Business Ethics Journal Article

Aviation Psychology and Applied Human Factors Testing the Compliance Behavior Model in General Aviation: A Pilot Study Anthony A. Stanton, Sidney W. Dekker, Patrick S. Murray, and Gui Lohmann Online First Publication, December 2, 2020. http://dx.doi.org/10.1027/2192-0923/a000200

CITATION Stanton, A. A., Dekker, S. W., Murray, P. S., & Lohmann, G. (2020, December 2). Testing the Compliance Behavior Model in General Aviation: A Pilot Study. Aviation Psychology and Applied Human Factors. Advance online publication. http://dx.doi.org/10.1027/2192-0923/a000200

Original Article

Testing the Compliance Behavior Model in General Aviation A Pilot Study

Anthony A. Stanton1 , Sidney W. Dekker1, Patrick S. Murray2, and Gui Lohmann3

1Safety Science Innovation Laboratory, Griffith University, Brisbane, QLD Australia 2Aviation and Logistics, University of Southern Queensland, Australia 3School of Engineering and Built Environment, Griffith University, Brisbane, QLD, Australia

Abstract: Australian general aviation accident data show pilots who conduct operations into adverse weather, when against the rules, remain as a significant cause of fatal accidents. This paper presents the background, methodology, and results of a theory of planned behavior (TPB) elicitation study, which extracted key psychological beliefs of aircraft pilots in such circumstances. The present study established a psychometric survey instrument with items that are valid and reliable, to then further explore the TPB psychological constructs concerning the intentions of pilots when presented with adverse weather. Given the principled deliberations associated with rule-related behavior, the project explores an extension of the TPB by investigating the addition of two psychological constructs – personal norms and anticipated affect and their power to provide a discrete contribution and improved explanation of variance.

Keywords: pilot violations, theory of planned behavior, rule-related behavior, general aviation compliance, personal normative influences

Australia aviation accident data between 2008 and 2017 show that pilots conducting operations into adverse weather, when against the rules, resulted in 109 reported safety occurrences (Australian Transport Safety Bureau, 2018). Safety data show the lethality of venturing into instrument meteorological conditions (IMC) for those pilots who are limited to operations under the visual flight rules (VFR). Reviewing US and Canadian data, Batt and O’Hare (2005) have shown that pilots operating VFR into IMC, when against the rules, is around four times more likely to prove fatal than any other general aviation safety occur- rence. The causation is often resolved to general statements around a pilot’s desire to just press on, or what is often called “get-home-itis.” Is this behavior really that simple? Or are there much more complex influences affecting the pilot’s behavior? While some researchers have attempted to understand the problem, sometimes examined in the context of plan continuation error, the research particularly within general aviation is limited. Much of the research has focused on occurrence statistics and related contexts, rather than an attempt to uncover an understanding of people and the latent psychological factors that influence their safety-related behaviors (decision-making). The present study explores the latent psychological beliefs of general aviation pilots when faced with adverse weather, by adapt- ing an expectancy–value psychosocial behavior theory – the

theory of planned behavior (TPB) – as a theoretical framework.

The objective of this article is to outline the theoretical framework of the research project, including a conceptual extension of the TPB and to present the results of an elicitation study. The paper begins with a description of the TPB, its sufficiency assumptions, and evidence of its appropriate application here. A discussion of a conceptual compliance behavior model then follows, which is put forward as an extension of the TPB for empirical testing in an attempt to improve the explained variance in the present rule-related behavior context. The paper then articulates the methods and results of the elicitation study, which has identified modal salient beliefs (i.e., those commonly held among the target population) and provided formative research for a following study.

The TPB is an expectancy–value behavior model that theorizes an explanation for the formation of behavioral intentions (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). As shown in Figure 1, the TPB posits that a particular behavior follows reasonably and consistently from an indi- vidual’s salient psychological beliefs associated with per- forming that behavior. Further, it is suggested that only a limited number of these beliefs dominate and hence predict an individual’s intention to perform the behavior. More specifically, the theory posits that a person’s behavioral

� 2020 Hogrefe Publishing Aviation Psychology and Applied Human Factors (2020) https://doi.org/10.1027/2192-0923/a000200

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

intention (BI) is the most immediate antecedent of performing the behavior (B) and, further, these intentions are formulated as a result of three key psychological deter- minants:

(1) An individual’s attitude toward the specified behavior (Ab);

(2) An individual’s perceived social pressures in relation to performing or not performing the behavior (SN); and

(3) Perceptions of behavioral control (PBC) or self- efficacy.

The relationship between each of the model’s direct psy- chological constructs (B, BI, Ab, SN, and PBC) is illustrated by the equation

B � BI ¼ Ab β1ð Þ þ SN β2ð Þ þ PBC β3ð Þ: ð1Þ

The equation suggests that each of the direct constructs is separated by a beta weight (β) that reflects the influence of that particular construct (e.g., Ab) toward the formation of behavioral intentions. Different behaviors have been shown empirically to be influenced to varying extents by each of these constructs. The TPB allows this to be explored through a well-established methodology. As a result of identifying these influences and importantly their respective dominance on the specified behavior, interven- tion strategies can be developed that target the latent influ- ences that have the greatest leverage on the behavior. For example, if a particular behavior, in a given context, was identified to be most influenced by subjective norms (social influence), a behavior change program may focus its efforts on injunctive (what is perceived others might expect) and descriptive (what others are seen to be doing) normative beliefs, rather than distributing information about the dis- advantages or disadvantages of performing the particular behavior (attitude toward the behavior).

By further review of Figure 1, we can see that each of these direct constructs are said to be formulated from underlying latent beliefs, that is, behavioral beliefs, norma- tive beliefs, and control beliefs. As an expectancy–value

model, each of these beliefs is characterized by a two-factor composite form, which is shown empirically to correlate with the corresponding direct construct. These composite forms are often referred to as “indirect measures.” The behavioral beliefs composite is expressed by

Pn i¼1 biei

where b represents the subjective probability that outcome i exists when performing the behavior and e represents an evaluation of that outcome i. Concerning outcome i, these two composites are then multiplied (i.e., expectancy–value). The total set of the salient belief composites is then summed. Likewise, the normative beliefs are expressed by

Pn i¼1 nimi where n represents the normative belief about

referent i and m represents the motivation to comply with referent i. Finally, control beliefs are expressed by Pn

i¼1 cipI where c represents the subjective probability that the control factor i will be present when performing the behavior and p represents the perceived power of that factor to make performing the behavior easier or more difficult. Empirically, each aspect of these equations is explored by asking respondents a series of questions that explore respective subjective probabilities and evaluative aspects using 5- or 7-point scales.

Reason and coworkers (1990) have shown the likely heritage of violation or rule-related behavior to be within social (social norms) and motivational (intention) founda- tions, as opposed to errors, which have origins within human information-processing limitations. Because of this, and as argued by Fogarty and Shaw (2010), the TPB is therefore an ideal theoretical framework from which to examine rule-related behavior, since the TPB encompasses these influences and others. In their research, Fogarty and Shaw (2010) specifically explored the usefulness of the TPB to understand violation behavior within an aircraft mainte- nance setting. Their study identified the TPB was highly successful in explaining the variance in behavior and resulted in the development of a behavior model appropri- ate for that setting. The present study applies a similar approach, although within a very different context with quite different influences and it also explores the suffi- ciency of the model.

While the TPB has been used extensively as a concep- tual framework for behavioral science investigations

Figure 1. The theory of planned behavior (adapted from Ajzen, 2006).

Aviation Psychology and Applied Human Factors (2020) � 2020 Hogrefe Publishing

2 A. A. Stanton et al., Compliance Behavior Model: A Pilot Study

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

(Barber, 2011; Conner & Armitage, 1998; Rivis et al., 2009), some researchers have explored the sufficiency of the model (Conner & Armitage, 1998; Parker et al., 1995) to find additional gains in variance explanation, through inclu- sion and adaption of psychological constructs. The TPB suf- ficiency assumption states that “additional variables should not improve prediction of either intention or behavior” (Fishbein & Ajzen, 2010, p. 281), although Ajzen (1991, p. 199) has stated that the TPB is “in principle, open to the inclusion of additional predictors if it can be shown that they capture a significant proportion of the variance in intention or behavior after the theory’s current variables have been taken into account.” The present research explores this sufficiency assumption. That is, the addition of other variables to improve the explained variance in rule-related behavior intention.

The subjective norm construct has been cited as the TPB weakest predictor (Armitage & Conner, 2001; Sheppard et al., 1988). As discussed earlier, this construct considers the influence of significant others. Specifically, the subjec- tive norm construct evaluates perceived expectations of what ought to be done (injunctive social norms) and whether significant others are themselves performing the behavior (descriptive social norms). Injunctive social norms influence behavior by reflecting the patterns of the collec- tive group, enticing reward or threatening sanction by the group for acquiescence (Cialdini et al., 1991). Descriptive norms influence differently, where observations of effective and adaptive action provide information-processing advan- tages (Cialdini et al., 1991). Neither of these two normative components cogitates self-expectations, being internalized personal values, which it is argued may independently influence rule-related decision-making.

The TPB does not explicitly include consideration of personal normative influence; instead, such influences are considered to be embedded within behavioral beliefs. A review of the theory of propositional control (Dulany, 1961, 1968), from which the TPB evolved, and the early adaptions by Fishbein (1967), show a personal normative component in those models. For example, the personal normative component can be identified clearly in the theory of propositional control equation [(NBp)(MCp)]w1 + [(NBs) (MCs)]w2 where subscript p represents beliefs of a personal nature, and subscript s represents the beliefs of a social nat- ure (Fishbein, 1967).

Personal norms have been investigated extensively by Schwartz (1973, 1977) in relation to altruism, resulting in the norm activation model. In this model, pro-social envi- ronmental behavior is hypothesized to result from three determinants: awareness of consequences, the ascription of responsibility, and personal norms. Schwartz (1977, p. 227) defines personal norms as self-expectations or an internalized sense of duty constructed from general norms

and personal values. Schwartz (1977) hypothesizes that con- formity, or otherwise, with personal norms are experienced in a state of subjective self-awareness (Duval & Wicklund, 1972), as opposed to more conscious intellectual judgments of right and wrong. Schwartz (1977) suggests individuals experience personal norms as feelings of moral obligation, such as pride and guilt.

In the domain of illegal, antisocial, and dishonest behav- iors, Manstead (2000) provides an extensive review of researchers who have sought to improve the TPB through the inclusion of a personal norm or moral norm construct. Gorsuch and Ortberg (1983) included a single item, a direct measure of perceived moral obligation, and demonstrated within moral settings that the construct provided an additional 20% explanation of the variation in behavioral intentions. The moral obligation construct provided an independent contribution and was more highly correlated with behavioral intentions than either the attitude or subjec- tive norm constructs.

In a TPB investigation of driving violations, Parker et al. (1995) included items representing a personal normative construct: a single moral norm item and two anticipated regret items. The additional measures resulted in an improved variance explanation of behavioral intentions of up to 15%. In the Parker et al. (1995) study, the personal norm constructs contributed more to the variance than did the extant constructs. This suggests, in relevant con- texts, that the personal norm construct has a considerable influence on behavior. Akin to the investigation of driving violations by Parker et al. (1995), the situational violation behavior of general aviation pilots is one likely to be influ- enced by internalized self-expectations, and, therefore, the personal norm construct is expected to improve the expla- nation of behavioral intentions. As identified in the Huntzinger (1997) study, it is plausible that when forming a behavioral intention to commit a situational violation, a general aviation pilot may hold beliefs of anticipated affect, activated as a result of their own internalized values or per- sonal norms and the resulting moral dilemma (Seligman et al., 2002). This anticipated negative affect could include regret, apprehension, anxiety, shame, guilt, anger, or fear and influence behavioral intention (Moan et al., 2005).

While researchers have, to some extent, tested the inclu- sion of anticipated affect and personal norm constructs to the TPB, this has never been in the context of pilot rule- related behavior. Each domain and each behavior have unique influences, hence a need for empirical validation in each. The compliance behavior model (CBM) shown in Figure 2 is a conceptual model that is unique since it con- cedes anticipated affect to be a behavioral belief construct and postulates, based on the work of Schwartz (1977), that anticipated affect is an experiential (affective) attitude asso- ciated with a personal normative construct. The model,

� 2020 Hogrefe Publishing Aviation Psychology and Applied Human Factors (2020)

A. A. Stanton et al., Compliance Behavior Model: A Pilot Study 3

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

if proven, would enable a methodology for uncovering and evaluating personal normative influences on rule-related behavior in this domain. This separation of the constructs facilitates practical distinctions and interventional interests between the influences. Noting an observation by Ajzen and Sheikh (2013), unlike other researchers, the compliance behavior model would measure anticipated affect to per- forming the behavior itself, consistent with other measures (i.e., not the alternate course of action). In Figure 2 the com- pliance behavior model shows the addition of the personal norm and anticipated affect, constructs including the expected covariance. The compliance behavior model re- labels extant TPB constructs more relevantly to the present application. It is hypothesized that: The compliance behav- ior model will predict the intentions of general aviation pilots to conduct a situational violation associated with adverse weather, and that the addition of the personal norms and anticipated affect (from the TPB) will improve the explained variance of violation intentions after existing TPB measures have been considered.

Method

TPB methodology requires that researchers conduct two separate studies. Initially, researchers conduct an elicitation or pilot study, where the primary objective is to elicit from the target population’s latent beliefs in relation to the behavior. In the second (main) study, researchers leverage the elicitation study results in order to explore potential associations between the TPB constructs and the behavior. The following sections articulate the method and results of the elicitation study.

The elicitation study also has two secondary objectives: to formulate and then test the internal reliability of items for the measurement of each of the TPB direct constructs, and to evaluate the suitability of a set of TPB background measures. The methodological and analytical frameworks for TPB studies are well documented elsewhere (Fishbein & Ajzen, 2010; Francis et al., 2004). Approval to com- mence the elicitation and main studies was obtained from

a university Human Research Ethics Committee after demonstrating compliance to specified conditions and guidelines.

The elicitation study survey was published by using the Survey Monkey web tool with invitations to participate primarily generated by social media posts from several gen- eral aviation flying organizations, such as flying clubs and flying schools. The survey was open for 2 days with the average respondent taking 6 min to complete all items. Of the 42 respondents, 30 answered all questions providing a completed response rate of 71%. A total of 47 items were employed as part of the survey. Prior to publishing the Survey Monkey questionnaire, five respondents were used to construct content for the questions that elicited beliefs in the Survey Monkey web tool. This smaller group was asked open-ended questions exploring belief themes. Each belief theme was then included in the survey.

Participants

The target population was defined as any licensed aircraft pilot, or student pilot, who is currently operating, or has ever previously operated, as a pilot within the general avia- tion sector. The target population was not limited to respon- dents within Australia, nor pilots who are operating within the general aviation sector, since any licensed pilot or trai- nee is reasonably able to contemplate their influences and reactions to the behavior under examination and then pro- vide a considered response. Responses would be influenced by their particular past and present background factors.

A total of 42 participants provided responses, which is a representative sample of the target population as recom- mended by Fishbein and Ajzen (2010) for this formative research stage. To validate respondents were within the tar- get population, the elicitation study asked two validation questions (1) regarding the level of pilot license held and (2) the country in which they conducted the majority of their general aviation flying. Each question had a response option allowing them to identify themselves as being out- side the defined target population. Nil respondent exclu- sions were required.

Figure 2. The compliance behavior model is a conceptual model, adding two additional psychological constructs to the TPB and labeling the behavior.

Aviation Psychology and Applied Human Factors (2020) � 2020 Hogrefe Publishing

4 A. A. Stanton et al., Compliance Behavior Model: A Pilot Study

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

Background Factors

The elicitation study included seven items as trial back- ground measures. The items related to age, gender, flying hours (experience), country, past or present employment as a pilot, self-rated skill compared with others with the same experience, and the level of pilot license held. Table 1 summarizes the key characteristics of the respondents’ background information. Interestingly, the majority of respondents (92.86%) indicated that they rated their flying skill as either the same or better than other pilots of the same experience level. These seven background measures were shown to provide meaningful information. The two questions relating to license level and country in which gen- eral aviation flying took place provided a simple test that the participant was within the target population. It is plausi- ble the background measures may have correlations with other measures. For example, the self-rated skill response is likely to be shown as correlated with PBC in a study with higher statistical power.

The Behavioral Criterion

The TPB requires the behavior that is under research exam- ination (the behavioral criterion) to be clearly defined by four specific elements. These elements are (1) the action, (2) the target, (3) the context, and (4) time. For the elicita- tion study, the behavioral criterion is broadly considered – a situational violation. Such a violation is where a person operates at a rule-based level of cognition (Rasmussen, 1983), applying predefined action to preconceived situa- tions. The behavioral criterion was described to respon- dents by way of a detailed scenario that incorporated each of these four elements. The scenario included a

photograph taken from the perspective of the pilot’s seat, illustrating the imagined weather conditions to support the text description and to ensure all respondents had the same perspective of the environmental context. The scenario depicted a hypothetical private, recreational flight in which the respondent operated an aircraft with five passengers in deteriorating weather conditions and, in doing so, committed a situational violation. The scenario depicts the respondent encountering adverse weather for which they are unable to lawfully operate within, 10 min from the destination having flown the aircraft for 50 min at that point toward the destination. This situational violation is known by pilots as operating visual flight rules (VFR) into instrument meteorological conditions (IMC).

Results

The IBM SPSS software package was used for statistical analysis of respondent data. Primarily, reliability analysis and descriptive statistical reporting were used. The results and statistical analysis are discussed in this section by way of indirect measures of background factors as well as direct and indirect measures of the model.

Elicited Behavioral, Normative and Control Beliefs (Indirect Measures)

As stated earlier, the typical TPB methodology requires the pilot of beliefs from the sample population through free text responses and subsequent content analysis. Modal salient beliefs are then identified and used in the principal study for the calculation of indirect constructs according to the belief equations mentioned earlier. As an alternative, before the elicitation study, a small group of participants were asked a series of open-ended questions to obtain lists of potential salient behavioral, normative, and control beliefs associated with performing the behavior. All of these responses were then included in the elicitation study for respondents to select those that readily and spontaneously came to mind (readily accessible beliefs). To supplement this, elicitation study respondents were also provided with a free text box to include additional salient beliefs if they were not identified in the available list. For example, to obtain behavioral beliefs, respondents were asked to list the advantages and disadvantages of performing the behav- ior. To obtain normative referents, respondents were asked who might approve or disapprove of you performing the behavior. To obtain control factors, respondents were asked what factors make performing the behavior easier or more difficult.

Table 1. Notable respondent characteristics

Characteristic Value

Respondents (n) 42

Male 90.48%

Female 9.52%

Median age 31–40

Min. age range Under 20

Max. age range 71–80

Country of flying (Australia) Australia

Flying hours (experience) < 500 hr 50%

Flying hours (experience) > 500 hr 50%

Employed as pilot (current or past) 50%

Own skill rating – same as average 47.62%

Own skill rating – slightly better than average 45.24%

License level – < Student, recreational, private 45.24%

License level – > Commercial, airline 44.24%

� 2020 Hogrefe Publishing Aviation Psychology and Applied Human Factors (2020)

A. A. Stanton et al., Compliance Behavior Model: A Pilot Study 5

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

To construct a set of modal salient beliefs for the princi- pal study, TPB methodology applies a 75% rule (Francis et al., 2004). That is, the salient beliefs that reflect at least 75% of the elicitation study respondents are considered modal and are adopted in the principal study. In other words, these are likely to be the majority of the readily accessible beliefs for the sample population. Table 2 shows the modal salient beliefs and their cumulative account.

Elicited Anticipated Affect (Conceptually an Indirect Measure of Personal Norms)

Typically, in TPB methodology, instrumental attitude (affect) is obtained during the elicitation study by asking respondents about the advantages and disadvantages of per- forming the particular behavior as a behavioral outcome (behavioral belief). The conceptual compliance behavior model advocates that such a methodology is unlikely to yield affective responses associated with performing the behavior. Rather, such pilot questioning usually directs a respondent to consider behavioral outcomes that are experiential or cognitive and hence affective responses are not exposed as outcomes associated with the behavior. As such, in this elicitation study, respondents were directed to an extensive list of affects that was potentially related to performing the behavior. This list was constructed by ask- ing a small focus group earlier to select affects that could be associated with performing the behavior. Nine items were tested as measures of the indirect construct – antici- pated affect.

Table 3 presents the statistical analysis that was con- ducted. The analysis identified two items that scored highly on the frequency of the mid-point score (i.e., the percentage of respondents who selected a mid-point neither score), indicating that a high proportion (40.5% and 81.3%) of the respondents did not associate these two particular types of affect (dull vs. exciting and fun vs. boring) with the behavioral criterion. As a result, these two items were dropped. Additionally, two other items were dropped to reduce the overall number of items; unpleasant–pleasant, and worried–unconcerned. Internal reliability is not a requirement of the indirect measures since different acces- sible beliefs may be inconsistent with each other (Fishbein & Ajzen, 2010).

A Test of Direct Measures

As shown in Table 4, a total of 21 items were formulated for testing as measures of the direct construct scales; attitude towards behavior (Ab), personal norms (PN), social norms

(SN), perceived behavioral control (PBC), and violation intention (I). Each item consisted of a question or statement stem (e.g., continuing would be against my principles) and a corresponding 7-point bipolar adjective scale, with positive and negative endpoints (e.g., agree vs. disagree, bad vs. good). Positive and negative endpoints were mixed from left to right to reduce “response set,” as recommended by Francis et al. (2004). Respondents were asked to select the score that best represented their opinion concerning the question stem. Each scale consisted of multiple items. Items were re-coded in SPSS to reflect a high rating as being a positive attitude toward performing the behavior (i.e., that they would perform the situational violation).

Reliability analysis was conducted to ascertain the level of internal consistency between items of the same scale (i.e., for each direct construct). The first of a series of reli- ability tests for the direct measures is shown in Table 4. Values for Cronbach’s α and the corrected item total corre- lation are reported. For each scale, items were removed in successive reliability tests until a Cronbach α of 0.70 was exceeded for the scale and the correlated item total corre- lations were above 0.50. Table 5 shows the final reliability results for the direct constructs with six items having been removed from the original.

Table 2. Indirect measures – other than anticipated affect

% of elicited responses

Cumulative % of responses

Elicited behavioral outcomes (modal)

Loss of control 19.7 –

Flight into terrain 19.7 39.4

Valued time and money invested 18.7 58.1

Keeps on schedule as committed 17.2 75.3

Spatial disorientation 1.5 76.8

Penalty from regulator 1.3 78.1

Avoids inconvenience to others 1.0 79.1

Elicited normative referents (modal)

Other pilots like me 22.0 –

Flight instructors 13.6 35.6

Passengers on board 13.3 48.9

Pilots much more senior than me 12.9 61.8

Regulator 8.7 70.5

My employer 5.9 76.4

Elicited control factors (modal)

Local area knowledge 20.2 –

Much longer distance flown so far 18.5 38.7

Safe terrain 17.1 55.8

More flying hours 14.5 70.3

Shorter distance remaining 5.7 76.0

Note. Elicited modal salient beliefs with cumulative responses greater than 75% for each construct.

Aviation Psychology and Applied Human Factors (2020) � 2020 Hogrefe Publishing

6 A. A. Stanton et al., Compliance Behavior Model: A Pilot Study

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

Discussion

The primary objective for this study was to uncover the sali- ent beliefs associated with each of the model’s constructs and to then construct a set of modal salient beliefs for the target population. Fishbein and Ajzen (2010) have con- tended that salient beliefs, which are those that are readily accessible in memory and activated spontaneously with limited cognitive effort, are the primary determinants of a person’s attitude toward performing the behavior. Further,

Fishbein and Ajzen (2010) argue that there are only five to nine of these beliefs that are the dominant influences of a given psychological construct. To identify these limited determinants of attitude toward the behavior, Fishbein and Ajzen (2010) advocate eliciting from respondents the advantages and disadvantages of performing the behavior (i.e., specify the behavioral outcomes of performing the behavior). The behavioral outcomes that are most commonly elicited from a representative sample are then considered as a modal set of salient beliefs for the target

Table 4. Direct measures – first reliability test

Scale Abbreviated stem description N

Cronbach α if item deleted

Corrected item total correlation SD

Ab Bad/good 36 0.423 0.535 0.64488

Ab Wise/foolish 36 0.387 0.533 0.72320

Ab Harmful/beneficial 36 0.800 0.426 1.83852

PN Against my principles 36 0.513 0.456 0.80277

PN Would be morally wrong 36 0.841 0.410 2.02122

PN Would be irresponsible 36 0.314 0.661 0.84092

SN Valued others would do 33 0.645 0.410 1.81586

SN People important think safe 33 0.625 0.605 0.86930

SN People valued would approve 33 0.618 0.564 1.05886

SN I would feel under pressure 33 0.701 0.181 2.32004

SN People important think safe 33 0.608 0.702 1.00284

SN Would be expected of me 33 0.671 0.172 1.95305

PBC Safe for me 31 0.733 0.441 1.07663

PBC Easy for me 31 0.655 0.676 1.85959

PBC Up to me 31 0.774 0.205 1.23393

PBC I am confident I could 31 0.676 0.617 1.77194

PBC Difficult for me 31 0.754 0.356 1.79904

PBC I have the ability 31 0.655 0.673 1.99731

I I would 32 0.688 0.733 1.13192

I I would not 32 0.820 0.620 1.54502

I Similar circumstances I intend 32 0.731 0.675 1.20775

Note. Shaded items were dropped to improve the scale reliability.

Table 3. Indirect measures (anticipated affect constructs)

Scale Abbreviated stem

description N Cronbach α if item deleted

Corrected item total correlation

Freq. mid-point score

a Dull/exciting 32 0.760 0.257 40.5%

a Stressful/relaxing 32 0.699 0.397 0.0%

a Restfulness/tension 32 0.704 0.378 0.0%

a Unpleasant/pleasant 32 0.687 0.569 0.0%

a Anxious/calm 32 0.696 0.431 5.6%

a Self-respect/guilt 32 0.700 0.398 16.7%

a Worried/unconcerned 32 0.676 0.602 3.0%

a Fun/boring 32 0.703 0.372 81.3%

a Regret/satisfied 32 0.653 0.596 18.2%

Note. Shaded items were dropped to improve the suitability of the anticipated affect scale.

� 2020 Hogrefe Publishing Aviation Psychology and Applied Human Factors (2020)

A. A. Stanton et al., Compliance Behavior Model: A Pilot Study 7

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

population. A similar concept applies to the beliefs associ- ated with the other direct constructs of the model. The elicitation study identified salient beliefs for each of the TPB direct constructs.

The modal salient behavioral beliefs were shown to reside within seven beliefs. These limited beliefs accounted for 79.1% of all behavioral outcomes obtained from respon- dents. The modal behavioral beliefs were: loss of control, spatial disorientation, regulatory penalty, flight into terrain, influences of time and money pressure, keeping the flight on schedule as committed to others, and avoiding an incon- venience to others. The last three beliefs provide for an enhanced description of what has been referred to as “get-home-itis.” The last two beliefs are of particular inter- est, as they may be more latent influences and hence less expected, although they are consistent with the explanation provided by Reason (2008) for situational violation motiva- tions. Modal anticipated affect was shown to reside within five sensations: stressfulness, tension, anxiousness, regret, and guilt. The association of these sensations with the behavior is consistent with the findings of Causse and coworkers (2013). Using a neuroergonomics approach, these researchers demonstrated a temporary impairment of decision-making when some pilots were faced with an adverse weather-related decision. Likewise, modal social influencers were identified as: other pilots like “me,” flight instructors, the passengers onboard the aircraft, pilots who are considered more senior, the regulator, and the pilot’s employer. Again, the influence of a person’s employer pro- vides additional context to get-home-itis and the perceived pressure to meet work commitments. Modal control beliefs were identified as: local area knowledge, the distance flown so far, whether the terrain was considered safe, experience in the form of flying hours, and the distance remaining to the destination. The most interesting in this set of beliefs

are references to the distance remaining and the distance from the departure. These two themes were evident in research by Batt and O’Hare (2005), where their analysis of 491 adverse weather-related events identified the major- ity of weather-related occurrences took place in the second half of the flight. Such decision-making may be associated with themes of sunk cost (Arkes & Blumer, 1985) and self-justification and escalating commitment (Staw, 1976).

The secondary objectives of the study were to: (1) formu- late and test scales for the measurement of the direct constructs of the model and (2) to evaluate potential back- ground measures. The study appraised 20 scale items, of which six were dropped to achieve an acceptable level of internal consistency for the respective constructs. The sample size within the present study is insufficient to have adequate statistical power to evaluate correlations among the model constructs. Such investigations of correlations and β are intended for a subsequent study, which leverages the findings here. The present study tested seven items as background measures and identified each of these were suitable for use in a principal study. The most interesting results here are related to the item that asked respondents to rate their own skill in comparison with other pilots of the same experience. Interestingly, 45.2% of respondents sug- gested their own self-rated skill was slightly better than a pilot with the very same level of experience. Comparatively, only 7.1% of respondents suggested their self-rated skill was slightly less than a pilot with the same level of experience. Such a statistic alludes to the target population having high perceptions of perceived behavioral control.

In summary, having identified the latent modal beliefs of general aviation pilots in relation to conducting VFR flight into IMC, it is these themes in particular that should be central in any intervention program that attempts to change pilots’ attitudes toward the behavior, their perceptions of

Table 5. Direct measures (Ab, PN, SN, PBC, and I) – final reliability test

Scale Stem abbreviated description Scale Cronbach α Mean Corrected item total correlation

Ab Bad/good 0.802 1.3784 0.673

Wise/foolish 0.673

PN Against my principles 0.841 0.727

Would be irresponsible 0.727

SN Valued others would do 0.790 0.546

People important think safe 0.675

People valued would approve 0.662

People important think safe 0.748

PBC Safe for me 0.799 0.537

Easy for me 0.652

I am confident I could 0.689

I have the ability 0.645

I Similar circumstances I would 0.822 0.700

Similar circumstances I intend 0.700

Aviation Psychology and Applied Human Factors (2020) � 2020 Hogrefe Publishing

8 A. A. Stanton et al., Compliance Behavior Model: A Pilot Study

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

what important others expect and do, and also a pilot’s own self-efficacy. For example, a persuasive safety education program is likely to be more effective if it leverages the social influencers listed in Table 2, who deliver messages that provide new information that underpins the beliefs also listed in that table. That is, providing pilots with new infor- mation that might vary these identified beliefs, leveraging those who provide social influence, or highlighting how the perceived control factors have influenced historical tragedy.

In conclusion, the elicitation study has achieved each of the set objectives and in doing so has provided aviation safety education experts lists by which to theme their messaging on the topic. A key limitation is that this study has not identified which of these beliefs, or even which of the direct constructs themselves, most strongly influence the intention to perform the behavior. Such an assessment requires statistical techniques such as path analysis, for which there is insufficient statistical power here. It is this more advanced analysis with a broader scale of respon- dents that takes place in the next stage of research, building on the essential findings here.

References

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T

Ajzen, I. (2006). Constructing a TPB questionnaire: Conceptual and methodological considerations. University of Massachusetts. https://people.umass.edu/aizen/pdf/tpb.measurement.pdf

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Prentice Hall.

Ajzen, I., & Sheikh, S. (2013). Action versus inaction: Anticipated affect in the theory of planned behavior. Journal of Applied Social Psychology, 43(1), 155–162. https://doi.org/10.1111/ j.1559-1816.2012.00989.x

Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140. https://doi.org/10.1016/0749-5978(85)90049-4

Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. The British Journal of Social Psychology/The British Psychological Society, 40(Pt 4), 471.

Australian Transport Safety Bureau. (2018). Aviation occurrence statistics 2008 to 2017. Australian Transport Safety Bureau.

Barber, J. S. (2011). The theory of planned behaviour: Considering drives, proximity and dynamics. Vienna Yearbook of Population Research, 9, 31–35.

Batt, R., & O’Hare, D. (2005). Pilot behaviors in the face of adverse weather: A new look at an old problem. Aviation, Space, and Environmental Medicine, 76(6), 552–559.

Causse, M., Dehais, F., Péran, P., Sabatini, U., & Pastor, J. (2013). The effects of emotion on pilot decision-making: A neu roergonomic approach to aviation safety. Transportation Research Part C, 33, 272–281. https://doi.org/10.1016/j.trc. 2012.04.005

Cialdini, R. B., Kallgren, C. A., & Reno, R. R. (1991). A focus theory of normative conduct: A theoretical refinement and reevaluation of the role of norms in human behavior. In M. P. Zanna (Ed.),

Advances in experimental social psychology (Vol. 24, pp. 201– 234). Academic Press. https://doi.org/10.1016/S0065-2601(08) 60330-5 (http://www.sciencedirect.com/science/article/pii/ S0065260108603305)

Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A review and avenues for further research. Journal of Applied Social Psychology, 28(15), 1429–1464. https://doi.org/10.1111/j.1559-1816.1998.tb01685.x

Dulany, D. E. (1961). Hypotheses and habits in verbal “operant conditioning”. The Journal of Abnormal and Social Psychology, 63(2), 251–263. https://doi.org/10.1037/h0047703

Dulany, D. E. (1968). Awareness, rules, and propositional control: A confrontation with S-R behavior theory. In T. Dixon & D. Horton (Eds.), Verbal behavior and general behavior theory (pp. 340–387). Prentice-Hall.

Duval, S., & Wicklund, R. A. (1972). A theory of objective self awareness. Academic Press.

Fishbein, M. (1967). Attitude and the prediction of behavior. In M. Fishbein (Ed.), Readings in attitude theory and measurement (pp. 477–492). John Wiley & Sons.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior. Addison-Wesley Publishing.

Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: The reasoned action approach. Taylor & Francis Group.

Fogarty, G. J., & Shaw, A. (2010). Safety climate and the theory of planned behavior: Towards the prediction of unsafe behavior. Accident Analysis and Prevention, 42(5), 1455–1459. https:// doi.org/10.1016/j.aap.2009.08.008

Francis, J., Johnston, M., Eccles, M., Walker, A., Grimshaw, J. M., Foy, R., Kaner, E. F. S., Smith, L., & Bonetti, D. (2004). Con- structing questionnaires based on the theory of planned behaviour: A manual for Health Services Researchers. Quality of life and management of living resources. Centre for Health Services Research. http://openaccess.city.ac.uk/id/eprint/1735

Gorsuch, R. L., & Ortberg, J. (1983). Moral obligation and attitudes: Their relation to behavioral intentions. Journal of Personality and Social Psychology, 44(5), 1025–1028. https:// doi.org/10.1037/0022-3514.44.5.1025

Huntzinger, D. (1997). Pilots behaving badly. Flight Safety Australia. (September 1997).

Manstead, A. S. R. (2000). The role of moral norm in the attitude- behavior relation. In D. J. Terry & M. A. Hogg (Eds.), Attitudes, behavior, and social context: The role of norms and group membership (pp. 14–25). Psychology Press.

Moan, I. S., Rise, J., & Andersen, M. (2005). Predicting parents’ intentions not to smoke indoors in the presence of their children using an extended version of the theory of planned behaviour. Psychology & Health, 20(3), 353–371. https://doi. org/10.1080/08870440512331317706

Parker, D., Manstead, A. S. R., & Stradling, S. G. (1995). Extending the theory of planned behaviour: The role of personal norm. British Journal of Social Psychology, 34, 127–137.

Rasmussen, J. (1983). Skills, rules, and knowledge; signals, signs, and symbols, and other distinctions in human performance models. IEEE Transactions on Systems, Man, and Cybernetics, SMC-13, 13(3), 257–266. https://doi.org/10.1109/TSMC.1983.6313160

Reason, J. T. (2008). The human contribution: Unsafe acts, accidents and heroic recoveries. Ashgate.

Reason, J. T., Manstead, A., Stradling, S., Baxter, J., & Campbell, K. (1990). Errors and violations on the road: A real distinction? Ergonomics, 33(10/11), 1315–1332.

Rivis, A., Sheeran, P., & Armitage, C. J. (2009). Expanding the affective and normative components of the theory of planned behavior: A meta-analysis of anticipated affect and moral norms. Journal of Applied Social Psychology, 39(12), 2985. https://doi.org/10.1111/j.1559-1816.2009.00558.x

� 2020 Hogrefe Publishing Aviation Psychology and Applied Human Factors (2020)

A. A. Stanton et al., Compliance Behavior Model: A Pilot Study 9

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .

Schwartz, S. H. (1973). Normative explanations of helping behavior: A critique, proposal, and empirical test. Journal of Experimental Social Psychology, 9(4), 349–364. https://doi.org/ 10.1016/0022-1031(73)90071-1

Schwartz, S. H. (1977). Normative Influences on altruism. In B. Leonard (Ed.), Advances in experimental social psychology (Vol. 10, pp. 221–279). Academic Press.

Seligman, C., Olson, J. M., & Zanna, M. P. (2002). The psychology of values: The Ontario (Vol. 8). Psychology Press.

Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Jour- nal of Consumer Research, 15(3), 325–343. https://doi.org/ 10.1086/209170

Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organi- zational Behavior and Human Performance, 16(1), 27–44. https://doi.org/10.1016/0030-5073(76)90005-2

History Received February 29, 2020 Revision received June 13, 2020 Accepted August 20, 2020 Published online December 2, 2020

Publication Ethics The views expressed by and attributable to this author are his own and do not necessarily reflect those of the Civil Aviation Safety Authority or the Australian Government.

ORCID Anthony A. Stanton

https://orcid.org/0000-0003-0206-1483

Anthony A. Stanton Safety Science Innovation Laboratory Griffith University 37 Sandpiper Avenue Brisbane, QLD 4509 Australia tony.stanton@griffith.edu.au

Anthony Stanton is a PhD candidate at the Griffith University Safety Sci- ence Innovation Laboratory, Bris- bane, QLD, Australia, with a research interest in the psychology of compli- ance behavior. Tony has held leader- ship roles such as CEO, Board Member, Chief Pilot and Flight Examiner. Today he holds a senior manager position with the Australian aviation regulator.

Sidney Dekker is Professor at Griffith University, Brisbane, QLD, Australia. He is best-selling author of, most recently, The Safety Anarchist (2018), The End of Heaven (2017), Just Cul- ture (2016), Safety Differently (2015), The Field Guide to Understanding ‘Human Error’ (2014), Second Victim (2013), Drift Into Failure (2012), and Patient Safety (2011).

Patrick Murray is Professor of Aviation and Logistics at the University of Southern Queensland, Australia. Pre- viously holding senior positions in the military, a major airline, and the Australian aviation regulator, he is an active flying instructor and researches in LOSA, evidence-based training, and airline safety.

Gui Lohmann is Professor in Aviation and Deputy Head of School of Engi- neering and Built Environment at Griffith University, Brisbane, QLD, Australia. His research expertise includes air transport studies, par- ticularly the interface with tourism, airline business models, transport geography, the regulatory regime of Australian airports, as well as airport passengers’ travel patterns and behaviors.

Aviation Psychology and Applied Human Factors (2020) � 2020 Hogrefe Publishing

10 A. A. Stanton et al., Compliance Behavior Model: A Pilot Study

T hi

s do

cu m

en t i

s co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. T

hi s

ar tic

le is

in te

nd ed

s ol

el y

fo r t

he p

er so

na l u

se o

f t he

in di

vi du

al u

se r a

nd is

n ot

to b

e di

ss em

in at

ed b

ro ad

ly .