Abstract

This study investigates intersectional disparities in experiencing three types of violence victimisation: stranger, acquaintance, and domestic, using five years of data from the Crime Survey for England and Wales (CSEW, N=165,661) and logistic intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA). Results show that young disabled men are more likely to experience stranger and acquaintance violence, while young disabled women are more prone to experiencing domestic violence. These results highlight the need for targeted interventions and policies to address the specific risks to and/or vulnerabilities of nuanced population groups. The study stresses the importance of incorporating intersectionality in understanding and addressing the systemic and structural marginalisation that contributes to the increased risk of violence faced by young disabled individuals.

Corresponding Author Information: Ferhat Tura, Bournemouth University. Email: [email protected]. ORCID: https://orcid.org/0000-0002-6683-2438

Introduction

According to Office for National Statistics (ONS, 2022b), violence victimisation rates in England and Wales peaked in 1995 (at 4.7%), then remained fairly flat between 2014 and 2020 (1.8-1.6%) before decreasing to 0.8% in 2022. Although the prevalence of violence victimisation is relatively low, there are many negative effects of it on victims1, ranging from emotional and/or physical impacts to isolation and withdrawal from social life (Shapland and Hall, 2007; Jackson and Gouseti, 2016; Semenza et al., 2021). Moreover, victims of violence suffer the effects for longer periods of time compared to victims of other crime types (Hanson et al., 2010). It is also well documented that members of some social groups are at higher risk for victimisation (O et al., 2017). Even while violence has decreased in many countries (Tseloni et al., 2010; Farrell et al., 2014), marginalised social groups have not benefited from this decline as much in England and Wales (Ganpat et al., 2022; Ignatans and Pease, 2016; Cooper and Obolenskaya, 2022), the United States (US; Thacher, 2004), and Sweden (Nilsson and Estrada, 2006). The economic cost of violence is also considerable. For example, the estimated total cost of violence with injury in 2015/16 was £15.5 billion (Heeks et al., 2018) and the total cost of violence in London alone in 2018-19 was £3 billion (Wieshmann et al., 2020; see also Jones et al., 2020). Therefore, the mapping, understanding, and prevention of violence are high priorities for researchers, the UK Government, police forces, and the general public, including potential victims and concerned community members.

A fundamental question for prevention purposes is: who is at risk of violence victimisation and how they are identified? Most scholars have investigated violence victimisation by focusing on individual-level risk factors (e.g., Brennan et al., 2010; Brennan, 2019; Bryant and Lightowlers, 2021; Blom and Gash, 2023) and applying lifestyle and routine activity theories (Cohen and Felson, 1979; Hindelang et al., 1978) as their theoretical frameworks (Henson et al., 2010; Tseloni and Pease, 2015). Theoretical frameworks like these emphasize individual-level interpretations of risk; an alternative is intersectionality theory (Crenshaw, 1989, 1991; Collins, 1990). Intersectionality is a critical theory that provides an important framework for understanding the root social drivers of victimisation risk, including sexism, racism, ableism, and other systems of oppression, and consequently how experiences of victimisation come to be patterned unequally across social groups.

Intersectionality also encourages a holistic understanding of the diverse drivers of these inequalities, and expansion of analyses to consider greater diversity. Although some studies have attempted to do so by including interaction terms in their single-level regression models, in order to estimate unique levels of victimisation experiences in multiple intersectional social groups (e.g., Cooper and Obolenskaya, 2022; Whitfield et al., 2021; Gonçalves and Matos, 2020; MacQueen, 2016), there are limitations to the number of terms one can practically accommodate before these models become prohibitively complex to interpret (Evans et al., 2024). Furthermore, the inclusion of interaction terms implicitly reinforces the idea that the reference groups (such as White, middle-aged men) are the “norm”. Such a practice can inadvertently centre these groups as the standard against which others are compared, potentially skewing the understanding of social dynamics and inequalities. However, individuals have intersecting identities which can result in increased or decreased advantage (e.g., young Black disabled women or old White non-disabled women), such as, in our case, the risk of experiencing stranger, acquaintance, or domestic violence victimisation. For example, Evans (2022) argues that multiply marginalised disabled people are especially disadvantaged and overlooked within current Criminological research because of an emphasis on unity in the disabled people’s movement in the UK. This echoes Healy’s (2019) suggestion that the dominant ethos of a homogenous disabled people’s movement, which has engendered success in campaigning and policy-influencing, and the recognition of disability as a meaningful category of identity, has led to a general lack of criminological intersectionality research looking at disability, and a lack of awareness broadly of the diversity of disability and impairments. Evans cites Goll (2020) who opines the lack of recognition and engagement with Black disabled people’s experiences as an example of this.

This study therefore investigates three separate types of violence victimisation as (1) there is limited literature on the intersectional nature of victimisation experiences, (2) most studies that consider intersectional identities have focused on intimate partner or sexual violence (Ivert et al., 2020; Gonçalves and Matos, 2020; MacQueen, 2016) and whether an intersectional group is high or low risk may depend on the type of victimisation, and (3) “violence provides an especially rich entry point for studying the theoretical and political contours of intersectionality” more generally (Collins, 2017: 1460). For this, the present study analyses a “gold-standard” victimisation survey (i.e., the Crime Survey for England and Wales; Flatley, 2014) using a “gold-standard” method (Merlo, 2018) that was proposed and developed by Evans et al. (2018; 2024): multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). Specifically, the current study (1) maps inequalities in the predicted probability of experiencing stranger, acquaintance, and domestic violence victimisation across intersectional social groups defined by socio-economic and demographic characteristics (see Methods section), and (2) provides estimates that are intersectionally diverse and more robust than previous analyses using conventional single-level regression with interactions. By doing so, we are better able to understand the dynamics behind violence victimisation in England and Wales, and thereby contribute to the evidence base needed to develop appropriate strategies to tackle intersecting systems of oppression such as sexism, racism, classism, and disablism2 to reduce violence victimisation.

Literature review

Intersectionality was first introduced as a theoretical framework and research method by Crenshaw (1989; 1991) in her ground-breaking work on black women’s experiences in legal, social, and justice systems in the US. Black women were marginalised and discriminated against within systems that failed to consider the combination of their identities, such that their experiences were framed either as women or as black people, but not both. Intersectional research provides a novel approach to “understanding and analysing the complexity in the world, in people, and in human experiences” (Collins and Bilge, 2020: 2) by considering the multiplicative ways in which identities come together, and the impact of social systems of power and oppression on them. It quickly became a “travelling” theory spreading to disciplines beyond black feminist and critical race studies (Knapp, 2005). Examining power relations is a necessary component of intersectionality work (Healy, 2024), without which intersectionality risks being nothing more than a “buzzword” (Davis, 2008). Despite criticisms of stripping intersectionality from its roots in black feminism (Alexander Floyd, 2012) as its popularity spread, Collins and Bilge (2020) encourage its use within a range of research areas related to social justice and social inequality.

Potter (2013) rightly contends that intersectionality is a natural fit for researching experiences of crime and victimisation, given its focus on interactive identities, hierarchies of power, oppression and inequalities. Healy and Colliver (2022) however note there is a lack of criminological intersectional research and analysis, despite substantial evidence that race, gender, and class substantially impacts on people’s experiences of the criminal justice system. Additionally, individual impairments and disability significantly increase risk of victimisation, particularly for disabled women (Healy 2021; Ballan et al 2014; Brownridge, 2006; Díaz-Faes et al., 2023). Hence, individual experiences as victims of crime can be understood better in terms of how ethnicity/race, gender, socio-economic status, disability, and other identities/positionalities intersect. The present paper therefore complements an emerging body of quantitative criminological research investigating intersectional inequalities, including trust in the police in England and Wales (Tura et al., 2024) and sentence severity in the US (Pina-Sánchez and Tura, 2024).

There is some evidence of the benefits of applying an intersectional lens to the study of violence victimisation from research into hate crimes and domestic abuse. Disabled women are at greater risk of victimisation than disabled men and non-disabled women, for example, and they are also at greater risk of sexual violence, sexual assault and stalking (Smith et al., 2011; Hughes et al., 2012; Balderston, 2013; Pettitt et al., 2013; Thiara and Hague, 2013; Williams and Tregidga 2014; McCarthy, 2017; Healy, 2021). In a quantitative intersectional study using national longitudinal data in the US, Bones (2013) found increased risk of violent and sexual victimisation for disabled women, particularly those of lower incomes. Ballan et al. (2014) conducted research on disabled women of colour and identified partners as most likely perpetrators, hence demonstrating the links between disability, gender, social class and risk of victimisation, as well as type of victimisation. There is less research on disabled men, though what exists suggest they are at higher risk of violent victimisation than non-disabled men (Cohen et al., 2006; Bones, 2013; Hahn et al., 2014; Sikweyiya et al., 2022). Mueller et al. (2019) explored how violence is patterned at the intersections of race and disability in the US. They identified how racism and ableism combined to increased risk of violence towards black and brown communities (see also Castillo, 2024). Further research by Kattari et al. (2021) found intersections of gender identity and disability significantly increased the risk of victimisation within healthcare settings in the US, with similar findings by Fredriksen-Goldsen et al. (2014). Díaz-Faes et al. (2023) found that individuals with intellectual disabilities living in Spain were four times more likely to become poly-victims (victimised repeatedly).

Fundamentally, any intersectional approach enables us to see how complex positionalities within interlocking social processes and systems of oppression create (potentially unique) risk profiles. While men are more likely to experience victimisation than women when we consider all types of victimisations together, differentiating by violence type/source shows that women are more likely to experience intimate partner violence and disabled women experience higher rates of victimisation than disabled men, highlighting the contribution of intersectionality to victim research (Bones, 2013). Notions of violence towards disabled people can also be considered along a continuum of violence, such as within a hate crime framework (Hollomotz, 2013). Furthermore, conceptualisations of violence are also dependent on social relationships, institutions and culture (Goodley and Runswick-Cole, 2011), including who is defined within group identities. Intersectionality provides an opportunity for understanding “the multiple ways in which prejudice and violence might be experienced” (Mason-Bish, 2015: 25) and hence also offers an opportunity for informing improved targeting of resources to address violent victimisation.

Criticisms of intersectionality in quantitative research have suggested that many researchers fail to define or explicate what intersectionality is, and its role in challenging social power (Bauer et al., 2021). Some critics suggest there is a lack of engagement on structural level constructs such as power and oppression (Beccia et al., 2021). Further, many studies are considered too simplistic, using basic statistical analytical techniques, relying on descriptive statistics, regression with interactions, or stratified analyses. This paper addresses these flaws by positioning intersectionality in theoretical and methodological terms. It recognises that social positions are hierarchical and reflective of power; hence experiences at the individual level are often reflected in systemic and structural levels, such as through racism, classism, sexism, and disablism. Furthermore, it utilises MAIHDA (Evans et al., 2024) as a method of examining quantitative data in an intersectional (as opposed to additive) manner (Hancock, 2007). Therefore, the present paper is the first study of its kind to use MAIHDA to examine intersectional inequalities in experiencing three types of violence: stranger, acquaintance, and domestic in England and Wales.

Methods

Data source

We merged five years of data from the Crime Survey for England and Wales (CSEW), 2015/16 (April 2015 – March 2016) to 2019/20 (April 2019 – March 2020; Office for National Statistics, 2021a, 2021b, 2021c, 2022a, 2023a). The reason for merging the datasets was to increase the sample sizes for the newly created intersectional social strata (see later the “Constructing the strata ID” section). The CSEW, formerly known as the British Crime Survey, is a nationwide face to face crime survey for England and Wales. It employs a stratified multistage cross-section sample design and asks residents of households about their experiences of a range of crimes in the 12 months immediately prior to the interview. This enables the recruitment of a representative sample of the adult population (16 years or older) living in private accommodation in England and Wales in each year. A limitation, however, is that “the CSEW does not cover the population living in group residences (for example, care homes or student halls of residence) or other institutions, nor does it cover crime against commercial or public sector bodies” (ONS, 2024d). Implications of this limitation are discussed in the “Discussion and Conclusion” section.

The stages of the multistage design of the CSEW involve the following geographical strata: selection of constituencies; within each selected constituency, selection of postcode sectors; within each postcode sector, selection of segments; within each selected segment, selection of addresses; and, in case of multi-occupancy selected addresses, selection of household. At the last stage, one adult member per household is randomly selected by the interviewers to undertake the survey (Tseloni and Tilley, 2016).

The datasets we analyse recorded victimisation experiences prior to the COVID-19 pandemic and are accredited official statistics (ONS, 2024c). No major disruptions (e.g., economic, political) that might affect the violence victimisation experiences examined here occurred in the UK during this period. Also, there were no major changes to the survey design during this period (ONS, 2024a). In the 10 years prior to the pandemic, the response rate for the CSEW fluctuated between 70% to 76% (see Figure 1 in ONS, 2024b). Implications of this are discussed in the “Discussion and Conclusion” section.

The CSEW questionnaire for the adult survey has a complex structure, consisting of a set of core modules (e.g., crime screener questionnaire) asked of the whole sample, a set of modules asked only of random sub-samples (e.g., performance and experiences of the Criminal Justice System), victimisation module completed only by victims, and self-completion modules completed by those who are 16 to 74 year olds (e.g., victims of domestic violence, sexual victimisation and stalking; 16 to 59 years old prior to October 2016). The crime screener questionnaire is administered to the entire sample and asks whether respondents have experienced certain types of crimes or incidents within the last 12 months. All incidents identified at the screener questionnaire are followed up in more detail in the victimisation module (Tilley and Tseloni, 2016; Ganpat et al., 2022). We use the screener questionnaire for the analysis because we are interested in the frequency of violence victimisation amongst the entire sample to investigate intersectional inequalities in experiencing violence.

Outcome variables

There are three binary dependent variables used in the analysis: stranger violence (0=no, 1=yes), acquaintance violence (0=no, 1=yes), and domestic violence (0=no, 1=yes). Stranger violence occurs in incidents where the victim did not know the offender(s) and had never seen them before. Acquaintance violence occurs when the victim knew one or more of the offenders at least by sight (for example, neighbours and local children, colleagues, clients or members of the public contacted through work, friends, and acquaintances), excluding household members or former intimate partners. Domestic violence occurs when the violence is perpetrated by a household member, within intimate relationships or by a partner or ex-partner. However, “it is important to note that sexual violence and domestic violence reported in the interviewer-led parts of the CSEW are prone to considerable under-reporting. This is because many victims will not be willing to disclose these incidents in an in-home, face-to-face personal interview” (ONS, 2024d). Therefore, the results for domestic violence in this paper should be read with caution.

Dimensions of social strata

To create contexts or intersectional social strata, we used five categorical socio-economic and demographic characteristics: sex, ethnicity, age, socio-economic status (SEC), and disability. Sex is a dichotomous variable with categories of (1) female and (2) male. Ethnicity is a categorical variable with four identity-options: (1) White, (2) Asian or Asian British, (3) Black or Black British, and (4) Mixed/Chinese or Other. Age is a categorical variable with three categories of (1) 16-24, (2) 25-44, and (3) 45-plus years. SEC is a categorical variable with four categorical levels: (1) higher managerial, administrative and professional occupations, (2) intermediate occupations, (3) routine and manual occupations, and (4) never worked and long-term unemployed. Finally, disability is a dichotomous variable with categories of (1) no long-standing illness, and (2) long-standing illness. The definition of disability used is consistent with the core definition of disability under the Equality Act 2010, whereby a person is considered to have a disability if they have a long-standing illness, impairment or condition that causes difficulty with day-to-day activities (ONS, 2024d). See Table 1 for descriptive statistics.

Table 1: Descriptive statistics (N= 165,661)

Constructing the strata ID

To conduct a MAIHDA analysis, we create a unique ID for each stratum or group. This method sees the data as having two levels: individual respondents (level 1) are treated as nested in the intersectional social categories (or “social strata”) they belong to (level 2). By treating social strata as the second level, we theorise them as proxies for positionalities or “contexts”, rather than just traits of individuals. This approach allows us to analyse intersectional social strata within a multilevel framework (Evans et al., 2024).

We created 192 different intersectional strata by combining the two categories of sex, the four categories of ethnicity, the three categories of age, the four categories of SEC, and the two categories of disability (i.e., 2x4x3x4x2=192 strata). In our data, one combination is empty (Male, Black or Black British, age 16-24 years, Intermediate SEC, Disabled), leaving 191 strata in the analysis. Strata were given unique 5-digit ID codes to make identification easier, with each digit corresponding to one of the five variables. For example, the stratum 23142 includes individuals who are “Female (sex=2), Black (ethnicity=3), aged 16-24 (age=1), never worked (SEC=4), with long-standing illness (disability =2)” (see Table 1).

Table 2: Number of individuals per intersectional social stratum, defined by respondent gender, race/ethnicity, age, SEC, and disability (n = 191).

To evaluate whether the sample size is sufficient across various intersections, we also calculated the percentage of strata that have at least X respondents, where X is 10, 20, 30, 50, or 100 respondents (Evans et al., 2024). Simulation studies suggest MAIHDA produces more reliable and accurate estimates than single-level regression models even when sample sizes are smaller than conventional wisdom would recommend (Mahendran et al., 2022a, 2022b; Bell et al., 2019; Van Dusen et al., 2024). In our case, 87% of the strata have at least ten respondents, 78% have at least twenty, and 71% have at least thirty, suggesting that most strata have a sufficient sample size for reliable estimates (see Table 2).

Statistical analysis

Following the MAIHDA methodology, we employed logistic multilevel regression techniques to assess inequities in experiencing each of the three types of violence victimisation across social strata. Individual observations (level 1, n=165,661) were clustered within intersectional social strata (level 2, n = 191) (Evans et al., 2018; 2024).

We first fitted a “null” or empty logistic multilevel model, which only included a random intercept to allow violence victimisation to vary by social strata. This model allows the calculation of the Variance Partition Coefficient (VPC; Goldstein et al., 2002; Leckie et al., 2020), which provides a measure of between-stratum inequality. To calculate the VPC for a logistic multilevel model, we used the latent variable method proposed by Goldstein et al. (2002).

Next, we fitted an additive main effects model (or “full” model), incorporating the categorical variables defining the strata as main effects predictors with fixed regression coefficients. This model captures two- and higher-way interaction effects between the predictors via the stratum random effect.

Additionally, the Proportional Change in Variance (PCV) was calculated to assess the relative contributions of additive versus interaction effects. Specifically, the PCV expresses the extent to which the main effects explain the overall variation in mean outcomes between the strata. A PCV meaningfully less than 100% suggests that interaction effects are crucial for accurately characterising observed inequalities between strata, indicating that at least some strata have predicted victimisation rates that substantially depart from those implied by “universal” or general patterns of additive inequalities (Evans et al., 2024).

Predicted rates of experiencing violence victimisation by each stratum were reported as percentages throughout (by multiplying with 100). We also calculated predicted counts, which represents the number of individuals within a stratum who are predicted to experience victimisation, by multiplying the predicted rate (converted to a proportion by dividing by 100) by the stratum sample size.

All data management and statistical analyses were performed using R (version 4.2.3). All data used in this paper can be accessed via the UK Data Service (Office for National Statistics, 2021a, 2021b, 2021c, 2022a, 2023a) and replication codes are available on GitHub (anonymised for peer-review: https://anonymous.4open.science/r/MAIHDA-violence-4B95/README.md).

Results

Table 3 presents results from the null and full multilevel logistic models fitted separately for stranger violence, acquaintance violence, and domestic violence. The general contextual effect of strata in the null logistic models, measured by the VPC, was 20% for stranger violence, 17% for acquaintance violence, and 26% for domestic violence. This suggests that the axes of intersectional identity and positionality included in this analysis capture a substantial amount of the variation in risk in the sample as a whole, for all three outcomes. In other words, there are substantial between-strata differences in risk of violence victimisation, even acknowledging that individual-level risk also varies within these strata.

After the social strata variables were included in the full models as additive main effects dummy variables, we compared the between strata-variance from full models to null models by calculating the PCV. We see that 98%, 93% and 96% of the between-stratum variances observed in the null models is accounted for by the additive main effects for stranger, acquaintance, and domestic violence, respectively. In other words, after accounting for the additive main effects, 2% (100% - 98%), 7% (100% - 93%), and 4% (100% - 96%) of the original between-stratum variance (or inequalities) are unexplained by the additive main effects, which indicates that meaningful interaction effects are required to adequately characterise the inequality patterns observed (Evans et al., 2024). In short, although the majority of between-strata differences were explained by the additive main effects, the remaining 2%, 7%, and 4% variance unexplained by the main effects shows us that some strata are predicted to be higher/lower than generally expected, above and beyond whatever typical, additive patterns of inequality we have already accounted for.

While we consistently found substantial inequalities across models, we also found substantial differences in who was at greater risk by violence type--stranger, acquaintance, and domestic violence victimisation. To see this, we focus next on the main effects patterns.

The results from the full logistic multilevel models for stranger, acquaintance, and domestic violence show significant differences in the odds ratios (OR) for various socio-economic and demographic factors. For gender, women have significantly lower odds of experiencing stranger violence (OR = 0.37, 95% CI: 0.32–0.43) and acquaintance violence (OR = 0.71, 95% CI: 0.59–0.85) compared to men, but substantially higher odds of experiencing domestic violence (OR = 2.27, 95% CI: 1.78–2.90). Ethnicity also plays a crucial role, with Asians having lower odds of experiencing all three types of violence compared to whites, particularly domestic violence (OR = 0.45, 95% CI: 0.26–0.79). Age is another significant factor; individuals aged 25-44 and 45 plus have substantially lower odds of experiencing all types of violence compared to those aged 16-24, with the most pronounced drop observed in domestic violence for those aged 45 plus (OR = 0.16, 95% CI: 0.11–0.22). For socio-economic status (SEC), those with routine occupations and those who have never worked have lower odds of experiencing stranger violence (OR = 0.79, 95% CI: 0.66–0.94) and (OR = 0.63, 95% CI: 0.44–0.92), respectively. Importantly, individuals with long-standing disabilities have significantly higher odds of experiencing all types of violence, with the highest odds observed in domestic violence (OR = 3.04, 95% CI: 2.40–3.86). We also visually investigate these findings in Figure 1, which suggests that regression relationships are basically consistent across violence types, except for gender and to a lesser extent disability.

Table 3: Results from MAIHDA logistic regressions predicting stranger, acquaintance, and domestic violence

* p<0.05   ** p<0.01   *** p<0.001. In the null model columns, the Likelihood Ratio Tests (LRTs) compare the null logistic MAIHDA models to their conventional logistic regression counterparts. In the full model columns, the LRTs compare the full logistic MAIHDA models to the null logistic MAIHDA models. VPC: Variance Partition Coefficient, PCV: Proportional Change in Variance, AIC: Akaike Information Criterion, AUC: Area Under the Curve. ~ Please note that the value 3.29 is the fixed variance of a standard logistic distribution. To calculate the VPC for a logistic multilevel model, we used the latent variable method proposed by Goldstein et al. (2002).

Figure 1: Comparison of coefficients and 95% CIs across models (full model – table 3)

To examine inequality patterns with greater attention to intersectional differences, we ranked the strata (low to high) for each outcome based on the predicted probabilities of experiencing stranger, acquaintance, and domestic violence (1=highest; 191=lowest) and compared these ranks across the three models (see Appendix figures 1-3 for a visual overview of the inequalities and Appendix table 1 – these figures and table can be downloaded via this link: https://anonymous.4open.science/r/MAIHDA-violence-4B95/README.md). The correlation analysis (see Figure 2) showed a strong positive correlation between the ranks for acquaintance and domestic violence (R = 0.80). In contrast, the correlation between the ranks for stranger violence and the other two types of violence was moderate (R = 0.49 for stranger vs. domestic) to strong (R = 0.86 for stranger vs. acquaintance). This suggests that while some strata may rank high for both stranger and acquaintance violence, others may rank high for one type and low for another, highlighting the complex and varied nature of crime victimisation across different groups. Importantly, Figure 2 shows that young disabled individuals are at a greater risk of experiencing violence.

Figure 2: Scatterplot matrix of the ranks of the predicted stratum means across models (full model)

The analysis then identified the top 10 strata with the highest and lowest predicted probabilities of victimisation for each type of violence (stranger, acquaintance, and domestic; see Table 4). The ranks of the top five strata with the highest predicted rates of victimisation and predicted counts were then visually analysed across the models (Figure 3). The results in the predicted rates figure showed that the top five strata that rank high for stranger violence tend to also rank high for acquaintance violence but not for domestic violence, reinforcing the findings from the correlations reported above. For example, stratum 14112 (men, mixed or Chinese/Other, aged 16-24, higher managerial, and disabled) rank high for stranger violence (rank=1) and acquaintance violence (rank=14) but far less high for domestic violence (rank= 44). The top five (and also 10, see Table 4) strata for stranger and acquaintance violence consist of young (aged 16-24) disabled men with different ethnic backgrounds and SEC status. Conversely, the top five strata that rank high for domestic violence tend to rank low for stranger violence but similarly for acquaintance violence, reinforcing the findings from the correlation matrix analysis above. For example, stratum 21132 (women, white, aged 16-24, routine and manual occupation, and disabled) ranks high for domestic violence (rank= 1) and acquaintance violence (rank= 11) but relatively low for stranger violence (rank= 55). Top 5 (and also 10, see Table 4) strata for domestic violence consist of young (aged 16-24) disabled women from mostly white ethnic background (but also black and mixed/Chinese/other) and with various SEC status.

When we look at the predicted count figure, which provides insight into the overall impact within each stratum, the story changes slightly. Although men and women still experience the highest rates of stranger and domestic violence, respectively, the strata that rank highest for predicted counts are different. This is because predicted counts consider the size of each stratum, leading to a higher total number of predicted victims in larger strata. The differences between these two metrics are crucial for planning interventions. Strata with high predicted rates of victimisation indicate where individuals are most at risk and may benefit from targeted interventions. On the other hand, strata with high predicted counts highlight where the overall impact is greatest, guiding resource allocation to manage the broader effects of violence.

Table 4: Inspection of 10 highest and lowest ranked strata for predicted percent violence victimisation by type

Stratum ID position and respective numbers are as follows: (1) Sex:1=Male, 2=Female;

(2) Ethnicity :1=White, 2=Asian or Asian British, 3=Black or Black British, 4=Mixed, Chinese or Other;

(3) Age:1=16-24, 2=25-44, 3=45+;

(4) SEC:1=Higher managerial, administrative and professional occupations, 2=Intermediate occupations, 3=Routine and manual occupations, 4=Never worked and long-term unemployed;

(5) Disability:1=No long-standing illness, 2=Long standing illness.

Figure 3: Stratum line plot of top 5 stratum with highest predicted percent and count across different violence models (full model)

We then identified outlier strata with respect to showing different rankings across the three types of violence. For this, we calculated the Mahalanobis distance for each stratum based on their ranks across the three violence models (stranger, acquaintance, and domestic). The Mahalanobis distance measures the distance between a point and the mean of a distribution, considering the correlations between variables. Strata with high Mahalanobis distances are considered outliers, as they rank very differently across the models. We identified the top 5% of strata with the highest Mahalanobis distances as outliers. We finally visualised the ranks of these outlier strata across the models to better understand their distribution. In short, the outlier strata exhibit significant variability in their ranks across the models and strata with a high probability of experiencing domestic violence tend to have a lower probability of experiencing stranger and acquaintance violence. However, while top strata with a high probability of experiencing domestic violence in Figure 4 tend to have a lower probability of experiencing stranger violence, stratum 22112 (women, Asian, aged 16-24, higher managerial, disabled) have a higher probability of experiencing stranger violence as well.

Figure 4: Outliers in stratum ranks across different violence models (full model)

We also plotted the predicted random effects from the full models across the strata to visualise the variability and identify patterns in victimisation risks that are not explained by the main effects alone (see Appendix figures 4-6 and Appendix table 2 – these figures and table can be downloaded via this link: https://anonymous.4open.science/r/MAIHDA-violence-4B95/README.md). Across all three violence types, the predicted random effects were rarely significant. We did not find any intersectional effect for stranger violence based on CIs not encompassing zero. However, stratum 21311 (women, white, aged 45 plus, higher managerial, no disability) experienced higher-than expected acquaintance violence. Similarly, stratum 21211 (women, white, aged 25-44, higher managerial, no long-standing illness) experienced lower-than expected domestic violence. These results suggest that only at a few intersections do we have statistical evidence that the percentage of individuals experiencing violence differs from what is expected based on the general patterns of additive main effects. Given that we have found strong statistical evidence of interactions at play (see Table 3 – row LRT), the difficulty in identifying the specific intersections where these interactions occur points to an issue of statistical power. With the characteristics of these data, larger samples of individuals per intersection are needed to achieve the precision required to pinpoint where the interactions are operating.

Discussion and conclusion

The primary aim of this study was to investigate the effect of intersecting socio-economic and demographic factors on the probability of experiencing different types of violence (i.e., stranger, acquaintance, and domestic) using intersectional MAIHDA. Our analysis resulted in several important findings. Firstly, and not unexpectedly, women are more likely to experience domestic violence compared to men. This aligns with existing literature that highlights the increased vulnerability of women to domestic violence (ONS, 2023b; Smith et al., 2011; Hughes et al., 2012; Balderston, 2013; Pettitt et al., 2013; Thiara and Hague, 2013; Williams and Tregidga 2014; McCarthy, 2017; Healy, 2021). Men, on the other hand, are more likely than women to experience violence from strangers and acquaintances—a clear reversal of who is “high risk” by type of violence experienced. Importantly, disabled people are more likely to experience all types of victimisations compared to non-disabled individuals, which again aligns with previous research (Cohen et al., 2006; Bones, 2013; Hahn et al., 2014; Sikweyiya et al., 2022). These results emphasise the need for targeted interventions to protect this at-risk group.

The intersectional strata analysis provided further insights into the specific groups at higher risk. All the top 10 strata with the highest predicted probability of experiencing stranger and acquaintance victimisation include young disabled men, suggesting that this group faces significant risks in public or unfamiliar settings. In addition, most of these strata include non-white individuals, which is a finding that aligns with previous research (Mueller et al., 2019). On the contrary, the strata including young disabled women ranked high for domestic violence, indicating a critical need for support and protection within domestic environments. These findings highlight the complex interaction between gender, age, disability, and the type of violence experienced. They highlight the importance of tailored prevention and intervention strategies that address the unique needs of different demographic groups for different victimisation types.

While this study provides valuable insights into the factors associated with different types of violence, several limitations should be acknowledged. The term “disability” encompasses a wide range of impairments, each with different associated risks. This study does not differentiate between types of disabilities, which may lead to an oversimplification of the risks faced by individuals with different impairments. For example, more visible disabilities can result in increased rates of victimisation (Action for Blind People, 2008) and people with mental health and/or learning difficulties are more likely to report hate crime victimisation than those with physical disabilities (e.g. Chakraborti et al., 2014). In addition, Harrell (2021, in Castillo 2024) reported that disabled people were four times more likely to experience violence victimisation and less likely to report it than non-disabled people. Future research should consider the specific types of disabilities to provide a more detailed understanding of victimisation risks and consider underreporting rates that will exhibit strong intersectional inequalities. Similarly, the gender variable used was binary, meaning we were not able to analyse the experiences of gender-minority individuals. Due to the stratum sample size issues, we had to merge mixed and Chinese or other ethnicity categories, and the categories of the SEC variable. The CSEW does not include populations living in group residences such as care homes or student halls of residence, nor does it cover crimes against commercial or public sector bodies. This exclusion means that the findings may not be generalisable to these populations, potentially underestimating the prevalence of victimisation in these settings. This limitation may affect our findings, particularly the observation that disabled individuals are at higher risk of victimization. Disabled people residing in care homes or other group settings might face different or increased risks compared to those in our sample. These individuals could be experiencing more severe forms of disability or impairment, potentially leading to greater vulnerability. Therefore, our results should be interpreted with caution, acknowledging that they may not fully represent the experiences of disabled people in group residences. Further research is needed to explore these differences and ensure our findings are accurately contextualised. In the decade prior to the COVID-19 pandemic, the response rate for the CSEW fluctuated between 70% and 76%. This variability, which is likely itself to be patterned by intersectional factors, could introduce response bias. Certain groups may be underrepresented in the survey, affecting the accuracy and representativeness of the findings. Finally, domestic violence reports in the interviewer-led sections of the CSEW are prone to significant under-reporting. Many victims may be unwilling to disclose such incidents in an in-home, face-to-face interview setting. This under-reporting can lead to an underestimation of the true prevalence of domestic violence, and there will be intersectional patterning to this too. These limitations highlight the need for cautious interpretation of the findings in relation to domestic violence victimisation and suggest areas for improvement in future research. Addressing these issues can help provide a more accurate and comprehensive understanding of the intersectional disparities in experiencing violence.

Our results have significant policy implications, suggesting the need for targeted interventions and resource allocation to address the specific risks of various groups. Policymakers could develop specialised programmes and legislation to protect young disabled individuals from the types of violence they are most at risk for, ensuring a more tailored and effective approach to violence prevention. In practice, these findings highlight the importance of training for professionals such as social workers, healthcare providers, and law enforcement to recognise and address the distinct types of violence faced by young disabled men and women. Additionally, the development of specialised support services and community programmes can provide the necessary resources and education to prevent violence and support victims. Future research should focus on understanding the underlying causes of these patterns, including hate crime motivations, and evaluating the effectiveness of targeted interventions over time, potentially through longitudinal and interdisciplinary studies. There is insufficient space to consider these findings in relation to extensive hate crime literature, suffice to say that existing research positions both domestic and stranger violence towards disabled women as a form of hate crime (Zempi and Smith, 2022; McCarthy, 2017).

Overall, our intersectional analysis highlights those at greatest risk of victimisation, enabling policy and policing to respond more effectively, and also draws attention to the differences and power dynamics that contribute to the increased targeting of marginalised groups in an unjust and unequal society. The systemic and structural marginalisation of disabled men and women is evident, but an intersectional methodology draws attention to the unique experiences of disabled women, who face multiple layers of oppression and occupy a liminal space within both categories of identity. Structural and symbolic forms of oppression and violence are thus reproduced and enacted in the everyday violence that disabled women and men experience, institutionalising disablism, whether perpetrated by strangers, acquaintances, or family members.

References

Action for Blind People (2008), Report on Verbal and Physical Abuse Towards Blind and Partially Sighted People Across the UK. London: Action for Blind People. Available at: https://www.communitycare.co.uk/2008/01/07/action-for-blind-people-visually-impaired-at-greater-abuse-risk/

Alexander-Floyd, N. G. (2012), ‘Disappearing Acts: Reclaiming Intersectionality in the Social Sciences in a Post-Black Feminist Era’, Feminist Formations, 24(1): 1-25.

Balderston, S. (2013), ‘Victimized Again? Intersectionality and Injustice in Disabled Women’s Lives After Hate Crime and Rape’, in M. Texler Segal and V. Demos, eds., Gendered Perspectives on Conflict and Violence: Part A (Advances in Gender Research, Volume 18A), 17-51. Emerald Group Publishing Limited.

Ballan, M.S., Freyer, M.B., Marti, C.N., Perkel, J., Webb, K.A., and Romanelli, M. (2014), ‘Looking Beyond Prevalence: A Demographic Profile of Survivors of Intimate Partner Violence with Disabilities’, Journal of Interpersonal Violence, 29(17): 3167-3179.

Bauer, G.R., Churchill, S.M., Mahendran, M., Walwyn, C., Lizotte, D., and Villa-Rueda, A.A. (2021), ‘Intersectionality in Quantitative Research: A Systematic Review of Its Emergence and Applications of Theory and Methods’, SSM-Population Health, 14: 100798.

Beccia, A.L., Baek, J., Austin, S.B., Jesdale, W.M., and Lapane, K.L. (2021), ‘Eating-Related Pathology at the Intersection of Gender Identity and Expression, Sexual Orientation, and Weight Status: An Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) of the Growing Up Today Study Cohorts’, Social Science & Medicine, 281: 114092.

Bell, A., Holman, D., and Jones, K. (2019), ‘Using Shrinkage in Multilevel Models to Understand Intersectionality: A Simulation Study and a Guide for Best Practice’, Methodology, 15(2): 88-96.

Blom, N., and Gash, V. (2023), ‘Measures of Violence within the United Kingdom Household Longitudinal Survey and the Crime Survey for England and Wales: An Empirical Assessment’, Social Sciences, 12(12): 649.

Bones, P. D. C. (2013), ‘Perceptions of Vulnerability: A Target Characteristics Approach to Disability, Gender, and Victimization’, Deviant Behavior, 34 (9): 727-750.

Brennan, I.R. (2019), Victims of Serious Violence in England and Wales, 2011–2017. Available at: https://library.college.police.uk/docs/college-of-policing/Victim_of_serious_violence.pdf

Brennan, I.R., Moore, S.C., and Shepherd, J.P. (2010), ‘Risk Factors for Violent Victimisation and Injury from Six Years of the British Crime Survey’, International Review of Victimology, 17(2): 209-229.

Brownridge, D. A. (2006), ‘Partner Violence Against Women with Disabilities: Prevalence, Risk, and Explanations’, Violence Against Women, 12(9): 805-822.

Bryant, L., and Lightowlers, C. (2021), ‘The Socioeconomic Distribution of Alcohol-Related Violence in England and Wales’, PLoS One, 16(2): e0243206.

Castillo, I. E. (2024), ‘Disability’s Dark Figure of Violence: Reporting Variations Across Intersections of Disability Type, Race, Ethnicity, and Sex’, Criminal Justice and Behavior, 51(10)

Chakraborti, N., Garland, J., and Hardy, S.J. (2014), The Leicester Hate Crime Project: Findings and Conclusions. University of Leicester. Available at: https://le.ac.uk/hate-studies/research/the-leicester-hate-crime-project

Cohen, L., and Felson, M. (1979), ‘Social Change and Crime Rate Trends: A Routine Activity Approach’, American Sociological Review, 44(August): 588-608.

Cohen, M.M., Forte, T., Du Mont, J., Hyman, I., and Romans, S. (2006), ‘Adding Insult to Injury: Intimate Partner Violence Among Women and Men Reporting Activity Limitations’, Annals of Epidemiology, 16(8): 644-651.

Collins, P.H. (1990), ‘Black Feminist Thought in the Matrix of Domination’ in P. H. Collins, Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment, 221-238. Boston: Unwin Hyman

Collins, P., and Bilge, S. (2020), Intersectionality, 2^nd ed. Cambridge: Polity Press.

Collins, P.H. (2017), ‘On Violence, Intersectionality and Transversal Politics’, Ethnic and Racial Studies, 40(9): 1460-1473.

Cooper, K., and Obolenskaya, P. (2022), Who Is at Risk of Experiencing Violence and Has It Changed Overtime? Social Policies and Distributional Outcomes in a Changing Britain, Research Note 5. London: LSE. Available at: https://sticerd.lse.ac.uk/dps/case/spdo/spdorn05.pdf

Crenshaw, K. (1989), ‘Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’, University of Chicago Legal Forum, 1(8): 139-67.

Crenshaw, K. (1991), ‘Mapping the Margins: Intersectionality, Identity Politics, and Violence Against Women of Color’, Stanford Law Review, 43(6): 1241-1299.

Davis, K. (2008), ‘Intersectionality as Buzzword: A Sociology of Science Perspective on What Makes a Feminist Theory Successful’, Feminist Theory, 9(1): 1464-7001.

Díaz-Faes, D. A., Codina, M., and Pereda, N. (2023), ‘Experiences of Bias Victimization Among People with Intellectual Disabilities’, Journal of Interpersonal Violence, 38(15-16): 9423-9437.

Evans, C.R., Leckie, G., Subramanian, S.V., Bell, A., and Merlo, J. (2024), ‘A Tutorial for Conducting Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA)’, SSM-Population Health, 26: 101664.

Evans, C.R., Williams, D.R., Onnela, J.P., and Subramanian, S.V. (2018), ‘A Multilevel Approach to Modeling Health Inequalities at the Intersection of Multiple Social Identities’, Social Science & Medicine, 203: 64-73.

Evans, E. (2022), ‘Political Intersectionality and Disability Activism: Approaching and Understanding Difference and Unit’, The Sociological Review, 70(5): 986-1004.

Farrell, G., Tilley, N., and Tseloni, A. (2014), ‘Why the Crime Drop?’, Crime and Justice, 43(1): 421-490.

Flatley, J. (2014), ‘British Crime Survey’, in G. Bruinsma and D. Weisburd, eds., Encyclopaedia of Criminology and Criminal Justice (ECCJ), 194-203. New York: Springer-Verlag.

Fredriksen-Goldsen, K.I., Cook-Daniels, L., Kim, H.J., Erosheva, E.A., Emlet, C.A., Hoy-Ellis, C.P., Goldsen, J., and Muraco, A. (2014), ‘Physical and Mental Health of Transgender Older Adults: An At-Risk and Underserved Population’, The Gerontologist, 54(3): 488-500.

Ganpat, S.M., Garius, L., Tseloni, A., and Tilley, N. (2022), ‘Violence and the Crime Drop’, European Journal of Criminology, 19(4): 767-790.

Goll, K. (2020), The Missing Piece of the Movement, Available at: www.lappthebrand.com/blogs/perspectives/the-missing-piece-of-the-movement-black-disabled-activism.

Goldstein, H., Browne, W., and Rasbash, J. (2002), ‘Partitioning Variation in Multilevel Models’. Understanding Statistics: Statistical Issues in Psychology, Education, and the Social Sciences, 1(4), 223-231.

Gonçalves, M., and Matos, M. (2020), ‘Interpersonal Violence in Immigrant Women in Portugal: An Intersectional Approach’, Journal of Immigrant and Refugee Studies, 18(1): 22-41.

Goodley, D., and Runswick-Cole, K. (2011), ‘The Violence of Disablism’, Sociology of Health and Illness, 33(4): 602-617.

Hahn, J.W., McCormick, M.C., Silverman, J.G., Robinson, E.B., and Koenen, K.C. (2014), ‘Examining the Impact of Disability Status on Intimate Partner Violence Victimization in a Population Sample’, Journal of Interpersonal Violence, 29(17): 3063-3085.

Hanson, R.F., Sawyer, G.K., Begle, A.M., and Hubel, G.S. (2010), ‘The Impact of Crime Victimization on Quality of Life’, Journal of Traumatic Stress: Official Publication of The International Society for Traumatic Stress Studies, 23(2): 189-197.

Harrell, E. (2020), Crime Against Persons with Disabilities, 2009-2019 – Statistical Tables. US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Available at: https://bjs.ojp.gov/library/publications/crime-against-persons-disabilities-2009-2019-statistical-tables

Healy, J. (2019), ‘Thinking Outside the Box: Intersectionality as a Hate Crime Research Framework’, Papers from the British Criminology Conference, 19: 60-83. Available at: https://www.britsoccrim.org/pbcc2019/

Healy, J., and Colliver, B. (eds.) (2022), Contemporary Intersectional Criminology: Examining the Boundaries of Intersectionality and Crime. Bristol: Bristol University Press.

Healy, J. C. (2024), ‘Revealing the Benefits, Barriers, and Prevalence of Intersectionality in Disability Hate Crime Research’, in L. Burch and D. Wilkin, eds., Disability Hate Crime: Perspectives for Change, London: Routledge.

Healy, J.C. (2021), ‘An Exposition of Sexual Violence as a Method of Disablist Hate Crime’, in I. Zempi and J. Smith, eds., Misogyny as Hate Crime, 178-194. London: Routledge.

Heeks, M., Reed, S., Tafsiri, M., and Prince, S. (2018), The Economic and Social Costs of Crime. London. Available at: assets.publishing.service.gov.uk/media/5b684f22e5274a14f45342c9/the-economic-and-social-costs-of-crime-horr99.pdf

Henson, B., Wilcox, P., Reyns, B.W., and Cullen, F.T. (2010), ‘Gender, Adolescent Lifestyles, and Violent Victimization: Implications for Routine Activity Theory’, Victims and Offenders, 5(4): 303-328.

Hindelang, M.J., Gottfredson, M.R., and Garofalo, J. (1978), Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, MA: Ballinger Publishing Company.

Hollomotz, A. (2013), ‘Disability and the Continuum of Violence’, in A. Roulstone and H. Mason-Bish, eds., Disability, Hate Crime and Violence, 52-63. London: Routledge.

Hughes, K., Bellis, M. A., Jones, L., Wood, S., Bates, G., Eckley, L., McCoy, E., Mikton, C., Shakespeare, T., and Officer, A. (2012), ‘Prevalence and Risk of Violence Against Adults with Disabilities: A Systematic Review and Meta-Analysis of Observational Studies’, The Lancet, 379(9826): 1621-1631.

Ignatans, D., and Pease, K. (2016), ‘On Whom Does the Burden of Crime Fall Now? Changes Over Time in Counts and Concentration’, International Review of Victimology, 22(1): 55-63.

Ivert, A.K., Gracia, E., Lila, M., Wemrell, M., and Merlo, J. (2020), ‘Does Country-Level Gender Equality Explain Individual Risk of Intimate Partner Violence Against Women? A Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) in the European Union’, European Journal of Public Health, 30(2): 293-299.

Jackson, J., and Gouseti, I. (2016), ‘Threatened by Violence: Affective and Cognitive Reactions to Violent Victimization’, Journal of Interpersonal Violence, 31(18): 2987-3016.

Jones, L., Bigland, C., and Quigg, Z. (2020), Costs of Violence. Available at: https://phwwhocc.co.uk/wp-content/uploads/2021/07/Costs-of-violence-to-the-healthcare-system-in-Wales.pdf

Kattari, S.K., Bakko, M., Langenderfer-Magruder, L., and Holloway, B.T. (2021), ‘Transgender and Nonbinary Experiences of Victimization in Health Care’, Journal of Interpersonal Violence, 36(23-24): NP13054-NP13076

Knapp, G. A. (2005), ‘Race, Class, Gender, Reclaiming Baggage in Fast Travelling Theories’, European Journal of Women’s Studies, 12(3): 249-265.

Leckie, G., Browne, W.J., Goldstein, H., Merlo, J., and Austin, P.C. (2020), ‘Partitioning Variation in Multilevel Models for Count Data’, Psychological Methods, 25 (6): 787-801.

MacQueen, S. (2016), ‘Domestic Abuse, Crime Surveys and the Fallacy of Risk: Exploring Partner and Domestic Abuse Using the Scottish Crime and Justice Survey’, Criminology and Criminal Justice, 16(4): 470-496.

Mahendran, M., Lizotte, D., and Bauer, G. R. (2022), ‘Quantitative Methods for Descriptive Intersectional Analysis with Binary Health Outcomes’, SSM - Population Health, 17: 101032.

Mahendran, M., Lizotte, D., and Bauer, G. R. (2022a), ‘Describing Intersectional Health Outcomes: An Evaluation of Data Analysis Methods’, Epidemiology, 33(3): 395-405.

Mason-Bish, H. (2015), ‘Beyond the Silo: Rethinking Hate Crime and Intersectionality’, in N. Hall, A. Corb, P. Giannasi, and J.G.D. Grieve, eds., The Routledge International Handbook on Hate Crime, 24-33. Oxon: Routledge.

McCarthy, M. (2017), ‘What Kind of Abuse is Him Spitting in My Food?’: Reflections on the Similarities Between Disability Hate Crime, So-Called ‘Mate’ Crime and Domestic Violence Against Women with Intellectual Disabilities’, Disability and Society, 32(4): 595-600.

Mueller, C. O., Forber-Pratt, A. J., and Sriken, J. (2019), ‘Disability: Missing from the Conversation of Violence’, Journal of Social Issues, 75(3): 707-725.

Nilsson, A., and Estrada, F. (2006), ‘The Inequality of Victimization: Trends in Exposure to Crime Among Rich and Poor’, European Journal of Criminology, 3(4): 387-412.

O, SooHyun, Martinez, N.N., Lee, Y., and Eck, J.E. (2017), ‘How Concentrated is Crime Among Victims? A Systematic Review from 1977 to 2014’, Crime Science, 6(1): 1-16.

Office for National Statistics (2021a), Crime Survey for England and Wales, 2015-2016, [data collection], UK Data Service, 2nd Edition, Accessed 3 March 2024. SN: 8140, DOI: http://doi.org/10.5255/UKDA-SN-8140-2.

Office for National Statistics (2021b), Crime Survey for England and Wales, 2018-2019, [data collection], UK Data Service, 2nd Edition, Accessed 3 March 2024. SN: 8608, DOI: http://doi.org/10.5255/UKDA-SN-8608-2.

Office for National Statistics (2021c), Crime Survey for England and Wales, 2019-2020, [data collection], UK Data Service, Accessed 3 March 2024. SN: 8812, DOI: http://doi.org/10.5255/UKDA-SN-8812-1.

Office for National Statistics (2022a), Crime Survey for England and Wales, 2016-2017, [data collection], UK Data Service, 3rd Edition, Accessed 3 March 2024. SN: 8321, DOI: http://doi.org/10.5255/UKDA-SN-8321-3.

Office for National Statistics (2022b), The Nature of Violent Crime in England and Wales: Year Ending March 2022. Available at: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/thenatureofviolentcrimeinenglandandwales/yearendingmarch2022.

Office for National Statistics (2023a), Crime Survey for England and Wales 2017-2018, [data collection], UK Data Service, 3rd Edition, Accessed 3 March 2024. SN: 8464, DOI: http://doi.org/10.5255/UKDA-SN-8464-3.

Office for National Statistics (2024a), Crime Survey for England and Wales: Update on Latest Data and Findings. Available at: https://ukdataservice.ac.uk/app/uploads/grant2024-02-06.pdf.

Office for National Statistics (2024b), Crime Survey for England and Wales Data Quality Review: June 2024. Available at: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/methodologies/crimesurveyforenglandandwalesdataqualityreviewjune2024#:~:text=Main%20findings,-The%20quality%20of&text=The%20CSEW%20response%20rate%20fell,a%20minor%20impact%20on%20estimates.

Office for National Statistics (2024c), Crime in England and Wales QMI. Available at: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/methodologies/crimeinenglandandwalesqmi#:~:text=Preliminary%20analysis%20indicates%20the%20response,in%20the%20survey%20response%20rates.

Office for National Statistics (2024d), User Guide to Crime Statistics for England and Wales: March 2024. Available at: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/methodologies/userguidetocrimestatisticsforenglandandwales#crime-survey-for-england-and-wales-csew-.

Pettitt, B., Greenhead, S., Khalifeh, H., Drennan, V., Hart, T., Hogg, J., Borschmann, R., Mamo, E., and Moran, P. (2013), At Risk, Yet Dismissed: The Criminal Victimisation of People with Mental Health Problems. London: Victim Support and Mind. Available at: https://www.mind.org.uk/media-a/4121/at-risk-yet-dismissed-report.pdf

Pina-Sánchez, J., and Tura, F. (2024), ‘Intersectional disparities in sentencing: An application of MAIHDA to the US Federal Courts’. Available at: https://assets.pubpub.org/mf9ryw8d/Intersectional%20disparities%20in%20sentencing-71721951120366.pdf

Potter, H. (2015), Intersectionality and Criminology: Disrupting and Revolutionizing Studies of Crime. London: Routledge.

Semenza, D.C., Testa, A., and Turanovic, J.J. (2021), ‘Trajectories of Violent Victimization Over the Life Course: Implications for Mental and Physical Health’, Advances in Life Course Research, 50: 100436.

Shapland, J., and Hall, M. (2007), ‘What Do We Know About the Effects of Crime on Victims?’, International Review of Victimology, 14(2): 175-217.

Sikweyiya, Y., Stern, E., Hanass-Hancock, J., van der Heijden, I., Myrttinen, H., Addo-Lartey, A.A., and Dunkle, K. (2022), ‘Intersections Between Disability, Masculinities, and Violence: Experiences and Insights from Men with Physical Disabilities from Three African Countries’, BMC Public Health, 22(1): 705.

Smith, K., Coleman, K., Eder, S., and Hall, P. (2011), Homicides, Firearm Offences and Intimate Violence 2009/10: Supplementary Volume 2 to Crime in England and Wales 2009/10. Home Office Statistical Bulletin: 01/11. London: Crown Copyright.

Thacher, D. (2004), ‘The Rich Get Richer and the Poor Get Robbed: Inequality in US Criminal Victimization, 1974–2000’, Journal of Quantitative Criminology, 20 (2): 89-116.

Thiara, R. K., and Hague, G. (2013), ‘Disabled Women and Domestic Violence: Increased Risk but Fewer Services’, in A. Roulstone and H. Mason-Bish, eds., Disability, Hate Crime and Violence, 106-117. London: Routledge.

Tseloni, A., and Pease, K. (2015), ‘Area and Individual Differences in Personal Crime Victimization Incidence: The Role of Individual, Lifestyle/Routine Activities and Contextual Predictors’, International Review of Victimology, 21(1): 3-29.

Tseloni, A., and Tilley, N. (2016), ‘Choosing and Using Statistical Sources in Criminology: What Can the Crime Survey for England and Wales Tell Us?’, Legal Information Management, 16(2): 78-90.

Tseloni, A., Mailley, J., Farrell, G., and Tilley, N. (2010), ‘Exploring the International Decline in Crime Rates’, European Journal of Criminology, 7(5): 375-394.

Tura, F., Pickering, S., Hansen, M. E., and Hunter, J. (2024), ‘Intersectional inequalities in trust in the police in England’. Available at: https://www.crimrxiv.com/pub/k7ilnmd0/release/1

van Dusen, B., Cian, H., Nissen, J., Arellano, L., and Woods, A. D. (2024), ‘Comparing the Efficacy of Fixed-Effects and MAIHDA Models in Predicting Outcomes for Intersectional Social Strata’, Sociology of Education, 97(4): 342-362

Whitfield, D.L., Coulter, R.W., Langenderfer-Magruder, L., and Jacobson, D. (2021), ‘Experiences of Intimate Partner Violence Among Lesbian, Gay, Bisexual, and Transgender College Students: The Intersection of Gender, Race, and Sexual Orientation’, Journal of Interpersonal Violence, 36(11-12): NP6040-NP6064.

Wieshmann, H., Davies, M., Sugg, O., Davis, S., and Ruda, S. (2020), Violence in London: What We Know and How to Respond. Behavioural Insights Team. Available at: https://www.bi.team/wp-content/uploads/2020/02/BIT-London-Violence-Reduction.pdf.

Williams, M. L., and Tregidga, J. (2014), ‘Hate Crime Victimization in Wales: Psychological and Physical Impacts Across Seven Hate Crime Victim Types’, British Journal of Criminology, 54(5): 946-967.

Supplementary materials

Appendix tables, figures and replication codes can be found here: https://anonymous.4open.science/r/MAIHDA-violence-4B95/README.md