et al. (2021). Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress. https://www.faa.gov/sites/faa.gov/files/2022-04/Effects_of_Airplane_Cabin_Interiors_on_Egress_I.pdf.

project’s study design; data collection, analyses of results, and conclusions; and limitations.

STUDY DESIGN AND DATA COLLECTION

The research plan that was approved by CAMI’s Institutional Review Board (IRB) (see Appendix B) presents the study’s hypothesis as follows:

Aircraft evacuation should be considered as a system. Currently, the bottleneck within that system is the egress rates of the doors combined with evacuation slides/assist means. As long as ergonomic minimums at the passenger seats are respected, that is, people can get into and out of the seat, the rate at which the passengers can move from their seats into the main aisle is ultimately immaterial to the evacuation flow as a whole. Manipulation of the usable egress space between seats may provide a statistically significant difference in evacuation speed, but the practical significance of this difference will be determined by the consideration of overall egress speeds and the expected flow rates of the door/slide evacuation system.

In essence, the hypothesis posits that as long as passengers can get into and out of their seats, the time it takes them to move from their seats, through their row, and into the main aisle will not affect the evacuation flow as a whole because of the queuing created at the exit doors and down the center aisle. The rationale is that if a seat has enough space for a person to sit in it—that is, it does not violate the person’s ergonomic minimum seat size—then the person should be able to exit the seat during an evacuation and the time required to exit the seat row will not negatively impact the evacuation performance.

In designing its study, therefore, CAMI wanted to be sure that the seat widths and pitches that would be tested in the evacuation time trials represented the practical minimum in the sense that a significant majority of air travelers would still be able to sit in them. A starting point for doing this was to identify the currently narrowest seat widths and pitches in the transport airplane fleet since, demonstrably, they now accommodate enough travelers for the airlines to be willing to deploy them. The next step, then, would be to determine whether any further reductions in seat width and pitch could be practical from the standpoint of airlines, who have an interest in ensuring an ample base of customers so as not to narrow seat dimensions so much that this base would be significantly eroded. Evacuation trials involving seat dimensions that are not likely to be deployed in the fleet would have minimal utility, so it would be important for CAMI to know the practical minimums, whether represented by existing minimums or additional reductions.

This interest in establishing minimum practical seat dimensions (i.e., seat width and pitch) drove CAMI’s interest in recruiting a study test group that would be representative of the flying public in body size and other anthropometric and demographic characteristics and then to conduct pretrial experiments testing the ability of participants to sit in seats configured with the lowest currently available seat pitch (28 inches) and an even lower one (26 inches). Once this practical minimum seat pitch was established, by judging when a critical number of participants in the study test group—and thus presumably, in the general population—could no longer be accommodated by further reductions, then CAMI would be able to judge the relevance of the 28-inch seat pitch that was planned for the evacuation trials. These pre-trial experiments used a 17-inch seat width and did not test a 16-inch seat width, since this dimension is already found in the fleet (although rare) and it would need to be part of the trials regardless. CAMI did not test an even lower width (e.g., 15 inches) believing that 16 inches, as found in some cabins, is the practical minimum.

CAMI therefore designed and executed three distinct, albeit connected, data collection efforts using the same recruited study test group. As explained in Chapter 2, the data collection entailed (1) taking measurements of study participants to obtain their anthropometric data considered pertinent to airplane seat fit, (2) pre-trial testing of the ability of study participants to sit in seats configured with the lowest deployed seat pitch (28 inches) and one that is even lower (26 inches) using a single average seat width of 17 inches, and (3) conducting airplane evacuation trials involving the study test group to determine the effect on evacuation time of the lowest practical seat width (16 inches) and seat pitch (28 inches), relative to a larger width (18 inches) and larger pitches (32 and 34 inches), as derived by consulting surveys of the airline fleet and the pre-trial seat experiments (which ruled out the 26-inch pitch as a practical minimum that would deserve testing in the trials).

The design and execution of each of these data collection activities is critiqued below in more detail and with regard to their relevance to CAMI’s stated research objectives and hypothesis. Before doing so, however, it merits considering the implications of CAMI’s decision to recruit a single study test group for all three of the project’s data collection purposes.

Study Test Group Recruitment and Use

Chapter 2 describes the recruitment of the study test group of 775 individuals. This recruitment involved two primary and, to some degree, conflicting aims. First, CAMI wanted to recruit a study test group that could be reasonably construed as resembling the flying public, as they needed to establish the minimum practical seat pitch and width using the pre-trial seat

experiments (for reasons explained above). Second, CAMI intended to use this same group (excluding anyone who was determined based on pre-trial experiments to be unable to sit in a seat having a 28-inch pitch) for evacuation trials that would investigate whether those seat minimums would affect evacuation times. The study test group would therefore have to be not only representative of the flying public but also limited in composition due to safety concerns of the evacuation trials. The pre-trial seat experiments designed to estimate the minimum practical seat pitch for the trials, therefore, would leave out some demographics, such as participants representing older travelers, even though the experiments were not hazardous.

At the same time, however, CAMI’s decision to use the same study test group for the evacuation trials would necessarily involve an assumption that the demographics and anthropometrics of the passengers onboard actual transport airplanes (i.e., the flight profile) are reasonably close to those of the study test group, including the group when it was divided into the 12 60-participant trial cohorts. CAMI explained the ethical and safety reasons that compelled it to limit some demographics from the evacuation trials, but it offered little reason for aiming for an evacuation trial study test group that should be generally representative of the flying public. The mix of passenger demographic and anthropometric characteristics on individual airline flights may differ significantly from the mix of characteristics of the flying public as a whole. For example, some flights may have a larger proportion of passengers with large body sizes than others, even though the average distribution for all flights may resemble that of the flying public at large. In other words, the distribution of passenger demographics and anthropometrics on a given flight (i.e., the flight profile) may be different from the distribution for the flying public. Inasmuch as the concern over narrowing seat dimensions stems from the concern about the impact of passengers with larger body measurements, CAMI does not explain why it did not purposely skew the study test group so that it involved more passengers with larger body sizes to assess their impact on actual flights.

CAMI’s hypothesis was that even if seat dimensions and large body sizes interact to slow an individual’s ability to exit the seat row and enter the aisle, this should not matter for evacuation performance because the time they spend queuing at the exit door and in the main aisle is greater than the time required for the passenger to exit their seat row. In its attempt to test this hypothesis but constrained by the number of trials it could run, CAMI designed the trials to manipulate the variables of seat width and pitch but not body size variables. Nevertheless, a case can be made to include body size variables in the evacuation trials or in a combination of evacuation trials and computer modeling simulations, as will be discussed later. However, had CAMI decided to run the same evacuation trials that manipulated seat dimension variables but used a study test group having

many more participants with larger body size measurements, the project’s results would be more compelling and robust by seeking to represent a skewed passenger composition that can occur on at least some flights among thousands on any given day. The safe evacuation of such flights, of course, would be of interest to policymakers.

With regard to CAMI’s interest in creating a study test group that was generally representative of the flying public, there are reasons to question whether this desired outcome was in fact achieved. Even when accounting for the effects of the deliberate exclusion of older participants (people over 60 years of age), it appears that enrollees in the study test group skewed young. This outcome could be expected given the time commitment, limited compensation, and physically demanding nature of the trials, which can make it more challenging to recruit middle-aged participants. Indeed, the FAA regulations that govern evacuation tests for airplane certification (Appendix J of 14 CFR 25.803 [see excerpt in Box 3-1]) recognize this

challenge by requiring that certification test populations be at least 40% female, 35% older than 50, and 15% female and over the age of 50. Given the reported demographics of study participants, it is clear that CAMI did not follow this scheme, even as guidance. While CAMI’s recruiters were instructed to ensure that the enrolled group had a roughly even balance of men and women and fewer than 40% of enrollees in any one age cohort, the recruitment fell short of the population demographics specified in the FAA certification test regulations. Apart from having an even balance of men and women, the study test group’s youngest cohort (18–30 years old) comprised 38% of participants while the oldest cohort (51 years and older) comprised only 14%. Women older than 50 accounted for only 9% of the study population.

For insight into whether its study test group provided a fair representation of the general public, CAMI compared the group’s demographic and anthropometric data with demographic and body measurement files from

the Centers for Disease Control and Prevention’s National Health and Nutrition Examination Survey (NHANES) for 2017–2018. NHANES data are intended to be representative of the U.S. population. After filtering the NHANES records to exclude people younger than age 18 and older than age 60, the CAMI researchers observed that the study test group participants were, on average, larger and heavier than the general U.S. population, as compared to the NHANES data (see Table 15, p. 33). For this reason, the CAMI report infers that its study population “can be assumed to be slower than the average U.S. population.”⁵

However, this inference is open to question because of the study test group’s relative youth and the absence of age comparisons with the NHANES data. Whereas larger and heavier people may be slower on average, youth may compensate for this effect. An important aspect of evacuation testing is ensuring that participants in the trials represent a wide range of characteristics of the flying public, including age–body size combinations.

Finally, while 775 participants were recruited for the seat experiments, only 718 participated in the evacuation experiments. Given that six were rejected from participating in the evacuation trials based on observations by the researchers concerning the participants’ ability to fit in the seats (see below), it is not clear on what basis the other 49 participants were excluded. On page 24, the CAMI report simply states that these 49 were “selected” for exclusion without providing the exclusion criteria. Following the exclusion of 55 recruited participants, it then becomes unclear whether the remaining group of 720 has the same degree of commonality with the NHANES data (a point returned to later).

Anthropometric Data Collection

In addition to demographic data (age, sex, education level), participants in the study test group were measured for height, weight, girth, shoulder width (bideltoid breadth), sitting hip width, sitting buttock-to-knee length, and sitting knee-to-floor length (KtF). The measurements were collected for assessing results from the seat experiments and evacuation time trials. Although more complicated, the collection of dynamic anthropometric data, such as the participant’s range of motion, balance, flexibility, and mobility—for example, tests like Timed Up & Go (TUG)—might have been helpful for analyzing the results of the project’s experiments (a point returned to later).

___________________

⁵ See p. 41 in Weed, D. B., et al. (2021). Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress. https://www.faa.gov/sites/faa.gov/files/2022-04/Effects_of_Airplane_Cabin_Interiors_on_Egress_I.pdf.

Another stated objective in collecting anthropometric data was to build a database for future research and computer-based evacuation modeling. The committee cannot ascertain the eventual value of these data for future uses although state-of-the-art computer-based evacuation modeling does not currently take into consideration many of the collected anthropometric measurements. Missing from the collected data, however, are measurements of people representing traveler demographics not represented within the study test group because of the safety risks of the evacuation trials. Supplemental data collection from a wider range of people might have been helpful for understanding the characteristics and capabilities of these missing demographics from the seat pitch and evacuation experiments.

Pre-Trial Seat Experiments

The seat experiments were conducted on the study test group prior to the evacuation trials. As explained in Chapter 2, participants were asked to sit in a seat mock-up consisting of three rows of decommissioned 1990s-era seats, all with a spacing of 17 inches between the armrests (i.e., seat width). Researchers set up the back row to match the 28-inch seat pitch that participants would experience in the evacuation trials. Participants were also asked to sit in the middle row consisting of seats configured at a 26-inch pitch. Because only a single seat width was tested (17 inches), it was not possible in this experiment to observe the smallest seat width that a participant could sit in.

As has been noted, one purpose of the experiments was to determine the percentage of participants who would be able to sit in a 26-inch seat (which is not currently deployed). By doing so, CAMI would be able to ascertain whether the 28-inch seat pitch planned for its trials was sufficiently representative of the likely minimum to be deployed in airline service. Testing a study test group that generally resembles the flying public, or airline customer base, would serve this purpose. Testing a seat with a 28-inch pitch would also serve the purpose of identifying study test group participants, in advance of the trials, who could not sit in a seat of this dimension and therefore would need to be excluded from the trials. The experiments would therefore be used exclusively to confirm that the evacuation trials would involve the lowest viable seat pitch, even though it was expected that some participants would not be able to sit in the seat configured with a 28-inch pitch so that they would be excluded from the trials under the premise that they would not be flying.

In the committee’s view, the seat experiments represented a missed opportunity to learn more about the effects that seat pitch and width could have on a person’s ability to exit a seat and seat row. The emphasis was placed on observing and judging whether the participant could sit in the

seat, presumably because of the research project’s hypothesis that the rate a person can exit a seat and row should not matter to the evacuation flow. Furthermore, the use of a study test group that was purposely formed to be representative of the U.S. population reduced the opportunity to collect a large amount of data on the interaction of body size variable with seat dimensions and configurations. The experiment’s reliance on a study test group that would be used for evacuation trials, and thus necessarily limited in demographic representation for safety reasons, further limited the potential for collecting data on seat and body size interactions across a wider demographic, even though the seat experiments themselves did not present a safety hazard.

The committee notes with concern that six members of the study test group were reported as not being able to get into the mock-up of a 28-inch seat pitch and so were excluded from the evacuation trials (with one exception when the participant only participated in the trials involving seat pitches greater than 28 inches). The six were judged by research staff to be unable to sit based on criteria discussed in Chapter 2. Given their judged inability to sit in the mock-up seat (17-inch seat width with a 28-inch seat pitch), their exclusion may have been considered necessary as a practical matter because one of the seat widths in the evacuation trials was even narrower (16 inches) and all these trial seats would have immovable armrests. A review of the videos, however, appears to show that all six managed to enter and exit the 28-inch pitch seat, albeit with apparent difficulty. The committee’s review of their answers to the post-experiment questionnaires indicates that three of the six had flown at least once in the past 12 months and five reported having some experience or being very experienced at flying on commercial airplanes. While it is possible that these individuals had been flying in the past with more spacious seating, their answers might also suggest that travelers will find ways to maneuver into and out of an airline seat even when doing so can be very difficult (such as by lifting the armrests before attempting to sit in or leave a seat). Furthermore, excluding those participants who could clearly enter and exit the seats in the mock-up, albeit with difficulty, excludes the very participants that are required to test CAMI’s hypothesis that the time required to exit the seat row is immaterial.

Another committee concern is that experiments designed to determine the share of the study test group that could sit in the seat configured with a 26-inch pitch required participants to self-judge whether they could sit in the seat. They were not given specific criteria for making this judgment, nor were their answers confirmed through a review of the video recordings in the case of the 56 participants who reported that they could not sit. The results from this experiment were used to justify the project’s use in the evacuation trials of the 28-inch seat pitch as the lowest practical pitch warranting assessment. However, the committee has low confidence

in self-reporting as the sole means for making this determination. Finally, it is noted that the seats used in the trials are older style seats no longer in commercial use within the United States. Newer style seats offer more “occupiable” space for a given seat pitch than the older style seats, and so the findings from the study may not be strictly relevant to the airplane interior configurations currently in service within the United States.

Evacuation Trials

While the anthropometric measurements and seat experiments had their own purposes, they were designed in large part to support the project’s evacuation trials. The study’s Statement of Task (SOT) calls for an assessment of the trials, including their design, experimental methods, and the analyses of the collected data, including how the study test group and seat dimensions tested were selected given the anthropometric makeup of air travelers and actual seat dimensions in commercial airplanes. Before turning to these matters, the next section turns briefly to the SOT’s question about whether CAMI followed accepted scientific principles for human subjects research.

Eligibility Limits for Trial Participation

As discussed in Chapter 1, CAMI has come under criticism for conducting its evacuation trials using a study test group that excluded individuals over the age of 60, pre-adults, and people with disabilities and other physical limitations, all commonly found among the passengers on airplanes. In the committee’s view, the safety reasons for limiting participation in the study test group are understandable, and CAMI’s report and IRB proposal make a well-documented case for them. While it is possible that the use of ramps as exits could have allowed some easing of the restrictions on participants over the age of 60, this is a judgment call based on reasoning and risk assessments the committee did not review. As an ethical matter, however, the decision to conduct live evacuation trials would necessarily require the recruitment of a study test group with significant gaps in demographic coverage.

Experimental Setup and Methods

The CAMI researchers had to balance realism with other interests, especially participant safety, across a number of dimensions. Both the use of trained flight attendants and a compensation scheme to motivate performance added realism but also some additional safety risk in the latter case. Pointing to learnings from previous evacuation research, CAMI offered

strong justification for both features for realism. CAMI used a narrow-body cabin simulator fitted with ramps at the exits as opposed to more realistic slides. Narrow-body airplanes account for a large majority airline operations in the United States. A slide was not an option because of the significant safety risks (as discussed in Chapter 2) and the simulator’s design. CAMI’s explanation for not using a slide at the exit is that the setup was designed to eliminate the variable of hesitation at the exit to allow for a better look at how seat configurations influenced the participants’ ability to move from a sitting position to the main aisle and toward the exit (see Appendix B, “Risk Assessment and Reasoning” and “Egress Assist Means”).

The use of a ramp, however, raises questions about CAMI’s decision to use an exit door having standard Type-I dimensions. A ramp fitted at the door, rather than a slide, would make exit flow higher than is realistic when passengers would be expected to hesitate before exiting to a slide. While CAMI acknowledges this point in the IRB proposal (see Appendix B), it does not provide an explanation on how this may impact the efficacy of the experiment. Because of shorter queues and reduced congestion at an exit door fitted with a ramp, individual and group evacuation times would be shorter, which could impact overall evacuation dynamics, such as conditions within the cabin that may affect evacuation behavior that are not representative of an emergency. A setup with a reduced door width such that the exit flow onto the ramp is roughly equivalent to that of an actual Type-I exit could have compensated for the substitution of a ramp for a slide. However, if seat pitch/width have a significant negative impact on evacuation performance, then faster clearing of the main aisle (through the use of ramps and not restricting the exit width) could make this negative impact more apparent. Thus, the door and ramp setup used in the trials, while significantly changing the expected exit flow dynamics, is not expected to compromise the validity of the findings with regard to the impact of seat pitch/width on evacuation performance.

Regarding the setup of the seats used in the trials, the use of seat designs that are no longer in service and armrest spacers to manipulate seat width detracted from realism. It should have been acknowledged that because seat widths were restricted only at the armrests, the manipulation did not reduce the room available for a person’s arms and shoulders as would be the case for actual seats when widths are reduced across the entire seat. For some people, the widest part of their body is either bideltoid breadth (measured at the upper arms just below the shoulder) or elbow-to-elbow breadth, which can be key measurements for seat width accommodation. Furthermore, by preventing the armrests from moving, the use of spacers would have made it more difficult for some participants with larger girths to sit in and exit the seats used in the trials. This anticipated difficulty may have been a factor in CAMI’s decision to exclude participants from evacuation trials who

were judged to be unable to sit in the pre-trial experimental seat configured with a 28-inch pitch.

Concerned again about realism, CAMI limited participants to four trials each based on findings from previous evacuation research indicating that repetition leads to learning effects by participants and the loss of real-world naiveté. McLean et al. is one of the reports of previous evacuation trials cited for CAMI’s decision to limit participants to four trials.⁶ It merits noting, however, that the trials run in that earlier project were for evacuations from an overwing Type-III hatch exit, which is a more complex and physically demanding operation than an evacuation from a Type-I door onto a ramp. Learning requires greater repetition when the tasks are more complex and unusual. Complexity was not a factor in CAMI’s seat pitch/width evacuation trials, as participants were simply asked to get out of their seat, walk along the aisle to an exit, and walk through the exit just as they would a normal door. Furthermore, the same exit was used for each trial, and so uncertainty in exit location was also not a factor. The simplicity of CAMI’s evacuation scenario causes the committee to question whether learning effects could have occurred earlier in the trials and had effects on evacuation performance as the trials progressed.

Data Analyses

Turning to CAMI’s methods for analyzing and presenting the results of the data collected from the evacuation trials, several matters deserve attention. Given the aforementioned possibility of learning effects, it may have behooved CAMI to analyze the data from the earliest trials separately because that is when participants were most naive. In Appendix C, for instance, the committee presents results from its analyses of the first and second trials (first trial only, second trial only, and combined first and second trials only) that suggest seat pitch may have had a greater influence on group evacuation time than seat width. This preliminary result would need to be confirmed and tested for significance, but the intended point is that additional analyses of the data may have yielded more meaningful and reportable insights.

With regard to the data analyses that CAMI did perform, the committee is puzzled by some of the choices that researchers made to exclude records. Notably, 34 individual evacuation times were identified as outliers because the times were three standard deviations above the average individual

___________________

⁶ McLean, G. A., et al. (2002). Access-to-Egress I: Interactive Effects of Factors That Control the Emergency Evacuation of Naïve Passengers Through the Transport Airplane Type-III Overwing Exit. DOT/FAA/AM-02/16. Washington, DC: U.S. Department of Transportation. https://libraryonline.erau.edu/online-full-text/faa-aviation-medicine-reports/AM02-16.pdf.

evacuation time. While removal of such times can be common practice in data analysis if the aberrations are attributed to influences not related to the study, it is important to give the justification for removal of these data (e.g., measurement error, inappropriate study participant behavior). As a minimum, the report should have provided information about the participants responsible for these 34 times, such as their physical dimensions, where they were seated, where in the exit sequence they appeared, which specific trials they were excluded from (i.e., trial day and seat configuration), and descriptions of their exiting behavior that potentially justifies their exclusion.

In addition to excluding these 34 records from the analyses, researchers also excluded the times of the first participants to evacuate during each of the 48 trials. They were removed on the grounds that first-out times are a measure of reaction time, only applicable to the first evacuee and not influential of group evacuation time. Another way to think about these 48 data points, however, is that they may provide the most direct indication of the impact of seat width and pitch of individual evacuation time. The person who is first out was probably not hindered by others, resulting in no queue time and minimal aisle travel time. The exit time is thus representative of the seat exit time (probably for a passenger located in an aisle seat). It would have been interesting, for instance, to know, for each trial, which seat was first out, the body size dimensions of the person, seat pitch and width, and individual evacuation time.

In determining the group evacuation time for each of the trials, it deserves noting that CAMI determined the evacuation time based on the time for the 54th participant in each trial, thus excluding the time for the last six participants. The given reason is that only 54 participants reported for a trial on day 1, and thus, for consistency, the total evacuation time for all trials was determined based on the time for the 54th participant rather than the 60th participant. While this adjustment may have been necessary, it means that the evacuation performance of 64 participants (note, there were only 58 participants on day 5) was excluded when determining the total evacuation time for the trials. It is conceivable that these last evacuees may have been the ones for which seat configurations and body size impacted evacuation time the most to affect the overall evacuation time, which should have been of direct interest to the researchers so as to warrant a sensitivity analysis. Furthermore, given that it is reasonable to assume that body size can affect evacuation performance (indeed CAMI did find that girth and KtF affected average individual evacuation time), the demographics of the participants whose evacuation times were excluded should have been reported along with the location of the seats left empty on the days when too few participants reported.

CAMI’s presentation of evacuation times also warrants consideration. Evacuation time was defined as “the time from one participant’s complete exit from the simulator to the next participant’s complete exit from the simulator.” The report notes that this metric, which is actually indicative of the flow rate of evacuees, has been used in previous studies, is considered a good gauge of evacuation flow, and can demonstrate problems that participants experience in “reaching, traversing, or using an egress assist (e.g., evacuation slides) at an airplane exit.”⁷ However, if the project’s aim was to obtain insights on the effect that seat pitch and width could have on an individual’s evacuation time, then reporting the flow rate may not be sufficient for revealing patterns of potential interest.

By way of example, the cumulative evacuation curves shown in Figure 3-1 are intended to reveal patterns warranting further evaluation. Derived

___________________

⁷ See p. 22 in Weed, D. B., et al. (2021). Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress. https://www.faa.gov/sites/faa.gov/files/2022-04/Effects_of_Airplane_Cabin_Interiors_on_Egress_I.pdf.

from an FAA-sponsored luggage retrieval project,⁸ each curve represents a specific evacuation scenario where carry-on luggage is being retrieved by evacuees from overhead bins by different percentages of the passengers (e.g., 0%, 25%). The cumulative number of passengers that have evacuated is plotted against the elapsed time in seconds. While the curves indicate when all the passengers have evacuated, they also identify the time when a specific number of passengers have evacuated.

The tight spread in the group evacuation times in the CAMI trials could mean that such data presentations will yield few, if any, meaningful insights. Here again, however, further explorations of this type may be helpful for observing if evacuation patterns differed for the seat configurations tested, whether shown individually for each trial or when using the averages for each scenario.

PERTINENCE TO POLICY INTERESTS

To conclude this chapter and transition to the next, it merits considering the applicability of the CAMI research project to Congress’s directives to FAA to issue safety regulations specifying minimum dimensions for passenger seats and to assess the procedures used for evacuation certification demonstration, including when passenger seat width and pitch are changed. The concern that underlies these directives is that seating space may be getting smaller as airline travelers are becoming larger in ways that may negatively interact with the seats to hinder evacuations.

Of concern is the unsubstantiated claim made in several places within the report, including in the conclusions (on page 44) and abstract, that the study “found that the experimental seat pitches, which are similar to the seat pitches currently found on flying commercial large transport category aircraft, should provide protection and not impede egress for 99% of the general U.S. population.”⁹ It is not clear how CAMI makes this claim based on the outcome of the seat experiments and evacuation trials. If this is based on the six participants who CAMI claims could not fit in the seats, they would need to identify the body size of those excluded (e.g., girth, KtF) and then identify what proportion of the American public this represents based on the NHANES dataset, and this would not apply to seats with widths less than 17 inches. Before they could make this claim on basis of the outcome

___________________

⁸ Galea, E., Lawrence, P., Cooney, D., Blackshields, D., and Filippidis, L. (2024). Investigating the Impact of Retrieval of Carry-on Luggage by Passengers on Aircraft Evacuation Using the airEXODUS Aircraft Evacuation Simulation Software. Technical Report: DOT/FAA/AM-24/19. Federal Aviation Administration, U.S. Department of Transportation. https://rosap.ntl.bts.gov/view/dot/79094.

⁹ See p. 44 in Weed, D. B., et al. (2021). Effects of Airplane Cabin Interiors on Egress I: Assessment of Anthropometrics, Seat Pitch, and Seat Width on Egress. https://www.faa.gov/sites/faa.gov/files/2022-04/Effects_of_Airplane_Cabin_Interiors_on_Egress_I.pdf.

of the evacuation trials, they would need to address the evacuation trial shortcomings identified in this report and undertake trials that specifically take body size into account as an independent variable. Furthermore, the trials did not consider widebody aircraft, where different seating configurations (e.g., 2×4×2, 3×4×3, 2×5×2) and twin aisles may also interact with body size and seat pitch/width to impact evacuation performance.

Based on its experiments, CAMI did find that body size (girth and KtF) does significantly impact the average individual evacuation time. While this finding does not indicate that seat dimensions and configuration (pitch and width) and body size (girth and KtF) together impact evacuation time, it suggests that further exploration may be warranted. For example, it could be that the time to exit a seat row is affected by a person’s girth in combination with the seat pitch and width, which, in turn, impacts their individual evacuation time. However, in discussing these findings from the trials, the CAMI researchers downplay their importance. Instead, they discuss prior study results of evacuation through a Type-III overwing exit by McLean et al.,¹⁰ in which seat width and pitch were not examined, that indicated participant gender, girth, and age were significant predictors of average individual evacuation times. It is important to keep in mind that evacuations from an overwing hatch exit are more physically demanding such that the variables of gender, girth, and age could be expected to have an impact on individual evacuation times. The same is not true for an evacuation from a Type-I door onto a ramp, such that the CAMI report’s findings of significance for these variables deserve more attention than is currently given.

As discussed above, CAMI collected a study sample that was designed with the goal of resembling the U.S. population or flying public, excepting the need to limit some demographics due to the safety hazards inherent in the trials. The rationale for that goal is not stated in CAMI’s report but seems to have originated from the research project’s interest in pre-trial testing of seat dimensions (i.e., a 26-inch seat pitch) that are not currently in airline service to determine if they may be practical and deserving of evaluation in trials. Seat pitch and width experiments on a study test group resembling the flying public generally (and thus the airline customer base) would provide a reasonable gauge for such determinations. However, given concerns about passengers with larger body sizes being slowed by narrow seat configurations, that same study test group would not necessarily be the most appropriate one for evacuation trials intended to inform decisions about whether to regulate seat configurations. The passenger composition of individual flights will vary such that even if CAMI’s study test group

___________________

¹⁰ McLean, G. A., et al. (2002). Access-to-Egress I: Interactive Effects of Factors That Control the Emergency Evacuation of Naïve Passengers Through the Transport Airplane Type-III Overwing Exit. DOT/FAA/AM-02/16. Washington, DC: U.S. Department of Transportation. https://libraryonline.erau.edu/online-full-text/faa-aviation-medicine-reports/AM02-16.pdf.

represented the flying public as a whole, it would certainly not be representative of the composition of the passenger loads on all flights. Indeed, one would expect that some flights will have disproportionately large numbers of passengers with large body sizes, and, of course, any assessment of the need for regulatory standards to ensure safety must account for that possibility.

CAMI’s own findings about body size variables affecting individual evacuation times raise questions about why the researchers did not make a greater effort to ensure that the study test group for the evacuation trials was not representative of the flying public generally but composed instead of participants who could represent a more plausible extreme in body sizes. Running the same evacuation trials but using a study test group having this composition would have provided a more compelling and robust test of CAMI’s hypothesis that seat dimensions should not influence evacuation performance. Had the evacuation trials yielded the same results, then CAMI could have been more confident about the project’s relevance to policymakers and their interest in knowing whether passenger seat configurations should be regulated. As it stands, the results of the CAMI project, designed to test an average rather than a plausible but more extreme passenger load composition, cannot provide that confidence.

Problematic as well is that CAMI cannot even be sure how its study test group cohorts, each consisting of smaller groups, resembled the broader NHANES demographic it had sought to achieve. Only 60 participants were required for each evacuation trial, and these participants, while meeting the age and gender selection requirements (pre-culling), do not necessarily satisfy the NHANES demographic. Moreover, in addition to excluding 55 participants (six based on the outcomes of the seat experiments and 49 for other reasons), the demographic and anthropometric makeup of the cohorts used to calculate evacuation time was clouded further by the removal of the times of the last six participants in most trials (i.e., a total of another 64 participants). At the very least, these issues and their potential implications should have been discussed in CAMI’s report.

While modifying the study test group recruitment to include more participants with larger body sizes would have been preferable to CAMI’s decision to run its trials on a study test group intended to represent an average mix of airline travelers, it would have been better still for CAMI to have designed its experiments to more directly assess the potential for body size variables to interact with seat configurations to affect evacuation performance. Measures for body size, for instance, could have been treated as independent variables and manipulated during the trials (e.g., by varying the number of people with a given body size or their seating location), which would have been possible if recruitment had emphasized body size. In the next chapter, the committee offers some ideas for CAMI to further this interest and address other concerns raised in this chapter.