Chapter 3.6: Gender Statistics

1. Understanding users' needs and translating them into measurable statistical concepts

When the need for new or improved statistics has been identified, statistical organizations should typically engage users to identify their detailed statistics requirements, propose high-level solution options, and prepare business cases to meet these needs. In this phase, the organization

• identifies the need for the statistics in question
• confirms, in more detail, the statistics needs of stakeholders
• establishes the high-level objectives of the statistical outputs
• identifies the relevant concepts and variables for which data are required.

The need for gender statistics is driven by major policy issues. In Canada, gender equality is a guaranteed right under the Canadian Charter of Rights and Freedoms and the Canadian Human Rights Act. In addition, since 1995, all federal departments and agencies have been required to implement gender-based analysis (GBA), including GBA frameworks, and to include GBA in all Treasury Board submissions and Memoranda to Cabinet. Finally, in 2010, the federal government developed the Departmental Action Plan on Gender-based Analysis, which provides the structure for federal departments and agencies to create a gender-based analysis framework and to apply gender-based practice in their service delivery.

Topics of interest to understand and address gender inequalities include the following:

• Labour force participation, workplace decision-making, and entrepreneurial opportunities
• Income, unpaid work, and paid work/family balance
• Access to assets, education, and health services
• Gender-based attitudes and violence
• Social exclusion and treatment of minority groups

There must be differentiation between the concept of gender and the simple measure of phenomena by sex. The starting point in such an analysis is to make the distinction between two terms: "sex" and "gender." "Sex" is a reference to the relatively fixed biological and physiological characteristics that define men and women. "Gender" is a reference to the socially constructed roles, behaviours, activities, and attributes that a given society considers appropriate for men and women, respectively.

Sex-disaggregated data are needed to show the differences between women and men in a given society. Data must be disaggregated by sex to allow for analysis of gender issues. However, this alone is not sufficient for gender analysis. In this sense, gender statistics are much more than the simple disaggregation of data by sex: they also require policy-oriented approaches that will help identify the information required to inform and understand the problems.^{Endnote 1} Gender-based analysis can help tailor policies, programs, initiatives and services for the purpose of responding to the differing realities of women and men, and can help advance the evaluation of existing legislation and initiatives.

Gender statistics is not a discrete or isolated field. This field relates to all fields of statistics, cutting across traditional fields to identify, produce and disseminate statistics that reflect the realities of the lives of women and men. It is crucial to recognize that gender intersects with other identity factors, such as age, education, language, geography, culture and income. For full gender sensitivity, statistics must respect all the key gender issues in the country, and cover issues, such as maternal mortality or prostate disorders, which might affect only one sex.

2. Consulting and engaging data providers, policy makers and stakeholders

Engagement and consultations with data providers, policy makers and other stakeholders will increase the availability, relevance and quality of the information being produced. For example, in the context of gender statistics, data producers should interact more closely with data providers and users to allow improvements in the following areas:

• Identification of known and possible gender issues in society and the resulting need for greater knowledge and understanding
• Production of gender statistics with emphasis on determinants of outcomes
• Marketing and dissemination of gender statistics.
• Gender and policy analysis
• Use of gender analysis for policy actions
• Identification of data gaps and of instruments (concepts, questions, etc.) to help fill those gaps.

The experience of Statistics Canada's first Violence Against Women Survey (VAWS) (see box 3.6.1) also taught the organization that projects in this area require developing or enhancing customized and innovative solutions and that heretofore unmet or new data needs should be supported by experts—to rapidly develop the subject-matter knowledge and adopt the right strategies for responding to critics.

3. Developing the subject-matter knowledge and capacity to produce statistical outputs relevant to users' needs

Strategies to develop and enhance specific subject-matter knowledge and statistical capacity within an NSO include targeted recruitment, informal and formal training, in-depth literature review, consultations with external experts, and participation in working groups.

Because gender statistics is a cross-cutting theme, experience has shown that it is important to focus on awareness, sensitivity and training. Gender-based analysis training throughout the organization can be more effective than establishing bureaucratic checkpoints or vetting offices.

Awareness and sensitivity campaigns need to put emphasis on departmental and individual responsibilities and accountabilities with respect to gender. Such campaigns foster understanding of, and support for, gender-based analysis, and are therefore very valuable.

Training areas could be related to specific functions (i.e., for policy development, research, evaluation, communications, consultations, program delivery); other training deemed important through a needs assessment could also be provided. Networking and sharing knowledge can also be seen as important informal training opportunities. At the international level, for example, Statistics Canada participates in international working groups and conferences, including the Interagency Expert Group on Gender Statistics, to share information and network.

In Canada, the federal government has developed an online gender-based analysis and training tool available to all federal public servants, as well as those outside the federal service.^{Endnote 2} As part of the training tools provided by the government, the Gender-based Analysis Plus (GBA+) is an analytical tool that the federal government uses to advance gender equality in Canada. The "plus" in the name highlights that gender-based analysis goes beyond gender, and includes the examination of a range of other intersecting identity factors (such as age, education, Aboriginal person, member of a visible minority, language, geography, culture and income). This tool is used to assess the potential effects of policies, programs or initiatives on diverse groups of women and men, girls and boys, taking into account gender and other identity factors. GBA+ helps recognize and respond to the different situations and needs of the Canadian population.

4. Assessing the adequacy of data available to study the phenomena and identify the data gaps

Once users' needs are defined and statistical organizations understand what should be measured, the next step is to check the extent to which current data sources can meet these needs and, if necessary, to prepare the business case (i.e., to produce the statistics) for approvals. This requires not only an assessment of the data included in several sources, but also the condition under which they were produced, to ensure that they comply with minimum standards of quality.

Data in support of gender-based analysis come from several sources, including, but not limited to, census, administrative data and surveys. Different types of data have different strengths and weaknesses. The main challenge of using administrative data for gender-based analysis relates to the reliability of existing databases—data are often pulled from existing administrative databases, which may or may not capture information on sex.  If they do, the data are not contextualized in a manner that permits analysis of outcome determinants from a gender perspective. For example, in Canada, for justice statistics, the sex of the offender is collected while the sex of the victim is not. This information gap limits the gender-based analysis on victimization.

On the other hand, survey or population-based data often exclude certain population groups in Canada as a result of

• language barriers experienced by immigrants, especially new immigrants, when surveys are conducted in English or French; and;
• the exclusion of institutionalized populations – homeless persons, persons in shelters, persons in prisons, and persons in long-term care facilities.

Integrating different data sources through record linkage can strengthen gender-based analysis and address information gaps; however, doing so requires some advanced technical capacity and reliable matching variables.

Addressing the data gaps through the introduction of a new survey or the redesign of an existing one does not come without a cost. However, in some cases, this is the only option for analyzing phenomena that are otherwise not measured, as the experience of Statistics Canada's VAWS showed (see box 3.1).

5. Establishing the right methodology to collect the missing information and the right governance to support this methodology

When NSOs are trying to measure new social or economic phenomena, a literature review of the available statistical frameworks recognized internationally should be one of the first steps to consider, in order to determine the right methodology to conceptualize, collect and analyze the information. These frameworks not only usually provide a certain assurance of international comparability but also can greatly contribute to improving the quality of the data being produced (see Chapter 1.3: Following international standards).

The process of producing gender statistics is similar to that for other fields of statistics. This process typically involves a number of key steps included in the Generic Statistical Business Process Model that integrate a gender perspective.^{Endnote 3} The following steps are specific activities that help improve the overall availability and quality of gender statistics:

• Selection of topics that need to be investigated, and identification of the data needed to understand gender differentials and women's and men's roles and contributions in various spheres of life;
• Evaluation of existing concepts, definitions, and methods for producing unbiased gender-relevant information;
• Methodology and production techniques that take into account stereotypes and social and cultural factors that might produce gender-based biases;
• Choice of appropriate unit of enumeration about which to collect information and appropriate survey design, such as the definition of a sampling plan and sample size. That is best achieved by first having a clear idea of the types of analysis that the data must support. In that sense, establishing in advance a broad analytical framework that will combine sex-disaggregated with qualitative policy-relevant questions is recommended;
• Development of the appropriate data collection instruments, approaches and mechanisms, given the sensitivity of the subject matter (e.g., sensitivity training, sex of interviewers), that will also ensure a standard approach in data collection and avoid sources of gender bias;
• Processing of data using practices that will deliver reliable results; and
• Regular analysis and presentation of sex-disaggregated statistics in easy-to-use formats; dissemination of gender-sensitive statistical products to a wide range of users, including policy makers and planners; development of analyses and presentation of data that can reach policy makers and the largest audience possible.

The governance structure overseeing the development, production and dissemination of new or improved statistical outputs or programs must also be conducive to favorable change. For example, given the cross-cutting nature of gender statistics, there are different options for an NSO to consider when establishing its governance structure.

The creation of a gender focal point or centre of expertise within an organization is ideal to champion the integration of gender statistics in every subject-matter division. This gender focal point may be located within a specific organizational unit (e.g., a division), or this may involve reporting directly to the Chief Statistician. The organizational positioning within the NSO, however, is less important than the real ability to influence the entire organization.

Statistics Canada aims to incorporate consideration of gender issues into the operations of each subject-matter division by developing gender-sensitivity awareness and expertise. In the overall organization, there is no formal unit specifically dedicated to gender statistics, but there is a focal point included under the Social and Aboriginal Statistics Division. This focal point is responsible for producing gender-based analytical reports and for compiling the Women in Canada compendium publication, a collection of gender-based analysis articles published every five years. This collection requires subject-matter contributions from all across the organization to highlight the improvements in the area of, and the challenges to, gender equality.^{Endnote 4}

6. Make data and analytical outputs accessible to users

As shown in section 4 of this chapter, there are various ways to make statistical outputs available to users (e.g., official dissemination, use of various communication channels, and use of microdata access facilities). NSOs must ensure that users are acquainted with the type of information available to them and know how to access this information. The participation of NSOs in various networks, including networks of experts, academics and policy makers, can help further disseminate what is available, and can contribute to NSOs taking full advantage of the data holdings.

Statistics Canada's gender statistics program, for example, includes a broad range of sex-disaggregated tabulations, regular gender-based reports and analytical papers, and a vast collection of microdata files available to the public. Over time, the organization has participated in various task forces and working groups to share knowledge, experience and information. From time to time, Statistics Canada's work with some of these groups has led to a publication; e.g., Finding data on women: a Guide to major sources at Statistics Canada.^{Endnote 5} This document, produced by Statistics Canada for Status of Women Canada, gives policy analysts, women's groups, organizations, researchers and other data users a comprehensive overview of the scope and diversity of data available on Canadian women and men, as well as an indication of the ways in which these data can be used. It has certainly contributed to increasing the use of Statistics Canada's data in gender-based policy research.