Matrix > Toolkit: Recruitment and Promotion > Data Collection Basics
Data Collection Basics
On this page, you will find practical guidance, insights and experiences from real employers, and simple examples for collecting and categorising data on the basic ethnic demography of your workforce. Visit our section on setting targets and population data for benchmarking, and our pay gap calculation section for guidance on more in-depth analysis of your data. Click here for Doing Data Right, the Toolkit’s landing page for all data advice.
Contents
Establish trust in ethnic data collection
Collecting ethnic data from applicants and employees works best and is only ethical if they know why the data is being gathered, and what it will be used for. Communication about how it will support equity and equality is key. Employers should make it clear the organisation is asking for this information so that it can establish current levels of inequality and whether there is a need for improvement. Employers should be committed to making improvements if the data shows clear inequities and ethnic pay gaps, and this should form a part of a wider strategy with annual timeframes, that will be shared with employees. This allows employees to see the purpose of disclosing their ethnicity, and therefore have a tangible stake in doing so. An example statement may be: “We are an equal opportunity employer. This data is being collected so we can make sure people are being paid fairly and not being discriminated against, through regular pay audits. We promise to protect your privacy and confidentiality.”
GEM Insights Snapshot: Compare data collection pitfalls, approaches and enablers in very different types of organisations
There are clear contrasts between ethnic data-gathering experiences in blue-collar and heavily Māori and Pasifika frontline workplaces, and white-collar organisations where Māori and Pasifika are very underrepresented.
When to collect ethnicity data
Ethnic data should be collected:
At the application stage – anonymous data to measure how well the recruitment process is attracting diverse candidates.
During onboarding – employee data, used to measure potential pay gaps and progression.
Monitoring the ethnic group of job applicants is an important part of seeing how successful an organisation is at engaging underrepresented groups when recruiting. However, it can only be gathered and held for equity monitoring of recruitment processes, not as employee data.
The ‘gold standard’ way to gather this data at recruitment is for People and Culture to collect it (on a voluntary basis) on a separate form, prevent hiring panels from accessing it, and be clear with applicants that the data is only for recruitment monitoring purposes, and will be kept separate. For example, DiversityWorks notes: “It is crucial that this data is kept entirely separate from the application and it should never be seen by anyone involved in the recruitment decisions and outcomes – it should be kept securely by designated HR personnel instead. Recruitment processes should include measures to safeguard this confidentiality. This allows the ethnic data to be separated for ‘ethnic-blind’ hiring. It should be reviewed during the recruitment process by HR staff independent from decision making to advise whether there is a representative spread of applicants, and whether recruitment needs to be extended to ensure an adequate range of candidates. If applicants are employed, this data should only be entered into an employee’s records with their permission.
GEM Snapshot: Pitfalls in collecting ethnic data at recruitment stage
Organisations with high levels of Māori and Pasifika frontline staff
GEM research has found that it is not unusual for large employers of Māori and Pasifika to ask for ethnic information as part of general job application forms, but do not necessarily collect or hold this data separately according to best practice. This assumes a high level of trust from prospective employees that they will not be discriminated against. It also suggests some blurred lines in employer understanding of what should be collected, when, and for what purposes. The ethical way to collect ethnic data for general recruitment (i.e. not a specific ethnically targeted programme) is to assure applicants that it will not impact their application, and is being held separately for equity monitoring purposes only.
Professional services organisations with low levels of Māori and Pasifika staff
Organisations with historically low levels of Māori and Pasifika staff may have more internal concern about institutionalised discrimination in hiring. In GEM testing these organisations were more likely to be gathering and keeping ethnic data separately at the recruitment stage, in accordance with best practice. However, sometimes data that is kept separate is left unused. In GEM research, one employer discovered that their business had collected ethnic data at the recruitment stage (and held it separately) for many years without any DEI-focused staff being aware or having been able to analyse the data to inform recruitment policy and practice. It is not ethical to collect ethnic data at recruitment if it is not being used to monitor equity in recruitment processes.
What type of ethnic data should I be collecting?
GEM supports the Kia Toipoto Public Service Commission recommendations that employees should be able to:
Report multiple ethnicities. The same multi-ticked or multi-coded data can be recoded, summarised and analysed in different ways as needed. This gives the most flexibility and is the best representation of people’s identities in a workforce and allows organisations to state their specific ethnic group(s) without employees having to identify themselves in a more general category.
In practice, a mix of drop-down menus and write-in boxes for the most flexibility, e.g. Rather than simply having a “Pacific” category, employees would be able to choose from a drop-down menu with “Samoan”, “Tongan”, “Niuean”, “Cook Islander” etc.
Change the information employers hold about them to reflect that a person’s ethnic identity may change over time (e.g. they may have reported a single ethnic group initially but realised that they want to acknowledge another part of their whakapapa not mentioned earlier). This is likely to be in such small numbers that it will not impact planning.
A data system should be able to store a minimum of three ethnicity responses per person.
What are Level 1, 2, 3 and 4 ethnic groups?
The Stats NZ ethnic classifications become more detailed as levels increase – from least detailed at Level 1 (e.g. ‘Pacific’, ‘Asian’ or ‘European’), to the smallest and most detailed subgroups at Level 4 (e.g. ‘Pitcairn Islander’, ‘South African Indian’, ‘Welsh’).
Employers’ forms are likely currently collecting ethnic data at Levels 1, 2 or 3, with write-ins to specify further details. If data is backcoded to Level 1 categories for summarising and analysis, care must be taken to make sure the more detailed Level 2 or 3 data is not discarded including write-ins that may be captured.
Multi-coded collection and back-coding example: The Example Factory
An employer has five employees in total working at the Example Factory, who all fill in their ethnic data, using Stats NZ Level 3 options provided. This is a ‘multi-coded’ or ‘multi-tick’ system where people can choose multiple ethnic groups that apply to them. Here’s what they ticked and wrote in.
In the data collection system or spreadsheet, this would look like:
How do I categorise people with multiple ethnic groups?
Data on detailed and multiple ethnic groups should be retained, and new back coded columns created for different purposes. For example, GEM recommends prioritising ethnic group according to the pay gap being calculated.
Click here for further explanation of total and prioritised ethnic group.
There are other reasons why an organisation might backcode or recategorise ethnic data, such as describing the demography and diversity of their workforce, or answering questions about combined Māori and Pasifika workforce. The Example Factory’s workforce can sound very different depending on how the ethnic data is summarised: is it 80% European, or 80% Māori and Pasifika?
What am I analysing?
Typically, a company would want to know:
What proportion of employees or candidates are from which ethnic group/s. Choose one of the methods above to describe this.
What is the success rate of candidates from different ethnic groups. Click here to see Kia Toipoto summary of assessing the current state of recruitment (see page 10).
What is the average (or median) pay of these different ethnic groups and the gaps between them. Click here for pay gap calculation guidance.
Retention/turnover, training uptake and wellbeing. These measures should use the same method of defining who to include in different groups being compared as a workplace’s pay gap analysis.
See our section on analysing pay gaps for more information and guidance.
Gather qualitative data too
It is important to realise that gathering qualitative data on the experiences of Māori and Pasifika staff is just as important as getting quantitative data right. This enables an organisation to not only see the scale of inequities, but to also understand the realities of why they exist and how they can be understood in ways that make sense for employees of your organisation. For more on this, see our section on creating spaces and channels for Māori and Pasifika staff to review and shape what happens in the workplace, including through unions.
Quick links for short overviews
Diversity Works Diversity Data Series provides user-friendly introductions, guidance and tips to shift from gathering to acting on equalities data. A series of short, user-friendly primers includes:
Introduction to diversity data
The art of asking
Engaging people
Understanding the data
Taking positive action
The Price Waterhouse Cooper Diversity Data collection guide is another toolkit on data collection, based on UK-based legal and cultural frameworks, but much of it is applicable to the Aotearoa New Zealand context.