Public Sector Equality Duty Implementation of Scottish Specific Duties: Views from public authorities

Chapter 3 employee information, equal pay and gender pay gap

Introduction

3.1 This chapter sets out the experiences of authorities in meeting the specific duties in relation to:

• gathering information on the composition of their employees in relation to their relevant protected characteristics;
• gathering information on the recruitment, development and retention of employees in relation to their protected characteristics;
• publishing their gender pay gap and an equal pay statement (if they have over 150 employees; and
• using this information to better perform the general duty.

3.2 The chapter explores:

• who led the process of producing employee information, and whether they worked with others;
• the processes that authorities used to gather the information;
• the (broad) level of resources dedicated to producing employee information, equal pay and pay gap information;
• key success factors, barriers and challenges;
• views on the early impact of publishing employee information - and the potential longer term impact; and
• views on approaches to meeting these duties in the future.

Lead responsibility

3.3 We explored the process that authorities went through to produce the employee information required. Generally, across all types of authority, the employee information was produced jointly by HR or organisational development staff, working with policy staff or staff with a specific equalities focus. In most cases, the HR team had responsibility for extracting and producing the information, and a staff member with responsibility for equality policy advised on how this information should be produced, and assisted with analysing it.

3.4 In some instances, different individuals within the HR team took responsibility for drawing together the information required across different themes - such as gender pay gap information; equal pay information; and employee characteristics. Some, particularly NHS Boards, set up steering groups involving key staff members from across departments.

3.5 In small organisations, some policy officers led on collating and gathering employee information. Some small organisations indicated that they did not produce employee information, as the number of staff they had was too small (generally less than ten).

3.6 Generally, local authorities took the lead for producing employee information for Licensing Boards - as they had a small staff team, and were local authority employees. While education authorities were not included in this review, some local authority consultees highlighted that they produced information for the education authority.

Working with other bodies

3.7 We asked authorities whether they worked jointly with any other public authorities to produce the employee information. The clearest instance of a co-ordinated joint approach came from NHS Boards. NHS Boards worked closely together to develop a shared approach to producing employee information. A national group was established with a remit of producing a template for the employee information, and this template was distributed to all NHS Boards. Many felt that this made the process clearer and easier, as the Boards simply had to populate the templates rather than develop these in each area. However, some NHS Boards did say that they developed their own approaches to producing the employee information.

Example: Creating templates for consistency

The divisional HR director at NHS Lanarkshire was invited to participate in a national group representing the whole of the NHS. This group was established one year before the data on employee information had to be published. This group met and spent time looking over what was being asked of them, and what their current systems were capable of producing. This group created a template for collecting employee information which was sent to all NHS Boards for their use - ensuring consistency from every NHS Board. This meant that all the NHS Boards had to do was populate the fields - making it a much easier process.

3.8 Some local authorities indicated that they worked jointly with neighbouring authorities or with local authorities across Scotland through the Scottish Councils Equality Network. Those who had worked together jointly had found this very valuable. Some colleges and universities had shared practice. Many indicated that they had received support through the Equality Challenge Unit, which works to advance equality and diversity in colleges and universities. Some highlighted that the Equality Challenge Unit had suggested looking to Wales for good practice examples. However, others felt that practice was very diverse across colleges and universities in terms of producing employee information.

"Universities are doing different things."

(Education Body)

Gathering information

3.9 Experience of gathering employee information varied significantly between authorities. Generally, organisations already held some information on the protected characteristics of their employees - particularly in relation to age and sex. Most began the process of gathering information by reviewing what they had already, and identifying gaps. Almost all found that they needed to either adapt the categories that they used to produce information on employees, or collect new information. However, some - particularly NHS Boards - indicated that they already collected and reported on most of the information required, so this did not result in significant additional work.

3.10 A minority of authorities said that they simply published what they had available, and acknowledged the need to develop information around key areas in the future. One very small organisation simply decided not to produce any employee information, as it did not hold the information required. However, most identified gaps and then introduced new systems to enable them to gather information in advance of publishing it.

3.11 Most identified gaps in relation to data around the composition of staff, by protected characteristic. Many undertook initial research to ensure that they were clear about the definition of the protected characteristics, and many spent considerable time reading and understanding the guidance produced by the EHRC about how to meet the duties around employee information.

3.12 In most cases, authorities modified the electronic systems they had in place for storing and analysing data about employees. Often this simply involved adding new fields to the system - and most said that this was not complex and did not take up much time. However, authorities then had to gather the information about employees, which was more time consuming.

3.13 In some small organisations a new survey was issued to all staff. In small organisations this tended to result in a high response rate - with some achieving 100% response rates. However, some small organisations - and particularly organisations where the level of staff diversity was perceived to be low - indicated that with the small numbers involved staff could be concerned about being identified. Large organisations, however, also had challenges gathering staff information. Often information was held for newer staff across a wider range of characteristics, as monitoring systems had been updated. But for longer standing staff, often limited information was held.

3.14 Generally, organisations which undertook exercises to gather further information about employees reported a response rate of around 70 to 80 per cent. Some organisations, particularly those which had not previously gathered employee information, reported lower response rates of around 50 per cent. Some authorities said that response rates to requests for information about protected characteristics had increased in recent years. Some felt that this was due to increased awareness of equalities issues in recent years, while others suggested that a shift from paper to electronic systems meant that staff felt more confident that their data would be treated confidentially.

3.15 Across all types of authority, organisations suggested that the most difficult areas to produce employee information related to sexual orientation, gender reassignment and pregnancy and maternity. Disability was also felt by many to present some challenges. There were different reasons for this.

3.16 In relation to sexual orientation, many organisations had not gathered this information until recently (or until the introduction of the specific equalities duties). Many organisations felt that there was some nervousness among staff in disclosing sexual orientation, and that this was an issue which required to be approached in a sensitive manner.

3.17 Organisations had different concerns about gathering and publishing data around gender reassignment. Many expressed concerns about the number of staff involved being so small that there was a danger of breaching data protection, through potentially identifying individuals. However, some felt that it was not clear what organisations were expected to do where numbers were very small.

"On gender reassignment, we needed to have the confidence not to publish this information."

(NHS Board)

3.18 In relation to pregnancy and maternity, many organisations indicated that this information was not routinely collected. Some felt that it was unfair that this information needed to be produced so quickly, with little warning, given that it had not previously been collected. Generally, there were some concerns about the short timescales as this meant that it was difficult to build staff confidence and give reassurance about why this information was required - which many highlighted was a long term process.

"The challenge is to reassure staff why we need this information and what we will do with it."

(Local Government)

3.19 In relation to disability, a number of concerns were raised. Some authorities felt that staff often indicated that they were not disabled when they could have been considered disabled under the law. Others indicated that there were some reservations about staff identifying as disabled, as this was not something they wished to discuss.

"Massive issues around collecting disability information in a healthcare setting... it is not spoken about widely among staff."

(NHS Board)

3.20 Authorities also identified some gaps in relation to recruitment, development and retention. However, while some organisations worked to address gaps, most found that they had to publish the data they had available - as it was more difficult to gather this type of information retrospectively.

3.21 Some particular issues were identified in producing recruitment information. Many local authorities indicated that their recruitment information was gathered on MyJobsScotland.com - an online recruitment site for local government in Scotland. This site gathers and retains information about the characteristics of applicants. However, many local authorities were concerned with the format in which information was provided. It was provided in raw form to all local authorities using the site, which meant that all local authorities required to analyse this - and could use different categories and interpretations. Some local authorities reported experiencing issues with receiving corrupt data from the site. Some felt that there should have been more national co-ordination - perhaps by COSLA - to ensure that local authorities had access to data analysed in a similar format across all areas.

3.22 One authority indicated that it was part of a larger organisation, which collected recruitment data and did not share it at a local level. This meant that data on recruitment was not available, and data on staff characteristics had to be gathered again.

3.23 Authorities which had been created through merging other public authorities reported significant issues gathering and collating information on employee characteristics across different organisations, as information was gathered in different ways.

3.24 In relation to staff development, some authorities indicated that this information was held by different teams. This meant that information could be collected in different ways, and some work was required to align the data and ensure that it could be produced by the relevant protected characteristics. Some found it difficult to decide what information should be reported under staff development, and others indicated that it could be challenging to keep track of external development opportunities.

3.25 Both large and small organisations experienced some challenges around gathering and publishing staff retention information. Some large organisations reported that it could be difficult to keep track of employees, as many were moving on. However, small organisations said that it was difficult to publish the characteristics of those who leave the organisation, given that very few leave in any one year - and there were concerns over data protection. This was also an issue for larger but stable organisations, with few staff leaving. Some had taken the decision not to publish information on retention, but were not sure whether this would be accepted by the EHRC.

3.26 Some - particularly NHS Boards - highlighted that information on equal pay, pay gap and occupational segregation had not previously been produced in the format required to meet the duties, so this required considerable work. This was also an issue for some local authorities. Many larger authorities required to produce this information had lengthy discussions about the way in which it should be produced - for example, which categories should be used to report on occupational segregation. Organisations highlighted that they wanted to be consistent and transparent, and ensure that appropriate categories or bandings were used to report on occupational segregation, and many were very keen for more guidance and support on this issue. Some also highlighted that often the production of gender pay gap, equal pay and occupational segregation information required more manual manipulation of the figures, compared to the production of employee characteristics which was often produced using electronic systems.

Resources

3.27 We asked authorities to provide a broad estimate of the resources invested in producing the employee information. Most authorities found this very difficult to estimate, and provided very broad answers. Generally, most authorities said that the process of producing the employee information was less resource intensive (often considerably less) than developing the equality outcomes. However, the level of resources invested largely depended on the existing systems in place, and the requirement to gather new information.

3.28 Organisations which already had comprehensive systems in place to gather and analyse information about employees often indicated that the resources required were minimal - ranging from a day to five days. These organisations - many of them NHS Boards and NDPBs - indicated that they were often asked to produce this type of information and it was readily accessible. Generally organisations felt that this level of resources was "about right" for the size of the task.

Example: Role of software in producing employee information The organisational planning and improvement manager for Dumfries and Galloway Council explained its approach to gathering and publishing employee information to meet the equality duties. Equal pay review software was installed in advance of the introduction of the duties, and training was implemented on this software. This allowed data from the council's Human Resources Management Information System (HRMIS) to be analysed and used to generate gender pay gap, equal pay and occupational segregation reports easily. Dumfries and Galloway Council felt that investment in the new software made analysis of the existing employee information more efficient and allowed the organisation to produce the required reports quickly and reliably. It was believed that the new system enabled the organisation to work smarter, saving time and resources as without it this information would have had to be extracted and analysed manually. HR confidence in its dataset is high and information is seen as more reliable. "The software has been a big success and a worthwhile investment. We wouldn't have met the deadlines without it."

3.29 However, other organisations indicated that they spent significantly more time producing the employee information. The process was particularly time consuming if:

• the organisation required to gather additional information from staff, and analyse this;
• the organisation did not have an electronic system in place for analysing employee information;
• some information required to be analysed manually or manipulated into different categories than normally used by the organisation; and/ or
• the organisation required to produce gender pay gap, equal pay and employee composition information - which they had not produced before, or not produced in the format required.

3.30 Some organisations had worked on producing the employee information for up to six months, full time. However, most indicated that it took between two and six weeks of a staff member's time. Many indicated that senior staff were involved in discussing and agreeing how to produce information - particularly on pay gap and occupational segregation. Others indicated that relatively senior staff had to undertake manual tasks, such as data collection and analysis, because there was no-one else in the organisation who had this within their remit.

"It was pretty time intensive, especially for the new information."

(Local Government)

"It is everybody's interest, but no-one's responsibility."

(Local Government)

3.31 A minority of organisations indicated that they had dedicated resources to installing new software (such as Equal Pay Review software) directly as a result of the requirement to meet the specific duties. Others had previously invested in software which made the process easier, in advance of the introduction of the duties.

3.32 Some authorities which had dedicated significant resources to producing the employee information questioned whether this was a good use of resources - particularly in a financial climate where public authorities are expected to "do more for less". Some - particularly local authorities - felt that relatively senior staff had to invest time in producing employee information, as it did not naturally fall into the remit of any other staff member.

Success factors

3.33 Authorities were asked what success factors enabled them to produce the employee information effectively. The most commonly mentioned success factor was previous investment in systems and software to gather and analyse employee data. Organisations which had invested in gathering employee information in previous years, and in software which enabled them to analyse this quickly and effectively found it much easier to meet their duties. Some felt that their software, and the fact that systems had been in place for a while, gave them real confidence in the data that they produced.

"We couldn't have met the deadlines without it."

(Local Government)

"This is something we have been producing for a while."

(NHS Board)

3.34 For example, one organisation had introduced new systems to gather employee information on protected characteristics at the same time as further employee information was required to the HMRC Real Time Information (RTI) payroll information requirements. When approaching staff to update their personal information for the RTI system, they also asked for information on protected characteristics. This organisation found that response rates were good for many characteristics, but many employees still chose not to respond on sexual orientation, gender reassignment and religion and belief.

3.35 Authorities which already collected some employee information also found it easier to meet their duties. Many commented that because information on some protected characteristics had been gathered previously, there was very little resistance to gathering additional employee information. They felt that staff were familiar with the process, and confident that information provided would be treated appropriately. Some indicated that including an option of "prefer not to say" also helped to increase employee response rates for some questions - meaning that employees could choose to answer certain questions without answering others.

3.36 Senior level commitment was also highlighted as a key success factor. Many authorities indicated that high level support from senior management made the process easier. Some felt it meant that different departments and individuals took the tasks they had been allocated seriously, due to the process being clearly visible and accountable.

"They took the agenda seriously."

(Education Body)

3.37 Finally, organisations which had shared practice with other organisations found this extremely helpful. Many NHS Boards highlighted the value of a shared approach across the NHS, with templates provided and support from a national working group. Others were pleased with the opportunity to share experience at workshops run by the EHRC. Those who had had regular dialogue with the EHRC (a minority) generally indicated that this helped them to meet their duties more effectively.

"Joining up [with other organisations] was a good thing to do."

(NHS Board)

Barriers and challenges

3.38 We also explored the barriers that authorities experienced in meeting their duties to publish employee information.

Lack of detailed guidance

3.39 The most commonly mentioned barrier was the challenge in interpreting the written guidance produced by the EHRC. Many found that it was vague - with some saying it was "broadly worded" and "a bit woolly". Some, particularly authorities which were not direct service providers, or which were small and specialist organisations, indicated that it was difficult to interpret how the guidance should be applied to their organisation.

3.40 Authorities mentioned a number of specific issues, on which they would have liked more guidance. Firstly, and most commonly, authorities suggested that it would have been very useful to provide a template for employee information - which could be adapted by each organisation. This was suggested across all types of authority.

"We had difficulty in defining the requirements... We need clarity around what employee information should look like."

(Education Body)

3.41 Others felt that there should be more specific guidance about which organisations were not required to produce employee information, due to their small size. Similarly, larger organisations sought clearer guidance about the level of detail to be reported for very small categories - for example where the number of staff going through gender reassignment is extremely small, or where the number of staff leaving means that reporting across all characteristics will involve very small numbers. Some organisations indicated that they had to take in-house legal advice to agree the level at which numbers should be reported. Others said that they simply had to have the confidence not to produce information, and would have to see what the response from the EHRC was in this regard.

3.42 Local authorities had specific issues with interpreting what was meant by education authority staff, and in extracting these figures to report on them separately. Similarly, while interviewees understood that legally Licensing Boards were discrete entities, most felt that for this purpose they should be included within the local authority employee information - as staff were employed by the local authority.

"Is it just teachers, or is it people working in schools? There is a difficulty in taking this out."

(Local Government)

Employee response rates

3.43 Another common barrier was the extent to which employees respond to data gathering exercises around protected characteristics. Many authorities reported that a low response rate resulted in incomplete figures, which was a major barrier not just to publishing the information but to understanding the trends and issues which needed to be addressed. A range of reasons for low response rates were suggested:

• apathy due to financial cuts, restructuring and redundancies (particularly within local authorities);
• concern about providing personal information when redundancies are happening (again particularly within local authorities);
• lack of diversity within authority employees resulting in a concern about being identified as having a particular characteristic;
• concern about providing certain types of information - including sexual orientation and religion and belief - due to worries about why this was needed and how it would be used.

"If there are gaps in the data, it's because people didn't provide it."

(Other Body)

Timescales

3.44 Some authorities felt that the duty to publish this information came about very quickly, and that time was required to build up confidence in gathering employee information - particularly around sexual orientation and religion and belief. There was some concern that the numbers published were low and un-representative, due to the need to build confidence in disclosing personal information. Some highlighted that this was a big cultural shift and would take time. Others suggested that the deadline of 30 April for producing this information did not fit well with other reporting deadlines within their organisation. Importantly, some authorities said that because the timescales were tight, they did not have time to think about what the figures meant - and what actions needed to be taken to address them - before publication.

"We know that not all staff who are gay disclosed this."

(Other Body)

Systems for collecting and analysing data

3.45 Some reported that they had particular difficulties analysing and producing the employee information because their systems were "old and cumbersome". This meant that it was challenging and time consuming to extract information, and it had to be done manually in many instances. Some noted that information had to be "cobbled together" from different sources. This was an issue across all sectors.

"It is a systems issue... It costs money and time to get a new system, but it is in our plan."

(Education Body)

"We believe we are compliant, but we could do more with better systems."

(NHS Board)

Agreeing categories for reporting

3.46 Authorities also suggested that it was difficult to decide what sub-categories to use for reporting, within protected characteristics. Most felt that the sex category was relatively straightforward, but categories like age required detailed consideration of the categories to use for reporting, in order to make information as meaningful as possible.

Other barriers

3.47 There were also tensions (in a small minority of large authorities) between policy and HR staff. Some policy staff felt that HR staff, who were responsible for producing the figures, did not understand the specific and general duties - resulting in the data not being as useful as it could be.

"There are weaknesses in some of the published information... There is not an adequate knowledge base in HR around the wider equalities agenda... It is not just about basic data."

(NHS Board)

3.48 Finally, one authority working on an international basis highlighted that it could be difficult to take a corporate approach to equality due to different laws in different countries in relation to equality.

"There can be cultural issues and tensions... It can be hard to align."

(Education Body)

Impact of publishing employee information

3.49 We asked authorities about the early changes they had seen in their organisation as a result of publishing the employee information, and about the long term changes they expect to see.

Short term benefits

3.50 Many authorities indicated that it was too early to see whether the publication of employee information had resulted in any changes. Some, across all sectors, felt that there was a greater awareness of equalities, and more awareness of the representation of people who share protected characteristics in their organisation. In many organisations, this had resulted in a commitment to improve the quality of data they gather - and a greater understanding of the importance of gathering this data among staff, managers and governors.

"It will help to challenge reluctance."

(Local Government)

3.51 In some organisations, the publication and analysis of employee information had resulted in practical actions based on the findings, such as designating Equalities Champions to address specific inequalities; or advertising senior posts as flexible (with a part time option) to address inequalities identified.

"We have already started to do things differently."

(Local Government)

3.52 A minority of organisations felt that the publication of employee information had resulted in a greater focus on equality more generally, and a very small number felt that it had resulted in greater confidence in employees in talking about their needs and experiences. Some felt that it added an impetus to work on equalities which was already happening.

"People are more willing to say I have a disability and this is the adjustment or support that I need. People are more open in terms of discussing information about protected characteristics, and I am heartened by the high return to our staff survey."

(Education Body)

"It keeps it [equalities] in your mind and gives it a sense check. Are you being a fair employer and are you behaving as a fair employee?"

(Education Body)

3.53 However, many authorities indicated that the publication of the data was only the beginning. The "hard part" was taking action to address inequalities. Some indicated that they would like to undertake more detailed studies to understand some of the trends they were seeing in the employee information. Others were unsure that the publication of information had resulted in any change in the short term.

"It is still largely viewed as an afterthought."

(Local Government)

Long term benefits

3.54 Almost all authorities felt that the publication of employee information would result in longer term change. Many felt that the overall aim of producing and analysing employee information was very positive in terms of tackling inequality. Many hoped that in future years their processes would improve and they would become more sophisticated in how they used data, resulting in a greater understanding of the inequalities experienced.

"The more we do this, the more sophisticated it will become."

(Education Body)

3.55 Many authorities hoped that a better understanding of employee characteristics would result in action which would enhance their reputation as an employer. However, many felt that changes in the profile of their workforce were a long term aim, and that the production of employee information was just one part of a much bigger task to enhance equality in the workplace.

"Without major societal changes, these won't change the big figures [in relation to gender pay gap]."

(Local Government)

"I would like to say that the long term benefits will be that the data will be used to achieve equality outcomes for [our] workforce. However, the tangible benefits of this will be difficult to measure in the first few years."

(Local Government)

Future

3.56 Authorities commented on how they would approach the task of producing employee information in the future. Many focused on their plans to collect more robust data from employees, and to introduce better systems and processes to do so. Many suggested that they would be focusing strongly on improving their data around pregnancy and maternity; gender reassignment; and sexual orientation. Some suggested that they would develop or improve guidance for staff on the purpose of gathering this information. And many highlighted that they would work on gathering better data from existing or longer serving employees - as often data gathering was better for newer employees and applicants.

Example: Plans to collect more robust data The Highland Council held good information on sex and age, but had incomplete data from staff on disability and ethnicity. No other data on protected characteristics was collected at the time of the review. The issue was with "long-term" staff who were recruited years ago and who were not asked for this information at the time of their employment. The council have previously issued paper surveys to try to capture outstanding information - but invariably these are not always completed. New staff are recruited through a web-based portal which gathers information on all protected characteristics from people at an early stage. The Highland Council have plans to implement an integrated payroll and HR system which will operate online and provide a service called MyView. There are 10,000 council staff; but only 4,500 have access to a computer as others are not desk-based jobs. MyView allows staff to log on to see their online profile, and on first use they will be prompted to complete their personal information including equalities data. There will be an option for staff who "prefer not to say" but the Highland Council hopes this will fill in a lot of the gaps on protected characteristic information. The Highland Council have been working with Stonewall to prepare briefings for staff on why this information is required and what it will be used for.

Example: Identifying gaps in the data Aberdeen City Council will be using the employee information in its Mainstreaming Report to identify and address any areas of under-representation within protected characteristic groups. There is always a portion of the workforce which does not wish to declare sensitive personal information. In 2012/13, the council undertook a data capture exercise where 80% of staff submitted an equalities form; but there remained 20% who chose not to submit any information. "The challenge for the future will be to further explain and reassure staff why this information is needed, what specifically it will be used for, and the importance of declaring." Within HR, groups have been set up and assigned to each equalities strand, and are working towards action plans to meet the equality outcomes. In addition, a diversity and equality survey was recently undertaken asking employees whether they had experienced any unfair treatment in their employment on grounds of a protected characteristic. It also explored what the council was doing well and what could be improved on regarding equality and diversity. "We will be using the information gathered to inform our action plans."

3.57 In addition to talking of improvements to the data gathering process, many authorities said that "it shouldn't just be about ticking boxes". Many wanted to get under the skin of the numbers, and understand real experiences and how to improve these. Some wanted to encourage others - such as Trade Unions - to use the data to inform their work.

"It should be less about numbers, and more about integration, mainstreaming and changing cultures and attitudes."

(Education Body)

"[We plan to] commission an academic study of the dataset, using international comparisons... to get a bit of external perspective."

(NHS Board)

3.58 Many authorities suggested that they would plan in advance for the production of the next round of employee information, and embed this into their everyday work, to avoid a rush next time this information was required.

Summary of chapter three

3.59 In the vast majority of cases, HR colleagues were responsible for sourcing and producing the employee information published by the authorities. Most, particularly in large authorities, worked alongside a colleague with a specific equalities focus who advised on the type of information that should be gathered and how to analyse it for the purpose of meeting the duties.

3.60 Some, particularly NHS Boards, set up steering groups involving key staff members from across departments. In small organisations, policy officers sometimes led on collating and gathering employee information.

3.61 Almost all of the authorities reported having to adapt or update their existing monitoring systems in order to comply with the new legislation. Most authorities, across all sectors, reported difficulties in gathering robust data in relation to gender reassignment, sexual orientation and pregnancy and maternity.

3.62 In terms of the resources dedicated to the production of employee information, most authorities found that they had spent less time producing this information than they had spent on producing equality outcomes - particularly if they already had good data gathering systems and procedures in place.

3.63 However, some authorities found the production of employee information particularly time-intensive as they needed to redesign their existing monitoring systems in order to gather and analyse the relevant information. Small authorities that needed to do this often managed to do so quickly and easily, but it was more of a challenge for large authorities due to the volume of staff employed.

3.64 The vast majority of authorities felt that in the long-term, the process of gathering employee information would be a very positive first step in tackling inequality in the workplace.

3.65 With this in mind, authorities suggested that they would continue to work hard to improve their employee data, with a specific focus on gender reassignment, pregnancy and maternity and sexual orientation.