A synthesis of (non-)compliance theories with applications to small-scale fisheries research and practice

Non-compliance in fisheries is a persistent challenge for the conservation and sustainable management of the oceans and has particularly acute impacts in small-scale fisheries contexts. Small-scale fisheries often suffer from chronic overexploitation, poor management, lack of enforcement and non-compliance, but small-scale fishers are highly dependent on the ocean as a source of employment and food. Improving our understanding of the determinants of non-compliant behaviours in small-scale fisheries can help develop strategies to prevent and reduce its consequences. Here, we review two main approaches for the study of non-compliant behaviours and crimes more broadly, spanning criminology, economics and psychology. On the one hand, actor-based approaches address the underlying motivations for people to comply or not with regulations. Opportunity-based approaches, on the other hand, assume that non-compliance is not distributed randomly across space and time and focuses on the role that the immediate environment plays in the performance of non-compliant behaviours. We discuss potential applications of actor-based and opportunity-based approaches in guiding small-scale fisheries non-compliance research. Moreover, we provide guiding principles for integrating these approaches in a complementary way, highlighting opportunities and challenges for building a better non-compliance research agenda for fisheries and beyond. Addressing non-compliance is a common challenge for natural resource management in multiple ecosystems. Integrating these two perspectives has the potential to improve both research and practice.

. Second, the applicability of actor-based theories and approaches for non-compliance prevention is limited by the difficulty of crafting interventions that change the underlying motivations that drive behaviour (Cornish & Clarke, 1987).
For instance, whilst normative motivations have been identified as key predictors for compliance in fisheries Thomas et al., 2016), changing a group's normative beliefs is challenging or even unfeasible (Cialdini, 2003).
The shortcomings of actor-oriented approaches to studying and preventing non-compliance have fuelled alternative ways to think about non-compliance and crimes more broadly (Clarke, 2016). As such, there has been a growing effort in the criminological literature to examine the situational opportunities that affect the occurrence of non-compliant behaviours, with the underlying premise (whether explicit or not) that non-compliance is mostly a product of opportunity rather than underlying motivation (Brantingham & Brantingham, 1981;Clarke & Felson, 2004;Wortley & Townsley, 2016).
Opportunity-based approaches assume that non-compliance is not distributed randomly across space and time and focus on the role that the immediate environment plays in the performance of non-compliant behaviours (Wortley & Townsley, 2016). Evidence from different studies suggests that, in fact, non-compliant use of, and trade in, natural resources concentrates at specific places, facilities, times and products Kurland, Pires, McFann, & Moreto, 2017;Moreto & Lemieux, 2015). This presents a potential opportunity to apply opportunity-based theories and approaches for studying non-compliance in natural resources more generally, and fisheries specifically.
Several studies have applied opportunity-based approaches to guide the study of non-compliance in commercial and recreational fisheries (Davis & Harasti, 2020;Isaacs & Witbooi, 2015; Lindley & Techera, 2019; Thiault et al., 2020;Weekers & Zahnow, 2018). However, efforts to apply opportunity-based approaches in small-scale fisheries contexts are lacking. Including an opportunity-based approach into small-scale fisheries management research and practice has the potential to provide new insights, methods and approaches that can complement the predominant actor-based focus. Whilst the line dividing actor-based and opportunity-based approaches might at times blur, they historically come from different theoretical perspectives, and as such, their application differs. For instance, actor-based approaches may (unintentionally) place the "burden" on the individual fisher -whilst opportunity-based approaches seek to understand how situations create opportunities for non-compliance. Combining these approaches, therefore, can enable researchers to better understand non-compliance, by tackling both the individual theoretical drivers of behaviour and the situations that, in practice, bring opportunities for non-compliance in small-scale fisheries contexts.
Here, we aim to bridge the gap between opportunity-based and actor-based approaches to studying non-compliance in smallscale fisheries and other natural resource use contexts. We structure our paper according to the analytical focus. First, we consider actor-based approaches, which try to explain the underlying motivations for non-compliance. Next, we describe opportunity-based theories, models and frameworks that have been used to study non-compliance more broadly. We then discuss how opportunity-based and actor-based approaches to study non-compliance can be applied in the context of small-scale fisheries. We finish by providing guiding principles on how to bring these approaches together in a complementary way. By doing so, we hope to point to the most pressing opportunities for building a better small-scale fishing non-compliance research agenda and contribute to the study of non-compliance beyond fisheries.

| AC TO R-BA S ED A PPROACHE S
Several theories and models have been proposed to explain the underlying motivations for actors to comply or not with rules and F I G U R E 1 Actor-based and opportunity-based approaches for studying non-compliance in small-scale fisheries [Colour figure can be viewed at wileyonlinelibrary.com] regulations (Becker, 1968;Cialdini & Trost, 1998;Ishoy, 2016;Ostrom, 1990;Tyler, 1990). Consequently, fisheries scientists and conservationists have drawn from these theories in order to better understand why fishers comply or not with conservation and management regulations (Arias, Cinner, Jones, & Pressey, 2015;Bergseth & Roscher, 2018;Bova, Halse, Aswani, & Potts, 2017;Kuperan & Sutinen, 1998). Ideally, better understanding what motivates noncompliant behaviours can inform and guide targeted interventions aimed at reducing the incidence of non-compliance (Bergseth & Roscher, 2018;Mackay et al., 2018;Nielsen & Mathiesen, 2003).
Generally speaking, the behavioural, psychological and economic approaches that have been applied for understanding fishers' motivations for non-compliance assume decision-making is similar to the approaches used for compliance with rules more generally (Gezelius, 2002;Keane et al., 2008;Sutinen & Kuperan, 1999). As such, research efforts have been aimed at understanding the diversity of factors that influence decision-making in the context of fisheries, with the underlying premise that reductions in non-compliance can be obtained through manipulating these factors in favour of compliance (Bova et al., 2017;Oyanedel et al., 2020). Below, we describe three conventional approaches that have been used to assess and understand why people engage in non-compliant use of natural resources in general, and fisheries specifically, namely; the Instrumental Model, Compliance Framework and the Theory of Planned Behaviour (Table 1). This is by no means an exhaustive list, but provides parallel, although sometimes overlapping, ways of thinking about why people engage in non-compliant behaviours.

| Instrumental or deterrence model
The instrumental model (also known as the deterrence model) of compliance has its roots in the economic theory of law, first proposed by Becker (1968). It assumes that, as individuals, actors will calculate the potential costs and benefits of non-compliant behaviours, and will engage in non-compliance when benefits outweigh costs. This calculation is essentially the same than for any actor attempting to maximize utility subject to budget constraints (Sumaila et al., 2006). As such, the level of non-compliance in which a utilitymaximizer actor will engage is calculated from the expected reward from non-compliance minus the costs, computed as the probability of detection and sanction multiplied by the severity of the resulting punishment (Equation 1) (Becker, 1968).
where EU is expected utility, p is the probability of capture and punishment, U is utility, b is income if undetected, f is fine, and b − f income if punished (Garoupa, 1997). Sutinen and Andersen (1985) first adapted Becker's model to understand the effect of imperfect enforcement on fisher behaviour. From there, this model has been largely applied in fisheries management to understand how to increase compliance TA B L E 1 Actor-based approaches: Instrumental model, compliance framework and the theory of planned behaviour

References in other natural resources
Instrumental model Probability of detection An actor's compliance decision is based on the calculated potential costs and benefits of the noncompliant behaviour and will decide to engage in noncompliance when benefits outweigh costs Arias and Sutton (2013), King and Sutinen (2010), Kuperan and Sutinen (1998), Nielsen and Mathiesen (2003) Bulte and Van Kooten (1999), Damania, Milner-Gulland, and Crookes (2005) (Arias, 2015;Doumbouya et al., 2017;King & Sutinen, 2010;Sumaila et al., 2006). Two main mechanisms by which to increase compliance can be deduced from this model. The first involves increasing the actual probability of detection. This mechanism requires increases in law enforcer numbers, or patrol effort or effectiveness, which are usually costly and can prove ineffective in raising the probability of detection to significant levels if not well-funded (Paternoster, 2010). As such, increasing the real probability of detecting non-compliance can prove challenging or logistically unfeasible, especially in small-scale fisheries contexts that lack proper enforcement capacities or budgets (Muller, Oyanedel, & Monteferri, 2019

| Compliance framework
The compliance framework was first proposed for forestry con-

| Legitimacy-based motivations
Legitimacy-based motivations relate to how the acceptance of decision-making and its outcomes motivate actors to comply with regulations (Levi, Sacks, & Tyler, 2009;Ramcilovic-Suominen & Epstein, 2012). Legitimacy can play a crucial role in motivating compliance, and can also make governance easier and more effective (Jentoft, 1989). There are several and evolving ways to conceptualize and measure legitimacy, but these can be categorized into procedural legitimacy, legitimacy of authorities, and outcome legitimacy.
Procedural legitimacy deals with how collective decision-making processes affect individual motivations for compliance (Tyler, 1990).
When decision-making is participatory, transparent and accountable, individuals are more likely to comply (Levi et al., 2009;Ramcilovic-Suominen & Epstein, 2012). Legitimacy of authority has to do with how leaders are perceived, including their perceived capability as decision-makers, and in turn, how that affects individual compliance (Levi et al., 2009). Finally, outcome legitimacy considers the fairness and appropriateness of rules as perceived by those who are affected by them. Rules that are perceived as fair and effective are much more likely to be complied with (Jentoft, 1989;Kuperan & Sutinen, 1998;Nielsen, 2003).

| Normative motivations
The normative component of the framework emphasizes social and personal norms as motivations for compliance. Norms are prescriptions commonly accepted in a group, supporting desirable behaviours and forbidding undesirable ones (Gezelius, 2002;Ramcilovic-Suominen & Epstein, 2015). As such, norms can have a significant effect in strengthening adherence to fisheries rules or reinforcing non-compliance (de la Torre-Castro, 2006). The role of norms as a motivation for compliance has been a topic of increasing interest in literature around non-compliance in fisheries, especially in recreational (Arias & Sutton, 2013;Bergseth & Roscher, 2018;Bova et al., 2017;Thomas et al., 2016) and small-scale fisheries contexts (Arias & Pressey, 2016;Battista et al., 2018;Oyanedel et al., 2020).
Normative motivations and the way they affect compliance can be classified in three main categories: personal norms (e.g., individual values regarding the behaviour), injunctive norms (e.g., perceived moral values of a group) and descriptive norms (e.g., perception of what others do) (Cialdini & Trost, 1998;Hatcher et al., 2000;Thomas et al., 2016). Oyanedel et al. (2020) provide an example of how the application of the compliance framework can aid in understanding small-scale fisheries non-compliance. They assessed non-compliance rates and the motivations behind these behaviours in a small-scale fishery in Chile. They found that whilst 93%-100% of fishers complied with gear or temporal restrictions, only 3% did so for the fishery's quota limit. Legitimacy-based motivations were more important than other motivations in explaining this diversity of fishers' responses towards regulations. Similarly, they found that normative motivations best predicted the degree of non-compliance with the quota limit, and contextual factors such as the per-fisher quota level (which relates to the instrumental component) explained broader non-compliance patterns.

| Theory of planned behaviour
The Theory of Planned Behaviour (TPB) is an integrative model widely used in social psychology, which seeks to predict an individual's behaviour (Ajzen, 2011). It focuses on the individual's deliberative decision-making process by understanding their intention to perform a behaviour (Bergseth & Roscher, 2018). It assumes that the stronger the intention, the more likely it is that the individual will perform the behaviour (Ajzen, 2011

| OPP ORTUNIT Y-BA S ED APPROACHE S
Here, we review opportunity-based approaches that have been used to study non-compliance and illegality more broadly, which gather around the Environmental Criminology and Crime Analysis school of thought. Environmental criminologists focus on the environmental (contextual) factors that influence the immediate decision to perform a non-compliant behaviour (Brantingham & Brantingham, 1981;Clarke & Felson, 2004). Environmental criminologists have an applied mission, and they guide their studies towards the development of opportunity-reducing strategies, with the premise that by manipulating crime-causing situations, effective prevention and disruption of non-compliant activities can be obtained (Clarke, 1980(Clarke, , 2016. Environmental Criminology and Crime Analysis have three main operational models, described below: Rational Choice, Crime Pattern and Routine Activity. These models were conceived and initially developed in isolation, but they have similarities and overlaps. As such, they are not exclusive, and their application in practice involves convergence (Wortley & Townsley, 2016).

| Rational choice model
The rational choice model is built upon the principle that "specific crimes are chosen and committed for specific reasons" (Cornish & Clarke, 1987). In this theory-based model, the premise is that several factors are considered in the actor's decision to engage in crime. These factors are viewed as properties of the circumstances and include the possible payoff, perceived risk or skills needed in the context of the actor's motives, experience, expertise and ability (Cornish & Clarke, 1987). This model implies that the environmental or contextual data that the actor uses can be modified to change their decision to commit a crime.
Whilst this model is similar to Becker's (1968) deterrence model (see Section 2.1), in that it asserts that crimes occur when the anticipated benefits outweigh costs, there are two main differences between these models in how costs and rewards are calculated. First, the rational choice model defines rewards not only in economic terms but also considers the emotional or psychological benefits of a criminal act (Clarke, 1980). Second, the rational choice model does not consider costs only in terms of the probability of detection and sanction but also concerning the particular properties of the crime that can make it costly to perform (such as the level of skill or physical fitness required). The rational choice model is built on the evidence that societies, in general, are incredibly inefficient at delivering economic punishment and therefore making crimes costly (Cornish & Clarke, 1987). In this sense, Cornish and Clarke (1987) assert that Becker's model might be useful in some particular circumstances, but fails to explain most crimes as it does not consider the opportunistic nature of many kinds of crimes and the non-economic rewards and barriers that potential criminals face. Altogether, there is conceptual overlap between the rational choice model and Becker's model. However, the rational choice model is a critical component of several of the tools and methods used in Opportunitybased approaches (see Section 3.4), and as such, its application differs substantially from Becker's model.

| Routine activity model
The routine activity model has its empirically based roots in the evidence that crime rate trends and cycles are influenced by structural changes in routine activity patterns. This occurs when changes in routine activities affect the convergence in space and time of the three minimal elements for a crime: (a) motivated actors, (b) suitable targets, and (c) the absence of capable guardians (Cohen & Felson, 1979). If any of these elements is missing, crimes do not take place. This model implies that even when the number of motivated offenders is constant, crime rates can change due to changes in suitable targets or the absence of guardians (note: guardians not only means police but could also be regular citizens). Therefore, this model takes user motivations towards non-compliance as a given and studies how spatial-temporal factors of the organization of daily life can help convert criminal inclination into action (Cohen & Felson, 1979).
In their seminal paper, Cohen and Felson (1979) use the routine activity model to explain the paradox that "urban violent crime rates (in the US) increased substantially during the past decade when the conditions that are supposed to cause violent crime have not worsened-have indeed, generally, improved" (Cohen & Felson, 1979).
They hypothesized that the increase in the crime rates in the United States in the 60s was related to changes in the routine activity structures of everyday life in American society. These changes increased the suitability of targets and decreased guardian presence. To test this, they examined the relationship between household unattendance and crime rates. They found a strong and significant positive relationship between these two variables. This analysis suggests that routine activity changes may provide the opportunity for crimes to occur. Societal shifts such as increases in female labour participation, more vacations and higher enrolment in college changed routine activities. This meant that houses were less attended and therefore provided more suitable targets which, in the absence of capable guardians, enabled crimes to occur.

| Crime pattern model
The crime pattern model's objective is to empirically measure and account for the non-uniformity and non-randomness observed in crime patterns (Brantingham & Brantingham, 1984). This model is based on the idea that people develop a pattern of repetitive activity in their normal life. Understanding the factors that determine the specific spatial patterns of crime is, therefore, necessary to prevent it. This pattern includes nodes (such as the workplace, home, shopping, etc.) and routes between them. Offenders behave in this same way as everyone else and will be more comfortable committing crimes closer to the areas they frequent (Brantingham, Brantingham, & Andresen, 2017;Kinney, Brantingham, Wuschke, Kirk, & Brantingham, 2008). Therefore, the routes and nodes that shape non-criminal activities, influence how criminal activities are shaped as well.
Using the crime pattern model, Kinney et al. (2008) identified crime attractors, generators and detractors. They conceptualized these as nodes whose structure influenced crime patterns. They found that most land-use types acted as detractors, as no crime was detected in them. However, some types of crimes, like assault and motor vehicle theft, concentrated on specific land-use types, such as shopping centres, which therefore acted as crime generators and attractors. This occurred because these were high activity nodes, that attracted large concentrations of victims and offenders.
Further, they suggested that better understanding the distribution of the nodes which act as crime attractors, detractors or generators throughout urban areas can help reduce overall crime rates. As such, by better understanding daily life activities and patterns, opportunistic crime can be better prevented and disrupted (Brantingham & Brantingham, 1993).

| Methods and tools used in Environmental Criminology and Crime Analysis
The rational choice, routine activity and crime pattern models are the theoretical basis of opportunity-based approaches, which fall under the Environmental Criminology and Crime Analysis school of thought. Several analytical methods have been developed to operationalize and combine the abovementioned models, breaking crime down into specific analysable components in order to propose and design prevention measures. In this section, we review some of these methods, focusing on those that have been used in fisheries or other natural resource management contexts (Figure 2).

| Crime script analysis
Crime script analysis was first proposed by Cornish (1994), based on the premise that crimes are discrete events in space and time, but the realization of the crime itself takes place within a context of many other events. The crime itself is usually the object of study, typically overlooking certain other stages in the crime-commission process (e.g., getting the necessary tools or exiting the setting). The script refers to an "event" schema where there is a causal effect from early to later events; that is, one event in the script enables the occurrence of a later event (Cornish, 1994). By concentrating on the way that events unfold through time, the crime script analysis provides researchers and practitioners with an analytical tool to understand a series of rational, goal-oriented actions (Cornish, 1994).
Crime script can operate at different levels of analysis; from specific crime situations where rich information is available, to analysing larger-level scripts or more general crimes.
Crime script analysis has been applied to study non-compliance in fisheries and seafood fraud. Petrossian and Pezzella (2018)

| Risky facilities framework
The risky facilities framework builds on the premise that "for any group of similar facilities (e.g., taverns, parking lots or bus shelters), a small proportion of the group accounts for the majority of crime experienced by the entire group" (Eck, Clarke, & Guerette, 2007).
Whilst several authors have analysed hotspots and map them, the comparison between facilities of the same sort allows identification of specific characteristics that could explain their risk, providing the base to design preventive actions. More practically, this framework allows concentrating of efforts in certain facilities where most crime occurs, instead of targeting a large number of facilities where little crime occurs. Some of the variables identified to influence a facility's risk are size, number and quantity of "hot products" in the facility, location, management effectiveness and design and layout. Petrossian, Marteache, et al. (2015) applied this framework to study what characteristics make ports attractive for non-compliant vessels to land their catch. To do this, they analysed data on the ports used by vessels that were listed as performing non-compliant fishing by Regional Fisheries Management Organizations (RFMO).
They identified a total of 120 ports in 70 countries where these vessels had operated between 2004 and 2009. They found that larger ports that had higher vessel traffic were more visited by non-compliant vessels. Also, ports in countries where non-compliance is more common, corruption is higher and catch inspection schemes were less effective, were also more visited. This points out the variables that could be modified to disrupt non-compliant vessel fishing operations.

| CRAVED framework
The CRAVED framework was first proposed by Clarke and Webb (1999) to analyse what makes some products more attractive for theft than others. CRAVED stands for: Concealable, Removable, Available, Valuable, Enjoyable and Disposable. These six attributes of a product are hypothesized to make a product more attractive (Clarke & Webb, 1999). For the application of CRAVED, the indicators for each of the attributes must be fitted to the specific crime and product being studied. An indicator of Removable will vary dramatically depending on, for example, the species being studied (Petrossian & Clarke, 2014). This framework helps users to compare the attributes between products and explain changing patterns in crime targets. This framework has been used for studying wildlife products (Moreto & Lemieux, 2015) and fish more specifically (Petrossian & Clarke, 2014;Petrossian, Weis, & Pires, 2015).  for instance, found that crab and lobster species that were subject to higher levels of non-compliant fishing were those that were more Available (subcomponent abundant), Valuable and Enjoyable. From this, they provide guidance on how to reduce non-compliance through prioritizing and targeting those identified attributes.

| Situational crime prevention
Situational crime prevention (SCP) was first suggested by Clarke (1980). SCP is a framework that offers a suite of techniques that can help build solutions to prevent crime. SCP techniques are based on an understanding of the processes undertaken to commit a crime.
By disentangling the situational features that enable crimes, SCP techniques aim to influence an actor's choice to engage in it. SCP techniques are organized into five categories: increase the effort,

F I G U R E 2 Examples of the application of methods and tools from Environmental Criminology and Crime Analysis in the study of smallscale fisheries non-compliance [Colour figure can be viewed at wileyonlinelibrary.com]
increase the risk, reduce the reward, reduce provocations and remove excuses. Petrossian (2015) applied the SCP framework to study the relationship between non-compliant fishing in 53 countries and local situational factors. She found that non-compliance risk was higher in countries with more commercially important fisheries that were closer to ports of convenience. Similarly, she found that countries with higher management and enforcement capacities had lower levels of non-compliance.

| B RING ING AC TOR-BA S ED AND OPP ORTUNIT Y-BA S ED APPROACHE S TOG E THER TO ADVAN CE S MALL-SC ALE FIS HERIE S NON -COMPLIAN CE RE S E ARCH
Small-scale fisheries operate in diverse economic, social and cultural contexts (Cohen et al., 2019), preventing bullet-proof solutions to non-compliance problems (Boonstra et al., 2017;Mahon, McConney, & Roy, 2008;Oyanedel et al., 2020;Song et al., 2020). However, one characteristic that small-scale fisheries share is that their operation depends on both social and ecological factors (Basurto, Gelcich, & Ostrom, 2013;Lindkvist et al., 2017). Considering the socio-ecological nature of small-scale fisheries, we provide an overview of three challenges for researching non-compliance that can be better framed and tackled through bringing actor-and opportunity-based approaches together.
However, the ecological characteristics of small-scale fisheries are also critical determinants of the availability of opportunities for non-compliance. The inherent spatial and temporal variability of ocean ecosystems makes it highly likely that opportunities for noncompliance vary over a range of temporal and spatial scales. There is growing evidence that this variability results in non-randomly distributed opportunities for non-compliant fishing, which is concentrated in hotspots (Davis & Harasti, 2020;Thiault et al., 2020;Weekers, Zahnow, & Mazerolle, 2019). Identifying these hotspots, and how they vary over time and space, is of crucial importance to understand how environmental context-specific variables produce emergent opportunities for non-compliance. Ignoring the dynamic ecological features of small-scale fisheries contexts will result in an incomplete understanding of why, how and when non-compliant fishing could emerge (Petrossian, 2018). Focusing on the ecological elements of the system is important even though most research has focused on social elements. As such, combining actor-based and opportunity-based approaches for researching non-compliance in small-scale fisheries can produce more robust results by incorporating both the social and ecological features that determine noncompliant fishing dynamics.

| Choose your battles wisely: improving identification for prioritization in diverse small-scale fisheries systems
Small-scale fisheries management usually comes with budget constraints affecting the design, implementation and enforcement of rules, from which situations conducive to non-compliance can emerge Gelcich et al., 2017). In the low-governance, budget-limited situation of small-scale fisheries it is of utmost importance to prioritize efforts to reduce non-compliance effectively. Here, combining actor-based and opportunity-based approaches can aid in the identification of the most pressing facilities, resources and locations where non-compliance is likely to concentrate. For instance, the Risky Facilities Framework can help identify the ports where non-compliant vessels land their catches , or researchers can use the CRAVED model to understand which species are more attractive to non-compliant fishers and which attributes makes them so ( Figure 2). Further, research can focus on understanding the social characteristics of places where non-compliance develops. Social disorganization models, for instance, focus on the effectiveness of communities at preventing non-compliance through informal control mechanisms (Sampson & Groves, 1989). Understanding fishers' perception of the legitimacy of rules at the local scale can help predict compliance levels and informal control mechanism that might help prevent non-compliance Sampson & Groves, 1989).

| Neither the actor-nor opportunity-based approach on its own fully explains non-compliance in socio-ecological systems such as small-scale fisheries
Understanding the interaction of actor-and opportunity-based approaches when studying non-compliance in small-scale fisheries can provide useful insights into the socio-ecological nature of non-compliance. For instance, Oyanedel, Keim, Castilla, and Gelcich (2018) describe a quite unexpected form of non-compliance that could be better understood based on the interaction of actor-based and opportunity-based approaches. Using the randomised response technique (Fox & Tracy, 1986;Lensvelt-Mulders, Hox, van der Heijden, & Maas, 2005), they empirically assessed the proportion of divers that violated several rules of a territorial user right for fisheries (TURF) system in a small-scale context in Chile. They found that 46% of fishers who belonged to unions with user rights non-compliantly fished with the consent of their union leaders (they were non-compliant because the catch was not reported to authorities).
On the one hand, normative and legitimacy-based motivations for non-compliance were aligned for these fishers, as they were authorized by their leaders to do so. On the other hand, there is evidence that TURF areas are more prolific fishing grounds, and as such are more attractive for fishers, which provides an opportunity-based account for this behaviour (Gelcich et al., 2017;Gelcich, Godoy, Prado, & Castilla, 2008). Further, the authors provide an explanation for this behaviour that can help complement the understanding of this form of non-compliance: "Because fisher unions find it too complicated, costly, or useless to officially report their catches, they are not reporting to authorities, even if they fish within legal margins (respecting the minimum size and closures)" (Oyanedel et al., 2018). As such, this form of non-compliance behaviour could be prevented through opportunity-based approaches such as those found in situational crime prevention, more specifically, "removing the barrier."

| G U ID ING PRIN CIPLE S FOR APPLYING AC TOR-BA S ED AND OPP ORTUNIT Y-BA S ED APPROACHE S FOR ADVAN CING NON -COMPLIAN CE RE S E ARCH
Here, we propose three guiding principles to help bridge the gap between these two types of approaches to advance non-compliance research in small-scale fisheries and natural resource use contexts in general.

| Analyse which approach better suits what is being studied and the possible policy levers
One fundamental difference between actor-based and opportunitybased approaches is that the former puts the individual at the centre of the study whilst the latter does exactly the opposite (Clarke & Felson, 2004;Cornish & Clarke, 1987;Keane et al., 2008;. The socio-ecological setting of small-scale fisheries (and other natural resource use contexts) allows for a variety of actors, processes and circumstances to converge. As such, some of the vast arrays of research questions that can be framed around the problem of non-compliance could be better fitted to actor-based approaches and some to opportunity-based approaches. However, it is essential also to consider the applied consequences of the research questions;  (Petrossian & Pezzella, 2018). Insights from research on these topics can help inform policies that might reduce non-compliance without having to intervene with the actors involved, but instead, focus on the opportunities for non-compliance.

| Explicitly consider each approach's shortcomings and methodological challenges
Actor-based and opportunity-based approaches have different ways to study non-compliance, and each has limitations in how they could be applied in natural resource use contexts. On the one hand, opportunity-based approaches rely heavily on the ability to identify the products and discrete locations in time and space where non-compliance occurs (Brantingham et al., 2017;Brantingham & Brantingham, 1981, 1984Clarke & Webb, 1999). However, this can be challenging in natural resource management contexts, which poses an essential limitation for applying opportunity-based approaches (Gavin, Solomon, & Blank, 2010;Keane, Jones, & Milner-Gulland, 2011). Whilst enforcement and infringement records are sometimes available, they are not always good indicators of where and when non-compliance occurs (Critchlow et al., 2017;Keane et al., 2011;O'Kelly, Rowcliffe, Durant, & Milner-Gulland, 2018a). This is because enforcement is reactive and non-random in nature, therefore, data from this activity is inherently biased (Keane et al., 2008;O'Kelly, Rowcliffe, Durant, & Milner-Gulland, 2018b). A second source of bias arises because enforcement acts as a deterrent, subsequently changing resource user behaviours and further reducing the ability of enforcement records to detect true non-compliance trends (Keane et al., 2011). However, advances in encounter data analysis and modelling have proven useful to disentangle confounding factors and biases, leading to a better interpretation of infringement records (Critchlow et al., 2015;Underwood, Burn, & Milliken, 2013 (Underwood et al., 2013). Properly accounting for and dealing with these biases is critical for applying opportunity-based approaches in natural resource use contexts.
On the other side, one of the major limitations of actor-based approaches for studying non-compliance relates to the difficulty of approaching actors who are involved in non-compliant behaviours (Kuk, 1990;Oyanedel et al., 2018;Solomon, Gavin, & Gore, 2015).
Non-compliance is a sensitive topic, and it is to be expected that people involved in the activity will be reluctant to participate in research projects aimed at reducing its incidence. This poses a major challenge for applying actor-based approaches since for the Compliance Framework and the Theory of Planned Behaviour (and to a lesser extent the Instrumental Model), methods rely on surveys or questionnaires that require actor participation (Fairbrass et al., 2016;Oyanedel et al., 2020;Ramcilovic-Suominen & Epstein, 2012 (Bova et al., 2018).
Whilst none of these methods can assure full participation or sincere responses, they do increase responses rates and can provide more transparent assessments of non-compliant behaviour and its motivations. Further, by using these confidential methods, retaliation or negative consequences for research participants can be prevented if methods are appropriately applied and presented (e.g., not reporting port-level aggregate results that might cause fishers from a particular port to be targeted).

| Consider the appropriate timescales at which changes can be detected
The time scales at which research on actor-and opportunity-based approaches need to be conceptualized and performed differ. This is because the interventions proposed by actor-and opportunitybased approaches have different time horizons. As such, actor-and opportunity-based approaches can complement each other through the temporal scale of the interventions that are put in place to address non-compliance. By bringing together these two approaches, the underlying causes of non-compliance in fisheries can be tackled, whilst also providing shorter-term gains in compliance.
On the one hand, altering the underlying motivations that drive behaviour is a long-term effort (Clarke, 1980). As such, research aiming to understand trends in how actor-based approaches might affect non-compliance must incorporate into its design the time horizon at which some of these underlying motivations might start to change. For instance, the social norms approach (SNA) has been proposed as a way to increase compliance with recreational fisheries regulations (Bova et al., 2017). The SNA uses targeted advertizing campaigns to correct misconceptions of the proportion of people that engage with undesirable or non-compliant behaviours (Berkowitz, 2005). By doing so, it aims to change descriptive norms (e.g., perception of what others do) as a way to motivate compliance.
However, for this approach to be effective, at least half of the population should exhibit the appropriate (compliant) behaviour (Bova et al., 2017;Perkins & Berkowitz, 1986). As such, the time horizon needed for these advertizing campaigns to have the intended effect might be significant, because of the need to assess and bring compliance levels up to the point where the SNA can be applied.
On the other hand, from their inception, opportunity-based approaches have relied on short-term, trial and error assessment of interventions to prevent non-compliance Weisburd, 2018). As such, research guided by opportunity-based approaches can help to design interventions that can be implemented in short timeframes. Techniques from Situational Crime Prevention allow for empirically based analysis of potential changes in non-compliance that can be detected over short time periods. Petrossian and Marteache (2018) provides a good account of the type of interventions that can be informed by Situational Crime Prevention and its application in fisheries, its time frames and potential results.

| CON CLUS ION
Sustaining fisheries and other natural resources into the future requires the reduction of non-compliance. This is especially pressing in settings where the impacts of non-compliance are more acute because of poor management, lack of enforcement capacity and the high dependence of users on natural resources for employment and food. Addressing the non-compliant use of natural resources requires us to push research frontiers to provide frameworks and insights that translate into practical actions and plans. Understanding how the transition from theory to practice has been achieved in other disciplines dealing with non-compliant activities can make this easier to achieve.
Here, we have shown how integrating actor-based and opportunity-based approaches can trigger new ways to explore non-compliance in small-scale fisheries. Moreover, these principles and approaches are generalizable to other natural resources and contexts, such as the illegal wildlife trade. Illegal wildlife trade has similarities with small-scale fisheries in that they both operate in the intersection of social and ecological systems. The diversity of ways that natural resources are used and managed precludes simple solutions to curtailing non-compliant use ('t Sas-Rolfes, Challender, Hinsley, Veríssimo, & Milner-Gulland, 2019). However, acknowledging that non-compliance can be framed as the interaction of a motivated actor and an opportunity serves as a starting point for broader applicability of our approach to other contexts and settings.
As demonstrated here, building a better research agenda on non-compliance issues in small-scale fishing should include active engagement with experiences and approaches from other natural resource management settings. The theoretical underpinnings of actor-based and opportunity-based approaches, as well as their integrated application, are the same whether the social-ecological system is terrestrial or marine. As such, these approaches provide a bridge through which collaboration between researchers studying non-compliant use of natural resources in a range of settings can be promoted. The application of these approaches can provide cross-learning opportunities and better identification of knowledge gaps and biases. Thereby, it could unleash the potential of collaborative studies for advancing the theory and practice of non-compliance research in natural resource management contexts. Understanding the commonalities and specificities of contexts where non-compliance occurs could be a critical step towards better managing and maintaining the natural resources we depend on.

ACK N OWLED G EM ENTS
We acknowledge funding from ANID PIA/BASAL FB0002 and ANID-Becas Chile. We thank Millennium Nucleus Center for the Socioeconomic Impact of Environmental Policies (CESIEP) and the

Walton Family Foundation. SG and EJM-G were supported by Pew
Marine Fellowships.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data availability not applicable to this article as no datasets were generated or analysed during the current study.