This article was written by Rosie Almond, Senior Behaviourial Scientist
In recent times, more attention has been paid to diversity and inclusion initiatives as more firms understand the risks of a lack of diversity and the benefits of having an inclusive working environment. Initiatives have been put in place to increase the diversity of the workforce and improve inclusion.[1]
But how can we tell which initiatives are worthwhile and impactful? This is hard to do without robust data collection. Reliably measuring diversity is the first step to being able to say with confidence if your firm is moving in the right direction. With low rates of demographic disclosure, there can be little confidence in the overall levels of diversity in a firm. This problem is exacerbated when we look at intersectionality. [2]
When we talk about measuring diversity, we are usually looking at the following demographic characteristics: gender and sex, ethnicity, sexual orientation, disability, age and socioeconomic status.
These are all sensitive topics and not everyone is happy to share these characteristics with others, let alone their employers. There will always be some who choose not to disclose their information, and this should be respected – all disclosure should be voluntary. However, there are many people who might choose to disclose in certain circumstances.
To look into this, we have launched the Sharing Diversity Data research programme under our D&I stream of work. Find out more about the programme here. We have conducted a literature review of the research in the field of barriers and enablers to demographic disclosure, particularly in the workplace. Although reasons for nondisclosure vary by demographic characteristic, some common themes have emerged. We look at these through a behavioural lens.
Our literature review – along with consultation with member firms – has found that behavioural factors can influence willing or undecided individuals.
So, how can we encourage employees to share their diversity data?
Behavioural principle 1: removing ‘frictions’
Any point of ‘friction’ in a user journey could lead to non-completion of an activity. If something is just too difficult or awkward to do, people are much less likely to do it.[3] Some firms feel that their HR systems are too clunky and this can discourage employees from sharing their demographic information. It is important to try to make the process of sharing demographic characteristics as easy as possible.
This could be as simple as:
- Sending a direct link instead of asking staff to find it themselves
- Asking the right questions – following best practice guidance for asking sensitive questions so people feel they can submit an answer that is right for them
- Providing guidance for demographic questions: Scope for disability[4], Stonewall for sexual orientation[5], the Social Mobility Commission[6] for socioeconomic status or ONS for ethnicity.[7]
Behavioural principle 2: the messenger effect
People behave differently depending on who is doing the ‘asking’ – who is the ‘messenger’ of the intervention. People are more likely to change their behaviour if the request comes from someone with similarities to them.[8] People are also responsive to the perceived authority of the requestor.[9] An example from the literature is the chief people officer of a firm encouraging the completion of a demographic survey, which, along with several other initiatives, increased sharing of socioeconomic background by around 50%.[10]
Who might be the right person to ask employees to complete a survey?
- A senior leader from a minority background
- An employee resource group representative
- A line manager
Behavioural principle 3: incentives
For those with a characteristic that is discriminated against, disclosing this characteristic can feel risky. In addition, the employee may perceive there to be no benefit in sharing their sensitive information.
If employees spend time evaluating which choice to make, they may use system 2 thinking – a more deliberative mindset[11]. When weighing up whether to share their information, employees may therefore see a risk and no benefit. There can be two ways to influence this decision-making process: reduce the risk or increase the benefit.
Reduce the risk:
- Does the firm demonstrate a culture of diversity, inclusion and acceptance of differences? This could reduce fear of discrimination.
- How can the firm demonstrate that data has been kept safe in the past?
- Could the survey be done anonymously? Could the firm use an external provider? This could increase trust in the anonymity of results.
Increase the benefit:
- Is it clear what the data is being collected for and what will be done with it?
- Can the firm demonstrate that the data has been used for benefit in the past and will be again?
These are just three of the themes that have arisen from our literature review. The impacts of behavioural factors are highly context-specific and vary by demographic characteristic. A key point to take from the literature is that the impact of initiatives in this area have not been robustly measured – we cannot yet say which of these might be most effective at increasing disclosure rates. The next stage in the programme is to trial behavioural interventions with member firms and generate evidence on their effectiveness.
[1] Shahab, S., & Lades, L. K. (2021). Sludge and transaction costs. Behavioural Public Policy, 1-22.
[2] Intersectionality looks at several demographic factors at once – rather than separately – when assessing the effects of discrimination and advantage. This acknowledges individuals have multiple facets to their social identity, that together can affect their experiences differently than could be explained by looking at these factors in isolation. (Crenshaw, 1989) [1]
[3] Crenshaw, K. (2018). Demarginalizing the intersection of race and sex: A Black feminist critique of antidiscrimination doctrine, feminist theory, and antiracist politics [1989]. In Feminist legal theory (pp. 57-80). Routledge.
[4] Scope. (2019). A guide to reporting on disability employment. [accessed July 29 2022]; available at: https://www.scope.org.uk/scope/media/files/campaigns/scope-guide-reporting-disability-and-employment.pdf
[5] Stonewall. (2016). Do ask, do tell: capturing data on sexual orientation and gender identity globally. [accessed July 29, 2022]; available at: https://www.stonewall.org.uk/sites/default/files/do_ask_do_tell_guide_2016.pdf.
[6] Social Mobility Works. (2021). Socio-economic diversity and inclusion – Employers’ toolkit: Cross-industry edition. [accessed July 29 2022]; available at: Socio-economic diversity and inclusion – Employers’ toolkit: Cross-industry edition – July 2021 (socialmobilityworks.org)
[7] Office for National Statistics. [accessed July 29 2022]. available at: Ethnic group, national identity and religion – Office for National Statistics (ons.gov.uk)
[8] Dolan, P., Hallsworth, M., Halpern, D., King, D., & Vlaev, I. (2010). MINDSPACE: Influencing Behaviour through Public Policy. [accessed July 29 2022]; available at: http://www.instituteforgovernment.org.uk/publications/mindspace
[9] Durantini, Albarracín, Mitchell, Earl and Gillette (2006) Conceptualizing the influence of social agents of behavior change: A meta–analysis of the effectiveness of HIV–prevention interventionists for different groups. Psychological Bulletin 132: 212–248.
[10] Social Mobility Works. (2021). Socio-economic diversity and inclusion – Employers’ toolkit: Cross-industry edition. [accessed July 29 2022]; available at: Socio-economic diversity and inclusion – Employers’ toolkit: Cross-industry edition – July 2021 (socialmobilityworks.org)
[11] Daniel, K. (2017). Thinking, fast and slow.
