Seagull Syndrome: Relationships between Patrol Workers and Government Officials in NSW, Australia

This paper explores the nature of relations between public officials and community workers, drawing on empirical data from a study on Indigenous patrols in New South Wales, Australia. Patrol workers interact with public officials from various state entities who are tasked with overseeing funding, carrying out evaluations and, to varying degrees, monitoring the ‘effectiveness’ of local patrol operations. These interactions illuminate several issues regarding the ways in which knowledge about patrols is created, contested and communicated between Indigenous and non‐Indigenous domains. The emergent patterns of these relations can be described as ‘seagull syndrome’, which involves the privileging of some types of knowledge over others in decision‐making regarding Indigenous affairs, often with disastrous consequences for Indigenous organisations and communities. The paper documents the core features of seagull syndrome with respect to the discrete practices, everyday decision‐making and mundane communication between public officials and patrol workers in New South Wales. It considers the implications of seagull syndrome for policy‐makers and academics working in the Indigenous justice space and suggests ways to resist or challenge this tendency.

Instead, that day, we just talked. We talked about Charlie Perkins, 4 about Brewarrina, 5 about how I grew up in a 'nearby' town called Warren (287 kilometres south of 'Bre') and that I used to play netball against the Brewarrina team. We talked about everything from Aboriginal identity to our appreciation of the $5.00 Countrylink 'Devonshire tea' deal and, last but not least, we talked about Uncle Bill's experiences on the Redfern Streetbeat in the 1990s. The 13-hour journey from Sydney to Bourke had never passed so quickly! When the train came to a stop and it was time to go our different ways, I went to collect a book from the pouch in front of me; it was a book on crime and local governance. As William turned to say goodbye he looked down at my book and said to me: 'well, you didn't have time to read that!'. I understood this comment as an acknowledgement of the lively conversation we had just had but it was also more than this. In my opinion it was also an acknowledgement of the different ways of knowing and the plurality of learning experiences. Indeed, that day I likely learnt as much about crime and local governance from our discussions as I would from a book on the same topic. Uncle Bill's comment made me think not only of a certain disconnection between the criminological literature and the everyday happenings in local Aboriginal communities, but also the 'baggage' outsiders -government officials, researchers, including myself, and others -carry with them when doing research in Aboriginal communities. IJCJ&SD 37 Online version via www.crimejusticejournal.com © 2017 6(1) These concerns go to the heart of questions that are addressed by this paper: What are the ways in which knowledge is produced about Indigenous peoples, communities and governance structures? Why are some stories or knowledges taken up more readily than others, or given preference over others? How might one overcome these biases and preferences in the production of knowledge about Indigeneity?
In the words of the late Aboriginal poet and activist, Lionel Fogarty (1980), the research 'industry' has been one of non-Indigenous gain, often out of non-Indigenous oppression. To some, these words might seem surprising and to others perhaps unfair, but it is now widely acknowledged that the research of Indigenous peoples and communities has been implicated historically with the 'worst excesses of colonialism' (Moreton-Robinson and Walter 2009;Nakata 2007;Smith 1999: 1). Whether we like it or not, researchers have a reputation in communities as being more motivated by the prompt collection of data for their own career-advancing purposes than in the research benefits to Indigenous peoples. Government officials similarly carry a reputation as being motivated more by the compliant ticking of boxes than in building relationships or genuine community consultation. Importantly both are seen as outsiders to the community; sometimes arrogant or self-interested but nearly always impatient and, ultimately, unaccountable.
In colloquial terms, these individuals are sometimes referred to as 'seagulls', a derogatory expression afforded to an individual who, in the words of research participants, 'flies into a community, craps over everything and everyone and then leaves the community to pick up the pieces'. 'Seagull syndrome' attests to the frustration felt by community members towards outsider 'experts' in making generalisations and false diagnoses based on what is sometimes only a superficial understanding of local community dynamics, often with disastrous consequences for the community. A related expression is that of 'remote control communities', 6 which refers in essence to the perception of Indigenous communities as being controlled by distant and nameless outsiders with limited knowledge of how the consequences of their decision-making and research play out at the local level.
This article explores how Aboriginal patrols provide a lens through which to examine the nature of everyday relations between state officials and members of the local Aboriginal community. Drawing primarily on interviews conducted with policy-makers and local patrol workers, I consider the ways in which information about patrols is managed and produced by individuals and entities, the interplay between different types of knowledge about patrols, and the ways in which it is transferred, disputed and contested between parties.
The ethnographic methods brought to bear here not only trace the contours of the relationships between patrol workers and government officials but also provide an opportunity for me to reflect on my own regrettable implicitness within these broader dynamics and processes of knowledge production. As will become apparent, it is not simply a case of some types of knowledge ('local' or 'expert') being more equal than others. I am also interested in examining the relationship between lay and desktop knowledge; and in exploring why some types of information might be seen as less credible, less trustworthy. Further, the different ways in which different types of knowledge are dismissed or taken up in Aboriginal policy-making and the reasons why is a focus of my research. Thus my overall intent in this article is to elucidate these power relations and to reflect on ways to challenge seagull syndrome in everyday contexts, both within academia and policy-making.
Firstly I expand on the methods used to investigate these concerns and then present the data and analysis within three sections. The first of these sections explains the key characteristics of two apparent types of knowledge that emerge in practice: local and desktop knowledge. The second section examines the interplay between these types of knowledge with reference to two discrete examples: the use of statistics within government departments; and evaluations and evaluative IJCJ&SD 38 Online version via www.crimejusticejournal.com © 2017 6(1) processes. These threads are then drawn together in the final section which offers a broad discussion of southern theory, neo-colonialism and lay/expert knowledge production.

Indigenous patrols: A qualitative study
This paper forms part of a larger empirical study which documents the everyday operation and politics of Aboriginal patrols in New South Wales (NSW), Australia. Aboriginal patrols are locally run initiatives with formal agendas that focus on keeping young people safe and preventing antagonistic contact between Aboriginal young people and the state police. Harry Blagg (2008: 107) estimates that there are approximately 130 such patrols currently operating in urban, rural and remote settings across Australia. A thick descriptive account of night patrol activities can be found in Porter (2016); however, here I outline the essential components.
The core ethos of patrol work includes, firstly, connection to the local Aboriginal community; secondly, independence of state police; and, finally, a consensual basis of operations (Porter 2016). Owing to the local specificity of patrol services, there is an enormous degree of diversity among the functions, objectives, composition and style of each unique patrol. Despite these variations, broad unity can be seen at the level of key functions, which in NSW included providing transport, maximising safety, mentoring, and preventing community and individual harms (Porter 2016).
Patrols exist and operate independently of -and not infrequently in some degree of conflict with -the state police. This does not mean that the relations between patrols and the state police are essentially acrimonious, although this is sometimes the case. Indeed, most patrols rely on the state police to intervene in instances of violent crime (domestic violence, assault) and many patrols have positive relationships with the state police. Indeed, in some cases senior police officers have been present on the management committees of local patrol initiatives.
Patrols also operate with some degree of independence from the Australian government. Specifically, patrols rely on a combination of one-off and ongoing grants from multiple sources. The main sources of funding for patrols in NSW are the NSW Department of Justice and Attorney General; the Commonwealth Attorney General; the Department of Community Services; the NSW Police Force; the NSW Department of Health; Aboriginal corporations; local government councils; and local businesses. The grant of money to individual patrols ranges from AU$19,000-$72,000 per annum and requires compliance with reporting mechanisms and other terms and conditions as stipulated in the contract. This includes inter alia the collation and sharing of statistics in order to monitor patrols' activities vis-à-vis certain performance indicators, as discussed below. As this suggests, a great deal of time is spent applying for competitive grants and fulfilling the administrative requirements of existing contracts. 7 Little is known about these Indigenous patrols. While there has been considerable attention given to Aboriginal criminality and to the impact of the criminal justice system on Indigenous peoples, there has been limited research on Indigenous governance initiatives or justice mechanisms. This omission has persisted within a framework that reflects American and European academic origins and concern and the urban-centric nature of the discipline's taken-for granted world-view (Carrington, Hogg and Sozzo 2016;Hogg and Carrington 2006). The Australian Law Reform Commission's report (1986: 103-105) Recognition of Aboriginal Customary Laws contains one of the earliest, albeit very brief, written references to Aboriginal self-policing initiatives and related activities. The first academic paper on the subject of patrols was written by Marcia Langton (1990Langton ( , 1992 who, in the early 1990s, saw their potential for providing an effective alternative to state intervention. Criminologists did not take up the topic until at least a decade later. This current study sought to rectify this neglect by documenting the everyday operation of patrols in NSW, Australia. A corresponding aim was to examine the mundane details and everyday relationships between patrol workers and other individuals -public officials, evaluators, funding bodies -IJCJ&SD 39 Online version via www.crimejusticejournal.com © 2017 6(1) involved with their everyday operation in order to demonstrate the variety and complexity of alternative governance structures and to understand how patrols were perceived by these various entities.
Methodologically, the study's fieldwork involved sitting in on patrol operations; conducting interviews with patrol workers, public officials and others (83 were recorded and some were not); and having informal discussions (including with individuals, with groups and, in some cases, with friendship-based contacts). The ethnographic study was supplemented by an archival search of documents, reports, photos and related materials. This stage involved accessing records in relation to past and present patrol initiatives by contacting federal, state and local government agencies holding relevant archival material. The purpose of this aspect of the research project was primarily to obtain information relating to the development and operation of patrols in Australia and to gain insights into governmental perceptions of these matters.
In the course of completing this research project, I observed and recorded the details of several patrols in New South Wales but have focused in particular on the activities of three patrols which will be used as case studies. The in-depth case study sites were Redfern, Bourke and Dubbo, which were selected in part because they represent three different environments: respectively, metropolitan, 'remote' Australia, and a regional city. These locations were used as case studies for this research project to reflect the diversity of patrols operating within Australia and to give some idea of the variety and differences that exist between urban and rural/regional and remote towns. In addition to observing patrol activities, I conducted interviews with patrol staff and with staff from various funding entities (both Indigenous and non-Indigenous). Although I spoke with some people on the basis of their affiliation with a government department, organisation or representative body, some research participants emphasised that they were expressing their own views and not necessarily those of the organisation.
While completing this project I learnt a great deal, not just about Indigenous patrols but also about young people's experiences with the state police, relations between patrol workers and government officials, some of the blind sights within criminology as a discipline and also about some of my own biases and assumptions. The following section focuses on the everyday relations and internal processes of government officials interviewed as part of this study. While these processes are not representative of the totality of funding arrangements between Aboriginal organisations and government entities, they are reflective of the kind of themes and issues that arise with respect to funding.

Knowledge production and preferences: 'Desktop' versus 'local' knowledge
Almost every person with whom I spoke -whether in remote Bourke, urban Sydney, regional Dubbo or elsewhere -described communities run by 'remote control'. A related colloquialism was that of the 'seagull'.
There was a sense of frustration that government officials -funding body representatives, evaluators and other government officials sent to a community to communicate with a patroldid not pay sufficient attention to and lacked awareness of local knowledge, cultural protocols and community dynamics. There was a perception that government entities were more concerned with 'political posturing' and the rhetoric of 'consultation' and 'engagement' than with what was happening on the ground. Moreover limited time spent in a community meant there was simply no opportunity for the exchange of in-depth information or for building meaningful working relationships: 'by the time you get the four-hour journey up and four-hour journey back, you know, there is no time for community'. It comes as no surprise then, that many community members were disenchanted with consultants, regardless of their good intentions.
IJCJ&SD 40 Online version via www.crimejusticejournal.com © 2017 6(1) The comments are illustrative of the divide between 'local' or 'lay' knowledge on the one hand and 'expert' or 'desktop' knowledge on the other. What becomes clear is that these are different types of knowledge, pertaining to different questions and relating to different levels of analyses. While it is fruitless to debate which knowledge is 'superior' or 'more objective', what emerges very clearly is that these knowledge's are divergent, and align with broader lines of cultural and political fracture. Importantly and unfortunately, however, not all types of knowledges are equal. In other words, some types of knowledge carry more weight than others. Notably, numerous decisions have been made on the basis of desktop knowledge, such as the decision to defund a patrol in light of one unfavourable evaluation. Desktop knowledge has the tendency to be seen as more abstract, more 'objective', more expert, more credible whereas local knowledge may be perceived as 'different', irrational, inferior, subjective, unreliable or based on ignorance.
Patrol workers and management staff possessed a particular type so knowledge pertaining to specific aspects of patrol operations. Drivers and patrol workers were on first name terms with the young people or 'regulars' of the local service. They knew where they lived, where their extended family lived, and how the client or regular fitted within a web of social networks, kinship and social relationships. They knew the young person's interests, their dreams and aspirations, their challenges. I call this knowledge 'local' or 'lay' knowledge.
Local knowledge was essential to the daily operation of Indigenous patrols. Patrol workers' knowledge about a client group (their families, home situation) and local community happenings (for example, community dynamics, information about who has just been released from prison or juvenile detention centres) informed a local patrol's operations. Patrol workers draw on local knowledge in a diverse array of contexts including locating an appropriate place to drop off a young person who was 'living rough'; conversing and building rapport with the client group; and judging the appropriateness of intervening by requesting a referral to a third party. In a patrol worker's words: A lot of problems … were from home and the set up and the dynamics of the house and how it was running and stuff like that. I think, because they knew us, because they were related to us. Because [they were from] similar backgrounds, a majority of the children were Aboriginal, a majority of the CAP [Bourke Community Assistance Patrol] patrollers were Aboriginal workers, um, I think that that was the major thing. A lot of the reasons too, we were related by blood in a lot of situations so, there was that level of respect too, for the family. (Patrol worker #8, Bourke Community Assistance Patrol) Public officials, on the other hand, have knowledge of a different kind. In particular, they have knowledge about reporting procedures, performance indicators, funding, of how things appear on a grid of evaluation techniques and are geared towards implementing programs that produced a measurable outcome. Public officials possessed knowledge pertaining to broad-picture analysis of patrol operations at a state level of analysis (often involving drawing comparisons between towns). I call this knowledge 'desktop' or 'expert' knowledge.
Desktop knowledge was informed by the public policy literature, evaluations, monitoring reports, statistics, email and phone communication with patrol management staff and occasionally fieldwork, which involved either the public official visiting the township or inviting patrol workers to an annual event held by the funding body. Government officials' thoughts also reflected a framework of economic rationality; there is a need to justify government funds that are not being used to their maximum benefit.
Government knowledge was, in a sense, a different type of knowledge, oriented to different immediate problems, using different methods and skills and different forms of understanding (for example, statistical and bureaucratic), and concerned with a more generalised picture. This was Government knowledge's emphasis on statistics and the broader results could easily be misconstrued as suggesting that government knowledge was more encompassing and thus, in a sense, more objective or abstract. But elevating such knowledge to the status of 'more encompassing' is significantly misleading: as government officials themselves acknowledged, their generalised knowledge of the local situations was superficial, incomplete and even flawed.

Government departments and the everyday
Within government departments, everyday decisions about funding are made on the basis of information provided through these monitoring reports, evaluations and, on occasion, visits to community. The principal mode of communication transfer between government departments and patrol management staff occurs via the monthly written 'monitoring reports', based on statistics relating to the patrols' 'performance targets'. Field visits by government officials occur infrequently, due primarily to budget restraints, lack of personnel and the regional locations in which many patrols are located.
'Monitoring reports' typically require the patrol to collate statistics on everyday interactions with their client group such as how many people use the service, the age, gender and Aboriginality of the client group and, for some patrols, the geographical locations where the client group were picked up from and transported to. 'Performance targets' are monthly targets designed to supply the funding body with information about each patrol's local operation and its impact on the ground. For example, the current Dubbo Safe Aboriginal Youth Patrol is expected to meet a monthly 'target' of 300 interactions with local young people whereas a smaller town like Bourke is expected to meet a monthly 'target' of 150 interactions.
Failure to comply with reporting requirements or to fall short of 'performance targets' as stipulated in the contract may result in a decision to defund the patrol. The other main situation in which patrols could be defunded is based on findings brought to light by evaluative reports. For example, following negative findings reported in a 2005 evaluation, a decision was made by the NSW Department of Justice and Attorney General (NSWJAG) to defund the Brewarrina Patrol and the Nambucca Heads Night Patrol (Morgan, Sanber and Woolford 2005). Generally speaking, government departments showed a strong reliance on statistics in its everyday decision-making. The selection process of the NSWJAG, for example, used statistics about reported crime rates in various communities, provided by the NSW Bureau of Statistics and Crime Research (BOCSAR), the overarching logic being that the towns that most needed a Safe Aboriginal Youth Patrol Program were those with highest reported crime rates. The Department also took into consideration whether a patrol was already in operation in the given town. 8 The departmental processes of monitoring and collation of statistical data illustrate both the modes and substance of information transfer between Aboriginal communities and government departments. Having briefly outlined these processes, I will now critically analyse examples, comparing and contrasting the perspectives of patrol workers and government bureaucrats.

A The collection and interpretation of statistical data
Interviews with patrol workers and community members revealed a sense of disjuncture between the government values and reporting requirements, and those of the Aboriginal communities. The reliance on statistical data to measure the patrol functioning and assess the usefulness of patrols -such as police statistics, crime statistics, and data from patrols -highlights a mismatch between government and community understandings and expectations. From the viewpoint of the patrol workers, one of the problems with reliance on statistics, especially in the absence of in-depth qualitative data, is the potential for growth of distance between the policy boardroom and the ground level. The following quotes are fairly representative of this sentiment: They were just collecting stats and numbers, they were not getting down to the real issues, they were picking the same kids up. Instead of just it being a number, thinking why are we picking the same kids up? Why is the kid there, his mum is out fighting all the time or he is being kicked out, I mean we've gotta know that. (Patrol worker #7, Dubbo night patrol, emphasis added) Social welfare work is really hard to measure. Governments continue in their social welfare infrastructure … and they do a systems approach. So we're gonna have this particular project and it's gonna transport this many people. Social welfare it takes longer, sometimes a generation … They want the outcomes too quick and what I find with government is they don't reinvest in projects. They're like the one hit wonders of the world. 'Oh yeah we haven't done that for a while, how about we do this, oh no we did this last year'. (Patrol worker #4, Dubbo Gordon Centre Patrol, emphasis added) Patrol workers lamented the disjuncture between their own perceptions and those tasked with monitoring such services. For patrol workers, the value of their services related to specific effects or achievements that are not represented or made visible by looking at statistics or quantifiable 'overall' effects. For example, one female patrol worker commented that: The community don't care if we picked up 12 females between this age and that age. They care if we stopped their daughter or niece from a DV [domestic violence] situation, we stopped their young ones vandalising, or we prevented their kids from being locked up because they breached curfew. So the measurables [sic] for Aboriginal people as such, is very different, whereas like the funding agencies wanted stats … They need to know how many adults between this age there were, how many children, they wanna know how many … how many females, how many between the ages of 25 and 30, how many between 15 -you know that sort of stuff. And well, I don't understand how that makes it beneficial, because yes OK you've got the numbers so we've transported like, say, 300 people that week. Out of those 300 people, 20 of those were on curfews which avoided being locked up because they got home, 10 of them didn't get locked up for re-offending, 5 of those were the worst DV [domestic violence] victims I've ever seen in that state of violence. Like that's the stuff that's important. … that to me was the most powerful thing because that translated how it worked for the community. … And you know 'cause I'd say to my guys and they knew the community and that's the most powerful thing. If you brought someone from outside of the community to drive the bus, you may have had the same success. The fact that they were employed from within their own community, they knew everyone. Everyone. So I could say to my workers the next day, what was last night like? And they could tell me. And then I would know because of our relationship with them at the centre and our partnerships with DOCS [NSW Department of Community Services] and Juvy [NSW Department of IJCJ&SD 43 Online version via www.crimejusticejournal.com © 2017 6(1) Juvenile Justice] and stuff, who was involved with who. (Patrol worker #8, Dubbo Gordon Centre Patrol, emphasis added) Such stories were typical across all three locations. There were many other examples of issues, highly significant in the eyes of patrol workers, which monitoring reports and statistics tended to overlook. For example, in the community of Bourke, where employment opportunities are extremely limited and long-term unemployment was identified as a major problem, increasing the self-esteem of, and the creation of meaningful job opportunities for, the long-term unemployed was seen as particularly important.
The result was a strange disjunction. For public officials, there was a degree of bafflement and disbelief as to why local information and anecdotes of local issues did not make their way into performance reports or other written communications. For patrol workers, there was a strong perception that the funding body was interested only in overall effects and arbitrary indicators. This might be explained in part by the language of the contractual agreements and monitoring report forms, which are heavily charged in departmental terminology (such as 'performance indicators' and 'performance targets') with an emphasis on statistics and measurable results. In other words, it may be that a key problem is not one of interpersonal relations between workers and bureaucrats. Rather, and more problematically, it appears as an issue deeply embedded in the (state) governmental assumptions about what it is to be 'effective', how this can be demonstrated, and to whom accountability is owed. In this way, the problems confronting patrol workers go to the heart of the question of self-determination. From this point of view, they perceive themselves to be reliant on contractual arrangements based in 'alien' and imposed assumptions about good governance.
During interviews I was surprised by the candour with which many public officials would comment on their feelings of anxiety over what were often described as 'gaps in knowledge'. Others described their remoteness from what was happening on the ground and questioned the veracity of the statistical information they had before them, which, they emphasised, was all they had to go by. One public servant described feeling 'haunted' by not knowing how events transpired in the communities: I don't know how well it really worked to tell you the truth. And that is the truth, we don't even really know. You got little indicators here and there, little indicators in different places, but I don't know that any evaluation has ever captured that.
(Public official #4) Another lamented that, while they were indeed interested in hearing personalised stories of localised effects, such stories rarely made their way into the monthly written reports: And also it would be amazing when you go out on the field and they would incidentally tell you things that the patrol had done and you sort of say [theatrically]: 'why don't you put that in your report!' Because you're struggling to keep these things up, and I know you can't work on instinct but your frontline experience on the ground says when things work they work really well. (Public official #7) I now provide a more detailed critique of the process of evaluation. My intention in this section is not to provide alternative ways to evaluate patrols but rather to highlight how, in the process of analysing what works and what doesn't, evaluators overlook important subtleties in the culturally distinct assumptions behind the operation of patrols. In particular, I question from whose perspective should one evaluate such initiatives? Which criteria, methodology and forms of calculation should be adopted? Given that the process of evaluation is inescapably and necessarily a value-laden process, my interest here is to discuss what these evaluations have to IJCJ&SD 44 Online version via www.crimejusticejournal.com © 2017 6(1) say about the evaluator and the various ways in which patrols are understood both by evaluating bodies and by patrol workers themselves.

B
The evaluation process The question I have been asked most frequently during this study has been: 'do they work?' This question and its relative, 'are they effective?', represent a defining theme within the policy literature on Indigenous patrols (Australian National Audit Office 2011; Beacroft et al. 2012;Blagg 2003;Blanchard and Lui-Chivitze 2000;Close the Gap Clearing House 2013;Cooper et al. 2014;Higgins 1997;Morgan, Sanber and Woolford 2003;RPR Consulting 2007;Russell 1999;Walker and Forrester 2002). The need to evaluate such initiatives arises from a due concern on the part of government entities to ensure the efficient allocation of finite resources. Within a governmental framework, undertaking research of an evaluative nature is a necessary exercise in ensuring efficiency and accountability of government expenditure, and I do not dispute the validity of this approach. My concern, however, is that such normative questions assume that a value-free, culturally neutral and objective assessment is indeed possible.
The evaluation process bears considerable importance in determining the fate of many patrols. Evaluative reports are undertaken periodically in order to gather more in-depth information about the operation of patrols. Reports are typically based on a combination of statistics provided to government departments and fieldtrips of one and two weeks' duration. Few evaluators ever go into the field for long enough or are sufficiently embedded to gather these diverse perspectives and their implications for a meaningful evaluation. In addition to problems with the time and depth of analysis, existing evaluations are marked by a tendency to neglect the perspectives of patrol workers and service users, a consideration that is crucial in seeking to offer a more culturally inclusive or sensitive evaluation of patrols. Moreover, many of the evaluations conducted so far have tended to focus on crime prevention, whereas most patrol workers saw their contribution in terms of providing a wider and more encompassing community service, above and beyond western conceptions of 'crime prevention' and 'law enforcement' efforts.
Relatedly, evaluations typically measure patrol functioning with reference to crime rates for specific offences (namely, public order offences and reported incidences of domestic violence). One shortcoming with focusing on crime rates is that it provides narrow parameters through which to analyse the potential impact of patrols. The 'reach' or impact of patrols is intended to extend well beyond the scope of crime prevention. The focus on crime rates similarly overlooks the fact that many patrols have been set up precisely to address the fact that certain crimes were previously ignored or neglected by the state police. 9 Furthermore, it is important to note that crime rates themselves are not objective representations of actual crime; rather, they reflect processing of actions by the state police and others (Reiner 2013).
The 'success stories' of patrol workers revolved around localised effects of the patrol. So local community members may rate patrol performance highly if it manages to bring one or two really troublesome kids back into line, or bring a family into community engagement. In this respect patrol workers' specialist knowledge of local issues has nearly always been ignored. In light of this, patrol workers spoke of the need to develop subtle and grounded perspectives to measure local indicators which focus on relationships and individuals rather than purely quantitative indicators.
A patrol may, for example, increase the willingness of young people or community members to report offences, and influence state police decisions on how to deal with problems brought forward by community members. Research participants emphasised the need for a thorough and detailed investigation according to a broad range of criteria: IJCJ&SD 45 Online version via www.crimejusticejournal.com © 2017 6(1) It would be essential to conduct a 360 degree assessment where you're obviously coming from all angles -members that use the services, members whose children use the service, people who don't use the service but who see changes in town or from the perspective of the funding bodies, their perspective of return for investment, [but also] the police, community services, local councils, crime prevention offices these days -all these type of stakeholders. (Police officer #5, Central Darling Local Area Command) I think that it had an influence on people's lives at the time of engagement. Um, so, at any one time a child may have been picked up which may have otherwise been in mischief or in trouble at home, and that was either delayed or taken out altogether. That risk was eliminated potentially. And that while I don't think any of our reports would have ever had any hard evidence of it, I'm sure that that's the case. And I'm not sure … I'm not sure though that it led to any tangible social and or economic benefit that you could measure, which is the crux of the problem for a statistician. (Public official #3, Bourke Shire Council) A second and related problem with current evaluation is that it may obscure effects of the patrols which may be seen by the community and patrol members as providing a more encompassing social and cultural service. When patrol workers were asked how they perceived their work, very few saw their role in terms of preventing crime, but rather tended to perceive their function more in terms of mentoring, caretaking and intervening in relationships. 10 It was stressed that narrowly focused evaluation might overlook other factors important in a patrol's operation, such as the qualities of individuals involved in the operation of patrols: The thing is you see it with all sorts of programs and all sorts of sectors, and you evaluate all these programs and you say this is why it works and this is why doesn't, but so often it comes down to the drive of individuals -the drive and skill and compassion of individuals and it's not something you can kind of write policy for. (Patrol management staff #3, Redfern Streetbeat; emphasis added) Most patrol workers saw their contribution in terms of providing a wider community service above and beyond western conceptions of 'crime prevention' and 'law enforcement' efforts. In fact, when I interviewed the authors of the evaluations as part of this study, the researchers commented on a certain frustration they experienced in carrying out their research. Specifically, they felt the focus on crime prevention was 'myopic' and neglected many important themes related to patrol functions. Yet what emerged from the views of workers as the most striking shortcoming of evaluations, and one which has been common to the evaluations discussed above, 11 has been the lack of consultation with patrol workers in developing criteria to assess the effectiveness of local patrols. Few evaluations ever go into the field for long enough or are sufficiently embedded to gather these diverse perspectives and their implications for evaluation.
In addition to problems with the time and depth of analysis, existing evaluations are marked by a tendency to neglect the perspective of patrollers and service users, a consideration that is crucial in seeking to offer a more culturally inclusive or sensitive evaluation of patrols. In light of these perceived shortcomings, some patrols have sought independent evaluations to focus specifically on issues of concern to patrol workers and administrative staff (Ireland 2010;RPR Consulting 2007;Russell 1999). Putting aside the obvious problems with self-assessments of effectiveness (including inter alia the potential for bias, the compromising of subjectivity, and proximity to the subject matter), self-evaluation uncovers some important issues that are often masked or ignored by departmental evaluations. 12 Self-evaluation allowed for the inclusion of tailor-made criteria and consideration of the discrete local effects of the service.

Conclusion: Overcoming seagull syndrome and the role of local knowledge
This paper described how community organisations and government entities produce different types of knowledge on Aboriginal affairs and issues (local or lay knowledge versus expert or desktop knowledge), including how knowledge is transferred and contested between the two entities in everyday relations. Public officials are burdened with proving their programs are 'working' and are accountable in terms of government expenditure, yet they often lack essential knowledge about how events transpire within a given community. Public officials' perceived over-reliance on statistics and other seemingly arbitrary objective 'performance indicators' further distances policy-makers from communities and patrol workers. Throughout this process, local patrols remain a creature of contract which must conform to reporting requirements and with the threat of their funding being terminated at any time. 'Expert' knowledge based on BOCSAR statistics, desk-generated research, overall effects and other seemingly objective indicators, tends to be privileged over local forms of knowledge about everyday patrol operations. Importantly, this paper has described how, perhaps due to the appeal of statistics and highly generalised theory, 'desktop knowledge' is viewed as being more objective and is regularly given preference over other types of knowledge. The challenge is not only to explore and understand the ways in which certain types of knowledge are given preference over others but also to reflect on the ways the 'colonial' effects of knowledge control are perpetuated through government entities, the mainstream media and academic disciplinary knowledge.
There is a need to rethink how we value (or prejudice) certain types of knowledge. This research highlights the need for government departments to reconsider the processes and methods by which they gather information about community initiatives. In addition, there is a need to rethink the processes used to gather information and conduct relations with local community organisations so that it reflects relationship-building with key contacts rather than merely data extraction. Certainly both sets of parties were acutely aware of the nature of these issues, again often in somewhat different ways. As was repeated during interviews with officials, statistics mean very little without a correlating knowledge about community dynamics, personalities, relationships, and knowledge of customary everyday practices. As this paper argues, the seagull syndrome is part of more systemic problems related to the management of programs within and the production of knowledge about local Aboriginal communities.
In light of these issues, a number of suggestions to resist or challenge Seagull Syndrome in everyday situations are offered.
First, policy-makers and researchers must reflect critically on the limitations of sources whether in field or desktop analysis. Academics and policy-makers working in the Indigenous justice space have a duty to actively seek out a variety of local Indigenous perspectives, and to incorporate these within the design and implementation of Indigenous programs and research. And yet it is not enough to simply 'try one's best not be a seagull'. Rather, a radical rethink with regards to Indigenous research and policy-making is also required, not just in terms of how policies are conceived and funding implemented, but more generally in terms of what outcomes are to be IJCJ&SD 47 Online version via www.crimejusticejournal.com © 2017 6(1) measured, and to whom accountability in Aboriginal policy-making and affairs is ultimately owed.
We need to think about collaborative approaches to policy-making and implementation.
By analogy, the same arguments apply with equal force to research in the Indigenous justice space. Academics in this space should pay particular attention to the cultural baggage that disciplines carries, including the terms we use, the questions we ask, and the way we conduct research. The development of the emerging literature on Indigenous methodologies is timely in this regard (Cunneen and Tauri 2016;Moreton-Robinson and Walter 2009;Nakata 2007;Smith 1999;Tauri 2012Tauri , 2014Tauri and Porou 2014;Behrendt, Porter and Vivian (forthcoming); Walter 2016). Indigenous methodologies prioritises the importance of building relationships and a partnership approach to research and policy-making. In practical terms, this requires the privileging of Indigenous voices and localised Indigenous knowledge in the design, development and implementation of both research and policy on Indigenous affairs. Importantly, partnerships must be based on mutual respect and a humble acknowledgment of the limitations of each party's knowledge base and the socially constructed nature of 'expertise'. This necessarily involves the adoption of a more flexible approach to both research and policing-making than has conventionally been the case.
Finally, there needs to be acknowledgement of the role of local Indigenous knowledge in the design, development and implementation of programs that are more culturally appropriate and research that is more beneficial to community interests. This requires a better appreciation of the defining features of local knowledge and how this differs from other sources of knowledge. As with the examples offered above, this requires a sense of humility or awareness of both the limitations of our knowledge and the socially constructed nature of 'expertise'. on an Aboriginal girl from the local public school who excelled in the Higher School Certificate. The producer's instead decided to air a fairly sensationalist story on child sex abuse, which did untold damage to the town. 7 As this implies, patrols do not fit squarely within the rubric of autonomous or governmental spaces, but occupy what scholars (Cunneen 2001;Blagg 2008) have termed 'hybrid' or 'third spaces', somewhere between the two. Indigenous night patrols reflect a varied spectrum of control and ownership. The Brewarrina Granny Patrol, for example, was run on a voluntary basis by Elders and senior Indigenous women, with the provision of in-kind support from the local council. In contrast, the NSW Department of Justice and Attorney General runs a government program known as 'Safe Aboriginal Youth Patrols' which currently operate in 11 Aboriginal communities in NSW. These services run on a fairly modest budget ranging from AU$54,000-$75,000 per annum and are contractually bound to meet Key Performance Indicators and other reporting requirements, with a fairly limited scope for communities in negotiating the terms and conditions of the patrol service. Elsewhere, the Redfern Streetbeat exemplifies a complex kind of hybrid form of governance-supported by a number of public and private grants and managed by a team of Indigenous and non-Indigenous staff members (including paid staff and volunteers). 8 Unfortunately, this logic runs counter-grain to the principle of community empowerment, a key theme emerging from the interviews with patrol workers and community members. Research participants stressed that it was important for the impetus for the patrol to come from the community rather than from outside influences: I really think that the Aboriginal community needs to want it first and foremost, and I think they need to have all the same initial design and function and role and responsibilities and that sort of thing and after that's done, then you get government in to see where they fit around it. (Suzie Forell, Redfern Streetbeat) 9 This is in addition to well-documented problems in the 'under-policing' of domestic and family violence in relation to Aboriginal communities (Cunneen 1992) and the under-reporting of sexual assault more generally (Wild and Anderson 2007). 10 Interestingly, this is also the case for the state police. Much research has shown that comparatively little policing is simply crime control in practice (Cumming et al. 1964). 11 Note that both of the two most recent reports commissioned by the NSW Department of Justice and Attorney-General involve consultation in relation to opinions on the effectiveness of the patrol in reducing crime, though not in the development of criteria by which to assess the effectiveness of the patrol. 12 It could be argued that this has similarly been an oversight within evaluations and research about the conventional police force. In the United Kingdom in the 1990s the Audit Commission began assessing community satisfaction with the police. The reports produced had major implications for the police.