“Don’t Freak We’re Sikh”—A Study of the Extent to Which Australian Journalists and the Australian Public Wrongly Associate Sikhism with Islam

Abstract

This study emerged from an incidental, and somewhat surprising, finding that 15 percent of working journalists who attend training on improving the ways that mainstream new media report stories about Islam and Muslims, wrongly associated Sikhism with Islam. We wondered if this was indicative of the Australian population and, through a random stratified survey of the Australian population, found that it was. The question about the extent to which populations wrongly associate Sikhism with Islam is an important one. In Australia, Muslims and Sikhs are minorities. Ignorance of Islam and its religious diversity coupled with ignorance of Muslims and their ethnic and cultural diversity underpins the intolerance of Islam in the West and the concomitant animus directed at Muslims. Intolerance and violence directed at Muslims and people wrongly assumed to be Muslims (such as Sikhs) increased after the terrorist attacks on 11 September 2001 (9/11). This speaks to religious literacy, the treatment of religious minorities and raises important questions around educating various publics (including the news media) about both Islam and Sikhism. It also speaks to the role of the mainstream news media in perpetuating Islamophobia, and its detrimental flow-on effects to Muslims and Sikhs.

1. Introduction

The international body of research has long established that the mainstream news media is influential in shaping the opinions of various publics on a wide range of matters, including influencing and perpetuating negative opinions of Islam and its adherents (Ewart and O’Donnell 2018). This raises important questions about how well prepared journalists are to accurately report stories about the faith and its followers, and where precisely the responsibility for “getting it right” rests. This study arose from an incidental finding in a broader study focused on addressing (among other things) these very questions. This study compares the extent to which a sample of Australian journalists, journalism educators and journalism students, as well as a stratified random sample of the Australian population, wrongly associate Sikhism with Islam. It sits in the context Australia having a deep-seated history of Islamophobia and racism, where Muslim and Sikh communities are minorities, where anti-Muslim sentiment is directed at Muslims and Sikhs, and where the mainstream news media routinely negatively stereotypes Muslims (Akbarzadeh and Smith 2005; Aslan 2009; Australian Bureau of Statistics 2017a; Kabir 2006; Poynting and Perry 2007; Saeed 2003). It also sits in the context of Australian journalists’ knowledge of Islam being low, which is broadly reflective of the Australian population (Dunn 2005; Ewart et al. 2018; O’Donnell et al. 2017).

In this study we find, overall, that 17.54% (n = 87/496) of journalists (15%), journalism educators (18.18%) and journalism students (19.02%) (who participated in a training course designed to improve the ways the mainstream news media report stories about Muslims) and 12.10% (n = 150/1240) of the general comparative population, wrongly associate Sikhism with Islam. While overall this difference is statistically significant, when individual media cohort groups (journalists, journalism educators and students), are assessed against the general population, largely due to the sample size, the difference remains significant for students but not for journalists or journalism educators. This means that while knowledge of journalists and journalism educators who participated in this study is comparable to that of the general population, this is not the case for the journalism students who participated in this study.

In the general comparative population, we find that the core demographic factors of age, education level and employment status were statistically significant predictors of people wrongly associating Sikhism with Islam.

Researchers have identified that when wrongly identified as Muslims, animus, vitriol and violence is directed towards Sikhs (Poynting and Perry 2007; Sidhu and Gohil 2009). These sorts of attacks directed at Sikhs escalate following terrorist attacks that are associated with Muslim perpetrators (Ahluwalia and Pellettiere 2010). This is not in any way to suggest that the aggression and violence directed at Sikhs and others mistakenly thought to be Muslim (such as Arabs) ought to be re-directed to Muslims, but rather, that it highlights a lack of religious literacy and moreover ignorance of two of the world’s organised and peaceful religions. The news media are not passive actors in generating animus towards Muslims in non-Muslim majority countries (Ahmed and Matthes 2017; Said [1981] 1997; Shaver et al. 2017). This study highlights that, when it comes to wrongly associating Sikhism with Islam, working Australian journalists are starting with the same (and relatively low) knowledge base as the general population.

This study has important implications for policymakers and news media organisations. This is because researchers have identified that in Australia, the core demographic factor that has been identified as impacting knowledge of Islam and Muslims in the general population, is tertiary education, meaning religious literacy strategies aimed at boosting knowledge of Islam ought to be targeted (at the very least) at the non-tertiary educated (O’Donnell et al. 2017). What this study suggests is that programmes aimed at boosting Australians’ knowledge of Islam could opportunistically (when addressing the various branches of Islam) ensure Sikhism is distinguished from Islam. Further, strategies aimed at boosting Australians’ knowledge of Sikhism need to be very broadly targeted. Given their lack of baseline knowledge of these minority faiths, Australian journalists reporting on Islam or Sikhism must exercise considerable caution to ensure their coverage is underpinned by the norms of good journalism including accuracy, balance, objectivity and fairness. This is challenging in the context of resourcing of news rooms, the pressures of the 24-h news cycle and where the news value of conflict has been pervasive.

2. Literature

The news media plays a crucial and powerful role in how various publics understand ‘unequal social relations and the play of cultural power’ (Cottle 2000, p. 2). Yet the mainstream news media ‘all too often produce shocking examples of xenophobic reporting and racist portrayal, while often publicly committing to the ideals and practices of an inclusive multi-ethnic, multicultural society’ (Cottle 2000, p. 3). Such reporting acts to “other” minorities by reinforcing the binary “us and them” dichotomy where the dominant group (for example, white Christian majorities) are positioned as superior to a minority group (for example, brown non-Christian minorities such as Muslims and Sikhs). Minority groups are frequently marginalised by the mainstream news media, among them Muslims and Sikhs (Poole and Richardson 2006; Sidhu and Gohil 2009).

With approximately 25 million adherents worldwide, Sikhism is a monotheistic religion and the world’s fifth largest organised religion (Ahluwalia and Alimchandani 2013, p. 935; Alimardi 2013; Chilana 2005, p. 108). By way of comparison Muslims enumerate 1.6 billion people worldwide and, after Christianity, Islam (also a monotheistic religion) is the world’s second largest organised religion (Pew Research Center 2015). While both religions are misunderstood, Sikhism is considered to be the world’s newest and least understood major religion (Sidhu and Gohil 2009, p. 7). Founded by Guru Nanak, Sikhism emerged in the sixteenth century in the Punjab region of South Asia—an area now split between India and Pakistan (Chilana 2005; Lewis 2002; Sidhu and Gohil 2009; Sikh Missionary Center 1990). While India is home to the vast majority of the worlds Sikhs (approximately 95%), just 2% of India’s population is Sikh (Ahluwalia and Pellettiere 2010, p. 304; Chilana 2005, p. 109; Nesbitt 2016, p. 3; Sidhu and Gohil 2009, p. 33). Of the Sikhs that live outside of India, half live in North America (Nesbitt 2016, p. 86). Both in India and internationally, Sikhs are considered ethnic and religious minorities (Ahluwalia and Pellettiere 2010).

In a similar vein to Islam, it is difficult to fully appreciate Sikhism outside of its full ‘religious, social and historical context’ (Nesbitt 2016, p. 4). Sikhs believe in a ‘universal brotherhood and the oneness of humanity’ (Ahluwalia and Alimchandani 2013, p. 934). Prayer, reflection, meditation, service to the community and social equity are central tenets of the faith as is a rejection of class and caste systems (Chilana 2005; Lewis 2002; Sikh Missionary Center 1990). The key outward manifestations of Sikhism are its symbols—the 5Ks: ‘kes (uncut hair), a kangha (comb), a kirpn (dagger or short sword), a kara (wrist-ring of either iron or steel), and kachh (shorts that must not reach below the knee)’ (McLeod 2008, pp. 326–27). Sikh men that adopt the 5Ks are recognisable by their beards and the turbans used to cover their topknots (Ahluwalia and Alimchandani 2013, p. 934). The turban ‘is an essential part of the Sikh identity’ (Sidhu and Gohil 2009, p. 203). Sikh boys cover their hair although not with a turban and many Sikh women may also cover their hair (Ahluwalia and Alimchandani 2013; McLeod 2008). While not all Sikhs observe the 5Ks (McLeod 2008), those that do are highly visible. Similarly, while not all Muslims observe some of the outward signs of their religion, those in non-Muslim majority countries that do (for example, women that veil) are also highly visible.

The anti-Muslim sentiment stirred in the wake of 9/11, has extended to hate crimes against those who “look” either Muslim or Arab (Ahluwalia and Pellettiere 2010; Poynting and Perry 2007). Sikhs, in particular Sikh men, have been incorrectly identified as Muslim by those ignorant of either faith (Ahluwalia and Pellettiere 2010; McLeod 2008). Like Islam, Sikhism has been conflated with terrorism and its adherents painted as dangerous “others” (Ahluwalia and Alimchandani 2013; Ahluwalia and Pellettiere 2010). In Western liberal democracies, discrimination and hate crimes directed at Sikhs have escalated since the terrorist attacks in the United States of America (US) on 9/11 (Birk et al. 2015; Chilana 2005, p. 108; Sidhu and Gohil 2009). One explanation proffered for the heightened level of discrimination and harassment faced by Sikhs following 9/11 was the repeated televising of images of Osama Bin Laden, the leader of the terrorist group Al-Qaeda, a turbaned, bearded man with dark skin (Ahluwalia and Pellettiere 2010; Sidhu and Gohil 2009). When it comes to both Islam and Sikhism, ignorance begets moral panic (Birk et al. 2015, p. 98).

Educating various publics ignorant of the Sikh faith has been challenging (Sidhu and Gohil 2009). Chilana (2005) explains the steps some government agencies in the US have taken to educate their staff about Sikh headdress specifically and religious minorities more generally. An inability to differentiate Sikhs and Muslims may also have been a factor in an attack on a gurdwara (place of worship) in the town of Oak Creek in the US state of Wisconsin, where in 2012 a gunman killed six Sikh worshippers and injured another four before taking his own life (Birk et al. 2015; Sidhu 2013). In a similar way, mosques are integral to Muslim lives and communities, for Sikhs, gurdwaras while places of worship, are also important social, cultural and community spaces (Ahluwalia and Alimchandani 2013). In the immediate aftermath of the Oak Creek shooting, the major news network CNN was unable to even explain to their audiences what a Sikh was (Birk et al. 2015, p. 100).

Sidhu (2013) points out that some critics have called for a reconstruction of the discourse about mistaken identity between Sikhs and Muslims. These critics believe this discourse suggests that it is fair to attack Muslims, while attacks on Sikhs are unfair. However, Sidhu (2013) also identifies that it is well documented, that post 9/11, there was an increase in the number of attacks on Muslims and others who may mistakenly be thought to be adherents of the Islamic faith including Sikhs. As Sidhu (2013, p. 86) puts so directly ‘Sikhs have been targeted specifically due to the perception that they may be Muslim’. It has driven Sikhs to seek to differentiate themselves from Muslims (Birk et al. 2015). In the Australian context, in seeking to address ignorance of the Sikh faith and Sikhs’ long association with and contributions to Australia, Sikh history in Australia now forms part of the school curriculum in the state of Western Australia (SBS 2016). This is in the context that Muslims’ long association with and influence in Australia has similarly largely been relegated to the annals of history.

In Australia, religious affiliation has long been recorded as part of official population counts (Wright 2010). Australia’s first national census of population and housing was conducted in 1911 (Wright 2010) and reveals a religiously diverse nation where Sikhs enumerated 102 people (just 0.002% of its population of 4.4 million) and where Muslims enumerated 3908 people (0.09% of the population) (Commonwealth of Australia 1911, p. 769). In contemporary Australia, there are now 125,000 Sikhs and 604,200 Muslims nationwide representing a meager 0.5% and 2.6% of its predominantly Christian population of 23.4 million people (Australian Bureau of Statistics 2017b). While there has been a rise in both Sikh and Muslim populations in Australia in the last decade (from 0.1% in 2006 to 0.5% in 2016 and from 1.7% in 2006 to 2.6% in 2016 respectively), both populations are, nevertheless, still minorities (Australian Bureau of Statistics 2017b). This expansion of Sikh and Muslim populations in Australia should be considered in the context of the “White Australia Policy” that since federation when the Immigration Restriction Act 1901 (C’wth) was enacted, until the 1970s when the policy was formally abandoned, acted to restrict immigration. Post-1970s Australia saw a gradual increase in immigration, including people from India (Bilimoria 2017, p. 7). Bilimoria (2017, p. 6) explains that Indian immigrants to Australia are not homogeneous and that early immigrants (between the late 1890s and the 1920s) were mainly Hindus and Sikhs. In contemporary Australia, the Indian diaspora bears little resemblance to earlier and typically unskilled immigrants, reflecting instead, respected professionals and students (Kaul 2009). With its religious, cultural and linguistic diversity, Indian immigration to Australia has been a ‘fairly recent phenomenon’ (Kaul 2009, p. 188).

Despite Australia’s Constitution mandating religious freedom and its accompanying long history of religious diversity, there is also an accompanying history of both racism and religious intolerance (Dunn et al. 2007). The news media are not passive actors in generating animus towards Muslims in non-Muslim majority countries including in Australia (Ahmed and Matthes 2017; Akbarzadeh and Smith 2005; Manning 2004; Said [1981] 1997; Shaver et al. 2017). Underpinned by ignorance, we know much about the ways Western news media negatively stereotype Islam and Muslims (Akbarzadeh and Smith 2005; Kabir 2006; Manning 2003, 2004; Poynting and Noble 2003; Rane and Ewart 2012). As Ruthven (2012, p. xii) explains, Islam is portrayed as a hostile force, and ‘the main ideological challenge to post-Enlightenment liberalism since the Soviet collapse’. Internationally the Western news media also portray Islam as a barbaric and backward religion, one that subjugates women and one where Muslims pose a threat to (insert relevant Western nationality here) and their (predominately white Christian) population’s way of life.

The US, the United Kingdom and Australia are at the forefront of research into construction of Muslim identify by the news media, collectively representing more than 60% of studies published during 2000 and 2015 (Ahmed and Matthes 2017, p. 9). In this period, discussing and analysing media content has been the most prevalent form of study (Ahmed and Matthes 2017, p. 10). Few studies put journalists (their knowledge, practices and processes) or Muslims at the centre of the research (Ewart et al. 2016). Recent Australian studies have identified that Australian journalists know little about Islam and Muslims and that this is broadly reflective of their audiences, the general Australian population (Ewart et al. 2018). In this study we seek to fill a small but important gap in the literature, namely, the extent to which Australian journalists mistakenly believe that Sikhism is associated with Islam, how widespread this misconception may be in Australia and what, if any key demographic factors may impact this. We do so in order to identify areas for policy and journalistic attention.

3. Methods

This study draws on two separate data sets to (1) assess the extent to which a convenience sample of Australian journalists, journalism educators and journalism students (the specific population) wrongly associate Sikhism with Islam and (2) assess the extent to which the general Australian population wrongly associates Sikhism with Islam to identify what, if any, key demographic factors are statistically significant influencers.

The first data set is a sub-set of responses to a survey administered as part of a broader project to training participants in Australia who attended sessions specifically focused on improving the ways the Australian news media report stories about Islam and Muslims. As part of this broader project (that commenced in 2014 and concludes in 2018), training for journalists, journalism educators and journalism students was conducted nationally. Training sessions were delivered by different members of the project team between late 2015 and mid-2016 (year two of the project, where 13 training sessions including two trials were conducted) and late September 2016 to mid-2017 (year three of the project, where 43 training sessions including one trial were conducted). The training sessions utilised different formats (depending on participant preferences and time constraints), were all conducted by experienced journalism educators and included a presentation by a Muslim trainer covering key facts about Islam and Muslims. Relevant to this study, at the beginning of each training session, participants were invited to complete an optional and anonymous questionnaire. It included a ten-question multiple choice quiz focused on establishing participants’ baseline knowledge of Islam and Muslims and on approaches to best-practice reporting of stories involving Muslims. The questionnaire developed and used in year two of the project was substantially refined for use in year three. This study draws on data collected in year three.

Relevant to this study, one of the multiple-choice questions included in the pre-training questionnaire for the purposes of assessing participants’ knowledge of Islam asked them, by selecting as many as appropriate, which of the response options is/are associated with Islam. The response options were—Sufi, Sunni, Sikh, Shi’ite, none of the above, or do not know/unsure. Also relevant to this study, the survey asked training participants to identify their job role. In the year three data set there are 575 completed pre-training questionnaires, and 515 responses could be mapped to the key cohort groups of journalists, journalism educators and journalism students (undergraduate and post-graduate). As we are interested in understanding non-Muslims’ knowledge of Islam and the extent to which participants believed Sikhism is associated with Islam, the responses from Muslim participants were removed (n = 17). Two participants held multiple roles and as we are also focused on understanding any differences specifically between journalists, journalism educators and students, for the purposes of this study they too were excluded. This data set for this study comprises 180 journalists, 11 journalism educators and 305 students.

We were surprised by the somewhat incidental finding (discussed further below) that in pre-training testing close to one fifth (17.54%) of the training participants wrongly identified Sikhism with Islam. Prima facie this raised serious questions for us including if this was a product of the questionnaire design, poor data entry (discounted by a close audit of the data and data repository), or perhaps a reflection of the Australian public’s knowledge. We decided to exactly replicate the question and response options in an upcoming national survey.

The second data set is the comparative data set and comprises responses from the 2017 National Social Survey undertaken by the Population Research Laboratory at CQUniversity in Australia. The National Social Survey is a cost-shared survey of a stratified random sample of people over 18 years of age resident in Australia (n = 1265). The survey was conducted between July and August 2017 and was administered through a twenty-station Computer-Assisted Telephone Interviewing system. Dual frame random digit dialling was used to capture the increasing proportion of people without landline telephones. The survey comprised a standardised introduction, questions submitted by participating researchers (including us) and a series of pre-determined demographic and core health questions. The survey results were also weighted to ensure representativeness.

Not all respondents answered all questions and this is reflected in the analysis. The demographic questions (among others) asked participants to nominate their religion. As we are interested in non-Muslims’ knowledge of Islam, we removed the 15 Muslim participants (n = 25 after weighting), resulting in an effective final sample of 1240 responses.

4. Data Analysis

A multi-step process was used to analyse the data. First, using frequency distributions and cross-tabulations observable differences were identified in both the specific and the general populations. This approach is suitable for examining single variables and multiple variables (Creswell 2009; Neuman 2011).

Then we compared the proportions observed from the two independent samples, using a two-sample z-test to identify if the proportion of responses overall in the specific population were statistically significantly different to those of the general population. Finally, working with the general populations’ responses, a series of bivariate and multivariate analysis were further undertaken to determine if any of the core demographic variables differed significant in people who wrongly associated Sikhism with Islam.

4.1. Observable Differences

Table 1. Responses linking branches of Islam and Sikhism with Islam—specific population.

Associated with Islam	Journalists		Journalism Educators		Students		Overall
	n = 180		n = 11		n = 305		n = 496
	n	Percent	n	Percent	n	Percent	N	Percent
Sufi	42	23.33%	5	45.45%	41	13.44%	88	17.74%
Sunni	135	75.00%	8	72.73%	157	51.48%	300	60.48%
Sikh	27	15.00%	2	18.18%	58	19.02%	87	17.54%
Shi’ite	140	77.28%	8	72.73%	136	44.59%	284	57.26%
None of the above	2	1.11%	1	9.09%	7	2.30%	10	2.02%
Do not know/Unsure	38	21.11%	1	9.09%	162	53.11%	201	40.52%

highlights the extent to which training participants in three key cohort groups of journalists, journalism educators and journalism students within the specific population wrongly associated Sikhism with Islam.

This shows the majority of journalists and journalism educators could correctly identify two of the key branches of Islam—Sunni and Shi’ite. Students were less likely to identify these. Less recognisable for all groups, is Islam’s more spiritual Sufi movement. It also shows that the majority of students did not know or were unsure about their responses. By contrast, journalism educators were the most confident in their responses. Of the three groups, students were the most likely to incorrectly identify Sikhism as a branch of Islam, followed by journalism educators and then journalists. As noted earlier, the overall finding that 17.54% of training participants wrongly associated Sikhism with Islam was somewhat surprising and acted as the catalyst to test this in the general population.

Table 2. Responses linking branches of Islam and Sikhism with Islam—Non-Muslim General Population.

Associated with Islam	2017 NSS Respondents
	n = 1240
	n	Percent
Sufi	170	13.71%
Sunni	465	37.44%
Sikh	150	12.10%
Shi’ite	644	51.94%
None of the above	54	4.35%
Do not know/Unsure	461	37.18%

highlights the extent to which respondents in the general population wrongly associated Sikhism with Islam, using weighted data from the 2017 National Social Survey.

This shows that 12.10% of the general population wrongly associated Sikhism with Islam compared with 17.54% of the specific population. It also shows that more than a third of participants were unsure about their response or did not know which of the response options were associated with Islam. The responses in wrongly associating Sikhism with Islam more closely resemble students’ responses than the responses of journalists or journalism educators.

4.2. Statistical Analysis

Table 3. Comparison of proportions observed within the specific population, to the general population.

	General Population: NSS	Specific: All	Specific: Journalists	Specific: Journalist Educators	Specific: Journalism Students
n	1240	496	180	11	305
% associating Sikhism with Islam	12.10%	17.54%	15.00%	18.18%	19.01%
Observed Difference	-	5.44%	2.90%	6.08%	6.91%
95% CI	-	1.86–9.02	−2.26–8.06	−13.31–25.47	2.63–11.19
Z value	-	3	1.1	0.6	3.2
p	-	0.0029	0.2711	0.5389	0.0015
Significant?	-	Yes	No	No	Yes

compares the observed responses obtained for the specific population, with the responses of those in the general population, and identifies that overall the responses from the specific population are statistically significantly different to those of the general population. It also highlights that when the key cohort groups are examined separately, the observed differences remain significant for students only.

The third step in the data analysis examined the demographic factors influencing people in the general population wrongly associating Sikhism with Islam.

Table 4. Count and Percentage—Demographic Factors by Wrongly Associated Sikhism with Islam1.

Demographic Factors	Total n	Number Associate Sikhism with Islam	Percentage	Chi-Squared	p = Value (* Sig at 0.05 Level)
Gender (n = 1240)
Male	604	77	12.75%	0.597	0.440
Female	636	72	11.32%
Age (n = 1231)
18–34 years	378	44	11.64%	2.627	0.453
35–44 years	212	20	9.43%
45–54 years	217	31	14.29%
55+ years	424	54	12.74%
Highest level of education (n = 1235)
Secondary school or less	361	32	8.9%
Technical or further education (TAFE)	230	33	14.3%	5.357	0.069
University	644	85	13.2%
Employment Type (n = 806)
Full time	519	51	9.83%	12.434	0.006 *
Part time	175	32	18.29%
Casual	112	8	7.14%
Not employed	430	59	13.72%
Individual Income (per annum) (n = 866)
Low = less than $65,000 p/a	532	65	12.2%	2.210	0.331
Medium = $65,000–$130,000 p/a	264	39	14.8%
High = More than $130,000 p/a	70	6	8.6%
Household Income (n = 689)
Low = less than $65,000 p/a	292	28	9.6%	1.414	0.493
Medium = $65,000–$130,000 p/a	184	18	9.8%
High = More than $130,000 p/a	213	27	12.7%
State/Territory of Residence (n = 1186)
New South Wales	398	42	10.6%	3.7972	0.434
Queensland	249	29	11.6%
South Australia	92	12	13.0%
Victoria	309	44	14.2%
Western Australia	138	12	8.7%
Australian Capital Territory	17	6	35.3%
Northern Territory	9	2	22.2%
Tasmania	23	1	4.3%
Settlement Hierarchy (n = 1234)
City	682	85	12.46%	0.225	0.893
Town	279	32	11.47%
Rural area	273	32	11.72%

represents the count and percentage data for the key demographic variables of gender, age, highest level of education, employment status (employed and the nature of the employment), individual income, household income, state or territory of residence and settlement hierarchy. This formed the basis of the final analysis. These demographic factors were considered for this study on the basis have been previously considered in assessing Australians’ knowledge of Islam (see Dunn 2005).

In undertaking the analysis, following on from the bi-variate descriptive analysis that was conducted, the decision was made to enter all identified predictors into a combined multivariate model, with the exception of those variables with considerable missing data (that is, household and individual income details), or where there were only limited numbers in the sub categories (that is, state or territory of residence). All of these variables were considered to be theoretically relevant to the question of interest, as to what demographic factors were consistent with the incorrect association of Sikhism with Islam. While the bivariate analysis presented in Table 4 identifies that only employment status was found to significantly impact a person in the general population, wrongly associating Sikhism with Islam, to explore this further, a multivariate logistic regression analysis was undertaken to determine the impact of each factor, whilst considering the presence of the other factors.3 The logistic regression model was found to be statistically significant (see

Table 5. Logistic Regression—model fit.

Test	χ2	df	p
Overall model evaluation: Likelihood ratio test	19.035	10	0.040
Goodness-of-fit test: Hosmer and Lemeshow	13.909	8	0.084
Percentage correctly classified: 87.9%

).

When considering each of the variables included in the analysis (see

Table 6. Logistic Regression analysis.

	Predictor	β	SEβ	Wald’s χ2	df	p	eβ (Odds Ratio)
Gender	female (v. male)	0.286	0.185	2.401	1	0.121	1.331
Education	uni (v. high school)	−0.490	0.231	4.486	1	0.034 *	0.613
	uni (v. tafe)	0.087	0.227	0.145	1	0.703	1.091
	tafe (v. high school)	−0.576	0.269	4.595	1	0.032 *	0.562
Age	18–34 years (v. 35–44 years)	−0.623	0.315	3.914	1	0.048 *	0.537
	18–34 years (v. 45–54 years)	0.266	0.261	1.042	1	0.307	1.305
	18–34 years (v. 55 yr or more)	0.039	0.242	0.027	1	0.871	1.040
	35–44 years (v. 45–54 years)	0.623	0.315	3.914	1	0.048 *	1.864
	35–44 years (v. 55 years or more)	0.396	0.305	1.679	1	0.195	1.485
	45–54 years (v. 55 years or more)	−0.227	0.263	0.744	1	0.388	0.797
Employment	Full Time (v. part time)	0.788	0.261	9.087	1	0.003 *	2.199
	Full Time (v. casual)	−0.228	0.412	0.305	1	0.581	0.796
	Full Time (v. not working)	0.486	0.229	4.492	1	0.034 *	1.626
	Part Time (v. casual)	−1.016	0.431	5.548	1	0.019 *	0.362
	Part Time (v. not working)	−0.302	0.265	1.300	1	0.254	0.739
	Casual (v. not working)	0.714	0.413	2.988	1	0.084	2.042
Settlement	City (v. rural)	−0.081	0.231	0.122	1	0.727	0.923
	City (v. town)	−0.050	0.225	0.050	1	0.823	0.951
	Town (v. rural)	−0.030	0.274	0.012	1	0.912	0.970

* Significant, p < 0.05.

), results revealed that the education level, age and employment status were all found o significantly predict a person wrongfully association of Sikhism with Islam.

In particular, those with an education of either University or Tafe (Technical or further education), were more likely to wrongly associate Sikhism with Islam. Those aged 35–44 years were less likely to incorrectly associate Sikhism with Islam as compared to those either 18–34 years or those 45–54 years. Finally, those working full time were less likely to incorrectly associate Sikhism with Islam as compared to either those working part time or those not working at all, and those working part time were also more likely to incorrectly associate Sikhism with Islam as compared to those working casually. Both gender and settlement location were not found to significantly predict a persons’ wrongful association of Sikhism with Islam.

While the model was able to correctly classify 87.9% of responses, it is acknowledged that there may be a range of variables beyond those captured in the current study that may further explain a person’s likelihood of wrongly associating Sikhism with Islam.

5. Discussion

This study emerged from a broader project aimed at improving the mainstream news media’s reportage of Islam and Muslims in Australia. The broader project itself is situated in the context of the Australian population’s general ignorance about Islam, its history of both Islamophobia and racism and the problematic news media coverage of stories involving Islam and its followers. A key focus of the broader project was the delivery of training sessions to journalists, journalism educators and journalism students nationally. Pre and post-training testing of participants’ knowledge of Islam and Muslims as well as best-practice approaches to reporting stories involving Muslims was a key (although not sole) method of evaluating the project’s overall effectiveness.

The context for this study is important. The study participants were professionals (or in the case of students in the process of being educated in the tertiary sector), Australia’s Constitution mandates religious freedom and while Christianity is the dominant faith group in Australia, it is nevertheless simultaneously religiously diverse. Further, both Muslims and Sikhs, while minorities, have long associations with Australia and its economic, cultural and social development. However, the deeper context that must also be considered is that Australia has an accompanying and very long history of Islamophobia and racism. It is in this context that anti-Muslim sentiment has been amplified (at least in part) by elites including some mainstream news media and politicians.

Despite Muslims’ long association with and contributions to Australia, Australians’ are largely ignorant of Islam and the religious and cultural practices of its diverse followers (Aslan 2009; Deen 2006; Dunn 2005; O’Donnell et al. 2017). The literature has highlighted that anti-Muslim sentiment has wrongly been directed at Sikhs and other (such as Arabs) who “look” Muslim. This manifests in a variety of ways including expressions and acts of hostility towards some Muslims by non-Muslims. Researchers have recently identified that tertiary education is the key demographic factor (of those considered including gender, age, employment status, income, state and territory of residence and settlement hierarchy) that positively impacts knowledge of Islam and its religious and cultural practices (O’Donnell et al. 2017). As O’Donnell et al. (2017, p. 49) point out in the context of Australia’s system of government, this means that ‘no local, or state or territory government bears a disproportionate burden of responsibility for educating their non-tertiary educated residents about Islam and its followers’. Ignorance of Islam is indeed very broadly based. This in itself poses significant challenges for policymakers whose remit must surely include the normalisation of Islam as a religion of peace and situating extremism in that context.

This study has found that in our sample of Australian journalists, journalism educators and students that just over 17% of this combined cohort misidentified Sikhism as being a branch of the Islamic faith. However, it is important to note that amongst that group students performed the worst. Amongst the general Australia population surveyed just over 12% wrongly identified Sikhism as being part of Islam. The general population’s responses are closer to students than they are to journalists and journalism educators. While our analysis identified a problem with journalists, journalism students were statistically more likely to wrongly associate Sikhism with Islam than the general Australian population and their news media colleagues. When demographic factors for the general population are examined it reveals that education level, age and employment status are significantly predict Australians’ wrongly associating Sikhism with Islam. It is interesting that a large number of the general Australian population and the training participants 40.52% (37.18% and 40.52% respectively) selected amongst their responses, chose the option of being unsure of or did not know the answer to the question. This may indicate that the problem is bigger than just mistaking Sikhism as a branch of Islam. Amongst the training participants students were more likely to choose the do not know/unsure option at 53.11%, while 21.11% of the journalists and 9.09% of journalism educators selected this option.

The prevalence of the do not know/unsure option suggests that in both the general population and amongst the training cohort three factors may be at work. Firstly, there could be confusion about the various branches of Islam, secondly people may not know that Sikhism is not a branch of Islam and thirdly some study participants may have been unwilling to commit to answering this question. In this latter respect because of the way the question was asked it is not possible to separate those who did not know which of the options was not a branch of Islam from those who were unsure and therefore unwilling to commit to an answer. However, the large number of people selecting this option indicates that there may be a more significant degree of ignorance or uncertainty about the religion of Islam amongst the general population.

The comparatively low rate of selection of the do not know/unsure option amongst journalists and journalism educators, compared to the general population, may indicate that they underestimate their ignorance of the difference between Sikhs and Muslims or their ignorance about the branches of Islam more generally. This may be due to a self-serving bias whereby journalists and journalism educators may be motivated to view themselves and their knowledge levels ‘in a positive light’ (Pronin 2006, p. 37). In an article about self-serving bias, Duval and Silvia (2002, p. 49) highlighted that the self-enhancement motivation ‘engenders a preference for cognitions and interpretations that foster a positive self-concept’. It may be that the journalists and journalism educators involved in our study underestimate their levels of ignorance to foster a positive image of themselves and the extent of their knowledge. For these cohorts, acknowledging that they do not know the answer to a question would not yield a ‘flattering result’ (Duval and Silvia 2002, p. 49). Importantly, as Pronin (2006, p. 37) points out ‘people are not always accurate and objective at perceiving themselves, their circumstances and those around them’. It is not possible to determine from our data whether the journalists and journalism educators in our study underestimate their ignorance in relation to the specific topic that is the different between Sikhs and Muslims, or whether they underestimate their ignorance across a range of issues and topics. The question of whether journalists underestimate their knowledge levels either in particular fields or more generally across the range of topics they report on, requires further exploration.

This study identifies that Australians’ ignorance of Islam extends (at least in part) to their knowledge of the Sikh faith and that, despite their clear differences as mainstream world religions, a sizeable percentage of Australians are unable to de-couple Islam from Sikhism. The same is also true for the journalists, journalism educators and more so journalism students involved in this study. The extent to which the journalists and journalism educators involved in this study wrongly associated Sikhism with Islam is reflective of the general population’s misidentification levels. This is concerning because, in an environment characterised by the politicisation of the presence of Muslims in Australia (Poynting and Noble 2003) and problematic mainstream news media coverage of Muslims (Ewart 2012), there is a greater onus on journalists to “get it right” when it comes to the basics of Islam and its adherents. This is critical in the current political and social environment where the presence of Muslims in Australia has been used as a tool in wedge politics. Journalists have a significant role in informing the general public and as public influencers particularly when it comes to shaping attitudes towards Muslims, as Shaver et al. (2017) identified in their study of news consumption habits and the impact it has on increased anger and decreased warmth towards Muslims in New Zealand. However, that is not to say that journalists need to be experts as the following quote highlights:

‘… journalists need to be better informed when reporting on stories about or involving Muslims we are not suggesting they need to be experts in the religion of Islam, but rather they return to the basics of good journalism. In this respect the focus should be on fact checking, balance and ensuring the voices affected by an issue are heard in news stories’.

(Ewart and O’Donnell 2018, p. 2)

The data analysis highlights that geographical location is not a significant factor in the general population’s knowledge levels about whether Sikhism is associated with Islam. Thus, developing ways of addressing the general Australian population’s knowledge deficits when it comes to Islam requires an “all of governments” approach. In other words, no particular level of government (that is local, state, territory or federal) in Australia bears the sole responsibility for education programmes. Importantly, such programmes should focus on the significant contributions Muslims have made to Australia’s economic development and their cultural contributions as they are currently and historically ethnically and culturally diverse. Efforts to educate the general population should focus on explaining the difference between Islam and Sikhism with the aim of decoupling Sikhism from Islam. Decoupling Sikhism from Islam may require similar educational programmes focusing on the historical links Sikhs have to Australia and their contributions to the country’s economy. This is important in addressing the confusion that exists for some people between the two faiths and their adherents. Importantly, some progress has been made in relation to educating some sections of the Australian population about Sikhs, for example in the Australian state of Western Australia the incorporation of the history of Sikhs into a special Sikh tourism trail (Government of Western Australia 2016). As mentioned earlier in this article the same government has included education in Western Australian schools about Sikhs, their religion and the history of their presence in the State. It would be instructive to examine the outcomes of the efforts of the Western Australian government in educating school students, the general public and tourists about Sikhism. In particular it would be useful to identify whether those programmes have resulted in the disentanglement of the two faiths and if there has been an improvement in program participants’ knowledge about Sikhism with attendant flow-on effects to their knowledge of Islam.

Another finding of this study is that there is a lack of knowledge about Islam and its adherents and a corresponding confusion as to whether Sikhism is a branch of Islam amongst some of the journalists who participated in our training. The journalists’ results in relation to the misidentification of Sikhism as a branch of Islam, and the number who included amongst their responses the do not know/unsure option, is cause for some concern. Mainstream news media has a key role to play in educating the general population about Australia’s religious and cultural diversity particularly as problematic news coverage can be a contributing factor to social division (Tahiri and Grossman 2013). As journalists have a significant responsibility for informing the general public, there is an increased onus on them to “get it right” when it comes to stories about Muslims and their faith and avoid mistaking Sikhs for Muslims. This is particularly important in an environment characterised by negative public attention towards Muslims and also towards Sikhs when they are mistaken for being Muslims. In particular, journalists should avoid making fundamental factual errors in their reporting.

A range of factors influence the way journalists cover stories about religion including Islam and Sikhism. In Australia and elsewhere, newsrooms have undergone immense change in the past two decades with the introduction of multi-platform journalism where journalists are often producing news for online and print formats and also adding audio and visual content to their stories. Combined with the 24 h a day, seven days a week news environment, these factors place significant time pressures and limitations on those gathering and writing news stories. Such pressures can result in unintended mistakes such as confusing Sikhs with Muslims and making basic factual errors when reporting stories about Muslims and their faith. Sitting alongside these considerations are the economic pressures many news organisations are facing including declining circulation and audience share, coupled with massive staff reductions in recent years all of which contribute to reporters who are overworked and under extreme stress. Add to this the networking of copy, where a news story is shared between multiple newsrooms and republished largely unedited or with minimal if any fact checking and the possibility of errors being replicated and sent out to new audiences increases exponentially.

The existence of codes of ethics within organisations that represent the interests of journalists and news organisations provide a potential remedy to some of the problems that occur when stories about Islam are reported or misreported. In Australia these industry and individual news organisations’ codes of ethics can assist reporters in their approaches to stories about Islam and other religions. For example, the Media Entertainment and Arts Alliance Code of Ethics for journalists (MEAA 2016) suggests that they should not ‘place unnecessary emphasis on personal characteristics, including race, ethnicity, nationality, gender, age, sexual orientation, family relationships, religious belief, or physical or intellectual disability’. The imperative behind this section of the code is to get journalists to consider whether it is necessary to include the religion of a source in a story but the test is in whether the source’s religion is of significance and relevance to the story. However, not all journalists follow the MEAA code and it is unclear as to how often individual news organisations’ codes of ethics are updated. It is uncertain how often the latter codes are updated in particular to reflect contemporary contexts in which the presence of Muslims in Australia is problematised, despite that they are a relatively low proportion of the population. Another important factor that journalists and media organisations should consider when covering stories about Islam and Sikhism, is that Muslims and Sikhs are members of the audience, and some may also be advertisers, and that alienating part of your audience, no matter how small it is, might be a problem.

While the project from which the data in this article was drawn focused on facilitating change in the way mainstream journalists covered stories about Islam and its adherents, it became obvious that some of the training participants were confused about whether Sikhs are Muslims. It is unclear as to why and how this confusion arises because amongst the training cohort are journalists who have an important and influential role in ensuring Sikhs are not confused with Muslims. It would therefore be instructive to undertake further research in this area to analyse Australian mainstream news media coverage of stories in which Sikhs have been confused with Muslims. This would assist in understanding the extent to which mainstream news media plays a role in this confusion. Additionally, further research might include interviews with the journalists who have written news stories where these types of errors have occurred as this could also shed further light on how and why this confusion arises. This may assist with the development of further strategies to address it and in turn further educate journalists about Islam. Additionally, it might provide some insights for journalism educators who are preparing students to be future journalists.

There are a number of limitations to this study that we need to highlight. While we relied on a convenience sample of journalists, journalism educators and journalism students, it did, however, provide a springboard to examine the issue nationally. We also relied on pre-determined demographic variables built into the 2017 National Social Survey. This was not conducive assessing responses from first- and/or second-generation non-Muslim and non-Sikh immigrants. Finally, a key part of this study has focused on assessing if and to what extent key demographic factors influenced the general Australian population wrongly associating Sikhism with Islam. A path for further research for both knowledge of Islam and Sikhism are some of the personal factors that may be influential on individuals’ knowledge levels. In addition, in their study set in religiously tolerant New Zealand, Shaver et al. (2017, p. 1) identify that news consumption is a predictor of anti-Muslim prejudice, more specifically that ‘greater news exposure is associated with both increased anger and reduced warmth toward Muslims’. This offers a clear pathway for future research and replication of their study in other Western contexts.

6. Conclusions

This article has quantified the extent to which a sample of Australian journalists, journalism educators and students wrongly associated Sikhism with Islam. It has also quantified the extent to which the same problem exists amongst the general Australian population. It has revealed that journalists and journalism educators wrongly associate Sikhism with Islam to the same extent that the general Australian public does. Our study has also highlighted the significance of the problem amongst our sample of journalism students. This is despite the fact that, in respect to key demographic factors, researchers have established that Australians’ ignorance of Islam and Muslims is associated with a lack of tertiary education, meaning education programmes focused on boosting Australians knowledge of Islam and Muslims ought to target, at the very least, the non-tertiary educated (O’Donnell et al. 2017). However, our findings in relation to journalism students may indicate that additional education/information programmes may need to be targeted at tertiary students studying journalism.

This article has emphasised the importance of understanding why journalists and the general population experience confusion over whether Sikhs are Muslims. We know from the research that this mistake can sometimes be related to appearance, but researchers have not explored the factors that contribute to journalists’ mistaking Sikhs for Muslims. It is unclear whether these same factors are at work with journalists, journalism educators and students. Finally quantifying a problem, as we have done in this article, that has been identified in the literature through largely anecdotal accounts of Sikhs being mistaken for Muslims, is an important first step in identifying and treating the problem. Our study provides evidence of the extent of the problem amongst two very important cohorts—journalists and the general Australian population. Additionally, it is a significant problem amongst journalism students who may in the future work as journalists. The cohort of journalism educators in our study was too small to assume that their lack of knowledge is reflected in the broader journalism educators’ cohort in Australia. In addition, the large number of the specific cohort in the study (more than a third) indicated they were unsure about their response or did not know which of the response options were associated with Islam, this reveals another group within the sample who may lack knowledge about aspects of Islam and Sikhism. The aforementioned factors all go towards establishing the extent of the problem. In doing so, this study provides a base from which educators and governments might address the problem of mistaken identity between Sikhs and Muslims and the lack of knowledge about Islam and its adherents. It also provides evidence that additional educative efforts are required for Australian journalists to address both their lack of knowledge and uncertainty about Islam and Sikhism.