Indigenous people display lower mental illness stigma in Aotearoa

Abstract

Background:

Mental illness stigma continues to be pervasive and problematic in society. Researchers have attempted to better understand this stigma through investigations into demographic factors that may predict stigma, focusing on factors such as age, ethnicity and education.

Method:

We investigated demographic factors in the context of Aotearoa New Zealand, with a particular focus on Māori, the Indigenous people of Aotearoa. We used data from the Health Promotion Agency, which collected representative samples from Aotearoa across three survey waves (total n = 3518). Assessment instruments were the Mental Health Knowledge Scale (MAKS), the Reported and Intended Behaviour Scale (RIBS) and the Community Mental Health Ideology subscale of the Community Attitudes towards the Mentally Ill (CAMI). Using linear mixed-effects model we controlled for several demographic variables (e.g. age, biological sex, education and socioeconomic status) and additional variables (e.g. having a psychological condition and whether participants knew someone with mental illness) across three models for each measure.

Results:

The results revealed that mental illness stigma was lower among both Māori and European participants. Additional variables and their associations with mental illness stigma are also discussed.

Conclusion:

Overall, this study illustrates mental illness stigma as lower among Indigenous people in Aotearoa, which prompts further research into ethnicity and mental illness stigma as well as non-Western understandings of mental illness.

Individuals with mental illness continue to face discrimination in interpersonal relationships, employment opportunities, obtaining healthy housing and service provision (Brouwers, 2020; Callard et al., 2012; Thornicroft et al., 2014). The psychological impacts of this prejudice are marked, including increased suicidal ideation and reduced quality of life (El-Badri and Mellsop, 2007; Oexle et al., 2017). Much of the discrimination faced by individuals with mental illness is the result of stigma. Goffman (1963) conceptualised stigma as an ‘attribute that is deeply discrediting’ and which makes one ‘tainted’ and socially undesirable (p. 3). Taking this process further, Link and Phelan (2001) argued that stigma is a process, with labelling followed by the formation of stereotypes and prejudice. Indeed, throughout the literature, there are consistently endorsed stereotypes of those with mental illness, including that people with mental illness are unpredictable, dangerous and incompetent (Corrigan and Bink, 2016; Kenny et al., 2018; Read and Law, 1999).

In attempts to further understand mental illness stigma, many researchers have investigated the contribution of demographic factors. Generally, this research has revealed that mental illness stigma is lower in females than in males (Corrigan and Watson, 2007; Yuan et al., 2016) and older generations relative to younger generations (Adewuya and Makanjuola, 2008; Aromaa et al., 2011; Yuan et al., 2016). Prejudice also tends to decrease as education (Griffiths et al., 2008; Yuan et al., 2016) and socioeconomic status (SES; Gonzales et al., 2017; Yuan et al., 2016) increase. With respect to ethnicity, much of the research is drawn from the United States (US), providing a limited picture. For example, greater mental illness stigma was observed among African American, Asian, Latinos, Native Hawaiians and Pacific populations compared with Caucasian or European populations (Corrigan and Watson, 2007; Misra et al., 2021; Rao et al., 2007; Subica et al., 2019).

Discussion of the link between ethnicity and higher mental illness stigma can reinforce views that all Indigenous peoples and ethnic minorities report higher mental illness prejudice and, relatedly, that non-Western beliefs regarding mental illness are inherently problematic (Read et al., 2013). This, in turn, has implications for treatment. For example, it may be used as a justification to further ignore, marginalise or pathologise cultural constructions of mental illness (Taitimu, 2008; Taitimu et al., 2018). Moreover, it ignores the fact that Western understandings of mental illness, which tend to favour biological explanations, can also be associated with higher levels of prejudice (Haslam and Kvaale, 2015; Read and Law, 1999; Zimmerman et al., 2021).

Māori (tangata whenua/the Indigenous people of Aotearoa New Zealand) understandings of mental illness include but are not limited to Western conceptualisations (Taitimu, 2008). For example, Taitimu (2008) interviewed Māori who had worked with or experienced psychosis or schizophrenia. Rather than simply pathologising experiences such as hearing voices, participants reported that hearing voices might reflect their whakapapa/tupuna (i.e. ancestors) trying to contact them or be a sign of matakite (i.e. giftedness). Importantly, these understandings stood alongside viewing these experiences as reflective of mate Māori (i.e. Māori illness), as well as experiencing a lack of balance between tapu (i.e. that which is restricted or sacred) and noa (i.e. without limitation). Thus, Māori understandings of mental illness are complex, combining Western conceptualisations with cultural constructions and the socio-political landscape (Taitimu, 2008; Taitimu et al., 2018).

Furthermore, it has been stated that Māori have always provided community care, rather than excluded, individuals suffering from problems relating to the mind (Kingi et al., 2018). Although the empirical evidence to support this statement is scarce, Baker (1988, as cited in Kingi et al., 2018) stated:

The Western psychiatric tradition of confining people with a mental health disability was foreign to Māoris, who had always cared for these people in their communities. The Mental Health system was originally established to cater for people to be taken out of society. Society had this fear of contamination from mental disease and also a massive denial that it even existed. These concepts were alien to Māori people whose whānau [family] members suffering from trauma were always included within the whānau, hapū [subtribe], iwi [tribe] boundaries and given special status. (p. 88)

The Western psychiatric tradition is further evidenced by the fact that the general public tends to distance themselves from those with mental illness (e.g. being less willing to work alongside, have as a neighbour, have as a friend, etc.) (Angermeyer et al., 2013; Cechnicki et al., 2011; Mehta et al., 2009).

In the context of mental illness stigma, one could argue that Māori constructions of mental illness (e.g. understanding that voices might reflect their whakapapa/tupuna) and collective approach (e.g. including rather than excluding people with mental illness) may indicate lower stigma and prejudice. To empirically test this hypothesis, we utilised data collected by the Health Promotion Agency (HPA) as part of their New Zealand Mental Health Monitor (NZMHM), which is a nationally representative survey with data collected over three waves (2015, 2016 and 2018). The NZMHM included three widely used measures of mental illness stigma, namely the Mental Health Knowledge Scale (MAKS) (Evans-Lacko et al., 2010), the Reported and Intended Behaviour Scale (RIBS) (Evans-Lacko et al., 2011) and the Community Mental Health Ideology (CMHI) subscale of the Community Attitudes towards the Mentally Ill (CAMI) scale (Taylor and Dear, 1981). Using linear mixed-effects models, we assessed the direct relationship between ethnicity and mental illness stigma, while controlling for the random effect of other demographics (e.g. age, biological sex, education, SES). We predicted Māori participants would display lower levels of stigma across all three measures, relative to all other major ethnic groups in Aotearoa.

Method

The New Zealand mental health monitor and participants

The HPA conducted the New Zealand Mental Health Monitor which is a nationally representative survey with data collected over three waves (2015, 2016 and 2018). Data collection was approved by the New Zealand Ethics Committee (NZEC Application Nos: NZEC2015#10 and NZEC2015#10_2). The Mental Health Monitor aims to assess mental distress, stigma towards those with mental illness, social connectedness and mental health knowledge (Health Promotion Agency, 2019). Participants were recruited through mesh blocks based on the 2013 census with booster samples utilised for Māori, Pacific and young people (15–24 years old). The survey was administered through interviews with the participants in their homes, with some items completed independently.

Measures

Explanatory variables

In this study, demographic information was collected using a written list of options for participants to choose from. Age was grouped into 10 age brackets, and we combined these together for a total of 5 age brackets to simplify our analyses. Biological sex only included two options: male and female.

Ethnicity was categorised into European, Māori, Pacific, Asian and Middle Eastern/Latin American/African (MELAA) (Statistics New Zealand, 2005). Participants could endorse multiple ethnicities to allow for more accurate comparisons across ethnicities. However, given it is still commonly employed, priority-coded ethnicity was also included in Supplementary Material, where participants were allocated a single ethnicity with the following priority: Māori, Pacific, Asian, MELAA or European (the standard defined by Statistics New Zealand).

SES was measured from a decile-based index constructed from SES variables in the Census 2013, known as the NZ Deprivation Index 2013 (Atkinson et al., 2014). Education level was obtained through an open-ended question of the participant’s highest qualification alongside a written list of options. This was coded into No formal school qualification, Secondary School, Trade/Professional Qualification, Undergraduate/College and Postgraduate.

Participants were asked whether they had a psychological condition which impacted on their ability to do everyday tasks and/or on their ability to interact with others. If participants said yes to either of these, they were categorised as having a psychological condition. Finally, participants were asked whether they knew someone with a mental illness.

Stigma and knowledge measures

The MAKS is a validated measure of stigma-related mental health knowledge as well as whether several conditions are recognised as types of mental illness (Ben Amor et al., 2023; Evans-Lacko et al., 2010; Pingani et al., 2019). We focused on the items relating to stigma-related mental health knowledge, as this was consistent across all three waves. There were five stigma-related mental health knowledge items, which were responded to on a five-point scale ranging from 1 (strongly agree) to 5 (strongly disagree). For example, ‘Most people with mental health problems want to have paid employment’. Lower scores were indicative of greater mental health knowledge and more positive views of those with mental health problems (Health Promotion Agency, 2019). The overall internal consistency of the MAKS in this study was 0.44 (Cronbach’s alpha). However, this low internal consistency reflects previous research using the MAKS as well as the range of topics the items assess (Henderson et al., 2020). Hence, it is noted that the MAKS was designed to test a variety of factors rather than to converge on a latent factor which explains the low internal consistency. See Supplemental Table S1 for items included.

The RIBS is considered a validated approach for assessing previous, current and future behaviours towards people with mental health problems (i.e. stigmatised behaviours) (Bitta et al., 2022; Evans-Lacko et al., 2011; Pingani et al., 2016). We focused on the items on participants’ willingness to engage with those with mental illness in the future. This involved asking about their willingness to live with, work with, live nearby and continue a relationship with someone with mental illness. These four items were rated on a five-point scale ranging from 1 (strongly agree) to 5 (strongly disagree). For example, ‘In the future, I would be willing to live with someone with a mental illness’. Lower scores indicated a greater willingness to interact with those with mental illness in the future and, thus, represented lower levels of mental illness stigma (Health Promotion Agency, 2019). The overall internal consistency among the items for the RIBS in this study was 0.84 (Cronbach’s alpha). See Supplemental Table S1 for items included.

The CAMI was developed to assess attitudes towards those with mental distress within the community (i.e. stigmatised attitudes) (Taylor and Dear, 1981) and is widely considered a validated measure (Kafami et al., 2023; Ochoa et al., 2016; Sanabria-Mazo et al., 2023). It contains 40 items and utilises a Likert-type scale ranging from 1 (strongly agree) to 5 (strongly disagree). There are four subscales (e.g. authoritarianism, benevolence, social restrictiveness and community mental health ideology), each containing 10 items split into pro (e.g. ‘As soon as a person shows signs of mental disturbance, he should be hospitalised’) and anti-items (e.g. ‘Mental illness is an illness like any other’). For the present study, only the ‘pro’ Community Mental Health Ideology (CMHI) subscale was used as this was the only subscale with consistent items across all three waves in the mental health monitor. Lower scores indicated more positive attitudes towards those with mental distress. The overall consistency of the subscale items of the CAMI in this study was 0.77 (Cronbach’s alpha). See Supplemental Table S1 for items included.

Analyses

Statistical analyses were conducted in Rstudio (version 4.0.3) (R Core Team, 2019). For descriptive statistics, only complete cases were used. In Supplemental Table S2 an overview of participants’ average scores across the MAKS, RIBS and CAMI is provided. For subsequent analyses, imputation was utilised for missing data (see below).

A linear mixed-effects model was constructed using the R package ‘lme4’ (Bates et al., 2015) to compute the relationship between each of the three mental illness measures (i.e. MAKS, RIBS and CAMI) and ethnicity. Within each model, we further controlled for the influence of additional variables. This included age, biological sex, education, SES, having a psychological condition and knowing someone with a mental illness. Time was added as a random effect for each relationship examined to account for any changes over the three waves. Finally, each ethnicity was entered as a separate fixed effect so participants with multiple ethnicities could be represented in our analyses. In doing so, this meant that each ethnicity was being compared with those who did not identify with that ethnicity (e.g. Māori vs non-Māori, European vs non-European, etc.). Prioritised ethnicity analyses were included as Supplementary Material (see Supplemental Table S3).

A confirmatory factor analysis (CFA) was performed on participants’ responses for each of the MAKS, RIBS and CAMI to produce a latent variable which represented the outcome variable for each model. Missing data were handled using the Maximum Likelihood (ML) imputation method. Analyses of the CAMI revealed a significant difference across each wave due to the changes in the wording of some of the items. To account for this, the confirmatory factor analysis used a random effect of time. Statistical significance was determined by an alpha value of less than 0.05 and a Comparative Fit Index (CFI) of 0.9 was used as a cut-off for the CFA.

Results

Overall, there were 4272 participants across the three waves. After removing 754 responses due to incomplete entries and missing data, 3518 participants (incl. n = 723 Māori participants) were left for data analysis. A summary of demographic information of participants is shown in Table 1.

Table 1.

Summary of demographic variables of participants used for regression analyses.

	Factor	Count (%)
Overall	Overall	3518
Time	2015	580 (16)
	2016	1646 (47)
	2018	1292 (37)
Ethnicity	European	2175 (62)
	Asian	367 (10)
	Māori	723 (21)
	Pacific	616 (18)
	MELAA	20 (0.6)
Age	15–19	477 (14)
	20–34	902 (26)
	35–54	1047 (30)
	55–69	630 (18)
	70+	460 (13)
Biological sex	Female	1685 (48)
	Male	1833 (52)
Education/qualification	None	657 (19)
	Secondary school	1322 (39)
	Trade/professional	618 (18)
	Undergraduate	551 (16)
	Postgraduate	283 (8.2)
Socioeconomic status	Most affluent	333 (9.5)
	2	332 (9.1)
	3	319 (9.1)
	4	310 (8.9)
	5	196 (5.6)
	6	403 (12)
	7	312 (8.9)
	8	296 (8.5)
	9	431 (12)
	Least affluent	564 (16)
Have psychological condition	Yes	590 (17)
	No	2921 (83)
	Yes	2212 (64)
Know someone with a mental illness	No	1266 (36)

CFA revealed a CFI of 0.88 for the MAKS, 0.98 for the RIBS and 0.98 for the CAMI. Hence, the CFI for both the RIBS and CAMI indicated a good fit, however, the CFI for the MAKS was below the cut-off of 0.9. Even so, despite being below the threshold, it is still considered acceptable within the context of this study.

Estimates calculated from the linear mixed-effects model between the MAKS and the explanatory variables are displayed in Table 2. Among the explanatory variables, lower mental health knowledge (represented by lower ratings on the MAKS) was observed for Māori (p < 0.05), while no further significant associations were observed for other ethnicities. Among the other explanatory variables, mental health knowledge was higher among females compared with males (p < 0.001), those with undergraduate/college (p < 0.001) and postgraduate (p < 0.001) qualifications compared with those with no formal qualification, and those with a psychological condition (p < 0.001). In contrast, lower mental health knowledge was observed among those who did not know someone with mental illness (p < 0.001) and higher SES (p < 0.05). Overall, the model revealed a marginal R² of 0.045, indicating that 4.5% of the variance for mental health knowledge was explained by the explanatory variables. In addition, the conditional R² was 0.047, indicating the 4.7% of the variance for mental health knowledge was explained when accounting for a random effect of time.

Table 2.

Linear mixed-effects model estimates between explanatory variables and each measure.

Predictors	MAKS			RIBS			CAMI
	Estimates	CI	p	Estimates	CI	p	Estimates	CI	p
(Intercept)	0.02	−0.01 to 0.05	0.223	−0.28	−0.40 to −0.16	<0.001	−0.08	−0.23 to 0.09	0.336
Ethnicity
European	−0.00	−0.02 to 0.01	0.554	−0.21	−0.28 to −0.13	<0.001	−0.00	−0.04 to 0.04	0.956
Māori	0.02	−0.00 to 0.03	0.037	−0.08	−0.14 to −0.01	0.018	0.01	−0.03 to 0.04	0.659
Pacific	−0.00	−0.02 to 0.02	0.777	0.07	−0.01 to 0.15	0.108	0.01	−0.03 to 0.05	0.685
Asian	0.01	−0.01 to 0.03	0.420	0.25	0.15 to 0.34	<0.001	0.06	0.01 to 0.11	0.022
MELAA	0.01	−0.05 to 0.08	0.662	−0.04	−0.33 to 0.25	0.796	−0.14	−0.30 to 0.01	0.070
Age	−0.00	−0.00 to 0.00	0.093	0.07	0.06 to 0.08	<0.001	0.02	0.02 to 0.03	<0.001
Sex
Male		Reference			Reference			Reference
Female	−0.03	−0.04 to −0.02	<0.001	0.00	−0.05 to 0.04	0.978	−0.03	−0.05 to −0.00	0.039
Education
None		Reference			Reference			Reference
Secondary school	−0.01	−0.03 to 0.00	0.053	−0.12	−0.19 to −0.06	<0.001	−0.02	−0.05 to 0.02	0.294
Trade/professional	−0.02	−0.03 to 0.00	0.061	−0.14	−0.21 to −0.07	<0.001	−0.02	−0.06 to 0.02	0.288
Undergraduate/college	−0.04	−0.05 to −0.02	<0.001	−0.25	−0.32 to −0.17	<0.001	−0.04	−0.08 to 0.00	0.061
Postgraduate	−0.05	−0.07 −to −0.03	<0.001	−0.27	−0.36 to −0.17	<0.001	−0.04	−0.09 to 0.01	0.094
Socioeconomic status	0.00	0.00 to 0.00	0.015	0.01	−0.00 to 0.02	0.028	−0.00	−0.01 to 0.00	0.368
Psychological condition	−0.01	−0.03 to −0.00	0.030	−0.05	−0.11 to 0.01	0.100	−0.04	−0.07 to −0.00	0.025
Not knowing someone with MI	0.03	0.02 to 0.04	<0.001	0.29	0.24 to 0.33	<0.001	0.06	0.04 to 0.09	<0.001
Marginal R2/conditional R2		0.045/0.047			0.190/0.190			0.042/0.155

Bold values indicate statistical significance as p < 0.05 as indicated in bold.

Estimates calculated from the linear mixed-effects model between the RIBS and the explanatory variables are displayed in Table 2. Lower stigmatising behaviours (represented by lower ratings on the RIBS) were observed among European (p < 0.001) and Māori (p < 0.05) ethnicities, while Asian (p < 0.001) populations reported higher stigma. Higher stigmatising behaviours towards those with mental illness were also associated with increasing age (p < 0.001), higher SES (p < 0.05) and not knowing someone with a mental illness (p < 0.001). Regarding education, there was a trend that lower levels of stigmatising behaviours were observed with greater education. Overall, the model revealed a marginal R² and conditional R² of 0.190, indicating that the explanatory variables explained 19.0% of the variance for stigma behaviours.

Estimates calculated from the linear mixed-effects model between the CAMI and the explanatory variables are displayed in Table 2. Higher stigmatising attitudes towards those with mental illness (represented by higher ratings on the CAMI) was greater with Asian ethnicity (p < 0.05), increasing age (p < 0.001) and among those who did not know someone with mental illness (p < 0.001). In contrast, stigma was lower among females (p < 0.05) and those with a psychological condition (p < 0.001). The model revealed a marginal R² of 0.042, indicating that 4.2% of the variance for community-based stigma was explained by the explanatory variables. Furthermore, the conditional R² was 0.155 when including time as a random effect and thus, increased the explanatory power of the model to 15.5%.

Discussion

The overall aim of this study was to explore whether Māori report lower levels of mental illness stigma in Aotearoa. Consistent with our hypotheses, Māori participants reported lower levels of stigma in terms of stigmatised behaviours (i.e. ratings on the RIBS) but not stigmatised attitudes (i.e. ratings of the CAMI) where no significant association was found. Even so, the findings suggest that Indigenous understandings of mental illness are not universally associated with high levels of mental illness stigma. In particular, Māori endorsing lower stigmatised behaviours towards those with mental illness reinforces literature suggesting that Māori are less likely to exclude those struggling mentally within the community (Kingi et al., 2018).

The finding in this study that mental health knowledge (as measured by the MAKS) was lower among Māori may be the result of Māori not perceiving mental illness through a Western lens. For example, the MAKS includes items such as ‘If a friend had a mental health problem, I know what advice to give them to get professional help’. For this item, Māori may be more inclined to believe that problems relating to mental health should be addressed through familial support rather than professional support, contributing to less mental health knowledge captured by the MAKS. Western constructs within the MAKS may also be less familiar to Māori and perhaps, in some sense, continue to be incongruent with their cultural beliefs and perspectives regarding mental illness. In addition, the low internal consistency of the MAKS in this study and lack of evidence of the MAKS being validated for Māori further highlight issues with this finding and caution with its interpretation is necessary.

Overall, these findings highlight how non-Western cultural heritage can be associated with lower stigmatising behaviours towards those with mental illness. This supports the idea that Māori health models, such as Te Whare Tapa Wha (Durie, 1984) and Te Wheke (Pere and Nicholson, 1991), as well as Māori understandings of mental illness may broaden Western understandings and, in doing so, reduce stigma. Furthermore, the findings challenge previous research linking Indigenous populations to higher mental illness stigma. That is, while mental illness is likely perceived and understood in a different manner that aligns with cultural beliefs among Indigenous populations, this does not mean their cultural beliefs are inherently problematic.

For Pacific populations, no significant associations were demonstrated. Even so, this is contrary to past research illustrating that Pacific populations tend to have higher levels of mental illness stigma (Ataera-Minster & Trowland, 2018; Subica et al., 2019). It is noted, however, that past research utilised priority-coded ethnicity (i.e. participants could only endorse one ethnicity), while this study used non-priority-coded ethnicity (i.e. participants could endorse multiple ethnicities). In supplementary analyses, when using priority-coded ethnicity, the only difference was that Pacific participants endorsed more stigmatised behaviours towards those with mental illness than European populations. This highlights the need for further research. In particular, research should focus on ethnicity-specific analyses given the diverse cultures representing the Pacific community.

Regarding other ethnicities, European and Asian participants reported lower and higher mental illness stigma, respectively. These data are consistent with a wealth of previous research (Ataera-Minster & Trowland, 2018; Corrigan and Watson, 2007; Misra et al., 2021; Rao et al., 2007; Subica et al., 2019). There were no significant findings for MELAA populations, which may be partly explained by the lack of representativeness of MELAA populations in the study.

The other explanatory variables were also consistent with previous work, with females reporting higher mental health knowledge and lower stigmatised attitudes (Corrigan and Watson, 2007; Yuan et al., 2016), higher education being associated with higher mental health knowledge and less discriminatory behaviours (Griffiths et al., 2008; Yuan et al., 2016), and older participants reporting higher discriminatory behaviours and stigmatised attitudes (Adewuya and Makanjuola, 2008; Aromaa et al., 2011; Yuan et al., 2016). There were mixed findings regarding SES, as high SES was associated with higher mental health knowledge but also higher stigmatised behaviours. In addition, the findings revealed that mental illness stigma was lower for individuals who had a psychological condition and higher among those who did not know someone with a mental illness.

Like all research, this study has both limitations and strengths. With respect to strengths, we utilised novel statistical methods that allowed us to use non-priority-coded ethnicity. Non-priority-coded ethnicity allows participants to endorse more than one ethnicity and, in doing so, produces a more accurate representation of ethnic identities and reduces bias (Cormack & Robson, 2010; Didham & Callister, 2012). Given that prioritised ethnicity continues to be widely employed in the literature, supplementary analyses are also provided. Limitations include the potential for social desirability bias and the absence of measures to detect this bias. Also, we did not specifically collect data on Māori conceptualisations of mental illness. We acknowledge that Māori are not a homogeneous group and that conceptualisations of mental illness likely vary widely across individuals. Furthermore, while the MAKS, RIBS and CAMI are each well-established measures, the sample populations each was validated on were predominantly white. While each has been validated beyond their initial sample population, there is limited evidence of these measures being validated with an Aotearoa sample. Given this, their applicability to Māori participants may be limited. Finally, despite the number of variables included in each model, there was still a considerable portion unexplained by the variables in the present studies.

Conclusion

Overall, this study illustrates that non-Western understandings of mental illness (in this case, Indigenous) are not universally associated with greater mental illness stigma (Misra et al., 2021). Specifically, Māori understandings in Aotearoa are linked with lower levels of stigmatised behaviours towards those with mental illness. This finding challenges how Indigenous understandings of mental illness are perceived from a Western perspective and encourages researchers to be more mindful of how ethnic associations and mental illness stigma are presented throughout the literature. Of note, given that most of our knowledge of mental illness tends to be derived from a Western perspective, researchers must consider non-Western understandings of mental illness before considering cultural beliefs as being innately synonymous with higher mental illness stigma.