Operant conditioning

Many of our behaviours are said to be emitted rather than elicited because they do not require an eliciting stimulus to occur. Behaviours such as walking, eating or talking are the function of a larger number of variables. In these cases, the organism operates on its surrounding environment encountering special kinds of stimuli – consequences - which alter the probability of repeating that behaviour in the future. Consequences that increase the probability of behaviour (or strengthen behaviour) are reinforcers; consequences that decrease the likelihood of behaviour (or weaken behaviour) are aversive stimuli or punishers. Reinforcers and punishers can both be positive or negative, depending on whether behaviour is strengthened or weakened by the addition or removal of a stimulus, respectively (Skinner, 1953). In casual discourse, positive reinforcement is usually implied by what we “want” or “like” to do, and negative reinforcement by the things that we “have to do”.

Larson et al. (2013) tested the effects of various reinforcers in the levels of moderate- to-vigorous PA (MVPA) in young children (n=2) (Larson et al., 2013) in four

conditions: alone, attention contingent on MVPA (e.g. praise), adult interaction contingent on MVPA (e.g. playing with the child), and escape from task demands contingent on MVPA (negative reinforcement). Results showed that the children were most active when attention and interactive play were contingent on MVPA, stressing the role of reinforcements derived from social interaction (Larson et al., 2013). The sample size in this study was very small but findings give some suggestion of the importance of companionship, for example in the context of the journey to school. The consequences of behaviour are typically correlated with other aspects of the environment or situation. A discriminative stimulus sets the occasion upon which responding will have a particular consequence; the relationship between a discriminative stimulus, the behaviour and its consequents form a three-term contingency, or contingency of reinforcement (or punishment) (Mazur, 2006). This notion is closely associated with that of antecedent-behaviour-consequents (ABCs) relationships. Antecedents include any contextual factors or conditions,

41

including discriminative stimuli, which alter the value of the reinforcer and increase likelihood of engaging behaviours to obtain the reinforcer (Kazdin, 2012). This

includes the extent to which the individual has gone with or without the reinforcer, as well as previous contingencies of reinforcement or punishment (Pierce, 2004). In other words, antecedents include momentary conditions and past experiences, all of which affect behaviour.

Various examples of ABCs are found in health promotion research. In sun-protection promotion, warmth, positive social comments and attractive tan were reinforcers for sun exposure, while parental approval and avoidance of criticism were reinforcing for sun-protective behaviours (Adams et al., 2009). Excessive heat, sweating, critical social comments and sunburns had a punishing effect on sun exposure; as did a greasy feeling from sunscreen, decreased time outdoors and unpleasant social attention for protective clothing, on sun protective action. Possible long-term reinforcers for sun-protection included healthy skin, avoidance of aged skin or skin cancer, and for sun-exposure, avoidance of criticism for pale colour (Adams et al., 2009).

PA interventions have also been classified having the ABC model as reference (Sallis and Owen, 1998). Examples of changing antecedents included planning specific times and locations for PA, keeping running shoes in the car, and living near attractive facilities. Changing behaviour by altering the consequences included socialisation with others while doing PA, and monetary incentives for active travel to work. The authors stressed that interventions needed not only to administrate

reinforcers, but also to consider ways of removing or reducing punishers, e.g. discomfort during exertion, or being laughed at because of poor sport skills. As an operant behaviour, an assessment of school travel behaviour requires

breaking down antecedents and consequents (Kirk et al., 2006). Table 2.2 shows a list of potential antecedents and potential reinforcers for AST, and potential

reinforcers for motorised travel. The list is non-exhaustive and suggested by the existing, mainly correlational, literature, and should be subject to further investigation. Antecedents are classed according to the various levels of the BEM. These include individual level (normative group and physical), local level (clinical services, built and social environment), community level (policies, laws, media), and social/cultural level (nationality and culture specific) (Adams et al., 2009). However, because Hovell et al.

42

provide little detail about each level, this division is only approximate, and for convenience, local and community levels are presented together here (e.g. speed limits exist locally, but reflect national laws and policies).

Most literature suggests that parents are the primary decision makers of their child’s travel modes to school (Davison et al., 2008; Panter et al., 2008a; Faulkner et al., 2010) but their decision may be affected by their child’s preferences (Pont et al., 2011a). Thus, an ABC assessment of school travel mode is likely to benefit from attending to both perspectives.

ABC assessments have been traditionally carried out at the individual level (e.g. (Cole and Bonem, 2000)). This tailoring reflects the fact that the same behaviour (e.g. smoking) can occur for quite different reasons, or in other words, be the function of different antecedents and consequents. However, because an individual

assessment or treatment is not at stake here, potential antecedents and consequents for AST are presented together, as previously reported in other contexts of health promotion (Gielen and Sleet, 2003; Adams et al., 2009).

‘Antecedents’ essentially means stimuli which increase opportunities for

reinforcement. As such, they increase the likelihood of behaviour, but always in combination with other contextual factors such as conditions favouring engagement in alternative activities and previous history.

Some correlates such as gender of the child or ethnic background are omitted

because their effects are likely to be indirect. For example, in the parent environment, parents are often more protective towards girls (e.g. (De Meester et al., 2014)) and safety concerns are reported more frequently for girls than for boys (Evenson et al., 2006). For many parents, protective behaviour towards girls may be more accepted (i.e. reinforced), or less resisted (i.e. punished), than towards boys (e.g. (Mayhew et al., 2004)). Likewise, in some social minorities, car ownership, and its daily use, is more common, and those who drive their children may be less often subject to unwanted attention than parents from other backgrounds who do the same (e.g. (Eyre et al., 2013)). We still need to ask how these contingencies can be tackled, but their specification is likely to allow a higher level of practicality. Distance, a key determinant of AST, is also likely to be a moderator for other variables (e.g. house location), rather than being itself amenable to direct manipulation.

43

Table 2.2 - Potential antecedents for AST, potential reinforcers for AST, and potential reinforcers for motorised travel

“Social encouragement and approval for active lifestyles” and “equality and

environmental awareness” are particularly difficult to operationalise, and may refer both to antecedents (e.g. laws, behaviour of other people) and to consequences (e.g. approval from others contingent upon behaviour). Interestingly, both may well be an umbrella term that encompasses many, if not all, of the other antecedent variables.

Potential antecedents for AST Potential reinforcers for AST

Potential reinforcers for motorised travel In di v id ua l

- No car available (Sirard and Slater, 2008; Pont et al., 2009)

- Lower socio-economic status (e.g. less money available) (Davison et al., 2008; Sirard and Slater, 2008; Pont et al., 2009) (Panter et al., 2013) - Independent mobility (allowed to be out alone) (Davison et al., 2008)

- Parents currently walk or cycle to work (Davison et al., 2008; Van Kann et al., 2015)

- Parents encourage AST and walking (Sirard and Slater, 2008) (Panter et al., 2010a)

- Parents encourage PA and social interaction during AST (Davison et al., 2008)

- Encouragement from others (Davison et al., 2008)

- AST more convenient, time available (Panter et al., 2013) (Trapp et al., 2012)

Immediate:

- Being active/fit (Fusco et al., 2013) (Davison et al., 2003) (Kirby and Inchley, 2009) - Having fun (Mitchell et al., 2007) (Hunter et al., 2015) (Romero, 2015)

- Listening to music (Kirby and Inchley, 2009)

- Gives energy (Mitchell et al., 2007)

- Interesting things to look at/nature (Mitchell et al., 2007) (Fusco et al., 2013)

- Interaction with other

children/make friends (Davison et al., 2008) (Kirby and Inchley, 2009) (Fusco et al., 2013) - Cycling is “cool” (attention and approval from others) (Trapp et al., 2011) (Baslington, 2009)

- Feeling the sun (Mitchell et al., 2007)

- Breathing fresh air (Kirby and Inchley, 2009)

- Faster than walking (cycling) (Mitchell et al., 2007)

- Feeling more alert at school (Mitchell et al., 2007) (Kirby and Inchley, 2009)

- Save money (Kirby and Inchley, 2009)

- Approval from parents and others (Sirard and Slater, 2008) (Panter et al., 2009) Non-naturally occurring: - Rewards/incentives (Kirby and Inchley, 2009) (Davies, 2012)

- Raising funds for charity (Hunter et al., 2015) Delayed:

- Better health (Kirby and Inchley, 2009; Fusco et al., 2013)

- Benefits to the environment (Davison et al., 2003) (Kirby and Inchley, 2009) Immediate: - Time saved (convenience) (Faulkner et al., 2010) - Less effort (Faulkner et al., 2010)

- Safer (Fyhri et al., 2011)

- Having a car is “cool” (attention and approval from others) (Lorenc et al., 2008) - Listening to music (Romero, 2015) - Cheaper (bus) (Pooley et al., 2010) - Interaction with friends (bus) (Mitchell et al., 2007)

- Good view from the bus (Mitchell et al., 2007)

- Avoid bullies (Ahlport et al., 2008) - Avoid cold/rain (Kirby and Inchley, 2009)

- Avoid carrying bag (Kirby and Inchley, 2009) - Avoid exposure to pollution (Pooley et al., 2010) Delayed: - Benefits to the environment (bus compared to car) (Baslington, 2009) Lo c a l/ c om m un ity

- Other children in the area walk/cycle to school (Davison et al., 2008)

- Shorter distance (Davison et al., 2008; Panter et al., 2008a; Saelens and Handy, 2008; Sirard and Slater, 2008; Pont et al., 2009; Wong et al., 2011b)

- State-funded schools (Davison et al., 2008; Sirard and Slater, 2008)

- Non-religiously affiliated (Davison et al., 2008) - Schools that encourage physical education and active travel initiatives (Sirard and Slater, 2008) - Proximity to shops (Panter et al., 2008b; Sirard and Slater, 2008)

- Walkability index (calculated by residential density, retail floor area ratio, intersection density and land use mix) (Sirard and Pate, 2001; Davison et al., 2008; Saelens and Handy, 2008)

- Urban areas (Davison et al., 2008; Sirard et al., 2008)

- Pavements, cycle lanes and cycle parking facilities (Davison et al., 2008; Saelens and Handy, 2008; Sirard and Slater, 2008; Pont et al., 2009)

- Parks and recreation facilities (Pont et al., 2009) - Aesthetics and more windows facing the street (Sirard and Slater, 2008)

- Good weather (Fraser and Lock, 2011) - Road and area safety (Davison et al., 2008; Panter et al., 2008a; Sirard and Slater, 2008; Pont et al., 2009) Soc ia l/ c ul tura

l - Social encouragement and approval for active lifestyles (McMillan et al., 2006)

- Equality and environmental awareness (Garrard, 2011)

44

Punishers for AST, not covered here, are worth investigating in view of devising strategies to minimise their impact.

Table 2.2 suggests that school travel behaviour can be tackled essentially in three ways: by changing the antecedents (stimulus control), by introducing (or

strengthening) reinforcers for AST, and/or by withdrawing reinforcers for motorised travel (extinction). This perspective and techniques are largely consistent with the ‘nudging approach’ adopted by the Behavioural Insights Team, a government institution dedicated to the application of behavioural science (BIT, 2016). Punishment of motorised travel (e.g. fines for driving near schools) is another possibility, but on its own, may be less effective and carries well-known undesirable side effects such as fear or anger which may interfere with performance and learning, or escape or retaliation by the punished (Lerman and Vorndran, 2002). When

applied, the resort to punishment is usually supplemented with the implementation of alternative behaviours using positive reinforcement, a technique known as differential reinforcement (Lerman and Vorndran, 2002; Miltenberger, 2011) (e.g. fining speedy drivers but rewarding exemplar ones).