You are on page 1of 9

Accident Analysis and Prevention 57 (2013) 131139

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention


journal homepage: www.elsevier.com/locate/aap

The effects of risk perception and ight experience on airline pilots


locus of control with regard to safety operation behaviors
Xuqun You a , Ming Ji a, , Haiyan Han b
a
School of Psychology, Shaanxi Normal University, Xian, Shaanxi 710062, China
b
Department of Education, Xian University of Arts and Science, Xian, Shaanxi 710065, China

a r t i c l e i n f o a b s t r a c t

Article history: The primary objective of this paper was to integrate two research traditions, social cognition approach
Received 14 February 2012 and individual state approach, and to understand the relationships between locus of control (LOC), risk
Received in revised form 30 March 2013 perception, ight time, and safety operation behavior (SOB) among Chinese airline pilots. The study
Accepted 31 March 2013
sample consisted of 193 commercial airline pilots from China Southern Airlines Ltd. The results showed
that internal locus of control directly affected pilot safety operation behavior. Risk perception seemed to
Keywords:
mediate the relationship between locus of control and safety operation behaviors, and total ight time
Locus of control
moderated internal locus of control. Thus, locus of control primarily inuences safety operation behavior
Risk perception
Flight experience
indirectly by affecting risk perception. The total effect of internal locus of control on safety behaviors is
Safety operation behavior larger than that of external locus of control. Furthermore, the safety benet of ight experience is more
pronounced among pilots with high internal loci of control in the early and middle ight building stages.
Practical implications for aviation safety and directions for future research are also discussed.
2013 Elsevier Ltd. All rights reserved.

1. Introduction behavioral control) inuence aviation SOB (OHare, 1990; Hunter,


2005; Stewart, 2006; Ji et al., 2011a). Based on this research, one
Safety operation behavior (SOB) refers to the non-technical of the most popular strategies of promoting aviation safety is to
skills or social psychological skills that pilots perform and exhibit hire the pilots with the most appropriate personality traits and
during ight tasks, including communication and collaboration, change current pilot attitudes and risk perceptions related to risky
decision making, workload management, situation awareness, operations. However, several literature reviews have concluded
and ight automation management, all of which help guarantee that the majority of these aviation safety campaigns have not had
modern aviation safety (OConnor et al., 2002; You et al., 2009). any effect on the number of accidents (ODea et al., 2010). On one
However, as electromechanical systems have improved, the hand, the role of individual state factors often has been ignored
relative contribution of human factors in aviation accidents has with regard to these campaigns. Personality traits are genetically
increased such that approximately 80% of aviation accidents have determined (Eysenck, 1967) or due to habits acquired via social
a component of human error (Jensen and Benel, 1977; Wiegmann learning, which suggests they are relatively enduring dispositions.
and Shappell, 1997). Consequently, a large body of research has However, situational factors inuence individual states and are
been conducted to identify the factors that affect SOB. Some therefore malleable across time and situations. In fact, the ight
studies have demonstrated that pilot personality traits inuence operating behaviors of crewmembers can be classied as a dynamic
SOB including the Big Five personality factors (i.e., openness, process that depends more on individual states than dispositional
conscientiousness, extraversion, agreeableness, and neuroticism), traits. Several studies have supported this hypothesis (Stewart,
sensation seeking, and risk tolerance (Poropat, 2009; Berg et al., 2006; Hunter and Stewart, 2009). On the other hand, the indirect
2002; Helmreich et al., 1997; Musson et al., 2004; Pauley et al., effect of these state factors might be underestimated with regard to
2008). Others have found that certain social cognition vari- accident involvement and risky operation behavior. In particular,
ables (e.g., attitudes, perceived risk, social norms, and perceived the role of these state factors in the relationship between social
cognition variables and SOB has rarely been studied. Theoreti-
cally, individual state might affect an individuals perception and
appraisal of the environment (McCrae and Costa, 1995). Such
This study was supported by the joint fund of Civil Aviation Research of China
appraisals affect behavior. A similar observation has been found
(61079004) and by the Fundamental Research Funds for the Central Universities of
China (GK201302051).
in several studies (Mischel, 1968; Stewart, 2008). However, no
Corresponding author. Tel.: +86 29 85 30 38 52; fax: +86 29 85 30 32 36. research has formally tested these relationships. To ll this gap, we
E-mail address: jiming@snnu.edu.cn (M. Ji). empirically examine two research questions: Are risk perception

0001-4575/$ see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.aap.2013.03.036
132 X. You et al. / Accident Analysis and Prevention 57 (2013) 131139

H4
Risk Perception

H3 H2
Locus of Control
Flight Experience
Internal
H5 Safety Operation
Behaviors
External
H1

Fig. 1. Model of the safety operation behaviors.

and ight experience associated with the SOB? and, Does LOC lead face emergency situations that require them to choose between
individuals to perceive different levels of risk in the same situa- several alternatives with little time, the consequence of which
tion? To answer these questions, we tested the model shown in might include a possible loss of the aircraft or even the crew
Fig. 1. The following section describes the theoretical background (Stewart, 2008). Thus, extensive research has dealt with the role
of the study, the constructs in the model, and the hypotheses. of LOC in accident causality and risky operation behavior. Previous
research ndings have shown that pilots exhibit more internality
1.1. Locus of control than externality (Hunter, 2002; Vallee, 2006; Joseph and Ganesh,
2006) and have found a signicant correlation (r = 0.20) between
Locus of control (LOC) is a concept from Rotters (1954) social the number of hazardous events (i.e., no-fatal accidents) experi-
learning theory. It refers to the degree to which people perceive enced by pilots and their internality score (Hunter, 2002). Pilots
that the outcomes of the situations that they experience are under who had lower perceived internality scores tended to experience
their control. This control consists of a set of expectancies that more hazardous aviation events compared with those with higher
are inuenced by ones own efforts (internal) or environmental scores. Wichman and Ball (1983) also found that internals tended
forces that are beyond ones control (external). Individuals with to hold stronger self-serving biases compared with externals. For
an internal LOC orientation perceive that they have control over instance, internal pilots believed that they were signicantly more
the outcome of a situation, whereas individuals with an exter- skilled and less likely to have an accident than did pilots with higher
nal LOC attribute outcomes to external factors such as chance, external LOC. Recently, researchers examined the relationships
luck or the actions of other people. Stewart (2006) suggested that between LOC, accidents, situation awareness, and workload among
one misconception of LOC is the assumption that it is a person- operators of ground vehicles. They found that LOC predicted many
ality construct. This misconception might be because the popular attitudes and behaviors that are consistent with aviation safety
dimension of internalityexternality (IE) is supercially similar to including risk taking, risk management, the management of multi-
that of introversion-extraversion (Jung, 1928), a personality con- ple tasks, distractibility and time management (Arthur et al., 1991;
struct with similar descriptors. In fact, LOC describes situational Guastello and Guastello, 1986; Ozkan and Lajunen, 2005; Stanton
states and not dispositional traits. However, Rotters social learn- and Young, 2005). In addition, Salminen and Klen (1994) found
ing orientation conceptualizes personality as a situational state a relationship between LOC and risk-taking behavior in a non-
composed of self-attributions and expectancies rather than a dispo- aviation sample that consisted of Finnish construction and forestry
sitional trait. Self-attribution theory includes situations that might workers. They found that forestry workers had more external LOCs
be positive or negative. Individual excess might be attributed to than construction workers. Regardless of these between-group dif-
external factors or an individual succumbing to external factors ferences, externals tended to take more risks than internals. Based
(Heider, 1958). Situational factors inuence expectancies and are on these arguments, our rst hypothesis is the following: H1: LOC
therefore malleable, whereas stable dispositions tend to be resis- predicts pilot SOB. Internal will be associated with a higher level of
tant to change across a variety of situations. Another reason that SOB, and external will be associated with a lower level of SOB.
LOC is a situational state factor is founded in methodology. The
Rotter IE scale does not establish a typology. However, LOC scale 1.2. Risk perception
score distributions are often dichotomized to create a between-
participants variable for typical mixed factorial designs. Obviously, Risk perception is an important skill to identify with regard
performing a median split to dichotomize the data sacrices varia- to hazardous ight situations. This perception allows pilots to
tion (i.e., information) in the dataset for simplicity and convenience. overcome the complex cognitive demands that the aviation envi-
Lefcourt (1982) emphasized that the practice of dichotomizing the ronment dictates. Hunter (2002) dened risk perception as the
IE continuum into internal and external factors has led to the com- essentially cognitive ability to discern the risk inherent in a sit-
mon erroneous assumption that LOC is a stable dispositional trait uation; this skill involves an accurate appraisal of the external
within individuals. This notion has been perpetuated by much of situation and ones personal capacities. Underestimating the exter-
the literature, which has referred to as a personality trait. Conse- nal situation or overestimating ones personal capacities leads to
quently, it is easy to fall into the trap of assuming that LOC is more a misperception of the risk. Compared with other professionals,
enduring and stable than it actually is. LOC scores are best inter- pilots are more likely to accurately appraise the probabilities of the
preted as momentary samples of ones beliefs about causality or specic ight risks. Thus, risk perception is frequently considered
rough expectancies under ones control. Thus, as a predictive vari- a factor that contributes to aviation accidents.
able of attitudes and behaviors, LOC has been applied to a host of
settings over the last three decades, including traditional classroom 1.2.1. Risk perception and SOB
learning situations, industrial safety, and medicine. Pilots (OHare, 1990) and drivers (Trankle et al., 1990) might
Compared with the aforementioned professional staff (i.e., stu- fail to accurately perceive the risks involved in ying and driving,
dents, construction and forestry workers, and doctors), pilots often respectively, which contributes to accidents. Pilots and drivers who
X. You et al. / Accident Analysis and Prevention 57 (2013) 131139 133

do not correctly perceive the risks inherent to the situation might increasing trend between internality and age. Thus, a related issue
not take avoidance or other risk-mitigating actions, thereby caus- concerns the point in the career cycle at which pilots become more
ing accidents or incidents. The relationship between risk perception cautious. Stewart (2006) argued that internality peaks mid-career.
and pilot SOB is clearly strong. Hunter (2006) found signicant, neg- In the early stages of ones career, early experience builds stages.
ative correlations between measures from the Risk Self-Perception In addition, experience allows an individual to quickly diagnose
Scale and previous hazardous events. This result indicated that a situation and identify a feasible course of action. Therefore,
participants who had been in more hazardous aviation events (a) experience allows the individual to overcome the effects of time
tended to rate the scenarios as lower in risk and (b) more inaccu- pressures because little need exists to compare the feasibility of
rately estimated the safety of general aviation (p. 142). Ji et al. different action alternatives. In fact, Burian et al. (2000) found that
(2011b) collected a larger sample of airline pilots (N = 257) and pilots who were at or below the 25th percentile of total ight hours
found that pilots who perceived risk as high exhibited higher SOBs were more likely to commit an in-ight plan continuation error
than those who perceived risk as low. In addition, Simon et al. in the context of weather encounters (i.e., decide to continue to
(2000) found a relationship between risk perception and starting a y in adverse weather) compared with those at or above the 75th
venture in an entrepreneur sample. They found that risk perception percentile. Experience often produces better diagnostic decision-
directly inuenced the decision to start a venture. Consequently, making skills; however, it may also reduce pilot perception of
our second hypothesis is the following: H2: Risk perception will risk, particularly as it relates to adverse weather. Therefore, more
positively predict the level of pilot SOB. experienced pilots might choose to divert a ight around adverse
weather or choose to continue. Furthermore, Li and Baker (1994)
1.2.2. LOC and risk perception found that the risk of a crash decreases in a nonlinear fashion
Numerous studies have shown that internals have an ability as total ight time increases in a case-control study conducted
to attend to relevant cues and ignore those that are not relevant among commuter air carrier and air taxi pilots. Thus, theory and
(DuCette and Wolk, 1973; Gregory and Nelson, 1978; Sanders et al., empirical evidence led to the following hypothesis:
1976). False cues are more likely to inuence externals than inter- H5: Flight experience will moderate the relationship between
nals because the former were not vigilant enough. Internals relied LOC and pilot SOB.
more on their sense of time, whereas externals time estimates In sum, the existing evidence suggests that both LOC and risk
depended on deliberately false cues presented by the experimenter perception are strongly related to SOB. The present study exam-
(Stewart and Moore, 1978). Internals were better than externals ined the relative importance of LOC, risk perception and ight
at detecting errors while proofreading a text and with regard to experience with regard to SOB in aviation. A central aim was to
perceiving subtle and incidental cues (Wolk and DuCette, 1974). investigate whether LOC has indirect effects on SOB via its inu-
This ability is germane to aviation settings in which malfunctions ence on hazardous attitudes. Furthermore, we examined whether
can occur under conditions of high workload. Internals should be ight experience signicantly moderates the relationship between
able to more quickly detect a system malfunction and concen- LOC and SOB.
trate only on the tasks that are critical to controlling the aircraft
(if the malfunction is serious). As a result, we believe that inter- 2. Methods
nals will show smaller latencies in detecting and responding to
possible emergency situations compared with externals. Similarly, 2.1. Participants
inaccurate risk perception can lead pilots to ignore or misinterpret
external cues that demand immediate and effective decisions to In total, 193 experienced male commercial airline pilots from
avoid hazards. This ability suggests that LOC is associated with risk China Southern Airlines Ltd. participated in this research. The pilots
perception. Based on these data, Vallee (2006) demonstrated that age ranged from 22 to 54 years (M = 35.84, SD = 10.62) and total
LOC had a sizable effect on risk perception in a sample of pilots from ight hours ranged from 367 to 3941 h (M = 2,912.38, SD = 775.36).
the United Kingdom. Pilots who believed that they could control In addition, 4 ight experts from China Southern Airlines Ltd.,
negative events (and prevent their occurrence) considered them- including 2 ight managers, 1 check pilot, and 1 ight instruc-
selves less at risk than others. Thus, theory and empirical evidence tor, were invited to rate the level of the participants SOB, These
led to the following hypothesis: H3: LOC will predict pilot risk per- experts were aged between 37 and 46 years (M = 41.05, SD = 3.67)
ception. Internal orientation will be associated with a higher level and had a total ight time range between 8597 and 12,147 hours
of risk perception, but external orientation will be associated with (M = 10,103.25, SD = 986.56). These experts had received human
a lower level of risk perception. factors training as part of the Crew Resources Management (CRM)
As noted earlier, LOC directly inuences risk perception, and risk syllabus. They rated pilot SOB performance based on their over-
perception directly inuences pilot SOB; thus, LOC should indirectly all performance. All participants volunteered for this survey and
affect pilot SOB through its effect on risk perception. Based on this answered the questionnaires.
theory, we extend the logic of Hypotheses 23 and posit the follow-
ing prediction: H4: Risk perception fully mediates the relationship 2.2. Measures
between LOC and pilot SOB.
Data were collected on LOC, risk perception and SOB. Appendix
1.3. Flight experience A presents specic measures including its items and (where appro-
priate) the relevant internal reliabilities, mean scores and standard
Given that LOC is based on expectancy theory and that experi- deviations of those items.
ence affects subjective expectancies, it is reasonable that a sense of
personal control should grow with experience. Harvey (1971) dis- 2.2.1. Aviation Safety Locus of Control (ASLOC) Scale
covered that internal LOC scores predicted the length a person had The ASLOC Scale developed by Hunter (2002) consists of 20
occupied an executive government position: Senior administrators items ( = 0.75) and includes 10-item internal subscale ( = 0.69)
with more years of experience had higher scores than those with and 10-item external subscale ( = 0.73). For example, Pilots can
fewer years of experience. Age is another factor. Penk (1969) found do very little to avoid minor incidents while working is an exam-
that internality is positively and signicantly correlated with age. ple of external subscale. Participants responded to the items using
Consistent with this nding, Hunter (2002) found a signicant a 5-point Likert scale (i.e., Strongly Agree = 1; Strongly Disagree = 5).
134 X. You et al. / Accident Analysis and Prevention 57 (2013) 131139

The internal and external subscale scores were computed as the hierarchical regression analysis using SPSS 15.0. A structural
sum of the responses to the 10 items for each subscale. Each equation analysis was performed to investigate whether LOC has
subscale score had a possible minimum value of 10 (i.e., the lowest a direct effect on SOB and whether risk perception mediated the
level of agreement) and a maximum value of 50 (i.e., the highest
hypothesized effect of LOC on SOB using LISREL 8.70 (Joreskog,
level of agreement). 2004). A conrmatory factor analysis (CFA) was used to evaluate the
measurement model and the t of the LOC and risk perception sub-
2.2.2. Risk Perception Scale scales to their respective latent constructs. The following t indices
The Risk Perception Scale, developed by Hunter (2006), consists were used to assess the t of the model: the goodness-of-t index
of 26 sentences that describe an event or situation. Seven of the 26 (GFI), the adjusted goodness-of-t index (AGFI), the comparative
sentences describe non-aviation events (e.g., driving a car), and the t index (CFI) and the root mean square error of approximation
remainder are concerned with aviation. This scale includes 5 fac- (RMSEA). GFI, AGFI and CFI values of 0.9 or above and an RMSEA of
tors: (a) general ight risk, which included 10 items ( = 0.87) that 0.08 or below indicated an acceptable t (Hair et al., 2006).
address normal and high-risk ight operations; (b) high ight risk,
which consists of 6 items ( = 0.74) that describe high-risk ight 3. Results
conditions; (c) altitude risk, which consists of 3 items ( = 0.78)
that assess the ight altitude as a risk element; (d) driving risk, Descriptive statistics and Cronbachs alpha coefcients were
which consists of 3 items ( = 0.83) that describe different driv- calculated for each instrument in the study prior to analysis. The
ing situations; and (e) everyday risk, which is dened by 4 items numbers of items, mean scores, standard deviations, and internal
( = 0.71) that deal with everyday life situations other than driving. consistencies for all measures are listed in Table 1. The reliability
Participants rated the risk as if he or she would be involved in that coefcients were acceptable in that they achieved the suggested
situation tomorrow. The measure used a common rating scale to minimum value.
assign a risk rating to each scenario. This rating scale ranges from Table 2 displays the correlations between the variables. To sim-
1 (low risk) to 100 (high risk). A high score on this scale indicated a
plify the correlation matrix, latent variable scores (Joreskog and
high level of risk perception. Sorbom, 2001) were computed for risk perception and SOB. The
measurement models for the latent variables are shown in Fig. 2. As
2.2.3. SOB Scale shown in the table, the LOC measures were signicantly correlated
The SOB Scale was developed as a research project of China Civil with risk perception and SOB. Risk perception was correlated with
Aviation Authority (CCAA) based on the Line/LOS Checklist Ver- SOB. Pilots who perceived the risk related to aviation accidents as
sion 4.0 (LLCv4, Helmreich et al., 1997). It consists of 27 sentences high were more likely to operate safely. In addition, both risk per-
( = 0.85) that describe a crews operational behaviors or situations. ception and SOB were correlated with ight time. Pilots with more
This scale includes 5 factors: (a) automation system understanding ight time tended to perceive the risk related to aviation accidents
(7 items, = 0.73), leadership and management (6 items, = 0.69), as high and operated more safely.
situation awareness and decision making (7 items, = 0.81), and A structural equation analysis was performed to investigate
communication and cooperation (7 items, = 0.78; You et al., whether risk perception mediated the hypothesized effect of LOC
2009). Flight experts synthesized an appraisal of participants on SOB. Fig. 1 shows the tested model with standardized path coef-
based on their ight simulation training results, their real-time cients. For clarity, only signicant paths (p < 0.01) are shown in the
line operation safety audit (LOSA) observation records, and their gure. The t measures indicated that the proposed model t the
airlines safety performance appraisals over the last two years. All data well: 2 (41, n = 193) = 72.24, GFI = 0.91, AGFI = 0.90, CFI = 0.93,
items were rated on a 4-point Likert scale (1 = poor, 2 = minimum NNFI = 0.90, and RMSEA = 0.063. The effect of risk perception on SOB
expectations, 3 = standard, or 4 = outstanding)1 for the following 4 was signicant and positive (i.e., = 0.38, t = 3.29), which indicates
ight phases: preparation/skidding, taking off/climbing, cruising, that high levels of risk perception increase the likelihood that the
and descending/approach/landing. Each ight expert rated all pilot will engage in SOBs. This nding is consistent with previous
pilots, and the Kandall coefcient of concordance for the four studies based on social cognitive theory. Thus, Hypothesis 2 was
appraisals was 0.874 (2 = 1027.56, df = 192, p < 0.01), which fully supported. The direct effects of LOC on risk perception were
showed that the pilot safety operation appraisals were signi- signicant. More specically, external negatively predicted risk
cantly consistent. A mean score on each subscale was computed perception ( = .15, t = 1.96), whereas internal ( = .17, t = 2.01)
based on the items within it. A high score on the scale indicated a showed a positive relationship. All linear relationships were in the
high degree of safety operation. expected direction. Thus, Hypothesis 3 was fully supported.
With regard to the direct effects of LOC on SOBs, only internal
2.3. Data analyses had a signicant coefcient ( = .14, t = 1.92), which indicates that
high levels of internal increases the likelihood that a pilot will show
Cronbachs alpha coefcient was used to evaluate the internal SOBs. Thus, Hypothesis 1 was partly supported. Although exter-
consistencies of the aforementioned measures (see Appendix nal did not demonstrate an effect on SOBs, it had indirect effects
A). The moderating effects of risk perception on the relation- on SOBs via the mediating effect of risk perception. As shown in
ship between ight experience and SOB were estimated via a the path model, 31% of the variance in risk perception is explained
by the difference in LOC. The lack of a direct effect of external on
SOBs implies that risk perception generally mediates the effects on
SOBs. That is, LOC indirectly affects SOBs by inuencing risk per-
1
Poor: Performance is signicantly below expectations. This category includes
ception. Thus, Hypothesis 4 was fully supported. To determine the
instances in which necessary behavior was not present and examples of inap-
propriate behavior that were detrimental to mission effectiveness. Minimum total effect of the LOC variables on SOBs, both the direct and indirect
Expectations: Performance meets minimum requirements but ample room for effects of LOC were computed (Table 3). The total effects showed
improvement remains. This level of performance is less than desired for effective a similar pattern found in the correlation analysis in Table 2. High
crew operations. Standard: This demonstrated behavior promotes and maintains internal scores were associated with safety operation in aviation.
crew effectiveness. This level of performance should normally occur during ight.
Outstanding: Performance represents exceptional skill in the application of specic
In contrast, the higher external scores were related to less safety
behaviors and serves as a model for teamwork. This performance is truly noteworthy operations. However, external showed small total effects on SOBs
and effective. (i.e., the sizes of the standardized total effects).
X. You et al. / Accident Analysis and Prevention 57 (2013) 131139 135

Table 1
Number of items, mean scores and Cronbachs alpha for all measures.

Measures Number of item Mean S.D. Cronbachs alpha

Locus of control
Internal 10 3.51a 0.43 0.69
External 10 2.54a 0.40 0.73
Risk perception
General ight risk 10 51.04b 12.73 0.87
High ight risk 6 67.88b 10.08 0.74
Altitude risk 3 72.16b 15.24 0.78
Driving risk 3 55.38b 12.51 0.83
Everyday risk 4 48.29b 7.56 0.71
Safety operation behaviors
Automation system understanding 7 2.99c 0.49 0.73
Leadership and management 6 3.31c 0.51 0.69
Situation awareness and decision-making 7 3.14c 0.52 0.81
Communication and cooperation 7 3.08c 0.48 0.78
a
Range 15.
b
Range 1100.
c
Range 14.

Table 2
Means, standard deviations, and correlations among measures of the study variables (n = 193).

Variables 1 2 3 4 5 6

1. Age 1.00
2. Total ight time 0.81*** 1.00
3. Internal .17** .05 1.00
4. External .15** .04 .26** 1.00
5. Risk perception a .04 .12** .19** .12** 1.00
6. Safety operation behavior a .06 .13** .16** .15** .28** 1.00
a
Latent variable score was computed for risk perception and safety operation behavior.
**
p < 0.01.
***
p < 0.001.

Table 3 was not a signicant predictor ( = 0.05, p < 0.01). The interaction
Direct, indirect and total effects of LOC on safety operation behaviors in aviation
of ight time and internal signicantly predicted the performance
(standardized coefcients).
of SOB ( = 0.11, p < 0.01). Consequently, ight time moderated
Direct effect Indirect effect Total effect the relationship between internal and SOB. In addition, the study
Internal 0.14 0.07 0.21 applied tolerance, VIF, and DW tests to examine multicollinearity
External .06 .06 (Haan, 2002; Edriss, 2003). The tolerance was between 0.31 and
0.17 (T > 0.1), the VIF was between 3.19 and 6.03 (VIF < 10), and the
DW = 1.87 (du < d < 4 du ), which are acceptable values.
This study tested the moderating effect of ight time on the rela- To understand total ight time as a moderator of the results and
tionship between SOB and LOC using a regression analysis, and the examine its inuence, internal scores and total ight time were
results are shown in Table 4. After controlling for age, total ight divided into three groups: high (M + 1 SD), medium (M), and low
time and internal, the overall explanation rate of SOBs signicantly (M 1 SD). Fig. 3 shows the interaction of internal and total ight
increased in the regression equation. Internal signicantly pre- time with regard to SOBs. At a certain level of internal, pilot SOB
dicted the performance of SOB ( = 0.14, p > 0.05), whereas external signicantly improves as total ight time increases. That is, total

Fig. 2. Path diagram of the relationship between LOC, risk perception and safety operation behavior.
136 X. You et al. / Accident Analysis and Prevention 57 (2013) 131139

Table 4
Results about the relationships among LOC, total ight time, and SOB.

Variables Model 1 (SOB) Model 2 (SOB) Model 3 (SOB)

t t t

Step 1: controlled variables


Age 0.04 0.21 0.04 0.21 0.04 0.21
Step 2: predictor variables
** ** **
Total ight time 0.11 2.73 0.11 2.73 0.11 2.73
Internal 0.15** 3.24 0.14** 3.12
External 0.07 0.36 0.06 0.34
Total ight time internal 0.11** 2.74
Total ight time external 0.05 0.31
F 0.12** 11.52*** 17.83***
R2 0.10 0.15 0.27
Adjust R2 0.09 0.12 0.25

All parameter coefcients are standardized estimates.


**
p < 0.01.
***
p < 0.001.

are most likely able to detect a system malfunction more quickly


and concentrate on the tasks that are critical to controlling the air-
craft when the malfunction becomes serious. These suppositions
are reected by their SOBs. In addition, the results indicate the
direct effect that an internal LOC has on SOB. Pilots with higher
levels of internal were more likely to exhibit SOBs. A possible expla-
nation for this result is that pilots who are more internal actively
restructure ambiguous situations and seek information that will
assist them to solve problems and make decisions. Thus, they tend
to take fewer risks than externals (Salminen and Klen, 1994). More-
over, their way of addressing danger is not to ignore them but act
to reduce the danger (Wichman and Ball, 1983). In addition, inter-
nal LOCs also showed indirect effects on SOBs via the inuence of
risk perception. Internal is an exogenous variable that inuences
Fig. 3. The interaction effect of total ight time and internal to safety operation
behaviors.
risk perception, which in turn affects SOBs. Internal individuals
tended to perceive the risk related to aviation accidents as high
and showed more SOBs. Although the results did not demonstrate
a direct effect of external on SOBs, they showed that, similar to
ight time has a signicant moderating effect. When the internal
internal LOCs, external LOCs had indirect effects on SOBs. Exter-
scores are low/medium, the pilots with more ight time exhibit
nal individuals perceived the risk of aviation accidents as low and
a higher level of SOBs than those with less ight time. However,
had fewer SOBs. In sum, internal showed a large total effect on
when the internal scores were high, the pilots with more ight
SOBs ( = 0.21), whereas external demonstrated a small total effect
time exhibited the lower levels of SOBs than those with less ight
( = 0.06). Based on these results, it is reasonable to believe that
time. These results show that many hours of total ight time reduce
LOC, especially internal LOC, could be used as a reliable predictor
the positive effects of internal on the SOBs of high internal pilots,
of future pilots SOBs. In other words, strategies intended to pro-
whereas medium and few total hours of ight time did not greatly
mote aviation safety might be aimed at changing pilots senses of
affect them.
personal control such as self-control, fatalism, expectancies, and
social alienation according to the situational state approach (Duke
4. Discussion and conclusions et al., 1977).
Risk perception had a direct effect on SOB in the path model.
The primary aim of this study was to integrate social cogni- Note that risk perception includes ve factors (i.e., general ight
tive and situational state approaches to explore the relationships risk, high ight risk, altitude risk, driving risk, and everyday
between LOC, risk perception, ight time, and SOBs among Chinese risk), whereas SOBs encompass elements such as automation sys-
airline pilots. tem understanding, leadership, management, situation awareness,
Our results revealed the direct effects of LOC on risk percep- decision making, communication and cooperation. The results
tion and the direct effect of internality on behavior. Specically, showed that different aspects of risk perception and different inu-
high internal scores were positively associated with risk percep- ences on SOBs exist. Furthermore, risk perception mediates the
tion and SOBs. Pilots who had high internal scores were more relationship between LOC and behavior. However, LOC accounted
likely to perceive the risk of aviation accidents as high and operate for 31% of the total variance of risk perception, which suggests
more safely in aviation. In contrast, pilots who had high external that risk perception also independently affected SOB. This result
scores were more likely to perceive the risk of aviation accidents is in accordance with OHare (1990), Hunter (2001), and Ji et al.
as low. A plausible explanation for these results is that internals (2011a,b). In other words, risk perception accounts for additional
are better at detecting errors as well as at perceiving subtle and variance in behavior. Therefore, more research is needed to identify
incidental cues while proofreading text than externals (Wolk and the factors that cause inaccurate perceptions of risk. Such factors
DuCette, 1974). Internals attended to relevant cues and ignored might include how pilots assess their personal levels of compe-
those that were irrelevant (Gregory and Nelson, 1978). This abil- tency, what causes these assessments to be inaccurate, how pilots
ity is likely germane to aviation settings in which malfunctions gather information from the environment, and how they identify
occur under conditions of high workload. Consequently, internals the cues associated with conditions of high risk. Only in this way
X. You et al. / Accident Analysis and Prevention 57 (2013) 131139 137

will the interventions of aviation safety management and airline short, high internals might perceive themselves as being at greater
pilot safety training be truly effective. risk than before when they discover that bad things can happen
The present study also examined the moderating effects of ight even when precautions are taken. Consequently, designing training
time on the relationship between LOC and SOBs. Flight experience programs to minimize the tendency for moderately experienced
plays an important role in the performance of ight-related tasks, pilots to feel in control and the illusion of unique invulnerability
especially because it compensates for age-related declines in cog- is possible.
nitive function (to some degree), and overlearned complex tasks To summarize, the results of the present study showed that LOC
such as piloting are less susceptible to age-related deterioration indirectly affects SOBs by inuencing risk perception, the inter-
than abilities performed during novel situations (Yesavage et al., nal total effect on safety behaviors is larger than the external total
1994; Tsang and Voss, 1996). Our results indicate that total ight effect, and ight time moderates the relationship between internal
time signicant moderated the relationship between internal LOC LOC and SOBs among Chinese airline pilots. Although one can-
and SOBs. Specically, a high total ight time reduced the posi- not expect to change a pilots LOC directly, one can minimize the
tive effects of internal LOC on SOBs among high internal pilots, tendency for moderately experienced pilots to perceive being in
whereas medium and low total ight times did not greatly affect control and change the risk perceptions related to SOBs to promote
this relationship. In other words, when pilot internal LOC reached a aviation safety. Future research should explore the potential effects
certain (high) threshold, the benet that ight experience has with of organizational-level variables such as safety culture (Cooper,
regard to behavior diminishes. Similar results were found in previ- 2000) and its relationships with LOC and SOBs.
ous research, which showed that ight experience has a protective
effect against the risk of crashes (Li et al., 2003) and hazardous Acknowledgments
attitudes in the early (but not the late) experience building stage
(Defense Research and Development Canada; DRDC, 2010). It is rea- The joint fund of Civil Aviation Research of China (61079004)
sonable to expect a curvilinear relationship between LOC and ight and the Fundamental Research Funds for the Central Universities
hours with high internals peaking in their mid-careers. A possible of China (Gk201302051) supported this study. The authors thank
explanation for this nding is that high internal pilots might be the four ight experts from China Southern Airlines, Ltd. and all the
more invulnerable during the early and middle experience build- airline pilots who answered our questionnaire. In addition, many
ing stages but more cautious in the later stages. In other words, thanks are due to the anonymous reviewers for their supportive
they begin to realize that an experienced pilot can still make errors comments.
and that he or she is not invulnerable. Therefore, an accident might
happen despite competency and prociency. Some evidence in the
Appendix A.
literature (Helweg-Larsen and Shepperd, 2001) suggests that prior
experience with negative outcomes moderates optimistic biases. In
See Table A1.

Table A1
Mean and standard deviation of the items measuring locus of control, risk perception, and safety operation behavior (N = 193).

Item Mean S.D.


a
Aviation Safety Locus of Control Scale
(I) If pilots follow all the rules and regulations, they can avoid many aviation accidents. 3.71 0.64
(E) Accidents are usually caused by unsafe equipment and poor safety regulations. 2.71 0.42
(I) Pilots should lose their license if they periodically neglect to use safety devices (for example, seat belts, checklists, etc.) that are 3.30 0.57
required by regulation.
(I) Accidents and injuries occur because pilots do not take enough interest in safety. 3.45 0.43
(E) Avoiding accidents is a matter of luck. 2.51 0.38
(I) Most accidents and incidents can be avoided if pilots use proper procedures. 3.70 0.42
(E) Most accidents and injuries cannot be avoided. 2.67 0.41
(I) Most accidents are due to pilot carelessness. 3.53 0.43
(E) Most pilots will be involved in accidents or incidents which result in aircraft damage or personal injury. 2.42 0.51
(I) Pilots should be ned if they have an accident or incident while horsing around. 3.41 0.42
(I) Most accidents that result in injuries are largely preventable. 3.73 0.81
(E) Pilots can do very little to avoid minor incidents while working. 2.72 0.42
(E) Whether people get injured or not is a matter of fate, chance, or luck. 2.48 0.39
(I) Pilots accidents and injuries result from the mistakes they make. 3.38 0.43
(E) Most accidents can be blamed on poor FAA oversight. 2.57 0.41
(E) Most injuries are caused by accidental happenings outside peoples control. 2.60 0.38
(I) People can avoid getting injured if they are careful and aware of potential dangers. 3.52 0.42
(E) It is more important to complete a ight than to follow a safety precaution that costs more time. 2.42 0.39
(I) There is a direct connection between how careful pilots are and the number of accidents they have. 3.12 0.43
(E) Most accidents are unavoidable. 2.29 0.38
Risk Perception Scaleb
Start a light aircraft with a dead battery by hand-propping it. 73.2 16.9
At night, y from your local airport to another airport about 150 miles away, in a well-maintained aircraft, when the weather is 75.1 21.3
marginal VFR (3 miles visibility and 2000 foot overcast).
At night, take a cross-country ight in which you land with 30 min of fuel remaining. 72.4 15.7
Make a 2-h cross country ight with friends, without checking your weight and balance. 69.7 14.8
Fly in clear air at 6500 feet between two thunderstorms about 25 miles apart. 76.2 18.5
Take a 2-h sightseeing ight over an area of wooded valleys and hills, at 1000 above ground level. 69.7 16.4
During the daytime, y from your local airport to another airport about 150 miles away, in a well-maintained aircraft, when the 71.5 20.9
weather is marginal VFR (3 miles visibility and 2000 foot overcast).
Make a trafc pattern so that you end up turning for nal with about a 45 bank. 68.7 13.4
Drive your car on a freeway near your home, during the day, at 65 MPH in moderate trafc, during heavy rain. 66.9 15.8
Fly across a large lake or inlet at 500 feet above ground level. 72.3 16.5
138 X. You et al. / Accident Analysis and Prevention 57 (2013) 131139

Table A1 (Continued)

Item Mean S.D.

During the daytime, take a cross-country ight in which you land with 30 min of fuel remaining. 64.3 12.1
Jaywalk (cross in the middle of the block) across a busy downtown street. 58.7 21.4
Fly across a large lake or inlet at 1500 feel above ground level. 55.6 17.6
Drive your car on a freeway near your home at night, at 65 MPH in moderate trafc. 54.7 13.0
At night, y from your local airport to another airport about 150 miles away, in clear weather, in a well-maintained aircraft. 52.8 18.3
At night, take a cross-country ight in which you land with over an hour of fuel remaining. 52.4 20.8
Take a 2-h sightseeing ight over an area of wooded valleys and hills, at 3000 above ground level. 53.7
Fly across a large lake or inlet at 3500 feet above ground level. 48.3 16.6
Drive your car on a freeway near your home during the day, at 65 MPH in moderate trafc. 44.8 14.2
Make a trafc pattern so that you end up turning for nal with about a 30 bank. 46.7 17.2
Make a 2-h cross-country ight with friends, after checking your weight and balance. 40.1 18.7
During the daytime, take a cross-country ight in which you land with over an hour of fuel remaining. 38.2 15.7
During the daytime, y from your local airport to another airport about 150 miles away, in clear weather, in a well-maintained 36.7 17.5
aircraft.
Climb up a 10-foot ladder to replace an outside light bulb. 33.4 16.3
Take a 2-h ight in a jet aircraft on a major US air carrier. 30.8 18.7
Ride an elevator from the ground oor to the 25th oor of an ofce building. 21.3 15.1
Safety Operation Behaviors Scalec
Environment for open communications established and/or maintained. Crew members listen with patience, make eye contact as 2.68 0.12
appropriate.
There is appropriate and good group climate. Crew members do not interrupt or talk over, do not rush through the brieng. 3.57 0.13
When new staff, lines, airports, and other conditions, the crew members can take the initiative to share operational knowledge 2.69 0.23
and experience.
Cabin crew is included as part of team in briengs, as appropriate, and guidelines are established for coordination between ight 3.15 1.03
deck and cabin. Passengers are briefed and updated as needed concerning delays, weather, etc.
Crew members speak up, and state their information with appropriate persistence, until there is some clear resolution and 2.57 0.69
decision.
Tasks and workload are clearly distributed and providing enough time, which was accepted by other crews. 3.25 0.36
Captain shows leadership and coordinates ight deck activities. Strikes a balance between authority and crew member 1.98 0.51
participation, yet acts decisively when necessary.
Crew members can identify and report on their own or others work overload situation 3.65 1.23
Crew members can note the appearance of fatigue, and take effective means to demonstrate high levels of vigilance, such as 3.08 1.10
conversation, activities, caffeine, and moving and so on.
Operational tasks are prioritized so as to allow sufcient resources for dealing effectively with primary ight duties, such as 2.68 0.75
dealing with passenger needs, crew meals, company communications.
Operational plans and decisions are clearly stated to other crew members and acknowledged, and include cabin crew and others 2.98 0.38
when appropriate.
Briengs are operationally thorough, interesting, and address crew coordination and planning for potential problems, such as 3.27 0.56
rejected T/O, engine failure after liftoff and go-around at destination.
Caption can make an effective brieng, and can foresee the bias to occur during normal operation. 3.61 0.57
Automated systems are used at appropriate levels. When programming demands could reduce situational awareness and create 2.86 1.01
work overloads, the level of automation is reduced or disengaged, or automation is effectively used to reduce workload.
Guidelines are followed for the operation of automated systems. When systems will be disabled, PF and PM [PNF] duties. 3.35 0.89
Crew members verbalize and acknowledge entries and changes to automated systems parameters. 3.23 1.20
Crew Members exchange information regarding the FMC capabilities, limitations, and operations. 2.77 0.37
Aircraft automation systems are reviewed and conrmed regularly. For example, the best sailing condition, correct runway 3.34 0.87
proles.
When the plane goes into the automation status and system parameters are modied, the crew members notice each other timely. 3.32 0.31
PF and PM [PNF] duties established and implemented. For example, input dates and interaction checked. 3.87 1.34
Crew members demonstrate high levels of vigilance in both high and low workload conditions. 2.54 0.80
Crew prepares for expected or contingency situations including approaches, weather, etc. 2.79 1.01
Crew members stated critical information and/or solutions with appropriate persistence. 3.54 0.56
Objectively, without disguise to accept the work of the feedback. 3.68 1.03
When disputes occur, crew can still focus on current problems or situations and actively listen to the suggestions and comments, 2.97 0.64
and can correct their mistakes, thereby enabling the disputed issues to reach consensus and resolved.
Crew members ask questions regarding crew actions and decisions, e.g. effective enquiry about uncertainty of clearance limits, 2.69 0.23
clarication of confusing. Unclear ATC instructions.
Given to conduct a positive or negative feedback at appropriate times, as a direct learning experience to whole crew, such as the 3.15 1.03
comments on the take-off or landing.

I, internal; E, external.
a
Range 15.
b
Range 1100.
c
Range 14.

References Cooper, M.D., 2000. Towards a model of safety culture. Safety Science 36, 111136.
Defence Research and Development Canada (DRDC), 2010. Flight Experience, Risk
Arthur, W., Barrett, G.V., Alexander, R.A., 1991. Prediction of vehicular accident Taking, and Hazardous Attitudes in Glider Instructors (Tech. Rep. No. 137).
involvement: a meta-analysis. Human Performance 4, 89105. Defence Research and Development Canada, Toronto, VA.
Berg, J.S., Moore, J.L., Retzlaff, P.D., King, R.E., 2002. Assessment of personality and DuCette, J., Wolk, S., 1973. Cognitive and motivational correlates of general-
crew interaction skills in successful naval aviators. Aviation Space and Environ- ized expectancies for control. Journal of Personality and Social Psychology 26,
mental Medicine 73 (6), 575659. 420426.
Burian, B., Orasanu, J., Hitt, J.,2000. Weather-related decision errors: differences Duke, M., Johnson, T.C., Nowicki, S., 1977. Effects of sports tness camp experience.
across ight types. In: Proceedings of the 14th Triennial Congress of the Inter- The Research Quarterly 48, 280283.
national Ergonomics Association/44th Annual Meeting of the Human Factors Edriss, A.K., 2003. A Passport to Research Methods. International Publishers and
and Ergonomics Society. HFES, Santa Monica, CA. Press, Las Vegas.
X. You et al. / Accident Analysis and Prevention 57 (2013) 131139 139

Eysenck, H.J., 1967. The Biological Basis of Personality. Charles C. Thomas Publisher, OConnor, P., Hrmann, H.J., Flin, R., Lodge, M., Goeters, K.M., 2002. Devel-
Springeld, IL. oping a method for evaluating crew resource management skills: a
Gregory, W.L., Nelson, D.V., 1978. Locus of control and cue explication. Personality european perspective. International Journal of Aviation Psychology 12 (3),
and Social Psychology Bulletin 4, 248251. 263285.
Guastello, S.J., Guastello, D., 1986. The relation between the locus of control ODea, A., OConnor, P., Kennedy, Q., Buttrey, S.L., 2010. A Review of the safety climate
construct and involvement in trafc accidents. Journal of Psychology 120, literature as it relates to naval aviation. NPS-OR-10-002, 20301-4000 and funded
293297. by the Defense Safety Oversight Council,Washington, DC.
Haan, C.T., 2002. Statistical Methods in Hydrology, second ed. Iowa State University OHare, D., 1990. Pilots perception of risks and hazards in general aviation. Aviation,
Press, Ames, IA. Space, and Environmental Medicine 61 (7), 599603.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., Tatham, R.L., 2006. Multivariate Data Ozkan, T., Lajunen, T., 2005. Multidimensional trafc locus of control scale (T-LOC):
Analysis, sixth ed. Pearson Education Inc., Upper Saddle River, NJ. factor structure and relationship to risky driving. Personality and Individual
Harvey, J.M., 1971. Locus of control shift in administrators. Perceptual and Motor S Differences 38, 533545.
33, 980982. Pauley, K., OHare, D., Wiggins, M., 2008. Risk tolerance and pilot involvement in
Helmreich, R.L., Butler, R., Taggart, W., Wilhelm, J., 1997. The NASA/University hazardous events and ight into adverse weather. Journal of Safety Research 39
of Texas/Federal Aviation Administration Line/LOS Checklist: A behavioral- (4), 403411.
based checklist for CRM skills assessment (Version 4.4) [Computer software]. Penk, W., 1969. Age changes and correlates of internalexternal locus of control
NASA/University of Texas/Federal Aviation, Administration Aerospace Group, scales. Psychological Reports 25, 856.
Austin, TX. Poropat, A.E., 2009. A meta-analysis of the ve-factor model of personality and
Helweg-Larsen, M., Shepperd, J.A., 2001. Do moderators of the optimistic bias affect academic performance. Psychological Bulletin 135 (2), 322338.
personal or target risk estimates? A review of the literature. Personality and Rotter, J.B., 1954. Social Learning and Clinical Psychology. Prentice-Hall, Englewood
Social Psychology Review 5, 7495. Cliffs, NJ.
Heider, F., 1958. The Psychology of Interpersonal Relationships. Wiley, New York. Salminen, S., Klen, T., 1994. Accident locus of control and risk taking among forestry
Hunter, D.R., 2001. Retrospective and prospective validity of aircraft accident risk and construction workers. Perceptual and Motor Skills 78, 852854.
indicators. Human Factors 43 (4), 509518. Sanders, M.G., Halcomb, C.G., Fray, J.M., Owens, J.M., 1976. Internal-external locus
Hunter, D.R., 2002. Development of an aviation safety locus of control scale. Aviation, of control and performance on a vigilance task. Perceptual and Motor Skills 42,
Space, and Environmental Medicine 73 (12), 11841188. 939943.
Hunter, D.R., 2005. Measurement of hazardous attitudes among pilots. International Simon, M., Houghton, S.M., Aquino, K., 2000. Cognitive biases, risk perception, and
Journal of Aviation Psychology 15 (1), 2343. venture formation: how individuals decide to start companies. Journal of Busi-
Hunter, D.R., 2006. Risk perception among general aviation pilots. International ness Venturing 15 (2), 113134.
Journal of Aviation Psychology 16 (2), 135144. Stanton, N.A., Young, M.S., 2005. Driver behaviour with adaptive cruise control.
Hunter, D.R., Stewart, J.E., 2009. Locus of control, risk orientation, and decision mak- Ergonomics 48, 12941313.
ing among U.S. army aviators (Tec. Rep. No. 1260). U.S. Army Research Institute Stewart, J.E., 2006. Locus of control, attribution theory, and the ve deadly sins of
for the Behavioral and Social Sciences (DTIC No. ADA452056), Arlington, VA. aviation (Tech. Rep. No. 1182). U.S. Army Research Institute for the Behavioral
Jensen, R.S., Benel, R.A., 1977. Judgment Evaluation and instruction in Civil Pilot and Social Sciences, Arlington, VA.
Training, FAA-RD-78-24. Federal Aviation Administration, Washington, DC. Stewart, J.E., 2008. Locus of control and self-attribution as mediators of hazardous
Ji, M., Yang, S.Y., Zhao, X.J., Bao, X.H., You, X.Q., 2011a. The role of risk perception and attitudes among aviators: a review and suggested applications. International
hazardous attitudes in the effects of risk tolerance on safety operation behaviors Journal of Applied Aviation Studies 8, 263279.
among airline pilots. Acta Psychologica Sinica 43 (11), 13081319. Stewart, J.E., Moore, K.P., 1978. Time perception as a function of locus of control.
Ji, M., You, X.Q., Lan, J.J., Yang, S.Y., 2011b. The impact of risk tolerance, risk percep- Personality and Social Psychology Bulletin 4, 5658.
tion and hazardous attitude on safety operation among airline pilots in China. Trankle, U., Gelau, C., Metker, T., 1990. Risk perception and age-specic acci-
Safety Science 49 (10), 14121420. dents of young drivers. Journal of Accident, Analysis and Prevention 22 (2),

Joreskog, K.G., 2004. On chi-squares for the independence model and t measures 119125.
in Lisrel. http://www.ssicentral.com/lisrel/techdocs/ftb.pdf (13.04.08). Tsang, P.S., Voss, D.T., 1996. Boundaries of cognitive performance as a function of age

Joreskog, K., Sorbom, D., 2001. LISREL 8: Users Reference Guide. Scientic Software and ight experience. International Journal of Aviation Psychology 6, 359377.
International Inc, Chicago. Vallee, G., 2006. Perception of risk and hazards among general aviation pilots. M.Sc.
Joseph, C., Ganesh, A., 2006. Aviation safety locus of control in Indian aviators. Indian tHesis. Craneld University, School of Engineering.
Journal of Aerospace Medicine 50, 1421. Wichman, H., Ball, J., 1983. Locus of control, self-serving biases, and attitudes
Jung, C.G., 1928. Two Essays on Analytical Psychology. Balliere Tindall, London. towards safety in general aviation pilots. Aviation, Space, and Environmental
Lefcourt, H.M., 1982. Locus of Control. Current Trends in Theory and Research. Medicine 54, 507510.
Erlbaum, Hillsdale, NJ. Wiegmann, D.A., Shappell, S.A., 1997. Human factors analysis of postaccident data:
Li, G.H., Baker, S.P., 1994. Prior crash and violation records of pilots in commuter and applying theoretical taxonomies of human error. International Journal of Avia-
air taxi crashes: a casecontrol study. Aviation Space and Environment Medicine tion Psychology 7 (1), 6781.
65, 979985. Wolk, S., DuCette, J., 1974. Intentional performance and incidental learning as a
Li, G.H., Baker, S.P., Grabowski, J.G., Qiang, Y.D., Melissa, L., McCarthy, M.L., function of personality and task dimensions. Journal of Personality and Social
Rebok, G.W., 2003. Age, ight experience, and risk of crash involvement in Psychology 29, 90101, in Lunenburg, Fred C. Locus of control, pupil control
a cohort of professional pilots. American Journal of Epidemiology 157 (10), ideology, and dimensions of teacher burnout. Journal of Instructional Psychology
874880. 19(1), 13. http://search.epnet.com/login.aspx?d
McCrae, R.R., Costa, P.T., 1995. Trait explanations in personality psychology. Euro- Yesavage, J.A., Dolhert, N., Taylor, J.L., 1994. Flight simulator performance of younger
pean Journal of Personality 9, 231252. and older aircraft pilots: effects of age and alcohol. Journal of the American
Mischel, W., 1968. Personality and Assessment. Wiley, New York. Geriatrics Society 42, 577582.
Musson, D.M., Sandal, G.M., Helmreich, R.L., 2004. Personality characteristics and You, X.Q., Ji, M., Dai, K., Yang, S.Y., Chang, M., 2009. Developing a multidimensional
trait clusters in nal stage astronaut selection. Aviation Space and Environmen- scale to assess safety behaviors in airline ight. Acta Psychologica Sinica 41 (12),
tal Medicine 75 (4), 342349. 12371251.