air safety attitudes: preliminary insights post mh370
DESCRIPTION
Presentation by Dr Matthew Beck, of the Institute of Transport and Logistics Studies, University of Sydney. Delivered as a seminar on 11 July 2014 at the Choice Modelling Centre, University Leeds: www.cmc.leeds.ac.uk The disappearance of Malaysian Air Flight MH370 on the 8th of March 2014 received worldwide media attention. Whilst air disasters resonant with the wider public more so than other transport disasters, the lack of information on why MH370 disappeared has the potential to make this incident particularly affective. With the exponential growth in international travel, along with a hyper-competitive marketplace for air service providers, understanding how travellers might react to such disasters will help inform the decisions of relevant policy makers, who are seeking to alleviate the concerns of the travelling public and minimise the potential for future issues. This study presents the preliminary results from a series of stated preference type experiments that examine the attitudes of potential travellers with regards to air travel and air travel safety. Dr Matthew Beck is a Senior Lecturer and Program Director for Transport and Infrastructure Management at the Institute of Transport and Logistics Studies, University of Sydney. Matthew recently completed a PhD investigating the role of group decision making in the purchase of household motor vehicles. Prior to that he completed a Master of Philosophy in Marketing, also from University of Sydney, where he examined the transition of casual sporting fans to fanatical supporters. Matthew is also extremely active in consulting, working on diverse brand management, media relations, fast moving consumer goods and pharmaceutical projects. In the rare periods when not working, you might find him working on his real objective; becoming the number one golfer in transportation and logistics. http://sydney.edu.au/business/staff/matthewbTRANSCRIPT
Air Safety: Preliminary Perceptions Post MH370Choice Seminar, University of Leeds, July 2014
DR. MATTHEW J. [email protected]
Co-Researchers
› Professor John Rose:
2
› Dr. Rico Merkert
Co-Researchers
3
Tourism and Economic Activity
4
5
Growth in International Travel
Air Disasters: A Lasting Impression
› 1985: Pan-Am Flight 103 (Lockerbie Bombing)
› 2000: Air France Flight 4590 (Concorde)
› 2001: American Airlines Flight 11 – One World Trade Center
United Airlines Flight 175 – Two World Trade Center
American Airlines Flight 77 – Pentagon
› 2009: Air France Flight 447 (Mid-Air Stall)
› 2009: US Airways Flight 154 (Hudson River)
6
Air Disasters: A Lasting Impression
7
Media Attention to MH370
8
Media Reports & Public Perceptions
› Technology deficiencies:
- You can track an iPhone anywhere in the world, but not a plane?
› Past performance is no guarantee of future safety:
- 2012 and 2013 two of the safest years in aviation history
› Airspace management:
- Is anyone tracking these planes in real time?
- Exactly who is watching who (NSA, Malaysia and Indonesia military radar)?
› Passport control, airport and airline security:
- 2 passengers travelling undetected on stolen passports
- Much speculation as to the mental state of the pilot of MH370
9
Economic Impact of an Air Disaster: 911
› Global long-haul travel grew 40% from 2000 to 2010:
- Overseas travel to the United States during this same timeframe rose 2%
› Calculated Cost to US Economy:
- 68.3 million lost visitors
- 441,000 lost jobs
- $509 billion in lost spending
- $270 billion in lost trade surplus
- $32 billion in lost tax revenue
10
MH370 and the Malaysian Economy
› In 2012, tourism accounted for approximately:
- 16% of Malaysian GDP
- 14% of employment
› 12% of tourists were from China
› Passenger numbers were down 60% for Malaysian Airlines post-MH370:
- Value of shares down to 22.5 Malaysian Sen ($0.04 / ₤0.04)
› Public relations disaster for the Malaysian government:
- Lacking basic empathy
- Slow release of information (at times inaccurate and contradictory)
11
12
› Interesting thought experiment!
› International travel to and from Australia strictly dominated by air
› Are people willing to choose more invasive safety measures?
› Are people willing allocate more time to safety measures?
› Are perceptions of destination safety influenced by carrier incidents?
› Do these attitudes persist or subside as time from critical event increases?
Research Objectives
13
› One survey, four instruments:
One Survey: Four Instruments
Likert scale measurement of general attitudes
SP experiment to examine preferences for safety procedures
Best-worst scaling experiment to examine attitudes towards aviation
safety / security
Best-worst scaling experiment to examine perceptions of route safety
14
› Likert scale analysis to identify general attitudinal concerns:
- 1 = Strongly Disagree to 7 = Strongly Agree
- I find that I have to spend too much time at airports waiting in lines
- So long as I don’t miss my fight, spending time in lines is not a problem for me
- If I am making a connecting flight and have already been screened I should not have to wait to be screened again
- Overall I feel safe when on-board an aircraft
- Mechanical issues are a threat to the safety of my flight
- Other passengers are a threat to the safety of my flight
- Airline staff (pilots/cabin crew/ground staff) are a threat to the safety of my flight
- I feel that airlines and authorities currently do enough to make air travel safe
- I feel that airlines and authorities do enough to minimise mechanical threats
- I feel that airlines and authorities do enough to minimise threats from passengers
- I feel that airlines and authorities do enough to minimise threats from staff (pilots/cabin crew/ground staff)
Survey Instrument: Attitudinal Scales
Results: Attitudinal Scales
15
Strongly Disagree
2 3 4 5 6 Strongly Agree
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0%
4%7%
13%
22%
40%
12%
Overall I feel safe on-board an aircraft
Results: Attitudinal Scales
16
Strongly Disagree
2 3 4 5 6 Strongly Agree
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
3%
12% 10%
25%22% 22%
4%
Airlines & authorities do enough to make air travel safe
Results: Attitudinal Scales
17
Airline Staff Threats
Don't Mind Waiting
Too Much Time
Authorities - Staff
Authorities - Mechanical
Authorites - Passengers
Other Passenger Threats
Transfer Screening
Mechanical Threats
1 2 3 4 5 6 7
5.2 (t = 5.30)
5.2 (t = 5.63)
5.2 (t = 0.81)
4.8 (t = 5.40)
4.8 (t = 5.01)
4.7 (t = 4.00)
4.7 (t = 3.97)
3.8 (t = -1.14)
3.5 (t = -2.13)
Survey Instrument: Route Safety
› Best-worst scaling experiment; 6 choice sets; 4 routes per choice:
› 26 routes/destination for comparison:
18
MOST Secure LEAST Secure
q France qq China qq Malaysia qq Canada q
Abu Dhabi Germany New Zealand Thailand
Bali Hong Kong Pakistan Turkey
Canada India Russia United Kingdom
China Italy Saudi Arabia United States
Dubai Japan Singapore Vietnam
Egypt Malaysia South Africa
France Mexico South Korea
Results: Route Safety (MNL)
› MNL Results:
- Best/Worst jointly estimated – preliminary results
19
Observations 408
Initial log-likelihood -1081.31
Final log-likelihood -633.22
0.414
Variable Value Robust S.E. Robust t
Scale (Worst) 0.898 0.134 -0.81
Alt 1 (ASC) 0.147 0.343 0.43
Alt 2 (ASC) 0.034 0.185 0.19
Alt 3 (ASC) -0.165 0.229 -0.72
Results: Route Safety (MNL)
Variable Value Robust S.E. Robust tAbu Dhabi -0.146 0.418 -0.35
Bali -1.520 0.490 -3.10
Canada 2.330 0.677 3.44
China 0.763 0.622 1.23
Dubai 1.170 0.507 2.30
Egypt -2.860 0.605 -4.73
France 1.800 0.510 3.52
Germany 1.340 0.509 2.63
Hong Kong 0.924 0.543 1.70
India -1.830 0.797 -2.29
Italy 1.410 0.450 3.12
Japan 1.740 0.590 2.95
Malaysia -1.860 0.608 -3.06
Mexico -1.260 0.820 -1.53
New Zealand 2.960 0.756 3.92
Pakistan -3.640 0.667 -5.46
Russia -1.350 0.481 -2.80
Saudi Arabia -1.450 0.553 -2.63
Singapore 2.580 0.572 4.52
South Africa -1.060 0.477 -2.22
Thailand -0.859 0.733 -1.17
Turkey -1.040 0.415 -2.51
United Kingdom 2.260 0.594 3.81
United States -0.116 0.536 -0.22
Vietnam -1.270 0.633 -2.00
20
Results: Route Safety (MNL)
21
PakistanEgypt
MalaysiaIndia
BaliSaudi Arabia
RussiaMexico
VietnamSouth Africa
TurkeyThailand
Abu DhabiUnited States
South KoreaChina
Hong KongDubai
GermanyItaly
JapanFrance
United KingdomCanada
SignaporeNew Zealand
-35% -25% -15% -5% 5% 15% 25% 35%
Results: Route Safety (LCM)
› Latent Class Model Results:
- Best/Worst jointly estimated – preliminary results
22
Observations 408
Initial log-likelihood -969.114
Final log-likelihood -571.586
0.410
Variable Value Robust S.E. Robust t
Class 1 -0.557 0.852 -0.65
Gender (F) 1.98 0.751 2.63
Class 1 – Scale (Worst) 1.132 0.261 0.47
Class 2 – Scale (Worst) 0.975 0.275 -0.09
Results: Route Safety (LCM)
23
CLASS 1Route
CLASS 2Value Robust S.E. Robust t Value Robust S.E. Robust t0.887 0.894 0.99 Abu Dhabi 0.140 0.890 0.16-0.886 0.909 -0.97 Bali -0.577 1.040 -0.562.900 0.870 3.34 Canada 2.120 0.609 3.480.384 0.931 0.41 China 4.360 1.610 2.722.610 0.946 2.76 Dubai 0.142 1.450 0.10-1.590 1.070 -1.49 Egypt -2.240 1.000 -2.233.390 0.848 4.00 France 0.643 1.040 0.622.440 0.902 2.70 Germany 0.688 0.725 0.952.130 0.936 2.28 Hong Kong 0.218 0.549 0.40-0.574 1.080 -0.53 India -2.060 1.020 -2.012.200 0.895 2.46 Italy 1.360 0.799 1.712.130 0.910 2.34 Japan 2.170 0.547 3.96-1.020 0.976 -1.04 Malaysia -1.690 1.180 -1.43-0.570 1.050 -0.54 Mexico -1.880 1.190 -1.584.200 0.992 4.23 New Zealand 2.060 1.240 1.65-2.860 1.040 -2.75 Pakistan -2.870 0.662 -4.34-0.635 0.905 -0.70 Russia -0.733 1.150 -0.64-0.811 0.893 -0.91 Saudi Arabia -0.975 1.130 -0.863.550 0.938 3.78 Singapore 2.940 1.520 1.93-0.325 0.915 -0.35 South Africa -0.986 1.030 -0.96-1.480 0.934 -1.59 Thailand 1.360 1.200 1.14-0.113 0.827 -0.14 Turkey -0.884 0.692 -1.283.850 0.964 4.00 United Kingdom 1.360 1.110 1.231.520 0.916 1.66 United States -1.110 1.290 -0.86-0.903 0.963 -0.94 Vietnam -0.298 0.712 -0.42
Results: Route Safety (LCM)
24
PakistanEgypt
ThailandMalaysiaVietnam
BaliSaudi Arabia
RussiaIndia
MexicoSouth Africa
TurkeySouth Korea
ChinaAbu Dhabi
United StatesHong Kong
JapanItaly
GermanyDubai
CanadaFrance
SingaporeUnited Kingdom
New Zealand
-55% -45% -35% -25% -15% -5% 5% 15% 25% 35% 45% 55%
PakistanEgypt
ThailandMalaysiaVietnam
BaliSaudi Arabia
RussiaIndia
MexicoSouth Africa
TurkeySouth Korea
ChinaAbu Dhabi
United StatesHong Kong
JapanItaly
GermanyDubai
CanadaFrance
SingaporeUnited Kingdom
New Zealand
-55% -45% -35% -25% -15% -5% 5% 15% 25% 35% 45% 55%
Class 1 ~ Female Class 2 ~ Male
Survey Instrument: Safety Attitudes
› Best-worst scaling experiment; 6 choice sets; 9 statements per choice:
25
Agree with MOST Agree with Least
q Airport security will never eliminate all threats to flights q
q I am willing to spend any amount of time in security if it will improve safety q
q People should be targeted for security screening by authorities q
qPat down and luggage x-ray are sufficient security
Measures for me to feel secure q
qThe collection of genetic material is not needed as
Part of security procedures q
q My privacy and dignity should be respected during the security process q
q Current allowances for liquids are acceptable q
qPilots / cabin crew / ground staff should have the same
security measures as passengers q
q CCTV should be installed at all airports q
Survey Instrument: Safety Attitudes
› Attitudes broken into 9 broad classes (29 statements in total):- Stop Threats: How able airport security is to stop threats to flights
- Time Spent: How willing passengers are to spend more time on security
- Scan Passengers: How people should be selected for security screening
- Scan Invasive: How invasive passengers are prepared to let security be
- Biometrics: What level of biometric data people will be prepared to allow
- Privacy: The importance of privacy and dignity in security screening
- Liquids: How much liquids passengers are allowed to take on flights
- Scan Staff: How rigorous security screening is for airline/airport staff
- Images: How images should be used/kept for security screening
› Within each class there are 3 to 4 differing levels of statements:
- Each level is “higher” than the previous
26
Survey Instrument: Safety Attitudes
27
Category Level Description
Stop Threats
1 Airport security is able to stop all threats to flights
2 Airport security significantly reduces the level of threats to flights
0 Airport security will never eliminate all threats to flights
Time Spent
1 Airport security should not be increased if it leads to delays regardless of safety/security
2 I am willing to spend a little more time in security than I currently do if it will improve safety
0 I am willing to spend any amount of time in security if it will improve safety
Scan Passengers
1 People should be randomly selected for security screening
2 People should be targeted for security screening by authorities
0 All people should be selected for security screening
Survey Instrument: Safety Attitudes
28
Category Level Description
ScanInvasive
1 Pat down and luggage x-ray are sufficient security measures for me to feel secure
2 Whole of body scans and luggage x-ray are required security measures for me to feel secure
0 I would allow any level security no matter how invasive in order to feel secure
Biometrics
1 The collection of genetic material is not needed as part of security procedures
2 I would permit authorities to collect finger prints as part of security procedures
3 I would permit authorities to take retinal scans as part of security procedures
0 I would permit authorities to collect any material needed for a DNA sample as part of security procedures
Privacy
1 My privacy and dignity should be respected during the security process
2 I would allow increased security if my privacy and dignity was respected
0 Privacy and dignity is irrelevant compared to guaranteeing airport security
Survey Instrument: Safety Attitudes
29
Category Level Description
Liquids
1 Passengers should be able to take more liquids onto flights
2 Current allowances for liquids are acceptable
0 All liquids should be confiscated no matter the amount
Scan Staff
1 Pilots / cabin crew / ground staff are adequately screened
2 Pilots / cabin crew / ground staff should have the same security measures as passengers
3 Pilots / cabin crew / ground staff should have increased security measures compared to passengers
0 Pilots / cabin crew / ground staff should have their mental state assessed before every flight
Images
1 There is no need for CCTV at airports
2 CCTV should be installed at all airports
0 CCTV should be installed at all airports and images of passengers boarding should be saved
Results: Security Attitudes (MNL)
› MNL Results:
- Best/Worst jointly estimated – preliminary results
30
Observations 408
Initial log-likelihood -1661.10
Final log-likelihood -1364.58
0.179
Variable Value Robust S.E. Robust t
Scale (Worst) 0.769 0.148 -1.77
Results: Security Attitudes (MNL)
Variable Value Robust S.E. Robust tStop Threats 0.179 0.375 0.48
Stop1 -3.520 0.573 -6.15Stop2 -0.027 0.391 -0.07
Time Spent -1.790 0.461 -3.89Time1 -0.131 0.589 -0.22Time2 1.950 0.487 4.01
Scan Pass. -0.591 0.431 -1.37ScanP1 -0.697 0.519 -1.34ScanP2 -0.376 0.410 -0.92
Scan Invasive -2.250 0.526 -4.29ScanInv1 1.110 0.534 2.08ScanInv2 1.640 0.607 2.7
31
Variable Value Robust S.E. Robust tBiometrics -1.860 0.510 -3.64
Bio1 0.885 0.636 1.39Bio2 2.490 0.481 5.17Bio3 1.220 0.542 2.24
Privacy -2.490 0.501 -4.96Priv1 2.670 0.576 4.63Priv2 3.220 0.576 5.59
Liquids -3.230 0.467 -6.92Liq1 1.460 0.599 2.44Liq2 1.910 0.441 4.34
Scan Staff -0.652 0.341 -1.91ScanS1 -0.794 0.463 -1.72ScanS2 0.442 0.358 1.23ScanS3 0.007 0.362 0.02Image1 -4.110 0.608 -6.76Image2 0.299 0.292 1.02
Results: Security Attitudes (MNL)
32
Image1Stop1
LiquidsPrivacy
Scan InvasiveBiometrics
Bio1Time Spent
Time1Liq1Liq2
ScanInv1ScanStaff
ScanS1ScanS2ScanS3
Bio3ScanInv2
Stop ThreatsStop2
ScanPassScanP1ScanP2Image2
ImageTime2Priv1Bio2
Priv2
-25% -20% -15% -10% -5% 0% 5% 10% 15% 20% 25%
Results: Security Attitudes (MNL)
› 5 statements agreed with most:- I would allow increased security if my privacy and dignity was respected
- I would permit authorities to collect finger prints as part of security procedures
- My privacy and dignity should be respected during the security process
- I am willing to spend a little more time in security than I currently do if it will improve safety
- CCTV should be installed at all airports and images of passengers boarding should be saved
› 5 statements agreed with least:- There is no need for CCTV at airports
- Airport security is able to stop all threats to flights
- All liquids should be confiscated no matter the amount
- Privacy and dignity is irrelevant compared to guaranteeing airport security
- I would allow any level security no matter how invasive in order to feel secure
- (I would permit authorities to collect any material needed for a DNA sample as part of security procedures)
33
Results: Security Attitudes (LCM)
› MNL Results:
- Best/Worst jointly estimated – preliminary results
34
Observations 408
Initial log-likelihood -1661.10
Final log-likelihood -1294.583
0.221
Variable Value Robust S.E. Robust t
Class 1 0.824 1.21 0.68
Gender (F) -1.59 0.738 -2.16
Results: Security Attitudes (LCM)
35
Stop1Privacy
Image1Image2Liquids
Scan InvasiveTime Spent
Time1Biometrics
Liq2ScanInv2
Liq1ScanInv1
Time2Stop Threats
Stop2ScanPass
ScanP1ScanP2
ScanStaffScanS1ScanS2ScanS3Image
Bio1Bio3Priv1Bio2Priv2
-50% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50%
Image1Liquids
Liq1Time1
ScanP1Liq2
ScanP2ScanStaff
ScanS1ScanS2ScanS3
Scan InvasiveScanInv1ScanInv2
Stop ThreatsStop1Stop2
Time SpentScanPass
BiometricsBio1Bio2Bio3
PrivacyPriv1Priv2
Image2ImageTime2
-50% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50%
Class 1 ~ Male Class 2 ~ Female
Results: Security Attitudes (LCM)
› Prominent Male Attitudes:
- My privacy and dignity should be respected during the security process
- Airport security is able to stop all threats to flights
- I would permit authorities to collect finger prints as part of security procedures
- I would allow increased security if my privacy and dignity was respected
› Prominent Female Attitudes:
- There is no need for CCTV at airports
- I am willing to spend a little more time in security than I currently do if it will improve safety
- All liquids should be confiscated no matter the amount
- Passengers should be able to take more liquids onto flights
36
Survey Instrument: Stated Preference
› Two alternatives security procedures describe by five attributes:
37
Level 1 Level 2 Level 3 Level 4
Level of Physical Interaction
Partial pat-down (current practice)
Thorough pat-down for targeted passengers only
Thorough pat-down forall passengers
Level of Luggage Screening
X-Ray with luggage opened for targeted cases only
(current practice)
X-Ray with luggage opened randomly
X-Ray with all luggage opened
Level of Physical Screening
Metal detector for all passengers
(current practice)
Whole of body scan for targeted passengers only
Whole of body scan forrandom passengers
Whole of body scan for all passengers
Identification Requirement
Passport(current practice)
Passport with finger printand/or retinal scan
Passport with DNA verification (e.g. hair or
saliva sample)
Waiting Time 10 minutes in security line 20 minutes in security line 40 minutes in security line 60 minutes in security line
Survey Instrument: Stated Preference
SECURITY PROCESS A SECURITY PROCESS B
The level of physical interactionpassengers have with security staff
Partial pat-down (Current Practice)
Thorough personal search of all passengers
The level of security screening for all luggage
X-Ray with luggage openedfor targeted cases only
(Current Practice)
X-Ray with some luggage opened randomly
The level of security screeningpassengers are required to complete
Whole of body scan for all passengers
Whole of body scan forrandom passengers
The level of identity verification required
Passport with finger print and/or retinal scan
Passport(Current Practice)
The average amount of time required to complete security checks only
10 minutes to complete security (including waiting)
20 minutes to complete security (including waiting)
Which of these two security procedures do you PREFER MOST?
(please choose ONE) q q
38
Results: Route Safety (LCM)
- Willing to give up 8.3 minutes for more scanning than metal detectors
- Will accept only passport only as identification if it saves 18.4 minutes
39
Sample Size 402
Initial log-likelihood -278.645
Final log-likelihood -231.627
0.166
Variable Value S.E. t
Metal detector for all passengers -0.340 0.166 -2.05
Passport 0.757 0.142 5.33
Waiting Time -0.041 0.006 -6.91
Results: Route Safety (LCM)
› Note that “Overall Safety” was dummy coded (6 or 7 = 1 vs 5 or less = 0)
- Class 1 Average Probability: 54.1%
- Class 2 Average Probability: 45.9%
40
Sample Size 402
Initial log-likelihood -278.645
Final log-likelihood -197.253
0.288
Variable Value S.E. t
Class 1 -0.340 0.166 -2.05
Overall Feel Safe 0.757 0.142 5.33
Results: Route Safety (LCM)
Class 1 (Feel Safe) Class 2 (Less Safe)
Variable Value S.E. t Value S.E. t
Thorough pat-down for targeted passengers only
-1.478 0.664 -2.22 0.463 0.256 1.81
X-Ray with luggage opened for targeted cases only
-0.394 0.299 -1.32 -0.379 0.212 -1.79
Metal detector for all passengers -2.172 0.722 -3.01 0.090 0.273 0.33
Passport 0.714 0.392 1.89 1.507 0.288 5.24
Passport with finger print and/or retinal scan
0.784 0.392 2.00 -0.355 0.276 -1.29
Waiting Time -0.151 0.037 -4.09 0.015 0.012 1.25
41
Results: Route Safety (LCM)
› Class 1 (Feel Safe):
- Time sensitive
- Will give up 9.6 minutes to avoid physical pat-downs for targeted passengers
- Will give up 14 minutes to avoid only scanning via metal detectors only
- Will accept only passport as identification if it saves them 4.7 minutes
- Will accept passport and retinal scan / finger printing if it saves 5.2 minutes
› Class 2 (Less Safe):
- Time insensitive
- Strong preference for passport only for ID purposes (avoid invasive data collection?)
- Weaker preference for thorough pat-downs for targeted passengers (not them?)
- Weaker preference for a more thorough luggage checks
42