LGBTI Variations in Crime Reporting

Research shows that people vary in their willingness to report crime to police depending on the type of crime experienced, their gender, age, and their race or ethnicity. Whether or not lesbian, gay, bisexual, transgender, and intersex (LGBTI) and heterosexual people vary in their willingness to report crime to the police is not well understood in the extant literature. In this article, I examine variations in LGBTI respondents’ attitudes, subjective norms, and perceived behavioral control on their intentions to report crimes to the police. Drawing on a survey of LGBTI individuals sampled from a Gay Pride community event and online LGBTI community forums (N = 329), I use quantitative statistical methods to examine whether LGBTI people’s beliefs in police homophobia are also directly associated with the behavioral intention to report crime. Overall, the results indicate that LGBTI and heterosexual people differ significantly in their intention to report crime to the police, and that a belief in police homophobia strongly influences LGBTI people’s intention to underreport crime to the police.


Introduction
Variations in crime reporting behaviors have consistently demonstrated that although people usually hold favorable views of the police and are willing to report crime to the police (Mastrofski, Parks, Reiss, & Worden, 1999), members of minority communities 1 are far more reticent to report crime (Webb & Marshall, 1995). Research focused on elements that effect police reporting has a long history in victimization literature. The issue of reporting crime to the police and its variation by age, gender, race/ethnicity, and type of crime has been widely researched (see Bosick, Rennison, Gover, & Dodge, 2012). Yet crime reporting variations due to differences in sexual identity is an area that has been underresearched, thereby resulting in a lack of complete understanding regarding how sexual identity impacts an individual's willingness to report crime to police.
One such theory that has been used to better understand the variation in willingness to engage with police and to report crime is the Theory of Planned Behavior (TPB; Ajzen, 1985). The TPB was initially conceptualized to link attitudes and beliefs to intention and behavior. The TPB followed on from the Theory of Reasoned Action (TRA; Fishbein & Ajzen, 1975) used to measure the predictive power of positive evaluation (attitude), and subjective norms or how significant others shape an individual's intention to perform behavior, and the motivation or intention of an individual to engage in a particular behavior. As crime reporting behavior is often reflective of attitudes toward the police as well as how easy or difficult a person perceives the behavior to be, the TPB can be a useful tool in understanding variations in crime reporting behaviors.
The theoretical components of the TPB (attitude, subjective norms, and perceived behavioral control [PBC]) can provide a useful framework for describing the psychological influences on crime reporting because previous studies have shown that the TPB is able to account for significant amounts of variance in intention to act in a particular way (Buchan, 2005;McMillan & Conner, 2003). By applying the TPB to measure willingness to report crime within the lesbian, gay, bisexual, transgender, and intersex (LGBTI) and wider community, questions such as "who is more willing to report a crime and why" and "does sexual identity make a difference in reporting crime" are among the many questions that can be examined by applying this framework (Connor & Armitage, 1998). However, an application of the TPB to understand crime reporting behaviors between LGBTI and 490707S GOXXX10.1177 1 The University of Queensland, Brisbane, Australia heterosexual people has not been conducted within Australia until now.
Research examining the relationship between variations in crime reporting behavior and sexual identity has largely been ignored in the Australian context. This is particularly troublesome because research indicates that members of the LGBTI community have far lower rates of reporting crime to the police than the general population (see Bernstein & Kostelac, 2002;Gerstenfeld, 2004;Herek, Cogan, & Gillis, 2003;M. Williams & Robinson, 2004). Hence, the present research aimed to determine whether the LGBTI community in Queensland underreports crime to the police and why.
Using data collected from a convenience sample of participants at the "Gay Day" Celebrations in Brisbane, Australia, and by online delivery (N = 329), the present study examines reasons why LGBTI people's beliefs in police homophobia are directly associated with the behavioral intention to report crime to the police. I begin this article with a review of the extant literature. I then describe the research method, sample, and analytic approach. Finally, I present the findings from the research, which indicates that heterosexual participants had significantly more positive attitudes toward the police than LGBTI participants; experienced stronger-positive social pressures to report crime to the police; and found the practice of reporting crime to the police to be an easier experience than LGBTI participants.

Research Questions
Specifically, the aim of the research was to address three research questions: Research Question 1: Are LGBTI participants less willing than heterosexual participants to report crime to the police? Research Question 2: Do participants' attitudes, subjective norms, PBC, and belief in police homophobia determine their intentions to report crime to the police? Research Question 3: Are there differences between LGBTI and heterosexual participants' attitudes, subjective norms, PBC, and belief in police homophobia?

Background Literature
Research has consistently shown that typically most people have positive opinions about the police (Merry, Power, McManus, & Alison, 2012). As such, the majority of people have favorable opinions regarding engagement with police when the need arises, for instance, when reporting a crime (Mastrofski et al., 1999). However, the willingness of residents to report crime varies depending on the type and severity of the crime (Kääriäinen & Sirén, 2011) and contextual factors (such as culture 2 ) that may influence crime reporting (Schaible & Hughes, 2012). In addition, it has also been found that regardless of the type and severity of the crime, members of minority groups (typically represented in previous research by racial or ethnic identifiers) are hesitant to report crime to the police due to negative perceptions of police interaction, particularly negative perceptions of police interaction that may result in further victimization (Beckett, Nyrop, & Pfingst, 2006;Browning, Cullen, Cao, Kopache, & Stevenson, 1994;Fagan & Davies, 2000;Kane, 2002Kane, , 2005Mastrofski, Reisig, & McCluskey, 2002;B. W. Smith & Holmes, 2003;D. A. Smith & Klein, 1984;D. A. Smith & Visher, 1981;Terrill, Paoline, & Manning, 2003;Terrill & Reisig, 2003;Weitzer & Tuch, 2006;Worden, 1996). Accordingly, members of minority groups (whose subordinate group status is defined due to external or other identifying features) differ significantly from other members of society in their willingness to interact with the police, regardless of whether the grounds for contact with police are positive or negative and or whether the outcome of police interaction may result in a constructive end to an adverse situation (see Webb & Marshall, 1995). Certainly, this has been the case for members of minority groups (also identified in this way) residing in Australia (Murphy & Cherney, 2010).
In 2007, the Australian Bureau of Statistics revealed that criminal activity in Australia is frequently unrecorded simply because it is not reported to the police. Australian research measuring variations in crime reporting behaviors have generally concentrated on ethnic and indigenous minority groups (Murphy & Cherney, 2010). However, research examining other minority groups (based on identifiers other than race or ethnicity) and their attitudes toward crime reporting in Australia have largely gone unnoticed. This is problematic because recent research suggests that members of marginalized minority groups whose identifiers are not based on race or ethnicity but other subjective factors are less likely than other members of society to call the police for help (see Carr, Napolitano, & Keating, 2007) and that most members of mainstream society have more reliance on the police (and therefore more positive expectations of police) than these types of minority group members when deliberating over whether to call the police in times of need.
Previous research has indicated that the decision to report or not report crime is typically the outcome of a complex decision-making process in which the victim will weigh the costs and benefits of each course of action (Tarling & Morris, 2010). It is also understood that a victim of crime may struggle with the conscience duty to report crime, and that such struggles may be linked to personal reasons such as the need of immediate help, protection and treatment (in relation to violent or sexual crime), or to obtain monetary redress in the form of compensation or insurance payments (in relation to property crime). 3 As such, the importance attached to the costs and benefits of crime reporting vary according to personal characteristics and experiences of the victim. For example, previous research indicates that women have been found to be more likely to report crime than men (see Baumer, 2002;Carcach, 1997;Felson, Messner, Hoskin, & Deane, 2002;Hart & Rennison, 2003;Rennison, 2007) because men tend to be bigger and stronger than women, more skillful in physical combat, and more willing to use violence (Felson et al., 2002). As a result, female victims of crime may be more likely to call the police for protection than male victims of crime (Felson et al., 2002) because the nature and circumstances surrounding an offence will also feature prominently in the decision to report crime to the police (Tarling & Morris, 2010). For example, it would be anticipated that a serious crime such as physical assault would be reported to the police by males and females. Yet there is no consistent evidence regarding crime reporting rates for males or females based on their differences in sexuality (e.g., heterosexual people in comparison with LGBTI people) or the factors influencing a LGBTI person to report or not report crime (such as a belief in police homophobia), particularly when LGBTI people are often the victims of serious and minor crime (Meyer, 2010(Meyer, , 2011. Unrecorded crime by the LGBTI community has several consequences: It contributes to the misallocation of police resources (thereby minimizing resources allocated to help and protect the LGBTI community), it prevents LGBTI victims from accessing public and private benefits, affects insurance costs, and does not help shape the police role in the LGBTI community (Tomsen & Mason, 2001). These factors impact community crime prevention and control strategies and decisions about the allocation of police resources (Australian Bureau of Statistics, 2007). Nonreporting of crime by the LGBTI community also limits the deterrent capacity of the criminal justice system, hinders the formation of an accurate picture of anti-LGBTI crime (thereby underestimating the extent of victimization), stands in the way of perpetrator convictions of anti-LGBTI crime, and affects the police mandate of fighting anti-LGBTI crime (Bohn, 1993;Herek & Berrill, 1992;Stonewall, 1994). Ajzen (2005) stated that by applying the TPB to examine a particular behavior, the proximal determinant of behavior "intention to engage in the behaviour" becomes the key concept of the research and is determined by three sets of variables: (a) attitude (the overall evaluation and the outcome expectancy of the behavior), (b) subjective norms (perceptions of social pressure from significant others to perform a particular behavior), and (c) PBC (a person's belief as to how easy or difficult performance of the behavior is likely to be). Under the theoretical framework of the TPB, to arrive at an overall attitude, it is important to distinguish between attitudes and beliefs because both mechanisms affect intention to behave in a particular way (Ajzen, 2005). A person's belief toward a particular object is the sum of all of that person's beliefs toward each attribute associated with an object (Ajzen, 2005). For instance, an LGBTI person's belief that police officers are homophobic links the object of opinion police to the consequence belief homophobia. Subsequently, beliefs in police homophobia could have an incremental influence on a lesbian, gay, bisexual, transgender, and intersex 4 (LGBTI) 5 individual's perceptions of the police over and above their general attitude toward the police and reporting crime.
If the psychological mechanisms that influence LGBTI victims to report or not report crime can be identified, then there is the potential to develop interventions that, by influencing those mechanisms, may lead to changes in crime reporting behavior (Viki, Culmer, Eller, & Abrams, 2006). Accordingly, the theoretical components of the TPB (attitude, subjective norms, and PBC) can provide a useful framework for describing the psychological influences on crime reporting because previous studies have shown that the TPB is able to account for significant amounts of variance in intention to act in a particular way (Buchan, 2005;McMillan & Conner, 2003). By applying the TPB to measure willingness to report crime within the LGBTI and wider community, questions such as "who is more willing to report a crime and why" and "does sexual identity make a difference in reporting crime" are among the many questions that can be examined by applying this framework (Connor & Armitage, 1998). However, an application of the TPB to understand crime reporting behaviors between LGBTI and heterosexual people has not been conducted within Australia until now.

LGBTI People's Attitudes to Crime Reporting
In a recent Australian survey by Leonard, Mitchell, Patel, and Fox (2008), the major barrier to LGBTI respondents reporting crime or seeking assistance from the police is the belief that the majority of police officers are homophobic. They also found that LGBTI people in Australia perceived that reporting crime to police will lead to further abuse from service providers, and that the majority of LGBTI respondents strongly believed that police officers would not treat LGBTI people fairly due to homophobic beliefs. They also found that almost all of the Australian LGBTI participants, who provided written responses to questions asking about the barriers preventing them from reporting crime, wrote about targeting the homophobic beliefs of mainstream police officers. However, specific data relating to underreporting of crime by the LGBTI community throughout Australia (and specifically in Queensland) are not readily available.
In Australia, the actual population size of the LGBTI community is unknown. 6 Yet the study of LGBTI crime reporting behavior has meaning, particularly because the Attorney General's Department of New South Wales (2003) found that the majority of LGBTI respondents who participated in their survey strongly believed that the police will not take LGBTI violence and harassment seriously, and indicated that they would be unwilling to report crime to the police. Yet the first contact that many victims of crime have with the criminal justice system is with the police and the decision to report or not report crime may be the most influential decision an individual makes in the criminal justice system, thereby emphasizing the role of the citizen as the gatekeeper for all that follows (Gottfredson & Gottfredson, 1980). In two separate studies, Tyler (2005) and B. Williams (1998) found that the less confidence citizens have in the police, the less willing they will be to cooperate with police, which includes reporting crime. This has been a particular concern with research in the LGBTI community because most LGBTI people underreport crime fearing "hostility and abuse" from the police (Comstock, 1989, p. 104). Moreover, in a recent study in the United Kingdom, it was determined that 78% of LGBTI respondents who had experienced physical assault did not report the crime to the police, with the majority of respondents indicating that they believed that the police would not treat their complaint seriously (Bourne, Reid, Hammond, & Weatherburn, 2010).
Historically, LGBTI people in Australia have experienced levels of social disadvantage that have resulted in decades of inequitable treatment (Butler, 2012). Similar to the experience of LGBTI in other parts of the world (such as in the United Kingdom and the United States), many Australian LGBTI people have suffered stigma, family rejection, and social isolation, and have had a life experience of fear of rejection and persecution, coupled with the impact of potential or actual discrimination from social institutions (Butler, 2012). According to Leonard et al. (2008), this is reflected in the way that many members of the LGBTI community purposefully avoid contact with institutions such as the police. Yet how this impacts on crime reporting behaviors is largely unknown.
Tarling and Morris (2010) argued that most of what is known about crime reporting behaviors has been obtained from international victim surveys distributed to the wider (heterosexual) public such as those conducted in the United Kingdom (MacDonald, 2001;Skogan, 1994) and in the United States (Baumer, 2002;Felson et al., 2002;Hart & Rennison, 2003;Rennison, 2007). However, research undertaken in other countries such as Australia have been few and far between, and have been typically based on a secondary analysis of state victim surveys (see Carcach, 1997). This raises questions whether crime reporting practices are different in Australia (or different in specific states within Australia) to what has been reported more universally in other countries or if Australia is unique in its crime reporting practices to the police. 7

LGBTI People's Subjective Norms and Crime Reporting
In Queensland, few minority groups defined by external behaviors or other features that distinguish them from the general population have voluntary contact or are involved in community partnership programs with the police (Cunneen, 2001). This is particularly true of the members of diverse minority groups such as the LGBTI community, who, in comparison with other members of society, purposefully avoid contact and interaction with the police (Herek & Berrill, 1992). This is not to suggest that the relationship between the police in Queensland and members of the LGBTI community has been static or that the police have not attempted to make significant changes in their policy and practice implemented toward LGBTI people (e.g., LGBTI police liaison officers and policing of hate crime). However, despite changes in the social, political, and legal history of the relationship between police and LGBTI people (analyses of which are beyond the scope of this article), the nature of the relationship between the police and LGBTI people in Queensland remains problematic (Crime and Misconduct Commission [CMC], 2009).
Herek (1990) stated that many LGBTI people are aware of a level of police hostility and prejudice against homosexual behavior and LGBTI people long before the need for crime reporting occurs, and that this awareness is learnt either through hearsay, the media, or cultural, familial, and societal influences. In addition, Myers, Forest, and Miller (2004) argued that it is typically a vicarious experience of police, and an awareness of the potential for police hostility that causes most people (particularly LGBTI people) to have negative beliefs and attitudes toward the police. To distinguish between LGBTI people's general attitudes toward reporting crime to the police, and a specific belief in police homophobia (which may be a particular influence on LGBTI people's crime reporting behavior), LGBTI people's beliefs in police homophobia need to be measured (Herek & Berrill, 1992).

LGBTI People's PBC and Crime Reporting
Perceptions of nonnormative sexualities (such as those expressed by the LGBTI community) challenge mainstream models and practices of policing (Moran, 2007). The majority of policing models and practices implemented toward the community are based on a heteronormative model of society and a White, masculine, heterosexual ethos (Myers et al., 2004). Subsequently, when police are confronted with a sexually diverse community (such as the LGBTI community), the breakdown in normative expectations of gendered behavior (which is situated in the context of heterosexuality) results in homophobic confrontations (Myers et al., 2004). As such, the difficulty with which LGBTI perceive interaction with the police (and the ease or difficulty of reporting crime to the police) coupled with the lack of confidence that LGBTI people have in the police has resulted in the underreporting of crime by members of this community (Chakraborti, 2009). For example, previous research indicates that LGBTI people are less likely than heterosexual people to enter a police station to report crime because many LGBTI people feel that the police view them as a deviant group (see Mason, 1993). Canales (2000) stated that the grouping of LGBTI sexuality into a homogeneous analytic framework may actually contribute to the "othering" of sexual identity associated with the LGBTI community by heteronormative agencies such as the police. As such, concerns were raised in this research about combining LGBTI sexual identity into one cluster or homogeneous group for use as an analytic framework due to its diversity; an analysis of the large body of sociological work examining the appropriateness of grouping the LGBTI community into a sexually homogeneous conceptual and analytic framework is beyond the scope of this research article. Yet grouped sexual identity (such as normative sexuality [heterosexuality] and nonnormative sexuality [such as LGBTI sexuality]) is one of the salient identity markers that many cultures use to categorize and judge others (Skeggs, 1999). For the purpose of this study, it was deemed appropriate to analyze LGBTI sexuality as a homogeneous group. 8 An in-depth analysis of LGBTI people's beliefs in police homophobia can provide a useful framework for determining whether this belief is based on personal or vicarious experience (Myers et al., 2004). Previous studies have also indicated that a belief in police homophobia is a strong negative psychological determinant, often influencing the amount of contact many LGBTI people have with police officers (Herek, 1990;Myers et al., 2004). M. Williams and Robinson (2004) also indicated that up to three quarters of LGBTI victims fail to report crime to the police primarily because they are fearful of secondary victimization from police officers as a result of perceived police homophobia (PPH). Ajzen (2005) argued that negative beliefs account for significant amounts of variance in salient beliefs (assumed to be the immediate influence of a person's attitude), which in turn persuade intention, the predictor determining different kinds of behavior. Subsequently, by examining LGBTI people's beliefs in police homophobia in relation to crime reporting, the current research could determine if it is in fact a negative belief such as PPH that is influencing LGBTI people to underreport crime. In addition, by also examining LGBTI people's attitudes, subjective norms, and PBC, it can also be determined whether LGBTI people are different from heterosexual people in their willingness to report crime to the police.

Site Selection
The research was conducted outside of a Brisbane nightclub (situated in an inner city area), and online (by online delivery). The nightclub was chosen for its involvement in the "Gay Day" celebrations (a festival for the LGBTI community, their family, and friends), its capacity to attract large numbers of patrons, and because it is known to be openly welcoming toward LGBTI and heterosexual people, although the nightclub is advertised as a Gay and Lesbian venue. The nightclub is situated within the metropolitan area of the city and has been established within the LGBTI (and wider) community as entertainment venues for more than 20 years. While it is not known whether the nightclub has a history of police concern regarding problems with patron assaults (the venue would disclose this information), the venues have maintained a positive relationship with the police regarding patron intoxication and drug use and/or drug dealing, and the nightclub is monitored by private security guards. Unlike nightclubs marketed for younger people (typically for patrons below 30 years of age), the venue attracts a wide age range of people, and is not recognized by police as a trouble zone. For ethical reasons, the nightclub has been de-identified in this research. The various online community groups (deidentified as part of the ethical agreement) were also chosen for their capacity to attract large numbers of LGBTI and heterosexual people, and because they are situated within a large online social networking site (again de-identified for ethical reasons).

Procedure
A face-to-face survey was administered to a nonprobability sample of visitors at the "Gay Day" celebrations. An online survey was administered to a nonprobability sample of members of an online community group (by online delivery) between March and April. It was anticipated that the convenience sample of visitors collected at the event and from the online community could provide results that would be suitable for the study because the LGBTI target population in Queensland is relatively diffuse and "hidden" and constitutes a "hard to get at population," 9 as a result recruiting a traditional probability sample of LGBTI people was deemed impractical (see Griffiths, Gossop, Powis, & Strang, 1993).
In addition, although members of the LGBTI community have differing lifestyles and sexual identities that may pose problematic when linking LGBTI people together as a collective group, it was determined that identity associations could be made between LGBTI people as they are primarily interconnected by their notions of sexual identity that are different from normative heterosexual identities (see Ghaziani, 2011). It was also recognized that collective grouping of LGBTI people would result in sample heterogeneity and therefore contribute some limitations to the study in terms of generalizability. However, it was anticipated that the results of this study would speak to the broader issues regarding intention to report crime to the police, specifically, how sexual identity difference shapes an individual's attitude toward crime reporting.
The online survey was posted on message boards within Queensland-based Internet community groups 10 that are visited (and participated in) by LGBTI and heterosexual people. The Internet link was also emailed to different community groups (de-identified for ethical reasons) participating within online community forums and redistributed throughout Queensland by email. Participants were given a choice between completing a paper-based survey or by completing the survey online at a later time. If a participant wanted to complete the survey online, they were provided with an information leaflet detailing the online web address and survey link. The online survey and the survey administered at the "Gay Day" celebrations were identical.

Participants
Using a nonproportional quota sampling technique to ensure that a minimum of 100 participants from the LGBTI and heterosexual community were represented in the study, 11 329 participants were recruited to participate in the research. Participants were randomly approached on the basis of gender 12 (male and female), their willingness to complete the survey, and on their intention to enter the nightclub (either by standing in-line to enter or waiting outside of the nightclub). 13 The paper surveys (40 items) were administered to the participants by a team of six volunteers (trained by the researcher to administer the survey and approved by an institutional ethics review board) and each of the surveys given to the participants was identical.
Although many patrons refused to participate in the research, 14 overall, the research team received positive (and polite) reactions from the patrons, and the acceptance rate to participate in the study was higher than expected. 15 While it is acknowledged that some patrons who attend nightclubs may be more predisposed to dislike the police due to the effects of intoxication, drug taking, and resulting incivility, there is no empirical research to suggest that patrons who frequent nightclubs will differ in their attitudes toward the police than patrons of other social venues. As such, it was anticipated that the convenience sample of visitors collected at the venue could provide results that would be suitable for the study. To avoid duplication of results, each respondent was asked if they had completed the survey prior to being approached. The online participants were selected on the basis of gender (male and female) and participation/membership within the online community forums. To avoid duplication of online results, each respondent was asked if they had previously completed the online survey.
The final sample comprised 147 participants (44.7%) recruited from visitors attending the Brisbane "Gay Day" Celebrations, and 182 participants (55.3%) obtained by online delivery (N = 329). A 10-page, 40-item, self-report survey was utilized to examine participants' willingness to report crime and to assess participant attitudes toward the police. 16 There was 100% completion rate and no missing data.
The ages of the participants ranged from 18 to 74 years (M = 35.32, SD = 12.03), and the majority of the participants were male (n = 173; 52.6%); with females comprising 44.1% (n = 145) of participants; transgender male to female participants comprising 2.1% (n = 7); transgender female to male participants comprising 0.3% (n = 1); and intersex participants comprising 0.9% (n = 3). More than half of the participants in the sample (64.1%) were identified as LGBTI (n = 211), and 35.9% of the participants were identified as heterosexual (n = 118). More than half of the participants were in a relationship (n = 201; 61.1%) and only 11 participants (3.3%) were identified as Aboriginal Australian or Torres Strait Islander. The majority of the participants were Australian citizens (n = 319; 97%) and all of the participants in the study were from Queensland. 17 Preliminary data screening was conducted to examine demographic differences (such as gender, sexual identity, age range, and area of residency) between the two different data collection methods: participants recruited from the Brisbane "Gay Day" Celebrations and participants obtained by online delivery. The analyses indicated that there were no statistically significant differences between the participants recruited from the Brisbane "Gay Day" celebrations and participants obtained by online delivery; therefore, it was determined that for all further analyses the two samples would be combined. The data were analyzed using univariate and multivariate approaches, as well as parametric and nonparametric statistics. 18

Measures
Willingness to Report Crime. To measure participants' willingness to report crime to the police, the participants were asked to respond to vignettes (see the appendix) depicting four specific crimes (vandalism, assault, break and entering, 19 and stalking). 20 These crimes were chosen because the CMC (2006) identified assault and stalking as the most feared (and typically experienced) forms of personal crime and vandalism, and break and entering as the most feared (and typically experienced) form of property crime in Australia. Participants were asked to respond on a 5-point Likerttype scale (1 = extremely likely and 5 = extremely unlikely) to the question "how willing would it be that they would report this crime to the police." Initial inspection of the data indicated that the responses to each of the four vignettes were bimodally distributed with few participants responding "don't know." Examination of participant responses to the four vignettes indicated that 45% of participants would report vandalism to the police as opposed to 49.9% of participants who would not report vandalism to the police; 49.6% of participants would report assault to the police as opposed to 47.7% of participants who would not report assault to the police; and 47.1% of participants would report stalking to the police as opposed to 43% of participants who would not report stalking to the police. Although 68.8% of participants would report a crime of break and entering to the police as opposed to 28.2% of participants who would not report break and entering to the police, examination of the data indicated that there did not seem to be a difference between participants reporting personal crime or reporting property crime.
Further inspection of the data also indicated that the responses to each of the four vignettes were highly intercorrelated ( Table 1). Examination of participant responses to the four vignettes indicated that vandalism and assault were highly correlated (r = .75, p < .01). Vandalism was also highly correlated with break and entering (r = .68, p < .01), and assault was highly correlated with stalking (r = .65, p < .01).This would also indicate that participant responses to each of the four crime vignettes were fairly consistent and that respondents would report on one type of crime to the police and on others.
For the final analyses, participant responses to four vignettes-vandalism, assault, break and entering, and stalking-were transformed into a single crime reporting variable. 21 The crime reporting variable had good internal consistency: α = .89 (DeVellis, 2003;George & Mallery, 2003). Dovidio, Kawakami, Johnson, Johnson, and Howard (1997) and Fazio, Jackson, Dunton, and Williams (1995) indicated that people are more likely to discriminate against group members for whom they have more negative attitudes. Thus, to distinguish between LGBTI people's general attitudes toward reporting crime to the police, and a specific belief in police homophobia (which may be a particular influence on LGBTI people's crime reporting behavior), LGBTI people's beliefs in police homophobia need to be measured (Fazio et al., 1995). According to Fazio and Olson (2003), Greenwald and Banaji (1995), and Greenwald, McGhee, and Schwartz (1998), implicit measures of beliefs assess automatic evaluations associated with attitude objects that the perceivers may not necessarily be aware of, or may not realize is influencing their overt behavior, or may not be able to control. However, Jellison, McConnell, and Gabriel (2004) determined that when studying prejudiced beliefs, explicit (or controlled) measures of beliefs are belief-object-evaluations that individuals can consciously express and differ in general attitudes that individuals may hold toward a specific group or object. Thus, explicit expressions of beliefs in police homophobia may be more likely to predict crime reporting behavior under conditions where social pressures have a strong influence (Ajzen, 2005). By applying the TPB to examine the likelihood of crime reporting by LGBTI people, the proximal determinant of behavior intention to engage in the behavior becomes the key concept of the research and is determined by three sets of variables: (a) attitude (the overall evaluation and the outcome expectancy of the behavior), (b) subjective norms (perceptions of social pressure from significant others to perform a particular behavior), and (c) PBC (a person's belief as to how easy or difficult performance of the behavior is likely to be).

Attitude, Subjective Norms, PBC, and PPH. Previous research by
To measure attitude, each participant was asked to respond to eight statements 22 on a 5-point, forced-choice Likert-type rating scale, ranging from 1 = strongly agree to 5 = strongly disagree. A composite Attitude score was calculated by summing across the eight items, 23 with a minimum score of eight indicating negative attitudes toward reporting to the police, and a maximum score of 40 indicating positive attitudes toward reporting to the police. The Attitude scale had good internal consistency 24 : α = .93 (DeVellis, 2003;George & Mallery, 2003). To measure subjective norms, each participant was asked to respond to seven statements on a 5-point, forced-choice Likert-type rating scale, ranging from 1 = strongly agree to 5 = strongly disagree. A composite subjective norms score was calculated by summing across the seven items, with a minimum score of seven indicating positive social pressure from others to engage in a behavior, and a maximum score of 35 indicating negative social pressure from others to not engage in a behavior. The Subjective Norms scale had acceptable internal consistency: α = .79 (DeVellis, 2003;George & Mallery, 2003). To measure PBC, each participant was asked to respond to 13 items on a semantic differential scale with 11 scales ranging from 0 = not confident at all to 10 = extremely confident. A composite PBC score was calculated, with a minimum score of zero indicating a negative evaluation of the possible effects of reporting to the police, and a maximum score of 130 indicating a positive evaluation of the possible effects of reporting to the police. The PBC scale had good internal consistency: α = .89 (DeVellis, 2003;George & Mallery, 2003). For each item measuring beliefs of police homophobia, participant's overall responses were recoded into two categories: Yes and No. It was determined that the PPH scale had acceptable internal consistency: α = .74 (DeVellis, 2003;George & Mallery, 2003).

Limitations
There were two specific limitations to the present study. First, it was determined that the research design may also have limited the research project as the use of the vignettes asked participants to respond to only four different types of scenarios depicting assault, break and entering, vandalism, and stalking. However, as previous research indicated that LGBTI people are less willing to report crime to the police than heterosexual people, careful consideration may need to be given to determine specific types of crime experienced by the LGBTI community. Second, participants recruited for this research were not obtained from a random sample of the population and therefore may not be representative of either the LGBTI or heterosexual communities. For example, the sample of respondents who identified as heterosexual male was small in comparison with the sample of respondents who identified as LGBTI male. As such, future research should attempt to select and survey a large representative group of LGBTI and heterosexual people in Australia to examine whether the results presented here can be replicated. Even with these limitations, however, the findings of the present study provide researchers and the police with insights into crime reporting behaviors of LGBTI and heterosexual people in Australia.

Research Question 1: Are LGBTI Participants Less Willing Than Heterosexual Participants to Report
Crime to the Police?
The differential behavior of females and males in reporting crime incidents (see Baumer, 2002;Carcach, 1997;Felson et al., 2002;Hart & Rennison, 2003;Rennison, 2007) suggested the need to control gender as an influence on the willingness to report crime to the police. To examine the relationship between the dependent variable Likelihood of reporting crime to the police and gender (male/female), a chi-square test for independence (with Yates Continuity Correction) was performed. 25 The chi-square test indicated that there was a significant association between gender and reporting of crime, χ 2 (1, N = 329) = 45.39, p < .001, ϕ = .38, with females being more likely to report crime than males (n = 112; 65.1%). The percentages for gender and the likelihood of reporting crime to the police are presented in Table 2.
To control for the differences between males and females in crime reporting, separate chi-square analyses were performed for males and females examining the relationship between sexuality and reporting crime to the police. Therefore, to examine the relationship between sexual identity and reporting crime to the police, chi-square analyses were performed for males and females identifying as LGBTI, and for males and females identifying as heterosexual. 26 For males, a significant relationship was found between sexual identity and willingness to report crime to police, χ 2 (1, N = 329) = 37.41, p < .001, ϕ = .48. Male LGBTI participants were less willing (25.3%) to state that they would report crime to police than male heterosexual participants (91.7%). Only two male heterosexual participants said they would not report crime to police. A similar significant relationship was found between sexual identity and willingness to report crime to police for female participants, χ 2 (1, N = 329) = 66.74, p < .001, ϕ = .68. Female LGBTI participants were less willing (34.4%) than female heterosexual participants (96.8%) to report crime to police. Only three female heterosexual participants said they would not report crime to police.
While there is a difference between males and females in reporting crime to police, there is a very strong relationship between sexual identity and reporting crime. Regardless of gender, almost all heterosexual participants stated they would report the crimes to police. The majority of LGBTI participants said that they were more unwilling to report crime to police than willing to report crime to police. However, this was more apparent for male LGBTI participants (74.7%) than female LGBTI participants (65.6%).

Research Question 2: Do Participants' Attitudes, Subjective Norms, PBC, and Belief in Police Homophobia Determine Their Intentions to Report Crime to the Police?
To determine if reporting crime to the police could be predicted from a specific set of measures under the TPB 27 (attitude, subjective norms, PBC, and PPH), a Mann-Whitney U test was performed. 28 The Mann-Whitney U test only revealed that there was only a significant difference in crime reporting behavior and levels of PBC for participants who were more willing to report crime to the police (median = 6, n = 172) and participants who were less willing to report crime to the police (median = 6, n = 157), U = 11,621.50, z = −2.18, p < .05, r = .12. As nonparametric tests tend to be less sensitive than parametric tests, it was decided that a series of independentsamples t test would also be performed. Although it was acknowledged that by using parametric tests with a convenient sample, assumptions about the populations from which the sample was drawn would not necessarily be generalizable to the wider public.
The t tests indicated that there was a significant difference between PBC and participants who were more willing to report crime to the police (M = 81.81, SD = 21.56) and participants who were less willing to report crime to the police (M = 87.31, SD = 18.26); t(327) = −2.51; p < .05, two-tailed. The magnitude of the differences in the means (mean difference = −5.50, 95% confidence interval [CI] = [−9.83, −1.18]) was very small (η 2 = .005). The t tests indicated that there was no significant difference between reporting crime to the police and attitude, subjective norms, and PPH. The results for the four t tests are presented in Table 3. Overall, the results from the series of independent t tests and the Mann-Whitney U tests each indicated that the TPB variable PBC (a person's belief as to how easy or difficult performance of the behavior is likely to be) could predict the willingness of high/low crime reporting.
Examination of Table 3 indicated that LGBTI people had significantly more negative attitudes than heterosexual people toward reporting crime to the police. Heterosexual participants scored lower on subjective norms than LGBTI participants indicating that LGBTI people were influenced by the social pressures put on them by significant others to report crime to the police. Inspection of the mean scores also indicated that LGBTI participants had lower levels of PBC than heterosexual participants indicating that LGBTI people felt that it was more difficult for them to report a crime to the police than heterosexual participants. In addition, inspection of the mean scores indicated that LGBTI people had stronger beliefs in police homophobia than heterosexual participants.
The results of the MANOVA test indicated that there are significant differences between LGBTI and heterosexual participants' attitudes, subjective norms, PBC, and beliefs in police homophobia. LGBTI participants were found to have significantly more negative attitudes, subjective norms, and PBCs than heterosexual participants. In addition, LGBTI participants were found to have stronger beliefs in police homophobia than heterosexual participants.

Discussion
The study aimed to understand whether LGBTI and heterosexual people vary in their willingness to report crime to the police. Specifically the first research question examined whether LGBTI participants are less willing than heterosexual participants to report crime to the police. When willingness to report crime to the police was examined by differences in sexual identity, a significant difference was found between the LGBTI and heterosexual communities.
The results of the chi-square test for independence indicated that controlling for gender differences in reporting crime behavior, LGBTI participants were significantly less willing than heterosexual participants to report crime to the police. Almost all heterosexual participants said that they would report crime to the police but only 25% of LGBTI participants stated that they would report crime to the police. This indicates that a person's sexual identity influences crime reporting behavior. Interestingly, a significant relationship was also found between female LGBTI participants and female heterosexual participants in their willingness to report crime to the police. Female participants were more willing than male participants to report a crime to the police. The results indicated that over and above gender, sexual identity impacted on willingness to report crime to the police.
The second research question examined whether participants' attitudes, subjective norms, PBC, and belief in police homophobia determine their intentions to report crime to the police. The results indicated that a participant's intention to report crime to the police was influenced by PBC (or a person's belief regarding how easy or difficult performance of the behavior is likely to be) particularly low levels of PBC indicating that if a person perceived crime reporting to be difficult, then they would be less willing to do it. Yet this raises interesting questions regarding the typical factors that may influence a person to not report crime to the police. However, an in-depth analysis of such factors (e.g., the characteristics of the victim, the nature of the offence, attitudes toward police, and the victim's relationship to the offender) was beyond the scope of this research due to ethical agreements regarding information gathering about prior victimization. 29 The results also indicated that as a group of variables, the components of the TPB and PPH were unable to collectively predict the willingness of crime reporting to the police. However, further investigation of the data suggested that on its own, behavioral control (particularly positive or negative levels of PBC) can strongly affect the nature of intention to report crime to the police. The third research question examined whether there are differences in sexual identity between LGBTI and heterosexual participant's attitudes, subjective norms, PBC, and belief in police homophobia. Examination of the data indicated that LGBTI participant attitudes, subjective norms, PBCs, and beliefs in police homophobia were significantly different from those conveyed by the heterosexual community.
LGBTI participants had more negative attitudes toward reporting crime to the police and more negative belief structures about police interaction than heterosexual participants. LGBTI participants had stronger perceptions than heterosexual participants of the social pressures put on them by significant others to perform a particular behavior. In addition, LGBTI participants indicated that it was less easy for them to report a crime to the police than heterosexual participants.
The results suggest that LGBTI participants have generally more negative belief structures than heterosexual participants in relation to the police and crime reporting. The results also indicated that while there is a significant gender difference in people's intention to report crime to the police, there is also a very strong relationship between sexual identity and reporting crime. Regardless of gender, almost all heterosexual participants stated they would report crime to the police. However, the majority of LGBTI participants said that they were more willing not to report a crime to the police than to report a crime to the police, although this was more apparent for male members of the LGBTI community than female members. Therefore, the research identified that the differences between LGBTI and heterosexual people's willingness to report crime to the police are sensitive to differences in sexuality and the ensuing perceptions of treatment from police such differences bring (in this case, negative perceptions of police homophobia by LGBTI people). However, given that LGBTI people in Queensland have had higher instances of homophobia directed toward them from social institutions such as the police, such negative perceptions of police homophobia by LGBTI people is unsurprising, and this will pose a long-term problem for the LGBTI community and the police service unless it is addressed.
Therefore, to change LGBTI people's negative attitudes toward reporting crime, previous research suggests that an individual's negative subjective norms can be transformed by the influence of a significant other (see Ajzen, 2005). Subsequently, by encouraging influential members of the LGBTI community to report crime to the police, it may also encourage other members of the LGBTI community to engage in the same behavior. In addition, by encouraging members of the LGBTI community to become involved in non-crime-related activities with the police, strong perceptions of police homophobia may also be dispelled. Although this strategy assumes that police officers are not homophobic and raises additional questions about how the attitudes of police officers who are homophobic can be erased. 30 Furthermore, by changing the way that LGBTI people perceive how easy or difficult it is to report a crime to the police, negative levels of PBCs could also be dispelled.

Conclusion
The present study demonstrated how an application of the TPB can be used to structure and interpret the psychological mechanisms that influence intention to report crime. Results indicated that there are differences between LGBTI and heterosexual participant's intentions to report crime, with LGBTI people being less willing to report crime than heterosexual people. In addition, the relationship between crime reporting rates is sensitive to differences in sexual identity. The results also indicated that under the theoretical components of the TPB there are differences in LGBTI and heterosexual people's psychological mechanisms that may account for LGBTI people's reluctance to report crime to the police. Examination of the data indicated that the LGBTI participant attitudes, subjective norms, PBCs, and beliefs in police homophobia were significantly different from those conveyed by the heterosexual community.
LGBTI participants had more negative attitudes toward reporting crime to the police and belief structures about the police than heterosexual participants. The results indicated that LGBTI and heterosexual people differ significantly in their intention to report crime to the police and that a belief in police homophobia strongly influences LGBTI people's intention to underreport crime to the police. To conclude, in Australia, underreporting of crime to police by LGBTI people poses long-term problems for the LGBTI community and the police service. Despite the extensive body of research examining crime reporting behavior, the empirical field is still in its infancy regarding variations in crime reporting behaviors due to differences in sexual identity. As such, specific attention needs to be focused on creating micro level strategies that will encourage LGBTI people to have better attitudes toward the police, thereby increasing the likelihood that LGBTI people will be willing to report crime to the police.
Australian Human Rights Commission (2012) definition of the community. 6. The Australian Census does not collect information on people's sexual orientation. 7. Although a comparative analysis of crime reporting practices between Australia and other countries was beyond the scope of this research, it is recognized that an integration of Australian crime reporting practices by the lesbian, gay, bisexual, transgender, and intersex (LGBTI) and heterosexual communities would place the research into a broader spectrum of crime reporting practices. 8. In addition, it should also be noted that as the participants from the bisexual, transgender, and intersex communities only comprised a very small part of the overall LGBTI sample (see "Participants" section), it also determined that they would be included with the lesbian and gay participants as a homogeneous group. As such, it was decided that an analysis of intragroup difference or variability between the factions of the LGBTI community would not be conducted in this research. 9. Although much has been written about the "end of the closet" in the United States (see Seidman, 2001), in Queenslandregardless of changes in the political climate about sexuality and sexual politics (Moore, 2001)-there is still a reluctance of LGBTI people to publicly disclose their sexual identity, with higher instances of homophobia being identified in Queensland unlike other Australian states such as New South Wales (see Barrett, Lewis, & Dwyer, 2011;Flood & Hamilton, 2005). 10. The message boards within Queensland-based Internet community groups were selected on the basis of being exclusively provided for (and tailored for) LGBTI and heterosexual Queensland residents. 11. It was determined that a minimum sample size of 100 participants from each population would minimize sampling error for each population, increase the confidence level of potential representation of the true population, and determine a degree of variability between each population included in the study (see Israel, 2009). 12. Although respondents were selected from both genders on the basis of their outward appearance (as either male or female), all the participants included in the study regardless of gender status (i.e., male, female, or transgender) were given the option to disclose their gender identity within the demographic part of the survey. 13. The survey was administered to patrons 2 hr before the "Gay day" celebrations began because patrons had begun queuing to enter the nightclub approximately 2 hr before the venue opened. Entry to the celebrations was via one entrance to the nightclub and by ticketed entrance only. 14. It was undetermined how many patrons in total refused to answer the survey. 15. It is acknowledged that the respondents who participated in the study had little privacy when completing the survey, and that this "open" administration may have influenced their responses to the survey. However, no patrons were seen to be visibly consuming alcohol while waiting to enter the venue and the general mood of the patrons waiting to enter the venue was positive and upbeat but not rowdy. As such, the patrons who did participate in the survey responded to the survey in a serious manner.
16. Due to the ethics requirements, the participants were not asked about prior victimization; therefore, items regarding incident type, crime occurrence/frequency, and location of incident were not included in the survey. 17. Although all the participants in the study resided in Queensland, the LGBTI sample included in the research may not be representative of the wider LGBTI population in Queensland (or the wider Australian LGBTI population) as it is impossible to estimate the number of LGBTI people living in Australia due to the absence of questions relating to sexual identity in the national census (Gay & Lesbian Community Health Alliance, 2012). 18. All statistical analyses were conducted using Predictive Analytics SoftWare (PASW) statistical analysis package version 17.0. 19. Other jurisdictions may term this crime "breaking and entering" or "break and enter"; however, Section 418 of the Criminal Code Queensland states that it is termed "Break and entering." 20. In accordance with the ethics agreement, none of the participants were asked if they had been a victim of crime. As such, prior victimization was not controlled for in the analyses. 21. Initially the items included in this measure suggested that there could be two measures: (a) property crime (vandalism, and break and entering) and (b) personal crime (assault and stalking) and that the items could be considered as two multiple-item composite scores rather than one. However, it was decided that for the purpose of this study the items would be combined into one single composite score. 22. For a list of the items used to operationalize each concept, please contact the author directly. 23. None of the included items in the additive Theory of Planned Behavior (TPB) measures had missing values; therefore, sums and not averages are presented in this research. 24. Although other statistical analyses could have been performed to evaluate each of the scales (such as a factor analysis to see how the respective items loaded on one factor), it was decided for the purposes of this study that Cronbach's alpha would be a sufficient tool to measure internal consistency of each scale. Although several authors have recommended that the coefficient alpha should be minimally .90, with an ideal value of .95 (see Nunnally & Bernstein, 1994), DeVellis (2003) and George and Mallery (2003) argued that a coefficient alpha of .70 is acceptable for new scales. 25. The same tests were also performed including age and race as an influence on the willingness to report crime to the police; however, there was no significant association between these variables. 26. It was determined that a chi-square analysis of independence rather than a chi-square test for goodness-of-fit test (or onetailed proportions test) would provide appropriate results for this research question because assumptions regarding which group would have the larger mean (or proportion of willingness to report crime to the police), a hypothesized value, was not made before data were collected. 27. To assist in understanding crime reporting behavior, it was determined that each of the components of the TPB would be measured as separate elements that influence intention to report crime to police. In this way, the current research was able to gain insight into the underlying cognitive foundation of the factors that influence people's intention to report crime to the police before an independent variable such as sexuality was introduced. 28. As a convenience sample was used in the research, it was determined that a Mann-Whitney U test could provide appropriate statistical analysis of the data. It was used to initially determine differences between the two independent groups as a nonparametric alternative to a t test. However, as nonparametric tests may fail to detect differences between groups, an independent-samples t test was also performed. 29. It is acknowledged that in future studies such factors could be statistically controlled for so that the estimated models could be better specified. 30. It is recommended that further research is needed to gain a more comprehensive understanding of the attitudes of police officers toward members of the LGBTI community.