Consultation outcome

Results and actions from the Ethnicity facts and figures website reform consultation

Updated 21 April 2023

In summer 2022, the Cabinet Office Equality Hub put a series of proposals to reform the Ethnicity facts and figures website out to public consultation.

We want to improve Ethnicity facts and figures, helping users to understand the drivers and factors behind disparities, and minimise the risk of misinterpretation and incorrect conclusions being drawn. For policy-makers we will be providing better evidence for targeting interventions and resources at the real point of need. We proposed to do this by streamlining measure pages while providing deeper analysis of the most notable disparities.

The consultation ran from 13 June to 7 August 2022 on the Citizen Space platform.

It was promoted via Twitter, LinkedIn, posts on Medium and the Statistics Authority blogs, as well as the Ethnicity facts and figures newsletter.

1. Misinterpretation of our proposals

During the final week, the consultation was shared extensively by the public via WhatsApp, increasing the response from 12 on 27 July 2022 to 498 by the closing date. The WhatsApp sharing was accompanied by a misinterpretation of 2 of the proposals in the consultation which led to some distortion in the data.

One proposal was to pause updating measures that are only available disaggregated into 2 ethnicity classifications (white and other than white) while we work with government departments to improve the disaggregation. The second proposal was to combine related measures into one measure page. The misinterpretation resulted in a widespread view that we proposed to combine the ethnicity classifications and only provide the binary classification of ‘white’ and ‘other than white’ for all measures. As a result, many respondents ‘disagreed’ with one or both of our proposals, while simultaneously commenting that the binary classification is not useful or helpful, is stigmatising, ‘othering’ and obfuscates the disparities between different ethnic groups.

The confusion is clearly illustrated by one respondent who ‘strongly disagreed’ with the proposal to pause updates of binary data in order to improve disaggregation, and then commented: “this proposal will hide information about sub groups who may be strongly impacted.”

Other comments include:

It will make it almost impossible for people to understand detail on the race disparity in this country.

This is a critical mistake - we need the breakdown by ethnicity otherwise no way to make improvements.

That [a binary classification] won’t explain the disparities. Disaggregated data will give us a better pictures. Having a binary system will create more otherness …. the white and the rest.

The table below also shows that, looking at comments from across the whole of the consultation, 40% of the respondents provided at least one comment stating that ethnicity classification should not be binary. This is a clear indication of the extent of the misinterpretation regarding the proposal around measures with binary classification.

Table: Has commented that data should not be binary

Response	Number	Percentage
Comment present	298	59.8%
No comment present	200	40.2%
Total	498	100%

The Equality Hub has no plans to reduce the ethnicity classification to a binary grouping, as this would be counter-productive to all the work we are undertaking. We welcome the comments from the respondents reiterating the importance of fully disaggregated ethnic groupings and will continue to work with departments across government to provide more granular and useful data.

2. Main findings from the consultation

Overall, the respondents agreed with the majority of our proposals. There were 3 areas where respondents disagreed with our proposals as presented:

1. Respondents suggested that the identification of criteria used to prioritise measures needs to be more responsive to developments in society.

2. The proposal to present data for some measures without any commentary was concerning for many users.

3. Respondents were also concerned that there were a number of measures relating to the criminal justice system in the medium and lower priority categories.

One additional finding was that a large proportion of the respondents had not been aware of the website. Almost one-third (32.7%) of respondents had not accessed Ethnicity facts and figures or only had a look but not used anything before responding to the consultation, with the main reason being they had not heard of it. Overall, of the full 498 respondents, around 1 in 9 were not aware of the website.

The full results of the consultation can be found in Annex A.

3. Actions for the Equality Hub

1. The Equality Hub will undertake formal annual reviews of the criteria used to prioritise work on the development of measure pages and also regularly ‘horizon scan’ to incorporate emerging priorities that may disproportionately impact ethnic minority groups.

2. Measures will be reviewed against the prioritisation criteria at least every 12 months. Where we identify an emerging priority, specific measures will be re-prioritised as needed. New measures will be published according to their priority against the criteria in use at that time.

3. The Equality Hub will create a template for ‘medium priority’ measures to show one table or chart with ethnicity breakdown and ‘Main facts and figures’ drawing out key points from other disaggregations. We will develop code to automate the writing of the ‘Main facts and figures’ commentary for these measures.

4. In addition, we will expand the automation of ‘Main facts and figures’ commentary to the lower priority measures.

5. We will review the priority of Crime and Justice measures to identify whether any should be high priority.

6. We will review the proposed groupings for combining measures; where measures are closely related (for example, key stage 2 attainment) the pages will be combined.

7. The Equality Hub will continue to work with departments across government to provide more granular and useful data.

8. The Equality Hub will work with the relevant departments to provide deeper analysis into the following priority topics: educational attainment, transition from education to the labour market, and maternal and perinatal health disparities.

9. The Equality Hub will take steps to publicise the Ethnicity facts and figures website more widely.

We began to develop new automated processes for updating the Ethnicity facts and figures website while the consultation was underway and we have developed an action plan to complete the streamlining of the website by April 2023.

The streamlining reforms, together with the automation, will allow us to redirect resources to provide deeper analyses and understanding of the most notable disparities in outcomes between ethnic groups. This move away from presenting only simple, descriptive statistics will help users identify the drivers of disparities and should ensure interventions and resources are targeted appropriately by government departments and third sector organisations alike. In doing so we will build on the approach documented in the Quarterly reports on progress to address COVID-19 health inequalities in which we described the impact of a number of explanatory variables (such as occupation, area, and underlying health conditions) on the risk of infection, hospitalisation and death from COVID-19.

4. Annex A: Results and actions

4.1 Organisations who responded

Response	Number	Percentage
Academic	48	9.6%
Central government	16	3.2%
Local authority	55	11.0%
NHS	46	9.2%
Private citizen (no organisation)	223	44.8%
Third sector	110	22.1%
Total	498	100.0%

The highest percentage of responses came from individuals responding as private citizens (44.8%) followed by responses from charities and other third sector organisations (22.1%). The smallest percentage of responses came from central government (3.2%) – however, these respondents will be using the data directly for government policy development.

4.2 Frequency of using Ethnicity facts and figures

Response	Number	Percentage
A few times a year	114	22.9%
Daily	21	4.2%
I haven’t accessed Ethnicity facts and figures	115	23.1%
I’ve had a look but not really used anything	48	9.6%
One or two times a month	76	15.3%
Only when relevant new data is published	85	17.1%
Weekly	39	7.8%
Total	498	100.0%

Almost one-third (32.7%) of respondents had not accessed Ethnicity facts and figures or only had a look but not used anything before responding to the consultation. These respondents were filtered away from the main consultation as the questions were only relevant to users of the website.

A small but significant percentage of respondents (primarily in the charity/third sector) used the website at least weekly (11.0%).

4.3 Sections used most

Response	Number	Percentage
Do not use the measure pages	114	22.9%
Main facts and figures	312	62.7%
Things you need to know	202	40.6%
Charts and tables	218	43.8%
Commentary	192	38.6%
Sources	158	31.7%
Downloads	119	23.9%
Links	120	24.1%

The section of the measure pages used most was the ‘Main facts and figures’ (62.7%). The data downloads section was used least (23.9%).

4.4 Main uses of the data

Response	Number	Percentage
Personal or general interest	250	50.2%
Fact checking	204	41.0%
Looking at trends	246	49.4%
Comparing geographies	182	36.5%
Comparing ethnic groups	296	59.4%
Policy development	187	37.6%
Communicating stats	234	47.0%
Other	57	11.4%

Other uses include Identifying disparities (3.2%), Targeting work (1.8%) and promoting equality (1.2%).

4.5 Frequency of reviewing criteria for prioritising

Response	Number	Percentage
Every 12 months	138	42.6%
Every 18 months	12	3.7%
Every 2 years	37	11.4%
Every 6 months	126	38.9%
Less frequently	2	0.6%
More frequently	4	1.2%
Emerging priorities	5	1.5%
Total	324	100.0%
Not answered	174

The majority of respondents thought the criteria against which measures are prioritised should be reviewed every 6 to 12 months (81.5%).

ACTION: Annual reviews of criteria.

4.6 Main suggestions for other prioritising criteria

Response	Number	Percentage
Measures should not be prioritised as all are priorities	8	1.6%
Are they of high interest to the public?	15	3.0%
Are they from community consultation?	10	2.0%
Which show the largest impacts?	6	1.2%

ACTION: Include ‘horizon scanning’ reviews to incorporate emerging priorities that may disproportionately impact ethnic minority groups.

4.7 Frequency of reviewing measures against criteria

Response	Number	Percentage
Every 12 months	145	44.8%
Every 18 months	8	2.5%
Every 6 months	130	40.1%
Every 2 years	31	9.6%
Less frequently	2	0.6%
More frequently	6	1.9%
Emerging priorities	2	0.6%
Total	324	100.0%
Not answered	174

The majority of respondents said the priority of measures should be reviewed every 1 to 2 years (87.4%).

ACTION: Measures will be reviewed against the prioritisation criteria at least every 12 months. Horizon scanning may bring a review forward for specific measures.

4.8 High priority measures that could be lower

Response	Number	Percentage
All should be high	17	35.4%
Community measures	3	6.3%
Crime and justice measures	6	12.5%
Education measures	8	16.7%
Health measures	8	16.7%
Work and pay measures	3	6.3%
Workforce and business measures	3	6.3%
Total	48	100.0%
Not answered	450

There were 48 suggestions for lowering priority, the highest percentage being Education measures and Health measures (16.7% each). However, comments suggest that respondents did not understand the question as some topics and data suggested are not currently available.

ACTION: No high priority measures will be deprioritised.

4.9 Usefulness of ‘ethnicity only’ pages

Response	Number	Percentage
Very useful	191	59.7%
Quite useful	47	14.7%
Neutral	29	9.1%
Not that useful	29	9.1%
Not at all useful	24	7.5%
Total	320	100.0%
Not answered	178

Three-quarters of those who responded to this question thought that measures with an ‘ethnicity only’ chart and ‘Main facts and figures’ section would be useful (75.4%).

ACTION: Create template for ‘ethnicity only’ measures to show one table/chart with ethnicity breakdown and ‘Main facts and figures’ drawing out key points from other disaggregations. Use R Markdown to write ‘Main facts and figures’.

4.10 Usefulness of ‘data only’ pages

Response	Number	Percentage
Very useful	104	32.1%
Quite useful	76	23.5%
Neutral	51	15.7%
Not that useful	53	16.4%
Not at all useful	40	12.3%
Total	324	100.0%
Not answered	174

Over half (55.6%) of those who responded to this question thought that measures pages with ‘data only’ would be useful. However, this sentiment was contradicted when respondents were asked how their work would be impacted by the reduction in commentary:

	Effect on work if no commentary: Very positively	Effect on work if no commentary: Positively	Effect on work if no commentary: Neutral	Effect on work if no commentary: Negatively	Effect on work if no commentary: Very negatively	Total
Usefulness of data-only measures: Very useful	4	4	16	35	41	100
Usefulness of data-only measures: Quite useful	0	5	26	31	14	76
Usefulness of data-only measures: Neutral	0	0	17	17	15	49
Usefulness of data-only measures: Not that useful	0	1	9	15	27	52
Usefulness of data-only measures: Not at all useful	1	0	3	12	24	40
Total	5	10	71	110	121	317

Three-quarters (76%) of respondents who said the data only pages would be very useful, then said the impact of no commentary would affect their work negatively.

ACTION: As R Markdown will be used to automate ‘Main facts and figures’ for medium priority measures, this could also be used for the lowest priority measures, mitigating the risks to users.

4.11 Lower priority measures that should be high

Response	Number	Percentage
All should be high priority	37	29.4%
Crime and justice measures	43	34.1%
Community measures	10	7.9%
Education measures	2	1.6%
Health measures	18	14.3%
Housing measures	2	1.6%
Those with biggest impacts	1	0.8%
Work and pay measures	6	4.8%
Workforce and business measures	7	5.6%
Total	126	100.0%
Not answered	372

Many more respondents identified lower priority measures that they thought should be high, over one-third of which related to Crime and Justice measures.

ACTION: We will review the priority of Crime and Justice measures.

4.12 Agree combining measures

Response	Number	Percentage
Strongly agree	21	6.5%
Agree	38	11.8%
Neither agree nor disagree	46	14.3%
Disagree	55	17.1%
Strongly disagree	161	50.2%
Total	321	100.0%
Not answered	177

Over half of those who responded to this question disagreed with the proposal to combine related measures onto a single page. Comments provided suggest that respondents conflated the ‘combination of measure pages’ with the misinterpreted ‘binary measures’ proposal (misinterpreted to propose we would combine all ethnic minority groups so that only White and Other than white is provided). This means that a number of those disagreeing in this question may be disagreeing with ‘combining ethnic groups’.

Having said that, a few respondents did comment that some of the proposed groupings may not be appropriate. There was also some concern that details and insights may be lost.

ACTION: We will review the proposed groupings; where measures are sensibly related (for example, key stage 2 attainment) the pages will be combined.

4.13 Agree different levels for new measures

Response	Number	Percentage
Strongly agree	31	9.6%
Agree	61	18.8%
Neither agree nor disagree	82	25.3%
Disagree	41	12.7%
Strongly disagree	109	33.6%
Total	324	100.0%
Not answered	174

Almost half (46.3%) of people responding to this question disagreed with the proposal. Comments suggest that the main concerns were around the possible lack of commentary, thus risking misinterpretation and lower use. With the amendment to the original proposal (2 levels of priority instead of 3), this risk is mitigated.

ACTION: New measures will be published according to their priority against the criteria in use at that time.

4.14 Impact of reduced commentary

See Usefulness of ‘data-only’ pages.

Response	Number	Percentage
Very positively	6	1.8%
Positively	11	3.4%
Neutral	72	22.2%
Negatively	114	35.1%
Very negatively	122	37.5%
Total	325	100.0%
Not answered	173

Contradicts Questions 11 and 12:

	Effect on work if no commentary: Very positively	Effect on work if no commentary: Positively	Effect on work if no commentary: Neutral	Effect on work if no commentary: Negatively	Effect on work if no commentary: Very negatively	Total
Usefulness of data only measures: Very useful	4	4	16	35	41	100
Usefulness of data only measures: Quite useful	0	5	26	31	14	76
Usefulness of data only measures: Neutral	0	0	17	17	15	49
Usefulness of data only measures: Not that useful	0	1	9	15	27	52
Usefulness of data only measures: Not at all useful	1	0	3	12	24	40
Total	5	10	71	110	121	317

	Effect on work if no commentary: Very positively	Effect on work if no commentary: Positively	Effect on work if no commentary: Neutral	Effect on work if no commentary: Negatively	Effect on work if no commentary: Very negatively	Total
Usefulness of high level breakdown measures: Very useful	5	6	35	68	74	188
Usefulness of high level breakdown measures: Quite useful	0	4	17	18	7	46
Usefulness of high level breakdown measures: Neutral	0	0	11	8	8	27
Usefulness of high level breakdown measures: Not that useful	0	0	5	10	14	29
Usefulness of high level breakdown measures: Not at all useful	0	0	3	5	16	24
Total	5	10	71	109	119	314

4.15 Binary measures for redevelopment

Most write-ins here related to the misinterpretation of the proposal: respondents thought the proposal was to make (all) measures binary, rather than to improve the disaggregation of ethnic groups.

4.16 Agree with not publishing binary data

		Has commented that data should not be binary: No comment present	Has commented that data should not be binary: Comment present	Total
Should we stop updating binary measures?: Not answered	Number	141	34	175
	% within Should we stop updating binary measures?	80.6%	19.4%	100%
Should we stop updating binary measures?: Strongly agree	Number	13	22	35
	% within Should we stop updating binary measures?	37.1%	62.9%	100%
Should we stop updating binary measures?: Agree	Number	16	13	29
	% within Should we stop updating binary measures?	55.2%	44.8%	100%
Should we stop updating binary measures?: Neither agree nor disagree	Number	19	14	33
	% within Should we stop updating binary measures?	57.6%	42.4%	100%
Should we stop updating binary measures?: Disagree	Number	30	19	49
	% within Should we stop updating binary measures?	61.2%	38.8%	100%
Should we stop updating binary measures?: Strongly disagree	Number	79	98	177
	% within Should we stop updating binary measures?	44.6%	55.4%	100%
Total	Number	298	200	498

Again, most write-ins here related to the misinterpretation of the proposal: respondents thought the proposal was to make (all) measures binary, rather than to improve the disaggregation of ethnic groups.

4.17 Impact of stopping binary data

It was difficult to analyse these results as they were affected by the misinterpretation regarding the binary classification.

ACTION: The Equality Hub has no plans to reduce the ethnicity classification to a binary grouping, as this would be counter-productive to all the work we are undertaking. We welcome the comments from the respondents reiterating the importance of fully disaggregated ethnic groupings and will continue to work with departments across government to provide more granular and useful data.

4.18 Agree with providing further analysis

Response	Number	Percentage
Strongly agree	155	48.0%
Agree	91	28.2%
Neither agree nor disagree	37	11.5%
Disagree	14	4.3%
Strongly disagree	26	8.0%
Total	323	100.0%
Not answered	175

Three-quarters of those who responded to this question agreed that more analysis should be provided for priority topics.

Top 3 topics of those listed:

School absence

Rank	Number	Percentage
1	35	33.7%
2	28	26.9%
3	41	39.4%
Total	104	100.0%
Not answered	394

Transition from education to the labour market

Rank	Number	Percentage
1	44	25.4%
2	67	38.7%
3	62	35.8%
Total	173	100.0%
Not answered	325

Maternal and perinatal health disparities

Rank	Number	Percentage
1	55	33.1%
2	41	24.7%
3	70	42.2%
Total	166	100.0%
Not answered	332

Stop and search geography

Rank	Number	Percentage
1	58	43.0%
2	40	29.6%
3	37	27.4%
Total	135	100.0%
Not answered	363

Poor mental health

Rank	Number	Percentage
1	51	35.7%
2	54	37.8%
3	38	26.6%
Total	143	100.0%
Not answered	355

Education attainment

Rank	Number	Percentage
1	62	31.0%
2	74	37.0%
3	64	32.0%
Total	200	100.0%
Not answered	298

Having inverted the scale so that first place scores 3 and third place scores 1, the sum of the scores reflects both the weighted score and the fact that more respondents voted for some subjects than others. Thus the top 3 topics are:

• educational attainment
• transition from education to the labour market
• maternal and perinatal health disparities

Topic	Sum
School absence	214
Transition from education to the labour market	364
Maternal and perinatal health disparities	347
Stop and search geography	249
Poor mental health	273
Educational attainment	402

ACTION: The Equality Hub will work with the relevant departments to provide deeper analysis into the following topics as a priority:

• educational attainment
• transition from education to the labour market
• maternal and perinatal health disparities

Other topics:

• education
• health disparities
• criminal justice
• housing
• income
• leadership in public sector

4.19 Further comments

Most comments related to the misunderstanding regarding binary classification.

4.20 Why not used Ethnicity facts and figures

It is not relevant to my work/research

Response	Number	Percentage
It is not relevant	42	8.4%
Not answered	456	91.6%
Total	498	100.0%

I wanted more analysis than descriptive statistics

Response	Number	Percentage
More analysis	14	2.8%
Not answered	484	97.2%
Total	498	100.0%

There’s no qualitative data or analysis

Response	Number	Percentage
No qualitative data or analysis	12	2.4%
Not answered	486	97.6%
Total	498	100.0%

Other (please state)

Response	Number	Percentage
Other (please state)	73	14.7%
Not answered	425	85.3%
Total	498	100.0%

Unaware of existence

Response	Number	Percentage
No	440	88.4%
Yes	58	11.6%
Total	498	100.0%

Approx one-third (32.7%) of respondents were filtered out of the main consultation because they answered “I haven’t accessed Ethnicity facts and figures” or “I’ve had a look but not really used anything”. These respondents were then asked why they had not used Ethnicity facts and figures. The response was multiple choice, so totals will not add up to 100%.

The largest percentage of responses was for Other

Of the 73 who selected ‘Other’, almost three-quarters were unaware of Ethnicity facts and figures before being informed of the consultation.

Of the full 498 respondents, around 1 in 9 were not aware of Ethnicity facts and figures.

4.21 Ideas for improvement

Of those comments that did not relate to the misinterpretation of the binary data proposal, the most common response was ‘Publicise more widely’ which is not surprising given findings above.

ACTION: The Equality Hub will look to publicise the Ethnicity facts and figures website more widely so that all interested parties are aware of it.