Segmenting anti-Muslim sentiment in Australia: Insights for the diverse project of countering Islamophobia

Abstract

Islamophobia is a global issue. Our nationwide survey of Australians, undertaken in 2015 and 2016, reveals the extent of this national calamity in Australia. But Islamophobia is not universal in its manifestation. Latent Class Analysis was used to develop a typology of Islamophobia in Australia. The results place Australians in four classes based on their perception of Islam: Islamophobes (13% of the population); those who are unsure about diversity and have some concerns about Muslims (24%), those with progressive attitudes about diversity but with concerns about Muslims (50%); and progressives who have no concerns about Muslims (13%). We offer a conceptual challenge to those who assert that there is a singular, legitimate approach to challenging Islamophobia. Our analysis offers a pragmatic enablement of the diverse work of those anti-racist practitioners who undertake the day-to-day work of challenging Islamophobia in our schools, workplaces, community and recreation venues.

Keywords

Islamophobia latent class analysis anti-Muslim sentiment race and ethnicity Muslims anti-racism action

Introduction

It has been appropriate to characterise Islamophobia as both a social malady and a national calamity. Linda Briskman (2015) evocatively referred to this calamity as a ‘creeping blight’. This evokes both the worsening of the issue, but also its expansion throughout the social classes of society. While Islamophobia may be championed by racial supremacists, research shows that it has burgeoned well beyond that group, and is now prevalent in a cross-section of western societies, as well as being normalized in areas of mainstream media (All Together Now and the University of Technology Sydney, 2017; One Path Network, 2018) and in political discourse (Briskman, 2015). This justifies urgent attention, and it is fair to say that policy responses across western societies could be most generously described as peripatetic. Following the Christchurch terror attacks, as well as those in Quebec and in England, there were spikes in public and policy attention to Islamophobia in some countries. This has raised the issue of how best to challenge Islamophobia.

We have argued elsewhere that racism is socially and spatially constructed (Dunn et al., 2004; Forrest and Dunn, 2010). Racism is varied in its nature and intensity, and it is dynamic (Bonnett, 1996; Dunn and McDonald, 2001; Jackson and Penrose, 1993; Kobayashi and Peake, 2000; Najib and Hopkins, 2020). That is not to say that there are not core and common characteristics of racism with respect to mechanisms, processes and outcomes (Dunn et al., 2007; Fredrickson, 2002). But the variation in manifestation and impact has been described as ‘everywhere but different’, and it provides a theoretical framework for varied and dynamic anti-racisms. Our argument in regard to Islamophobia is similar – and that there is a diversity of legitimate ways to challenge this form of racism; see also Modood and Thompson (2018) on the importance of context. Islamophobia may be ambient, but our aim in this paper is to show its diverse manifestation. This is a conceptual challenge to those who assert that there is a singular, legitimate approach to challenging Islamophobia. It is also a pragmatic enablement of the diverse work of those anti-racist practitioners who undertake the day-to-day work of challenging Islamophobia in our schools, workplaces, community and recreation venues. In this paper we pull apart the varied manifestation of Islamophobia, with the aim of providing justification for the diverse work of constructing remedies.

Islamophobia in the west: The prevalence and the concept

The Runnymede Trust’s concept of Islamophobia was made in a landmark statement in 1997. This has been incorporated within thousands of research articles and had profound relevance in our contemporary era (All-Party Parliamentary Group on British Muslims, 2018: 27). In western nations, a fear has been constructed of a Muslim ‘other’ that is a cultural and civil threat. This is a process which has been shown to be politically driven (Bliuc et al., 2019; Dunn and Kamp, 2012; Morgan and Poynting, 2012; Poynting, 2007). It is also exacerbated by media platforms that reap sales from the hyping of fear (Ewart and O’Donnell, 2018). As a consequence, Muslims are facing the sharp end of contemporary racism. Australian research shows that 62% of Muslims have experienced racism in the workplace and 55% in an education setting, these statistics are more than three times higher than the national average. Australian Muslims also experience race hate speech at a rate three times higher than other Australians (Dunn et al., 2016). Anti-Muslim sentiment is not novel to the contemporary era, it has colonial history in Australia (Rajkowski, 1987; Stevens, 1989), and Islamophobia has been accumulating particularly since the immigration programs of the late 1970s onwards (Collins, 1988). But the recent upsurge in Islamophobia in Australia has also tracked globally (Considine, 2017; Foroutan, 2011). According to the Southern Poverty Law Centre, anti-Muslim hate groups in the United States almost tripled over the years 2015–2018 and hate crimes against Muslims rose by 67% in 2015, the year Donald Trump was elected President. The ambience of Islamophobia now presents a threat to social order, and this was made apparent in a series of catastrophic events in western nations that have targeted Muslims, including the terror attacks on two mosques in Christchurch, New Zealand which killed 50 people and the attack on the City Islam Culture Centre in Quebec, Canada which killed six people.

The concept of Islamophobia has been debated as to its merits as a conceptual device. We acknowledge those debates (Bleich, 2012; Cheng, 2015; Garner and Selod, 2015; Halliday, 1999; Massoumi et al., 2017; Richardson, 2009; Sajid, 2005). The limits include its inference to an illness or disorder and exceptionalism. We affirm the benefit of conceiving Islamophobia as a social malady that affects people in nefarious ways. Unlike most psychological conditions, we assert that it is socially constructed, as inferred earlier in regard to political discourse and media. The strengths include the suggestion of anxiety, construction and contrivance, and the suggestion that it is a condition to which people are exposed. Islamophobia is a social scourge for which there are a set of remedies.

In 2018, in a second landmark statement on Islamophobia, the All-Party Parliamentary Group on British Muslims (2018: 11) moved on from the original Runnymede Trust statement to define Islamophobia as ‘rooted in racism’ that ‘targets expressions of Muslimness or perceived Muslimness’. Our view is that Islamophobia draws on the same core characteristics of racism, regarding mechanisms, processes and outcomes (Dunn et al., 2007; Fredrickson, 2002). The stereotyping of Muslims has all the essential characteristics of any form of racism – it ascribes essential characteristics, generates a hierarchy, sets out to disempower the target group, and has that effect (Dunn et al., 2007; Modood, 2020: 37). The All-Party Parliamentary Group on British Muslims cited the association between Islamophobia and racism that had been partially made in some previous United Nations and Council of Europe reports (2018: 13, 24–25), as well as in the Runnymede’s 2017 update (Elahi and Khan, 2017). British researchers were prominent among the experts that offered evidence to the All-Party Parliamentary Group inquiry on this matter (2018: 27–31, 36–44).

The inclusion of racism in the definition by the All-Party Parliamentary Group on British Muslims was also an important tactical characterisation in that it expressly confronted the assertion that anti-Muslim hate was not a form of racism. Research reports have shown the extent of this problem in countries like Australia (Akbarzadeh et al., 2009; Barkdull et al., 2011; Johns et al., 2015; Mansouri and Vergani, 2018). Too many Australians have stated that they have negative views about Muslims (Blair et al., 2017; Dunn et al., 2004), and this is not unusual within western nations (All-Party Parliamentary Group on British Muslims, 2018; Hopkins, 2004; Isakjee, 2016; Nagel, 2016; Najib and Hopkins, 2020). However, there have been attempts to deny or deflect the importance of Islamophobia. Without a public level of acknowledgment of Islamophobia there is no basis on which to mount a challenge to it. Denial is a fundamental weapon of new racism, in which Islamophobia is included (Babacan, 2008; Dunn and Nelson, 2011; Kobayashi, 2009; Van Dijk, 1992). There are many forms of racism denial (Nelson, 2013), and Islamophobia is associated with two particular forms. These include the assertion that anti-Islamic and anti-Muslim sentiment is not racism because it is not about ‘race’ but about religion. A second step in this logic is to assert that the critique of religious groups is fair political discourse in a culturally diverse democracy. Modood (2020: 45–46) has carefully outlined what the five conditions are for such discourse to be considered civil and when they are by definition racist, the main variation largely surrounding whether there is a dialogue. Islamophobia has also been deflected and diminished through the assertion that the effects of it are not substantial. In Australia, this was most infamously stated by the former Prime Minister Tony Abbott who, when asked for his view on the social importance of Islamophobia, stated that Islamophobia had never killed anyone (Coorey, 2017). This was erroneous in the context of anti-Muslim hate crimes. In the wake of the terror attacks on two mosques in Christchurch, New Zealand, the statement was especially forlorn. Abbott’s statement was also quite effective at inflaming anti-Muslim sentiment, by inferring that some other antipathies have hurt people, such as the discourses and terrorism of violent Jihadists.

Islamophobia is associated with organised racist groups, as both a characteristic and a tool. Racist political movements have been crystalising their community building and social identity around anti-Muslim sentiment (Bliuc et al., 2018, 2019; Peucker and Smith, 2019). But it has expanded far beyond these political extremes. Again the creeping blight metaphor is useful in grasping the trends in Islamophobia. We would expect Islamophobia to be associated with the organised racist groups, and to be associated with the one-in-ten or so Australians who are self-identified racial supremacists (Markus, 2019). Not all of the latter are extremists or members of organised racist groups, but we would expect them to share an antipathy towards Muslims and Islam. And, it is fortuitous for the extreme white supremacists that violent jihadists have also constructed Islam and the West as opposite (Dunn et al., 2016). These political discourses have helped expand the echo chambers that exist at such poles. Extremist groups and arguments have been able to edge into mainstream debate and media, and there has been a regrettable permissiveness of white supremacist discourse (Soutphommasane, 2018).

Segmenting Islamophobia: Enabling a pluralistic response

Researchers and commentators influenced by critical race theory have argued that associating Islamophobia with the extreme racist right neglects the Islamophobic nature of the mainstream citizenry in western societies. Research has found that Islamophobia can be found within many more than those who are racial supremacists. A focus on racist supremacists, as a deviant minority, allows mainstream Islamophobia to escape critical attention (Morsi, 2017). As many as half of the population in Australia have been found to have a negative disposition towards Muslims (Blair et al., 2017). This is manifest as discomfort with the idea of being in close proximity to Muslims, with more than 50% being concerned if a family member were to marry a Muslim (Bogardus social distance measures), and in having a negative attitude to Muslims (Blair et al., 2017; Markus, 2016). Islamophobia has become ambient. But, this does not mean that the nature of that Islamophobia is the same across different groups, as defined by their demography, or by their attitudes to diversity and difference. The remedies and means to challenge Islamophobia in different parts of societies would vary. The Islamophobia of organised white supremacist groups will require a different treatment to that of the Islamophobia of those who are unsettled about diversity, or for those who are otherwise comfortable with diversity.

There is debate as to the appropriate anti-racist response to the racism of Islamophobes. Ghassan Hage for example, is pessimistic about the merits of telling racists that their thinking is inaccurate (2017: 10–12). Using evidence or statistics to demonstrate the factual inaccuracies of racist statements will not necessarily be effective in convincing a racist to change their mind, or to no longer make Islamophobic statements. Contexts using truth look shaky in a post-truth world, where rhetoric and performatives are so powerful. In these circumstances, new anti-racist discourses, with associated rhetoric and performatives, may be more important. Discourses that are steeped in care, mutuality and transcendence may be the key to confronting Islamophobia. However, we should not dismiss the challenge to racist thinking too quickly. And we should not generalize and homogenize racist thinking. As we intend to show in this paper, Islamophobia is diverse and Islamophobic thinking is certainly segmented.

Some segments of society are characterized by thinking that can more easily be shifted. And modest shifts might be enough. Many of those with a low level of Islamophobia, once assured, may be helpful in the larger anti-racist effort. We need citizens to help change the behavior around them and to curtail Islamophobic acts. Consensus theory has shown how perceptions of what constitutes majority opinion can have a material effect on the level of racist behavior (Pedersen and Hartley, 2012; Pedersen et al., 2008). If racists feel that their views are the consensus position, perhaps encouraged by political statements or media comments, then they are more likely to say and do racist things. If they are constructed as a deviant minority then they are less empowered. The theory of Political Opportunity Structures (after Koopmans and Olzak, 2004; Meyer, 2004) also posits that extreme groups emerge to become more salient forces in circumstances where the political conditions favor that growth. Islamophobia and racism generally will fade or flourish depending on the political conditions set by leaders and prominent commentators.

Some people will have Islamophobic thinking and rhetoric that is well considered and purposeful. It will be difficult to convince these segments of society that their thinking is wrong, and the project to do so may be doomed, or will generate too low a return on the anti-racist investment. For those racists, and Islamophobes, it may be that behavior change and proscription are the most effective strategies. But even with this group there is worthy attitudinal work to be undertaken. Markus has shown how One Nation Voters in Australia are distinct in their anti-immigration and anti-Muslim sentiment, but they are also remarkably distrustful of public institutions (mainstream politics, media and universities, etc.). This disenchantment is one of the key markers of populism, and this distrust in formal politics and debate is a key justification for extreme action (Mudde, 2017). Challenging the asserted distrust of key institutions ought to be an important antidote to the worst behavior of intentional Islamophobes. Again, the theory of Political Opportunity Structures is relevant here in regard to the setting of political conditions.

Islamophobia is usefully seen as a creeping blight, and this infers a social malady that requires a remedy. It may be ambient, but it is unlikely to be universal in its manifestation, nor in its remedies. Some manifestations of Islamophobia will be challenged by better national leadership; some may require local initiatives that correct for dehumanization. Some hard core political manifestations of Islamophobia will be resistant to such interventions. For those, it may be that proscription and prohibition are the appropriate responses. In this paper we pull apart the manifestations of Islamophobia, with the aim of providing insight to the work of constructing remedies. We have put Islamophobia on the statistical slab, and we have undertaken a social dissection. Inspired by Modood’s (2020: 30) normative sociology, we are not satisfied with simply problematizing or deconstructing. Our aim is to point to ‘feasible, contextually sensitive solutions’ that can assist the diverse work of countering Islamophobia.

Method

Participants

The sample consisted of 6001 Australian respondents to a Challenging Racism Project national survey conducted in July–August 2015 (n = 5407) and November 2016 (n = 594). The sample was sourced by a commercial online panel provider, Dynata. The only criterion for the survey was that panelists were aged over 18. The survey was largely representative of the Australian population.¹ Just over half (52%) of the respondents were female and 48% male. Only 2.3% of the respondents were Aboriginal or Torres Strait Islander, which is aligned to the proportion of the population. The age of respondents ranged across 18–25 years (14.2%), 26–35 years (20.1%), 36–45 years (18.6%), 46–55 years (17.1%), 56–65 years (13.6%), and over 65 years (16.5%). Close to three-quarters (72%) of respondents were born in Australia. The most common countries of birth outside Australia were United Kingdom 7.6%, New Zealand 2.4%, India 2.4%, Malaysia 1.4%, and China (excluding SARs and Taiwan) 1.1%. The most common languages included Cantonese 1.6%, Mandarin 1.4%, Italian 1.3%, Greek 0.9% and Chinese 0.9%. The most common religious affiliations were Christianity at 50.2%, no religion, agnostic or atheist were 37.2%, Buddhism 2.8%, prefer not to say 2.7%, Hinduism 2.2%, and Islam 1.7%. The maximum level of educational attainment was: no formal qualifications at 2.1%; high school or equivalent 18%, trade or vocational qualifications 24.6%, university degree 23.8%, and postgraduate qualification 11.3%. Most respondents were in the labour force (61.1%) while 38.9% were not.

The survey included three questions that were used as the prime indicators of Islamophobia (see Table 1). The questions included a Bogardus social distance measure which asked respondents if they would be concerned if one of their closest relatives were to marry a Muslim, which we hereafter refer to as ‘Marry a Muslim Australian’. A second question borrowed the thermometer-type question used by the Scanlon Foundation Social Cohesion Surveys, in which respondents were asked whether their disposition was negative or positive towards Muslims, and whether the degree was somewhat or very much felt. We hereafter refer to this in short-hand as ‘Feelings towards Muslim Australians’. A third question was devised at the request of those who commissioned the survey, and it asked respondents if they would not object to a religious place of worship being built in their own community. Those who would disagree, and would oppose a place of worship, were speculated as most likely to be Islamophobic, and those who agreed that they would not oppose were seen as less likely to be Islamophobic. We hereafter refer to this in short-hand as ‘Place of worship’.

Table 1.

Primary indicators of Islamophobia.

Primary question	Second prompt	Response options	Source
In your opinion, how concerned would you feel if one of your closest relatives were to marry a person of: (B23)	Muslim Faith (B23.8)	Not at all concerned; Slightly concerned; Somewhat concerned; Very concerned; Extremely concerned.	Traditional Bogardus social distance measure
Would you say your feelings are positive, negative or neutral towards the following groups: (B27)	Muslim Australians (B27.3)	Very Positive; Somewhat positive; Neutral; Somewhat negative; Very Negative	Scanlon Foundation Social Cohesion Surveys
I would NOT object to a religious place of worship being built in my own community (B14)		Likert Scale: Strongly agree, Agree, Neither agree nor disagree, Strongly disagree)	Original for the SBS survey

Participants were also asked a range of questions relating to their attitudes and experiences of racism,² including: attitudes towards diversity, assimilation and immigration; ideology of nation; recognition of racial privilege; self-identification as racist; and discomfort or intolerance towards specific cultural groups. Participants were also asked a series of questions about any personal experiences of racism, including the nature of these experiences, the various spheres of life in which the incidents occur, frequency of incidents, and the impact of these experiences on victims. The above were helpful for identifying which attitudes were meaningful for understanding the different categories of Islamophobia among the sample.

Statistical analysis

To address the research questions, a series of statistical analysis were conducted. First, descriptive statistics were run to generate an outline on the extent of Islamophobia, and we present how that is manifest to different degrees across groups. These groups are defined by demography as well as by attitudes on diversity. Second, Latent Class analyses (LCA) with and without covariates were conducted in the structural equation modeling framework. LCA without covariates were conducted to determine the optimal number of classes that adequately represents the range of Islamophobia. Then, a series of bivariate analyses were conducted to select the covariates to include in further analyses. Finally, multinomial logistic regression was performed to characterize individuals’ membership in each class. The latent class variable was used as dependent variables while the covariates were used as independent variables.

LCA (Clogg, 1981; Heinen, 1996; Lazarsfeld and Henry, 1968; Uebersax, 1993) was used to group individuals into classes based on their perception of Islam in Australia. LCA is a model-based clustering technique in which the population is assumed to consist of k latent classes where the number of classes is not known a priori. The method classifies individuals into classes on the basis of responses made on a set of variables. A parametric statistical model was assumed and observed data were used to compute class membership and item response probabilities conditional on class membership. The method also allows the use of external variables to study the effect on class membership. This approach has several advantages over standard cluster-analytical techniques, including the use of fit statistics for deciding on the appropriate number of classes (Vermunt and Magidson, 2002).

In this study, we followed common practice, relying on the most widely used fit and test statistics: the Akaike’s Information Criterion (AIC; Akaike, 1987), Bayesian Information Criterion (BIC, Schwarz, 1978), sample size adjusted BIC (SBIC; Sclove, 1987), Lo-Mendell-Rubin likelihood ratio test (LMR; Lo et al., 2001). Recent simulation studies have, however, indicated that the BIC, SBIC, and LRT are more effective for identifying the correct number of classes (Diallo et al., 2016a, 2016b, 2017; Peugh and Fan, 2013). Therefore, we relied more heavily on these indices to determine the optimal number of classes in our final model.

To identify the optimum number of classes, a series of LCA models between one and six classes were estimated. Models were estimated using robust maximum likelihood estimator and Full Information Maximum Likelihood (FIML) procedure available in Mplus V8.1 (Muthén and Muthén, 2018) to handle missing data (1.06%). All LCAs were conducted using 800 sets of random starting values for 80 iterations each, and then we selected the 40 best sets of random starting values associated with the highest likelihood values for final optimization to minimize the risk of convergence to local maxima.

The extent and depth of Islamophobia

A significant minority of Australians have very negative views of Muslims. A tenth of respondents described their disposition towards Muslim Australians as very negative, and another one-fifth were somewhat negative. Putting that in the context of a specific form of cross-cultural contiguity, using a Bogardus test, one-quarter would be very concerned if a relative were to marry a Muslim. One-fifth also stated that they would oppose a place of worship in their locality, echoing the localized campaigns against mosque development throughout Australia and other western countries (Al-Natour, 2010; Dunn, 2001; Gale, 2005). There is also evidence of a group of Australians who are influenced by Islamophobia, but who may not be fixed on those racist views. Over a third of the respondents stated that they would be slightly or somewhat concerned about a relative marrying a Muslim, 38% stated that they were neutral towards Muslims, and a third could not say if they would oppose a place of worship. At the other pole, there were 37% of respondents who would not be concerned about marriage to a Muslim, 30% who were positive about Muslims and almost half said they would not oppose a place of worship. The below is a first and clear indicator that while Islamophobia is far too widespread it is not universal in its depth.

While there were similarities in the spread of respondents across the response options in Table 2, the distributions were not consistent. We undertook a series of cross-tabulations of the three questions to identify the clumping of respondents. The absolute distributions are presented in Table 3. There are some obvious and expected clumping of respondents. See for example the 470 (8%) respondents who would not be concerned at cross-marriage, who are very positive about Muslims, and would not oppose a place of worship. At the other end of the scale there were 309 (5%) respondents who would be extremely concerned at cross-marriage, who are very negative about Muslims and who might oppose a place of worship. Most respondents settled in the unsure area of this distribution. Fortunately, there are fewer respondents drifting towards opposition to places of worship, with most of those who are neutral on their feelings not disposed towards opposition. However, of concern is the drift of respondents into the ‘somewhat negative’ feelings towards Muslims.

Table 2.

Descriptive statistics on Islamophobia, Australia, 2015–2016.

	Not at all concerned %(n)	Slightly concerned %(n)	Somewhat concerned %(n)	Very concerned %(n)	Extremely concerned %(n)	Total %(n)
Marry a Muslim – concern	36.9 (2217)	19.7 (1185)	15.6 (939)	10.4 (627)	17.2 (1033)	100 (6001)
	Very positive %(n)	Somewhat positive %(n)	Neutral %(n)	Somewhat negative %(n)	Very negative %(n)	Total %(n)
Feelings towards Muslim Australians	12.3 (736)	17.9 (1073)	38.3 (2300)	21.1 (1268)	10.4 (624)	100 (6001)
	Agree	Neither	Disagree	Total
Place of worship	46.1 (2765)	33.4 (2003)	20.5 (1233)	100 (6001)

Table 3.

Cross-tabulation of the Islamophobia variables, Australia, 2015–2016.

Feelings towards Muslim Australians	Marry a Muslim Australian	Place of worship
Feelings towards Muslim Australians	Marry a Muslim Australian	Strongly agree	Agree	Neither agree nor disagree	Disagree	Strongly disagree
Very Positive	Not at all concerned	297	173	97	20	9
	Slightly concerned	22	25	21	6	0
	Somewhat concerned	11	9	8	2	0
	Very concerned	0	7	4	1	1
	Extremely concerned	9	7	3	2	2
Somewhat Positive	Not at all concerned	116	280	125	25	6
	Slightly concerned	42	133	80	30	3
	Somewhat concerned	16	57	43	17	2
	Very concerned	5	29	19	9	1
	Extremely concerned	5	7	10	9	4
Neutral	Not at all concerned	145	381	324	79	13
	Slightly concerned	48	215	257	48	8
	Somewhat concerned	31	141	212	46	19
	Very concerned	14	41	102	37	10
	Extremely concerned	9	30	45	23	22
Somewhat Negative	Not at all concerned	9	40	34	13	6
	Slightly concerned	21	64	105	35	10
	Somewhat concerned	14	79	120	60	22
	Very concerned	8	74	103	74	19
	Extremely concerned	18	53	133	89	65
Very Negative	Not at all concerned	3	1	10	5	6
	Slightly concerned	1	1	4	4	2
	Somewhat concerned	1	5	12	9	3
	Very concerned	2	4	15	33	15
	Extremely concerned	14	48	117	122	187

Twenty-three variables were tested for their association with the three Islamophobia indicators. Seven predictor variables were identified as having strength of association and were used in this study. The percentages of observations in each category for those variables are summarised in Table 4. It is worth observing at this point that some measures which we fully expected to have strong associations and predictive value were not shown to have those. These included gender, level of education, sector of employment as well as experiences of racism in different spheres of life. Past research had indicated that educational attainment, especially higher education, was significant in determining racist (Islamophobia) attitudes (Sides and Mogahed, 2018: 11). We did not find this here, which might suggest that Islamophobia has become more ambient across Australian society, and is not so easily confined to such a demographic. Religious affiliation did have an association, however, the relatively small numbers of non-Christians in the sample weakened the class analysis and could not be used further.

Table 4.

Predictor variables, Islamophobia, Australia, 2015–2016.

Variables	Descriptive Statistic
Demographic	n	%
Age	851	14.20
18–25 years	1204	20.10
26–35 years	1116	18.60
36–45 years	1024	17.10
46–55 years	817	13.60
56–65 years	988	16.50
Over 65 years
Employment	3310	55.20
Employed	356	5.90
Unemployed	2332	38.90
Not in Labour force
Attitude towards diversity
All races of people are equal
Agree	4862	81.00
Neither agree nor disagree	765	12.70
Disagree	374	6.30
It is a good thing for a society to be made up of different cultures
Agree	4522	80.40
Neither agree nor disagree	917	15.20
Disagree	262	4.40
Prosocial attitudes
I want to stand up for someone who was being discriminated against because of their culture, ethnicity or religion
Agree	4561	76.00
Neither agree nor disagree	1243	20.70
Disagree	197	3.30
Attitude towards other groups
In your opinion, how concerned would you feel if one of your closest relatives were to marry a person of Middle Eastern background?
Not or slightly concerned	4106	68.40
Somewhat concerned	857	14.30
Very or extremely concerned	1038	17.30
Would you say your feelings are positive, negative or neutral towards Refugees in Australia
Positive	2181	36.40
Neutral	2659	44.30
Negative	1161	19.30

Some of the key predictor variables deserve explanation at this point. A clear majority of the respondents believe that all races are equal, with only 6.3% disagreeing, and 12.7% neither agreeing or disagreeing. This 6% would constitute those with racial supremacist beliefs. A similar pattern emerges on respondent support for cultural diversity, with four-fifths seeing value in such diversity and 4% disagreeing. On the question of whether respondents would stand up for someone who was being racially discriminated against, there were slightly more who were unsure (21%). The intersectional relations between Islamophobia and anti-Middle-Eastern sentiment, as well as anti-refugee sentiment, were also tested. Social distance towards people from the Middle East was less stark than for Muslims, with 68.4% not concerned or slightly concerned (it was only 56.5% for Muslims), but it was still high with 17% very or extremely concerned about Australians of Middle Eastern background. One-fifth of respondents had negative feelings towards refugees (19.3%), but this was again less stark than for Muslims (31%), and many more were neutral (44.3%).

Latent classes of Islamophobia without covariates

As per common practice with LCA, only model results for latent classes between one and five were considered for further analysis for this data set (Connell and Frye, 2006; Magidson and Vermunt, 2001; Tofighi and Enders, 2007). The BIC and the SBIC pointed to the four-class solution. Entropy decreased with the number of latent classes going from 0.79 for the two-class solution to 0.59 for the five-class solution. Based on the BIC and SBIC fit statistics as well as the interpretability of the solution, the four-class model was retained in this study. This model resulted in a log-likelihood value of –24,379.07, an AIC of 48,860.14, a BIC of 49,201.83, an SBIC of 49,039.76, with 51 parameters, and an entropy value of 0.68 (see Table 5). The item endorsement probabilities for the four-class solution are presented in Table 6. As can be seen, the minimum class size was socially meaningful with each class comprising at least 1% of individuals within the overall sample. Latent classes are described in the paragraph that follows.

Table 5.

Fit statistics for Latent Class Model specifications, Islamophobia, Australia, 2015–2016.

Number of latent class	Loglikelihood	# Free parameters	AIC	BIC	SBIC	pLMR	Entropy
1	−26893.14	12	53810.29	53890.69	53852.55	NA	NA
2	−25081.43	25	50212.86	50380.35	50300.91	p <.001	0.76
3	−24533.96	38	49143.91	49398.50	49277.74	p <.001	0.69
4	−24379.07	51	48860.14	49201.83	49039.76	p <.001	0.68
5	−24348.05	64	48824.10	49252.88	49049.51	p <.001	0.59

: number; AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion; SBIC: Sample Size Adjusted BIC; p LMR: p-values for the Lo-Mendell-Rubin Likelihood ratio test for k versus k + 1classes.

Table 6.

Latent Class Analysis categories, Islamophobia, Australia, 2015–2016.

	Islamophobes	Unsure with concerns	Progressive with concerns	Progressive
n	792	1455	2982	772
%	13	24	50	13
I would not object to a religious place of worship being built in my own community (B14)
Strongly Agree	0.03	0.05	0.1	0.54
Agree	0.09	0.23	0.45	0.29
Neither Agree nor Disagree	0.24	0.46	0.36	0.13
Disagree	0.28	0.21	0.08	0.03
Strongly Disagree	0.36	0.06	0.01	0.02
How concerned would you feel if one of your closest relatives were to marry a person of Muslim Faith? (B23_8)
Not at all concerned	0.04	0.03	0.51	0.89
Slightly concerned	0.01	0.2	0.3	0.06
Somewhat concerned	0.02	0.32	0.15	0.02
Very concerned	0.1	0.27	0.04	0^a
Extremely concerned	0.04	0.18	0^a	0.03
Would you say your feelings are positive, negative or neutral towards Muslim Australians? (B27_3)
Very Positive	0.01	0.01	0.05	0.65
Somewhat Positive	0.02	0.05	0.3	0.19
Neutral	0.04	0.35	0.6	0.15
Somewhat Negative	0.26	0.56	0.06	0^a
Very Negative	0.68	0.03	0^a	0^a

Indicates a boundary solution.

In the four-class model, Class 1 represented 13% of the sample. Respondents in Class 1 tended to disagree, or strongly disagree on not opposing places of worship, with item-conditional probabilities of 0.28 and 0.36, respectively. Similarly, they tended to feel extremely concerned on marriage to a Muslim and were very negative on feelings towards Muslims. The item-conditional probabilities were 0.83 and 0.68, respectively. This class was labeled Islamophobes. These are a group within Australia who see Muslims as a threat and as a touchstone for political mobilization. The consistency of their views across these variables suggests some level of conviction, purpose and preparedness for action.

Class 2 represented 24% of the sample. Respondents in Class 2 neither agreed nor disagreed on places of worship opposition, and were somewhat concerned, or very concerned, about marriage to Muslims, and somewhat negative in their feelings. The corresponding item-conditional probabilities were 0.46, 0.32, 0.27 and 0.56, respectively. This class was labeled Unsure but with concern. This class of respondents does not have well considered positions, but they do have concerns nonetheless. They are likely the class among the Australian population who would have been influenced by the poor representations of Muslims in media and in politics (Ewart and O’Donnell, 2018). This quarter of the population are at risk of drifting toward the Islamophobes, they would be prime recruitment targets for the extreme right. If these classes were to merge they would constitute a significant minority of the population (37%). The creeping blight of Islamophobia is most starkly apparent in these first two classes.

Class 3 was the most prevalent class (50%). Individuals in Class 3 tended to agree that they could not see themselves opposing a place of worship. Class 3 respondents tended to be slightly concerned on marriage to a Muslim, but somewhat positive and neutral on their feelings towards Muslims. The corresponding item-conditional probabilities were 0.45, 0.30, 0.30 and 0.60, respectively. This class was labeled Progressives but with concern. It may well be that some of the misogyny stereotypes about Muslims that circulate within the West may be a major influence here (Dunn, 2009; Dunn et al., 2004; Sides and Mogahed, 2018). This class is marginally influenced by Islamophobia. It is critical that they are vigilant to the ambient Islamophobia that may be reproduced from classes 1 and 2, as the influences of Islamophobia could impede pro-social upstanding among this group.

Class 4 represented 13% of the sample. Individuals in Class 4 strongly agreed that they would not oppose a place of worship, were not at all concerned about a relative marrying someone of the Muslim faith, and were very positive about Muslims. The corresponding item-conditional probabilities were 0.54, 0.89, and 0.65, respectively. This class of respondents were labeled as Progressives. The class of respondents demonstrate consistency across the variables, and we could anticipate conviction on the issues of diversity and social cohesion. In the next section we analyse the demographic and attitude profiles of these four classes to seek further insight for challenging Islamophobia.

Latent classes of Islamophobia with covariates

A series of Chi-square tests were undertaken to determine which of the demographic variables, and the attitude or experience variables, were related to the latent classes. Then, a multivariate analysis was conducted with the selected variables. Here, a multinomial logistic regression analysis was used to test the relations between the predictors (the demographic, the attitude towards diversity, prosocial attitudes, and the attitude towards other groups) and the likelihood of latent class membership. Table 7 shows how latent class membership related to the predictors. Parameter estimates for each latent class membership are relative to the reference latent class, which was Class 4 (Progressives), adjusted for all other variables in the model.

Table 7.

Demographic and attitude profiles of Islamophobia classes, Australia, 2015–2016, using multinomial logistic regression estimates and odds ratios, against ‘Progressives’.

	Islamophobes		Unsure with Concerns		Progressive with Concerns
	B (SE)	OR	B (SE)	OR	B (SE)	OR
Age
18–25 years	–1.01 (0.32)^a	0.364	–0.64 (0.21)^a	0.529	–0.45 (0.17)^b	0.641
26–35 years	–0.32 (0.24)	0.723	–0.20 (0.18)	0.816	–0.06 (0.15)	0.941
46–55 years	0.29 (0.31)	1.335	0.32 (0.21)	1.373	0.17 (0.19)	1.186
56–65 years	0.80 (0.23)^a	2.218	0.70 (0.20)^c	2.021	0.43 (0.19)^b	1.529
Over 65 years	1.26 (0.27)^c	3.514	1.27 (0.20)^c	3.553	0.43 (0.17)^b	1.534
36–45 years	0^a		0^a		0^a
Employment
Unemployed	0.12 (0.29)	1.131	–0.20 (0.22)	0.82	–0.23 (0.20)	0.793
Not in Labour force	0.25 (0.13)	1.283	–0.08 (0.11)	0.928	–0.03 (0.09)	0.973
Employed	0^a		0^a		0^a
All races of people are equal
Agree	–0.94 (0.41)^b	0.392	–0.43 (0.38)	0.652	–0.24 (0.30)	0.786
Neither agree nor disagree	0.04 (0.46)	1.044	0.24 (0.43)	1.267	0.47 (0.39)	1.607
Disagree	0^a		0^a		0^a
It is a good thing for a society to be made up of different cultures
Neither agree nor disagree	1.26 (0.25)^c	3.52	0.72 (0.21)^a	2.047	0.52 (0.23)^b	1.678
Disagree	1.89 (0.54)^c	6.623	0.66 (0.54)	1.936	0.15 (0.46)	1.159
Agree	0^a		0^a		0^a
I would stand up for someone who was being discriminated against because of their culture, ethnicity or religion
Neither agree nor disagree	1.04 (0.23)^c	2.835	1.02 (0.21)^c	2.769	0.77 (0.21)^c	2.169
Disagree	0.81 (0.42)	2.252	0.46 (0.46)	1.583	–0.32 (0.41)	0.727
Agree	0^a		0^a		0^a
In your opinion, how concerned would you feel if one of your closest relatives were to marry a person of Middle Eastern background?
Somewhat concerned	3.30 (0.43)^c	27.049	3.17 (0.37)^c	23.88	1.85 (0.39)^c	6.351
Very or extremely concerned	4.15 (0.28)^c	63.244	2.73 (0.27)^c	15.283	0.10 (0.25)	1.104
Not at all or slightly concerned	0^a		0^a		0^a
Would you say your feelings are positive, negative or neutral towards Refugees in Australia
Neutral	3.70 (0.28)^c	40.595	3.81 (0.28)^c	44.937	3.34 (0.28)^c	28.086
Negative	4.96 (0.35)^c	142.513	4.34 (0.31)^c	76.506	2.26 (0.31)^c	9.611
Positive	0^a		0^a		0^a

Note. Parameter estimates for each latent class are relative to the reference latent class (Progressive), adjusted for all other variables in the model.

SE: standard Error of the estimates; OR: Odds Ratio. Odds-ratios are exponentiated parameter estimates.0^a: reference category. The coefficients and OR reflects the effects of the predictors on the likelihood of membership in the listed latent class relative to the reference class (Progressive).

p < .01.

p < .05.

p < .001.

The results showed that, compared to individuals aged between 36 and 45, respondents between 56 and 65, or over 65 years old, were more likely to be in Class 1 (Islamophobes), Class 2 (Unsure with concern) or Class 3 (Progressives with concern) than in Class 4 (Progressives). In contrast, respondents aged between 18 and 25 years were more likely to be in Class 4 (Progressives) than individuals aged between 36 and 45 years. It is older Australians who are positioned within the two most concerning classes. This association has been detected before, and is a reflection of the Australia that many in these age groups were raised in, with the narrow official identity and Christian-centrism (Nelson and Dunn, 2013). It is affirming that younger Australians are so heavily associated with the Progressives class. Only 7.4% of the Islamophobes class were aged 18 to 25, whereas that age span were 22.7% of the Progressives class. This reflects the official change to national identity, the rise of multiculturalism, and anti-racism education in schools (Forrest and Dunn, 2011). This contradicts the popular image of angry white youth at the Cronulla pogrom, or at far right rallies. Generational change is likely to have a slow, but nonetheless virtuous, challenge to Islamophobia.

Compared to Class 4 (Progressives), individuals who were not in the labor force were more likely to be in the Islamophobes class than those with employment. But the odds ratios were modest and not significant. This aligns with our findings on the dissipating effect of education and sector of employment on Islamophobia. We did not involve religion in the latent class analysis because of the relatively small sizes for some key faith groups (e.g. Muslims and other non-Christian groups). However, Christians did constitute 51% of the Islamophobes class but only 37% of the Progressives class. The opposite trend was apparent for those without a religion. This finding on Christian believers is important, as without an indicator of race or ethnicity, this is our main insight into cultural privilege as it pertains to Islamophobia. There is evidence here of some Christian privilege, and of it being attached to Islamophobia.

Respondents who agreed that all races of people are equal were negatively associated with the Islamophobes and the Concerned classes, and were more likely to be in Class 4 (Progressives). The strongest effect, and only statistically significant of those negative associations, was between racial equality and Islamophobes. The strongest and most significant of the links between attitudes and the classes was between those who disagreed that cultural diversity was a good thing and the Islamophobes class. Individuals who neither agreed nor disagreed also tended to be in the Islamophobes or Concerned classes. This reveals quite clearly the anti-diversity and pro-racial supremacism attitudinal profile of the Islamophobes. It is also concerning that those who are unsure about cultural diversity (15.2%) are also falling into the Concerned classes. This helps us finesse our sense of those in the Concerned group who are vulnerable to recruitment from the ideology of the Islamophobes class.

Those who were unsure if they would stand up for someone discriminated against because of their culture, ethnicity or religion were more likely to be classified as Islamophobes than those who would defend a person in that situation. As expected, those who disagreed they would take a pro-social stance were among the Islamophobes class (though not significantly). The concerning findings here are with those in the Concerned classes who are not sure they would help, and these effects are substantial and significant. This is another indicator of the potential for creeping blight, where a lack of conviction is mixing with ambient fear and loathing to produce a weak pro-sociality among a significant portion of the population.

Respondents who were somewhat concerned, or very or extremely concerned, with their closest relative marrying a person of Middle Eastern background were more likely to be classified as Class 1 (Islamophobes) than those who were not concerned (see Table 7). This cross-over between Islamophobia and anti-Middle-Eastern sentiment was to be expected (Human Rights and Equal Opportunity Commission, 2004). It is particularly notable that anti-Middle-Eastern sentiment was associated strongly with both of the Concerned classes, more so than it was with the Progressives class. Being somewhat concerned was associated with respondents in both of the Concerned classes, who constitute 74% of the population. The association was also strong for the Unsure with Concerns class, pointing to the impact of anti-Middle-Eastern sentiment on the group most exposed to the Islamophobes.

Individuals who felt neutral, or negative, towards refugees in Australia were much more likely to be Class 1 (Islamophobes) than those who felt positive towards refugees in Australia (Table 7). The associations between both classes of ‘Unsure with Concerns’ and ‘Progressives with Concerns’ with anti-refugee opinion was stronger as compared to Progressives. This is a definitive empirical link between Islamophobia and anti-refugee discourses and politics. Those who were neutral about refugees were also more likely to be associated with the Islamophobes and the Concerned classes. This also suggests that a pro-refugee disposition is heavily linked to a non-Islamophobia position.

Discussion and conclusion

Our data provide a more nuanced insight into the sections of non-Muslim society where Islamophobia has its greatest effects. There is a substantial cohort of Australians (13%) who have a consistent stance on Muslims that reveals them as Islamophobes. This covers their fears, dispositions and likely actions. They do not welcome cultural diversity and do not agree with racial equality. They have negative views about refugees and about people from Middle-Eastern backgrounds. They are the source for the membership of the far right white supremacist groups in Australia. Fortunately, this class is older rather than younger, and this must be generating a dwindling supply of activists in demographic terms. But they are a group whose opinions and attitudes will be difficult to turn. It may be that the best strategy is to overtly challenge the discourse and legitimacy of this group. It may be that anti-racist prohibition and proscription, of their more hurtful acts, ought to be the priorities. Consensus theory warns us that people with this sort of attitudinal profile will be more likely to say and do discriminatory things if they feel their views have majority support (Pedersen and Hartley, 2012; Pedersen et al., 2008). Furthermore, the theory of Political Opportunity Structures tells us that far right groups will flourish when the political settings are facilitative. One such facilitator is a burgeoning of distrust in key institutions and formal governance. Containing this Islamophobe class will be more easily progressed if their views are constructed as deviant and as a threat to public order, and this will be better progressed where any suggestion of non-formal political action (e.g. extremism, violence, prejudiced state acts, etc.) is roundly condemned by civil society.

There is a group of Progressives in Australia who almost mirror the Islamophobes in their size, and are the inverse in terms of attitudes and their age profile. This class is clearly able to mount a strong counter to the discourses of the Islamophobes. Empowering and legitimising this class would seem a sensible means for challenging Islamophobia. Making this group more visible could clearly help in delegitimizing the racist white supremacy behind the Islamophobes class. It is critical that such deviant voices are constructed as marginal, and that leaders and opinion makers do not generate any sense that the far right racists are a majority, or in possession of a consensus. The latter is known to facilitate the conversion of hate thinking into hate speech and acts.

Having socially dissected Islamophobia, we have found that in between these two groups is likely a spectrum of undecidedness, fear and positivity. From our analysis two classes emerged, both with concerns about Muslims, but with the larger class more generally assured about diversity in Australia. This larger group (50%) are exposed to the stereotyping of mainstream media and politics, and the concept of Islamaphobia as a condition works very well here. A better and more sensible public discourse would seem the appropriate means to shift them towards a more assured position on their Muslim fellow citizens. The smaller concerned group (24%) sits closer to the attitude and age profile of the Islamophobes class. The respondents in this class are more likely to be unsure if they would oppose a place of worship, and to not take action to help a fellow citizen suffering racism. It is in this class where ‘mainstream Australia’ is necessarily on the statistical slab. Action with this class will require a more concerted and holistic set of interventions. They are appropriately a priority group in order to construct social cohesion and to avoid public disorder and racist incivility. Inculcating antipathy to Islamophobia among ‘mainstream Australians’ is key. If more Australians can be shifted into the Progressive class it will generate the necessary up-standing, and challenge to Islamophobia, that can reset norms of what is permitted in public discourse. Calling out Islamophobia, reporting it, and showing solidarity with the victims of it, are all effective and legitimate forms of upstanding.

We have argued that a constructivist conceptual approach to Islamophobia, recognizing its spatially and socially constructed nature and impact, provides a more inclusive disposition towards the means to countering Islamophobia. This is a foil to those who would assert that the legitimate way to confront Islamophobia is singular. Further, we have deployed an analytical approach that has generated a segmentation of Australian Islamophobia. The method, and the set of empirics that were delivered, had the virtuous effect of justifying a diversity of response.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Rachel Sharples

Notes

References

Akaike

(1987) Factor analysis and AIC. Psychometrika 52: 317–332.

Akbarzadeh

Bouma

Woodlock

(2009) Muslim Voices: Hopes and Aspirations of Muslim Australians. Centre for Muslim Minorities and Islam Policy Studies. Melbourne, Australia: Monash University.

All-Party Parliamentary Group on British Muslims (2018) Report on the inquiry into a working definition of Islamophobia/anti-Muslim hatred. London. Available at: https://islamophobia-definition.com/#about (accessed 12 January 2021).

All Together Now and the University of Technology Sydney (2017) Who Watches the Media? Race-related reporting in Australian mainstream media. Available at: https://alltogethernow.org.au/wp-content/uploads/2017/12/ATN-Who-Watches-The-Media-FINAL.pdf (accessed 7 May 2019).

Al-Natour

(2010) Folk devils and the proposed Islamic school in Camden. Continuum 24(4): 573–585.

Babacan

(2008) Addressing denial: The first step in responding to racism. In: Babacan

Gopalkrishnan

(eds) The Complexities of Racism: Proceedings of the Second International Conference on ‘Racisms in the New World Order. Caloundra, Australia: University of the Sunshine Coast, pp.6–15.

Barkdull

Khaja

Queiro-Tajalli

, et al. (2011) Experiences of Muslims in four Western countries post—9/11. Affilia 26(2): 139–153.

Blair

Dunn

Kamp

, et al. (2017) Challenging Racism Project: 2015–16 National Survey. Report, Western Sydney University, Australia.

Bleich

(2012) Defining and researching Islamophobia. Review of Middle Eastern Studies 42(2): 179–188.

10.

Bliuc

A-M

Betts

Vergani

, et al. (2019) Collective identity changes in far-right online communities: The role of offline intergroup conflict. New Media and Society 21(8): 1770–1786.

11.

Bliuc

A-M

Faulkner

Jakubowicz

, et al. (2018) Online networks of racial hate: A systematic review of 10 years of research on cyber-racism. Computers in Human Behavior 87: 75–86.

12.

Bonnett

(1996) Constructions of ‘race’, place and discipline: Geographies of ‘racial’ identity and racism. Ethnic and Racial Studies 19(4): 864–883.

13.

Briskman

(2015) The creeping blight of Islamophobia in Australia. International Journal for Crime, Justice and Social Democracy 4(3): 112–121.

14.

Cheng

(2015) Islamophobia, Muslimophobia or racism? Parliamentary discourses on Islam and Muslims in debates on the minaret ban in Switzerland. Discourse & Society 26(5): 562–586.

15.

Clogg

(1981) New developments in latent structure analysis. In: Jackson

Borgotta

(eds) Factor Analysis and Measurement in Sociological Research. Thousand Oaks, CA: Sage, pp.215–246.

16.

Collins

(1988) Migrant Hands in a Distant Land: Australia's Post-War Immigration. London, UK: Pluto Press.

17.

Connell

Frye

(2006) Growth mixture modeling in developmental psychology: Overview and demonstration of examining heterogeneity in developmental trajectories of antisocial behavior across adolescence. Infant and Child Development 15(6): 609–621.

18.

Considine

(2017) The racialization of Islam in the United States: Islamophobia, hate crimes, and “flying while brown”. Religions 8(9): 165.

19.

Coorey

(2017) Tony Abbott: ‘Islamophobia hasn’t killed anyone’. Business Insider, 5 June.

20.

Diallo

TMO

Morin

AJS

(2016a) Impact of misspecifications of the latent variance–covariance and residual matrices on the class enumeration accuracy of growth mixture models. Structural Equation Modeling: A Multidisciplinary Journal 23(4): 507–531.

21.

Diallo

TMO

Morin

AJS

(2016b) Performance of growth mixture models in the presence of time-varying covariates. Behavior Research Methods 49(5): 1951–1965.

22.

Diallo

TMO

Morin

AJS

(2017) The impact of total and partial inclusion or exclusion of active and inactive time invariant covariates in growth mixture models. Psychological Methods 22(1): 166–190.

23.

Dunn

(2001) Representations of Islam in the politics of mosque development in sydney. Tijdschrift Voor Economische en Sociale Geografie 92(3): 291–308.

24.

Dunn

(2009) Public attitudes towards hijab-wearing in Australia. In: Dreher

(eds) Beyond the Hijab Debates: New Conversations on Gender, Race and Religion. Newcastle, Australia: Cambridge Scholars Press, pp.31–51.

25.

Dunn

Atie

Mapedzahama

(2016) Ordinary cosmopolitans: Sydney Muslims’ attitudes to diversity. Australian Geographer 47(3): 281–294.

26.

Dunn

Forrest

Burnley

, et al. (2004) Constructing racism in Australia. Australian Journal of Social Issues 39(4): 409–430.

27.

Dunn

Kamp

(2012) A failed political attempt to use global Islamophobia in Western Sydney: The ‘lindsay leaflet scandal’. In: Morgan

Poynting

(eds) Global Islamophobia: Muslims and Moral Panic in the West. Surrey, UK: Ashgate, pp.143–160 (Chapter 8).

28.

Dunn

Klocker

Salabay

(2007) Contemporary racism and Islamophobia in Australia: Racialising religion. Ethnicities 7(4): 564–589.

29.

Dunn

McDonald

(2001) The geography of racisms in NSW: A theoretical exploration and some preliminary findings. Australian Geographer 32(1): 29–44.

30.

Dunn

Nelson

(2011) Challenging the public denial of racism: A deeper multiculturalism. Journal of Intercultural Studies 32(6): 587–602.

31.

Elahi

Khan

(2017) Islamophobia, Still a Challenge for us All. London, UK: Runnymede Trust.

32.

Ewart

O’Donnell

(2018) Reporting Islam: International Best Practice for Journalists. London, UK: Routledge.

33.

Foroutan

(2011) Ethnic and religious discrimination? Multicultural analysis of Muslim minorities in the west. Journal of Muslim Minority Affairs 31(3): 327–338.

34.

Forrest

Dunn

(2010) Attitudes to multicultural values in diverse spaces in Australia’s immigrant cities, Sydney and Melbourne. Space and Polity 14(1): 81–102.

35.

Forrest

Dunn

(2011) Attitudes to diversity: New perspectives on the ethnic geography of Brisbane, Australia. Australian Geographer 42(4): 435–453.

36.

Fredrickson

(2002) Racism: A Short History. Princeton, NJ: Princeton University Press.

37.

Gale

(2005) Representing the city: Mosques and the planning process in Birmingham, UK. Journal of Ethnic and Migration Studies 31(6): 1161–1179.

38.

Garner

Selod

(2015) The racialization of Muslims: Empirical studies of Islamophobia. Critical Sociology 41(1): 9–19.

39.

Hage

(2017) Is Racism an Environmental Threat? Malden, MA: Polity Press.

40.

Halliday

(1999) Islamophobia’ reconsidered. Ethnic and Racial Studies 22(5): 892–902.

41.

Heinen

(1996) Latent Class and Discrete Latent Trait Models: Similarities and Differences. Thousand Oaks, CA: Sage.

42.

Hopkins

(2004) Young Muslim men in Scotland: Inclusions and exclusions. Children’s Geographies 2(2): 257–252.

43.

Human Rights and Equal Opportunity Commission (2004) IsmaεListen: National Consultations on Eliminating Prejudice Against Arab and Muslim Australians, Human Rights and Equal Opportunity Commission (HREOC), Sydney.

44.

Isakjee

(2016) Dissonant belongings: The evolving spatial identities of young Muslim men in the UK. Environment and Planning A: Economy and Space 48(7): 1337–1353.

45.

Jackson

Penrose

(eds) (1993) Constructions of Race, Place and Nation. London, UK: University College London Press.

46.

Johns

Mansouri

Lobo

(2015) Religiosity, citizenship and belonging: The everyday experiences of young Australian Muslims. Journal of Muslim Minority Affairs 35(2): 171–190.

47.

Kobayashi

(2009) Here we go again’: Christchurch’s antiracism rally as a discursive crisis. New Zealand Geographer 65(1): 59–72.

48.

Kobayashi

Peake

(2000) Racism out of place: Thoughts on whiteness and an antiracist geography in the new millennium. Annals of the Association of American Geographers 90(2): 392–403.

49.

Koopmans

Olzak

(2004) Discursive opportunities and the evolution of Right-Wing violence in Germany. American Journal of Sociology 110(1): 198–230.

50.

Lazarsfeld

Henry

(1968) Latent Structure Analysis. Boston, MA: Houghton Mill.

51.

Mendell

Rubin

(2001) Testing the number of components in a normal mixture. Biometrika 88(3): 767–778.

52.

Magidson

Vermunt

(2001) Latent class factor and cluster models, bi-plots and related graphical displays. Sociological Methodology 31(1): 223–264.

53.

Mansouri

Vergani

(2018) Intercultural contact, knowledge of Islam, and prejudice against Muslims in Australia. International Journal of Intercultural Relations 66: 85–94.

54.

Markus

(2016) Mapping social cohesion: The Scanlon Foundation Surveys. ACJC, Faculty of Arts, Monash University, Australia.

55.

Markus

(2019) The values of one nation supporters. In: Peucker

Smith

(eds) The Far-Right in Contemporary Australia. London, UK: Palgrave, pp.57–72.

56.

Massoumi

Mills

Miller

(2017) Islamophobia, social movements and the state: for a movement-centred approach. In: Massoumi

Mills

Miller

(eds) What is Islamophobia? Racism, Social Movements and the State. London, UK: Pluto Press, pp.3–32.

57.

Meyer

(2004) Protest and political opportunities. Annual Review of Sociology 30(1): 125–145.

58.

Modood

(2020) Islamophobia and normative sociology. Journal of the British Academy 8: 29–49.

59.

Modood

Thompson

(2018) Revisiting contextualism in political theory: Putting principles into context. Res Publica 24(3): 339–357.

60.

Morgan

Poynting

(2012) Introduction: The transnational folk devil. In: Morgan

Poynting

(eds) Global Islamophobia: Muslims and Moral Panic in the West. Farnham, UK: Ashgate, pp.1–14.

61.

Morsi

(2017) Radical Skin, Moderate Masks: De-Radicalising the Muslim and Racism in Post-Racial Societies. London, UK: Rowman & Littlefield International.

62.

Mudde

(2017) Introduction to the populist radical right. In: Mudde

(ed) The Populist Radical Right. London, UK: Routledge, pp.1–10.

63.

Muthén

(2018) Mplus User’s Guide. New York, UK: Author.

64.

Nagel

(2016) Southern hospitality? Islamophobia and the politicization of refugees in South Carolina during the 2016 election season. Southeastern Geographer 56(3): 283–290.

65.

Najib

Hopkins

(2020) Where does Islamophobia take place and who is involved? Reflections from Paris and London. Social & Cultural Geography 21(4): 458–1197.

66.

Nelson

(2013) Denial of racism and its implications for local actions. Discourse & Society 24(1): 89–109.

67.

Nelson

Dunn

(2013) Racism and anti-racism. In: Jakubowicz

A and

(eds) For Those Who've Come across the Seas…: Australian Multicultural Theory Policy and Practice. Melbourne, Australia: Australian Scholarly Publishing, pp.259–276 (Chapter 22).

68.

One Path Network (2018) Islam in the media. Available at: https://onepathnetwork.com/wp-content/uploads/2018/07/OnePath_Islam_In_The_Media_Report.pdf (accessed 7 May 2019).

69.

Pedersen

Griffiths

Watt

(2008) Attitudes toward outgroups and the perception of consensus: All feet do not wear one shoe. Journal of Community & Applied Social Psychology 18(6): 543–557.

70.

Pedersen

Hartley

(2012) Prejudice against Muslim Australians: The role of values, gender and consensus. Journal of Community & Applied Social Psychology 22(3): 239–255.

71.

Peucker

Smith

(2019) Not a monolithic movement: The diverse and shifting messaging of Australia’s far-right. In: Peucker

Smith

(eds) The Far-Right in Contemporary Australia. London, UK: Palgrave, pp. 73–100.

72.

Peugh

Fan

(2013) Modeling unobserved heterogeneity using latent profile analysis: A Monte Carlo simulation. Structural Equation Modeling: A Multidisciplinary Journal 20(4): 616–639.

73.

Poynting

(2007) Thugs’ and ‘grubs’ at Cronulla: From media Beat-Ups to beating up migrants. In: Poynting

Morgan

(eds) Outrageous! Moral Panics in Australia. Hobart, Australia: ACYS Press, pp.157–169.

74.

Rajkowski

(1987) In the Tracks of the Camelmen. North Ryde, Australia: Angus and Robertson.

75.

Richardson

(2009) Islamophobia or anti-Muslim racism – Or what? – Concepts and terms revisited. In: Unpublished conference paper presented at the Islamophobia and religious discrimination: New perspectives, policies and practices symposium, University of Birmingham, Birmingham, UK.

76.

Runnymede Trust (1997) Islamophobia. A Challenge for us All. London, UK: The Runnymede Trust.

77.

Sajid

(2005) Islamophobia: A new word for an old fear. Palestine-Israel Journal 12(2).

78.

Schwarz

(1978) Estimating the dimension of a model. The Annals of Statistics 6(2): 461–464.

79.

Sclove

(1987) Application of model-selection criteria to some problems in multivariate analysis. Psychometrika 52(3): 333–343.

80.

Sides

Mogahed

(2018) Muslims in America: Public Perceptions in the Trump Era. Washington DC: The Democracy Fund Voter Study Group.

81.

Soutphommasane

(2018) Confronting the return of race politics, lecture to the Whitlam Institute, Western Sydney University, 6 August 2018. Available at: www.humanrights.gov.au/news/speeches/confronting-return-race-politics (accessed 11 December 2018).

82.

Stevens

(1989) Tin Mosques & Ghantowns: A History of Afghan Camel Drivers in Australia. Melbourne, Australia: Oxford University Press.

83.

Tofighi

Enders

(2007) Identifying the correct number of classes in growth mixture models. In: Hancock

Samuelsen

(eds) Advances in Latent Variable Mixture Models. Charlotte, NC: Information Age, pp.317–341.

84.

Uebersax

(1993) Statistical modeling of expert ratings on medical treatment appropriateness. Journal of the American Statistical Association 88: 421–427.

85.

Van Dijk

(1992) Discourse and the denial of racism. Discourse & Society 3(1): 87–118.

86.

Vermunt

Magidson

(2002) Latent class cluster analysis. In: Hagenaars

McCutcheon

(eds) Applied Latent Class Analysis. Cambridge, UK: Cambridge University Press, pp.89–106.