Deciding how to collect

OSR’s expectations

The collection and reporting of statistics about sex and gender identity should support a legitimate public interest and be done in the least intrusive way. Producers should ensure that data are collected in a respectful way and should understand any risks to data quality or survey response when asking for sensitive information from a person. 

Individuals providing their personal data may have their own understanding of the terms sex and gender identity. For this reason, it is important that, where possible, producers provide guidance to respondents and interviewers on how these questions should be completed as well as explanations for why certain information is required. This guidance will help ensure accuracy and coherence across the data collection. In particular, producers should avoid collecting data or reporting on gender without clear guidance and explanation of what is meant by this term, as it is often unclear whether the aim is to collect information on an individual’s sex or their gender identity. 

The privacy and identity of individuals must be protected at all times during data collection, storage, analysis and reporting. This includes being clear and open with individuals who are providing information about how their data will be used and how they will be protected. Producers must apply relevant security standards to keep data secure. 

When designing data collection questions, there are many factors that producers should consider. Producers should avoid using response options that might cause confusion by ensuring that they undertake appropriate question testing of respondent understanding and acceptability. Producers should ensure that the response options align with the concept that they are aiming to collect data on. For example, we would expect a gender identity question to provide response options that reflect that gender identity is not a binary concept. 

Producers should consider whether they should provide additional response options which allow respondents to not provide a response, for example, where an individual would prefer not to say, or if an individual does not consider themselves to have a gender identity. Producers should also consider the design of the question, including the benefits and costs of using free-text boxes, as well as whether specific questions should be age restricted or be voluntary. These approaches may differ depending on different respondent groups or collection modes. 

Producers should consider the mode of data collection, including when there is a change of mode, or when collecting data on sensitive topics. It may be that a respondent would answer differently to a question depending on whether the data were collected by an interviewer face to face or if the questions were answered online. Responses may also vary depending on whether the data are self-reported or completed by another individual, such as a household representative. 

Producers who are collecting data through statistical surveys may have more involvement and control in making decisions on the inclusion of questions and question testing and design. If the data are collected through administrative systems, it may be more complex to add, adapt or improve questions, and there may be significant costs involved. This is because these data are usually collected for operational purposes and statistical reporting data requirements may be secondary. If administrative data collections cannot be changed, producers should assess whether this affects the suitability or quality of data for statistical reporting purposes. 

Where data are collected through an administrative system, producers should establish and maintain a constructive relationship with the data suppliers. They should seek to understand how the data are collected, discuss any data requirements and understand any quality assurance or adjustments that have been carried out on the data, all within the particular operational context. Producers should also have a mechanism for two-way communication on any issues such as data errors or other quality concerns and keep informed of any changes to data collections. 

Producers should consider how data about sex and gender identity will be recorded in administrative systems and how this may affect the responses received as well as the acceptability of questions, for example, whether the data are self-reported or completed by another individual such as an operational official. Producers should consider the impacts of this on the resultant data. 

Questions for statistics producers to consider