Understanding your force statistics: disparities in missing episodes based on ethnicity

Understanding your force statistics: disparities in missing episodes based on ethnicity

Statistics show that Black children and adults are disproportionately represented in missing reports to the police: Black people make up 4% of the population in England and Wales, but 14% of missing incidents.[1]

Research published by Missing People and Listen Up shows that Black children and adults are also more likely to be missing for longer than White people; less likely to be found by the police; and less likely to be identified as at risk due to mental health issues, or sexual and criminal exploitation.[2] This suggests there is bias or discrimination in the response when Black people go missing, and that they may be at higher risk, but getting less support.

The Police Race Action Plan[3] sets out a vision ‘for a police service that is anti-racist and trusted by Black people’. It acknowledges longstanding challenges in the police’s relationships with Black communities, and as consequence, the poorer levels of trust and confidence amongst those communities. The Plan, for the first time at a national level, also set out the specific need to consider the impact of ethnicity on risks of going missing, and on the response that people receive when they do. A key aim included in the plan is to “improve the police response and effectiveness in supporting missing persons from Black communities.”

While we have national data on these issues, it is vital that you understand the local picture in your force area. Recognising what is happening locally can be the first step in reducing any disparities, and ensuring that missing children and adults, as well as their family, get the support they need.

Your police force area

It is crucial to understand what is happening in your local police force area. The following guidance sets out two steps to begin this analysis:

A statistical review of your missing data
A deep dive case review

While statistics in and of themselves will not show all bias or discrimination, they are a helpful indicator of disparities in the response to missing people. They can show disproportionality in your missing population when considered against broader population data. They can also show disparities in factors such as length of time missing, how people are found, risk assessment level and risk factor disclosure or identification. They can also highlight issues that may be disproportionately affecting different communities.

A deep dive case review can aid in understanding potential bias and discrimination. It can enable comparisons to be made in response to missing people from different communities, by comparing the response in cases which were factually very similar.

Reviewing how risk factors are identified and how missing incidents are written up can help identify bias, including the use of problematic, discriminatory or inappropriate language.

Understanding your local statistics can be a first step in reflecting on practice and encouraging officers to consider their decision making in the context of race, and whether bias may be informing some of their decisions. It can aid in evidencing whether some issues are disproportionately affecting different communities, and should therefore trigger additional or specialist work to reduce risk or improve engagement.

How this tool should be used

This tool is a multi-step process:

The statistical analysis should be carried out internally, or if possible, by an independent external service[4].
Using the findings of the statistical analysis as a guide, a deep dive case review should be undertaken internally, or if possible, by your independent advisory group, or by an independent external service.
If the analyses are undertaken internally, the findings should be shared with senior leadership and your force’s Independent Advisory Group (IAG).
Be as transparent as possible: be bold and share publicly with a commitment to change if disparities are evidenced.
Create an action plan setting out how the force will explore the reasons behind any disparities, and how they will work towards eradicating them. This could include things like training, auditing processes for decision-making, increased supervision, engagement with the community, and a range of other tailored actions.
Monitor the action plan, with a clear timeline based on the specific findings, and report back to ensure that accountability is in place.

Internal understanding and support is a vital first step in trying to change culture around bias and discrimination. Once the review has been completed it should be shared internally with a recognition that understanding can be the first step in change. The findings should be shared with front line staff, alongside reflective questions (included at the end of this document) to encourage them to think about why the data shows what it does and what they can do to change. It is important that defensiveness is discouraged and the findings are engaged with in an open and supportive way.

Note: The process for this review will depend on your specific database, how the data is captured and how it is able to be exported and analysed. While these instructions will serve as a guide, you will need to adapt the details to reflect the specifics of your data.

Data to export

We would suggest exporting:

All missing incidents for at least a 12-month period
The data should be exported at an incident level, including an incident identifier and an individual identifier

Fields to export (this will vary depending on what you would like to review in your local area). We would suggest exporting all or some of the below for each missing incident:

– Ethnicity

– Length of time missing

Date last seen
Date returned

– Risk assessment level

Initial risk assessment level

– Risk assessment level at point of case closure

– Harm outcomes

– Age at missing

– Gender

– Missing from (type of location)

– Found by (police, family, returned of own accord, etc)

– Vulnerability markers

General vulnerability marker
Mental health
Child sexual exploitation
Child criminal exploitation
Trafficking
Other (depending on your system)

Comparing data by ethnicity

How you do this will depend on:

The population in your area
How you record ethnicity in your force, i.e. the ethnicity categories that your force uses

It is good practice to break down the data into ethnicity categories as specifically as possible – experiences of marginalised ethnic communities are not homogenous, and we should try to understand different patterns and experiences within each unique community. However, it is possible given the data involved that you may have to combine ethnicity categories in order to make any analysis meaningful. It may be that the data you have only enables analysing two broad groups, e.g. ‘people from ethnic minority communities’ and ‘white’, or you may be able to break down the data in a more detailed way. That will entirely depend on your own local population and your missing data.

If the numbers of missing individuals or incidents are really low (for example less than 20), it is unlikely that you can include them in the analysis in any meaningful way. The findings might not hold any ‘statistical significance’, meaning they could be due to chance rather than showing any consistent pattern. It is possible to calculate statistical significance – this would be best practice if you have the expertise within force to do so. However, we would otherwise recommend combining groups to ensure the opportunity for useful analysis, and to protect individual’s anonymity as small numbers might mean people are identifiable.

Analysis: What should you be looking for?

We suggest you use the data to explore:

Overall missing population:
- Are some communities over-represented in your missing population compared to your general population?

Notes: You can conduct this analysis on a person level and an incident level.

To understand proportionality you will need to look at the percentages of missing people and missing incidents compared to the demographics of your local population. You can find information about the demographic breakdown of the general population in your area through Census data[5].

Repeat missing:
- Are some communities more likely to go missing repeatedly than others?
Length of time missing:
- Are some communities more likely to be missing for longer than others?
How people are found / return:
- Are some communities more or less likely to be found by the police / found by another professional / return of their own accord / found by their family or friends?
Risk assessment level:
- Are some communities more or less likely to be assessed as high or low risk?
Vulnerability markers:
- Are some communities more or less likely to have different vulnerability markers, for example mental health, CSE or CCE?

Note: While the person conducting this analysis will need to have some data/analytic skill, it is entirely possible to do this analysis using Excel, you do not need more sophisticated software. This stage can be done entirely anonymously as no identifiable information needs to be used.

Here you can access a template Excel spreadsheet setting out how the data can be used for analysis. This Excel file includes example data which is entirely fake. It is also includes examples of ‘pivot tables’ a tool that can be useful in analysis large amounts of data. This template serves as an example of how you might analyse the data. It should not be used to directly copy data into for analysis.

A note on subjective data:

While some data fields will not be affected by subjectivity (e.g. length of time missing), some data is highly susceptible to subjectivity. This includes risk assessment levels and vulnerability marker identification and recording. Some data may suggest that individuals are less at risk, when in fact there might be a problem with risk identification or assessment.

For example, if you find that Black missing children and adults are less likely to have vulnerability markers for mental health, this does not necessarily mean that they are less likely to be experiencing mental health issues, and instead may show distrust amongst reporting persons in disclosing mental health issues when reporting their person as missing; issues in communication or understanding at the point of reporting – with warning signs going unrecognised due to cultural differences; or may show bias in the police identifying mental ill-health or vulnerability in Black people.

Any notable differences should be analysed, and the risk of subjectivity considered alongside what previous research and evidence shows about the experiences of people from marginalised communities.

Understanding the data in more depth: intersectional experiences

You can consider the above at a population level, but we would suggest also breaking down by:

Gender
Age: under 18s as compared to 18 and over

The extent to which you can do this will depend on the data you have and the population of missing in your local area. It may be the case that the population is too small to conduct meaningful analysis when broken down into these categories.

Intersectionality

People’s experiences are intersectional, they can be informed by their race or ethnicity but also by their gender or gender identity, having a disability/neurodiversity, their sexuality, their age, their socio-economic status, and a range of other factors. They can also experience bias or discrimination because of all of these factors.

Understanding whether there are any disparities in the response or outcomes in your force that are influenced by these additional factors could be helpful in identifying issues.

Depending on the data available in your force, it may be valuable to further breakdown the data in line with other protected characteristics to identify whether intersectional identities result in people facing additional disparities.

Once you have completed your statistical analysis, you should share the findings with senior and frontline officers who play a role in missing person investigations across your force. If any disparities are identified, you should ask them to consider the findings alongside the following self-reflective questions:

Why do you think these differences are happening?
What do you think you are doing in your practice or response to missing incidents that could be driving these disparities?
What can you do differently to change them?

An in-depth review can assist in identifying bias or discrimination in the response to Black missing people or their families.

How information is recorded at the point of reporting, following interactions with families or informants, in the risk assessments, and throughout the investigation, can be informed by assumptions or biases about the missing person or their loved ones. By reviewing records you can identify whether there is evidence of negative attitudes, stereotypes, or persistent assumptions being made about missing Black children, adults or families. Identifying and challenging negative language, or a lack of information, can help to shine a spotlight on biased decision-making, and can help to identify officers who need additional training, supervision, or ultimately disciplinary action.

Sampling

In order to do an in-depth case review you will need to sample cases. Depending on the demographics of your missing population, and the scale of your case record review, you should choose the most appropriate of the following options.

Random sampling:

You should do this if you want a broader sense of cases in your force. This involves randomly selecting a number of cases and doing a deep dive review of the ones that are selected.

Depending on your missing population however, it may be that using a random sample will not result in cases that enable you to identify any discrimination or bias, as it is likely that only a small number of cases relevant to Black missing people will be sampled.

Purposive sampling:

This does not rely on random selection, instead you use your knowledge and judgement to identify the cases you want to review.

In this case, it may be that you decide to review specifically cases that are related to Black people going missing, alongside a selection of cases related to White people going missing.

Within purposive sampling, you can then randomly sample within that population, e.g. if you have 100 cases that are related to Black people going missing, you may decide to randomly sample 20 of those cases and compare them to 20 randomly selected cases regarding White people.

Comparing like-for-like cases

We would suggest that part of your review should include comparing like-for-like cases to identify whether there are differences in actions taken, language used, and risk assessment where the circumstances are similar and the only difference is the ethnicity of the missing person. For example cases where the gender, age, and circumstances of missing are similar. However, it is important that you are not biased in the selection of these cases, for example choosing particularly high-profile cases that you know will have more activity than others, or selecting cases with the aim of specific outcomes that may not be more broadly representative.

The number of cases you review will depend on the size of your force and the numbers of missing records you hold. We would recommend including as large a sample as possible to ensure findings are meaningful.

Data to review

You should use the findings of the statistical analysis to determine areas of focus for the deep dive. For example, if the statistical analysis showed disparities in risk assessment levels you should ensure that the deep dive includes a review of risk assessment decision making.

In any case, we would suggest reviewing:

Risk level assignment and rationale
Resource level assigned to the case
Any variations in timeliness or urgency of actions
Risk marker / flag assignment
General data quality
Language used about the missing person and the reporting person

The review should look at whether the above factors are consistent across cases of missing people of different ethnicities. You should aim to identify any patterns in the response to Black missing people, and any disparities across all factors when compared to White missing people.

Promising practice: proactively exploring the issue

Following publication of the Ethnicity of Missing People report, senior leadership in one police force decided to be proactive in identifying what was happening in their area. They approached Missing People to help conduct a review of their missing person statistics, and carried out a deep-dive comparative review of missing incidents for both Black and White missing people. While the findings were challenging, the force engaged in constructive discussion about what this meant, and what they needed to do.

Learning from that review process has helped to shape this guidance.

Review of language

As well as reviewing data on elements such as risk assessment levels and risk markers, a key element of the review should be on reviewing the language used in case records in order to identify any potential bias, discrimination or problematic language.

Open text fields in the records should be reviewed to identify any concerning language.

Black people with lived experience of having been reported missing, or reporting a loved one missing, have collated the following examples of language that can either be racist, or might be disproportionately used in case records about Black missing people or their families, and may indicate bias or discrimination. This list is not exhaustive, but aims to inform the reviewer’s thought process around negative attitudes to look out for, and to act as a starting point for language to consider in the case review.

While some terms are obviously racist, or show clear bias, it can often be difficult to identify subtler indications of discrimination. Black people themselves will be best placed to recognise indicators of bias, so if possible this review should be conducted by Black members of your IAG, or a commissioned service who hold lived expertise.

The reviewer should not just be looking for ‘negative’ or racist language, but also for tone and any indication of the level of care, professional curiosity, or compassion shown throughout any records.

The following terms may be used for missing people and their families from all ethnic backgrounds. However, it is important to review whether they are disproportionately (more regularly) used for Black missing people and their families.

Term	Explanation
Gang-involved	While some missing people may be involved with gangs, if this term is used more regularly for Black missing people it often indicates stereotyping. Additionally, using the term ‘gang’ when referring to any group of young people without any evidence of organised crime indicates bias and can have an impact on safeguarding decisions. It is better whenever possible to use simple descriptors of the circumstances, for example, spending time with ‘peers’, ‘friends’ or ‘other young people’; was part of a ‘large group of people’, etc. Or if you are referring to unsafe people, be explicit, for example: ‘spending time with adults who are suspected to be involved with drug-supply’, there are concerns that ‘adults or peers may be grooming them with the aim of criminal exploitation’.
Promiscuous / Highly sexualised	Particularly when used for children, this term is victim-blaming and judgemental. It can result in a failure to acknowledge the nature of child sexual exploitation. In developing this toolkit we saw evidence of these terms used for children as young as 13. The term should not be used for adults either. It may be disproportionately used in relation to Black women and girls due to over-sexualisation and adultification. If someone is being exploited then name this so the risk and vulnerability is clear.
Drug-dealer	If this term is used without confirmed evidence, or when the person is known to be a victim of criminal exploitation, it can be an indicator and driver of stereotypes and can lead to reduced safeguarding responses.
Aggressive / Angry / Threatening	Black people are disproportionately likely to be considered or described as aggressive. The use of these terms, including in contexts like “spoke aggressively” should be closely examined to ensure they are not being applied more regularly to Black missing people or families, and not being used to try and justify a poorer response from professionals.
Stand-offish / disengaged / unresponsive	These terms may be used more often for Black people, where others might be described in kinder or more understanding terms like ‘quiet’, ‘reserved’ or ‘withdrawn’. This can be due to failures to identify or see vulnerability.
Frequent flyer / Repeat misper / serial offender (in relation to missing incidents)	These terms are unhelpful as they can minimise the risks of each missing incident, can reduce professional curiosity, and can result in a sense of inevitability about ongoing missing incidents, rather than efforts to address the underlying drivers. These terms may be used about all missing people, but may be used about Black missing people disproportionately regularly.
Gets themselves into trouble / causing trouble / seeking out trouble	These terms can be victim-blaming, and can be a result of adultification. They may also be linked to a risk of criminalisation.
Referring to a child persistently as a young person, or just as the ‘female’ or ‘male’.	While this may be done with no negative intent, or purely as a descriptor, if used disproportionately for Black children it may suggest adultification. An example of this was identified through this project where case records were found to include things like “Female left the building to meet a man for sex”, when the missing child was a 13 year old girl – referring to them as ‘Female’ risks failing to acknowledge their vulnerability and the exploitation that they are experiencing.
Victim-blaming language, particularly in relation to exploitation	Examples of this identified through this project included: “[Child] has gone to great lengths to organise [meeting up with adult men] despite having internet usage monitored” “[Child] has a habit of meeting older in men in the past” – used about a 13 year old. “[Child] made arrangements [to meet older men] weeks in advance of the day she went missing” All of these examples suggest the victim was the instigator, or had some power in the activity surrounding their exploitation. They do not lay blame with the perpetrator.

The importance of language

The language that professionals use about a member of the public can give insight into their attitude towards that person, and can potentially reveal bias. It is therefore important to review language as one element of driving accountability and challenging discriminatory responses. However, it is also important to acknowledge that problematic language doesn’t just reveal underlying bias, but can have very real, immediate consequences for safeguarding.

Victim-blaming language, language that minimises harm, or language that adultifies children, can inform safeguarding decisions. Other professionals will likely read what has already been written, and may in turn be informed by that original bias. Children and adults can live with the results of those assumptions or labels for many years. It is vital that we emphasise the importance of language to all professionals working with missing people and that there are opportunities for challenge and immediate redress when problematic records are identified.

Communication: tone and context

Police officers will sometimes need to communicate a difficult message, for example, that they can’t do something that a missing person or their family are asking for; or can’t stop a missing person from doing something their family think is unsafe. When this is the case, communication is extremely important to ensure that people do not feel let down, ignored, or discriminated against. Firstly, the police should be transparent and clear on why they’re deciding to make that decision, or to not take that action. Secondly, they must communicate this sensitively in a compassionate and kind way.

The police should avoid using judgemental language when responding to missing incidents. Going missing is not a crime, many people who go missing are experiencing serious harm, and families who report a loved one missing will be scared, and desperate for help. Blaming or judging people is unlikely to ever be helpful in these circumstances. Police should instead use objective language when discussing sensitive topics to help avoid victimisation, bias or blame.

Although effective assessment of someone’s communication can only really be conducted in real-time, it may be possible to review how officers have communicated by the case records detailing an interaction. Any indicators of good or poor communication, including how decisions are explained or justified, should be included in your case record review.

When conducting the review you should consider:

Would you want this written about your loved one if they were missing?
Do these records show any differences in decision-making that could be linked to race or ethnicity?
What would you expect to hear from the police if you were in this situation?