Radical
Statistics

The Journal

The Subjects

The Books

News

Links

About

Home

Developing gender statistics in the UK

Linda Murgatroyd*

Introduction

There have been a number of exciting developments in official statistics in relation to the statistics of gender issues in the last few years, and further work is planned for the future. This paper reviews recent and planned developments. To some extent future developments will depend on what users require, and on how successfully they can articulate their needs and work with producers of statistics to meet those needs.

What are gender statistics?

The term gender statistics itself is not very useful except as a shorthand to refer to statistics relating to people or activities with a significant gender dimension. Potentially these cover the majority of statistics with a social content, and some might argue even beyond. However, in some areas the gender dimension has been hidden - perhaps through measurement or conceptual difficulties. In addition, significant activities disproportionately affecting one gender have not been subject to statistical measurement, leading to an invisibility of these activities in statistical accounts of society. Where there is a lack of gendered statistics and/or statistics of highly gendered activities there is clearly a risk that the evidence to inform and monitor policy will be inadequate.

• Gender disaggregation. At a simple level, all social life is gendered. However, statistics may not be. Most social statistics include information on gender, and can therefore be analysed by gender. This is true of both administrative sources and of surveys of people or households. Some sources measure activities indirectly (e.g. money or a gender-neutral administrative source), or collect information about units other than individual people (for example households, businesses, schools, service utilisation). In these cases the gender dimension is likely to be hidden. More sophisticated (and expensive) alternative measures would be necessary to identify it. In practice official statistics are collected and used for a variety of purposes. There are costs at each stage and priorities for analysis and publication.
• Clearly particular topics may be of greater relevance to one gender than another. Gaps in statistics are not gender-neutral, given the different activities and characteristics of men and women. For example, gaps in information on some kinds of crime or health conditions affect the overall picture of men and women as victims or perpetrators of crime, and of the quality of their health. There may be a number of reasons for gaps in what is measured - including their location, reporting patterns, and cultural norms. Sexual assault, domestic violence and abortion are all examples of events whose measurement has been susceptible both to changing cultural norms and to measurement methods.
• These issues come together where statistics are needed to measure the impact of policy separately on women and on men. A modelling approach may be needed which brings together data from a range of sources, together with information about behaviour from yet other research. It is important that assumptions that lie behind the statistical measures and their analysis are well-founded. It may not be appropriate to assume that the behaviour of men under some circumstances will necessarily be the same as that of women. A diversity of tools is needed to inform and monitor policy. Qualitative and quantitative research projects are often needed to explore new topic areas and explore specific issues not subject to regular statistical measurement. In some cases such research may help make the case for more regular statistical measures; they may also help in developing measurement tools.

Recent achievements and current developments in government statistics

Improving communications about statistics and making them more accessible

The 1990s saw a wide range of developments in official social statistics, many of which directly affected statistics relating to gender. The general context for official statistics has moved away from the Rayner doctrine and towards wider consultation with users of government statistics and greater transparency of methodology, culminating in the White Paper in late 1999 and the establishment of National Statistics in 2000.

A number of specific initiatives have helped make the vast amount of information on women and men more easily accessible, both to end-users needing comparative information in summary form through publications and to those wanting to carry out their own analyses of existing data:

• In 1995 Social Focus on Women was published by the Office for National Statistics (ONS), to coincide with the United Nations (UN) World Conference on Women in Beijing;
• Following the Global Platform for Action arising from the Beijing conference, the Government Statistical Service's (GSS) Social Statistics Committee in 1997 published a policy statement on publishing statistics disaggregated by gender:

  The GSS aims always to collect and make available statistics disaggregated by gender, except where considerations of practicality or cost outweigh the identified need. (Hansard, 15 July 1998, col. 201)

• This was followed by closer working between the Office for National Statistics, the newly-established Women's Unit (WU) and the Equal Opportunities Commission (EOC). Social Focus on Women and Men, and A Brief Guide to Gender Statistics were published jointly by ONS and the EOC in 1998.
• In addition, ONS worked with the Statistics Users' Council, EOC, WU and others to hold a Conference in March 1998 on Making Gender Count and to establish a Gender Statistics User Group (see Irene Bruegel's article in this Issue). The aim of this was to provide a means of communication between users of statistics and producers, to inform future developments.
• The statistical service advised the Women's Unit as they developed their guidance on policy appraisal, to try and ensure an awareness of the gendered impact of policies and that the statistics needed to inform and monitor policy include a gender dimension.
• In June 2000 the United Nations reviewed progress on women's rights and the elimination of discrimination since the Beijing conference. A pocketbook of data on women and men is being produced, and the implementation across government of the policy statement on gender disaggregation will be reviewed (ONS, 2000 forthcoming).
• Meanwhile, work was also taking place on developing new data sources and ensuring existing sources were modified to better meet users' needs. A wide GSS consultation in the early 1990s had identified a need to fill gaps in measuring unpaid activities, and how people spent their time, and major developments have been taking place in these areas.

Measuring time-use

The measurement of unpaid activities was identified as a major gap in UK statistics. Because more of women's activities are unpaid, and more time is spent by men than by women in paid work, it is a gap which disproportionately affects women's activities. To fill this gap, ONS built on work done by academics in the UK and by colleagues internationally to develop two separate survey-based approaches to how people spent their time in the UK.

In May1995 a simple time-use module was included on the General Household Survey (GHS). The 2005 respondents ticked one of 30 pre-coded activities for each 15 minute time intervals a 24-hour period (see Gershuny and Smith, 1995; Koudra et. al., 1996). This was a cheap and effective way of gathering broad information about time-use, and showed some interesting patterns across the board. It has been valuable in providing overall estimates of unpaid activities (see below) and in raising the profile of these activities. This in itself is an important element of incorporating them into mainstream policy thinking. This time-use data has also been used to complement more detailed information from other sources on specific unpaid activities such as care and travel.

Time-use measures are, however, very sensitive to the particular methodology used. Very simple methods as used in the GHS naturally hid activities taking place in very short bursts, or concurrent with other activities. The number of categories and their labels also significantly influences the patterns of activity measured. The May 1999 GHS included a similar time-use module to that used in 1995 but added extra questions on travel, resulting in a substantial increase in the amount of travel measured. Such issues are not confined to time-use surveys, however; slight differences in questions on unpaid adult care gave rise to estimates ranging between 9% and 15% of adults providing such care (McConaghy, 1998; Thomas, 1998; Noble et al, 2000).

A more sophisticated approach to measuring time-use has been taken by Eurostat, who had been developing methodology in support of a European Time-Use Survey. A United Kingdom (UK) survey has built on the international work co-ordinated by Eurostat and extensive piloting has been carried out over the last two years. The survey went into the field for 12 months from June 2000 and some 5,000 households will be surveyed. This uses shorter time-slots of 10 minutes, and invites respondents to describe their activities in their own words and to record parallel activities and contextual information. This approach offers many ways of categorising and analysing activities and thus should be a very flexible research tool. The open approach to recording time-use (in respondents' own words), together with a wide range of other socio-economic information about them (using definitions harmonised with other surveys) will enable a range of different analytical concepts to be developed and used. There are also interesting questions around how different people perceive and report their use of time, which will be an interesting issue for further qualitative research. It would not be surprising to find systematic patterns including gender effects.

The main disadvantages of this survey method are its cost and the burden on respondents. The survey is complicated and takes a number of days to complete. This means response rates tend to be low: household level response rates at the pilot stages have been little more than 50%, and in addition there is partial response within households. Much effort has been put into exploring ways of boosting response, and appropriate weighting of results in analysis will be crucial.

These Time-Use Surveys represent a major innovation for UK official statistics. The main survey in particular will be a unique resource allowing a wide range of issues to be researched in quite new ways. In addition, the process of developing this data source is of interest:

• Throughout this project, user consultation and collaboration with others in the UK and abroad have been prominent features;
• Funding has been jointly from a number of government departments and from the ESRC (Economic and Social Research Council);
• Consultation gave rise to the work in the first place and continued throughout; ONS held a number of seminars for different audiences in the early stages of development, and also supported seminars organised by the ESRC in 1998 and the International Association of Time-Use Research in 1999; and
• ESRC and academic experts are on the steering group for the survey; further consultation will take place as the reporting strategy is developed, and the aim is to allow raw data to be promptly available for users, as well as producing some kind of overall report. The involvement of a range of academic and policy users will be crucial in developing understanding of data, and concepts and methods for using it.

Notwithstanding the methodological difficulties, the detailed measurement of people's unpaid activities together with rich contextual information about them will undoubtedly provide an exciting new perspective on a range of issues. The availability of good time-use data seems likely to raise awareness of unpaid activities, particularly those of women, and to offer a new and more gender-neutral perspective which will illuminate a wide range of policy issues.

Satellite accounts

Unpaid activities are not currently included in the standard internationally accepted system for measuring economic production: the National Accounts. As these accounts are the foundation stone for economic policy, this severely limits the analysis that can be carried out of the overall impact of policies. As well as limiting analysis, there are political issues associated with the invisibility of particular ranges of activity, such as housework, health and care, the environment, or travel. As the distribution of unpaid production activities (and between the paid and unpaid sectors) is gendered this also has implications for gender inequalities.

Because of international concern about this issue, the 1993 revisions to the System for National Accounts (SNA) made provision for satellite accounts to complement the national Accounts. These have the same form as the core national accounts but cover areas beyond the production boundary. They may often use units other than money. Exciting work is now under way to develop a range of satellite accounts, and an experimental household satellite account for the UK was published in 1997. It was a crude, inputs-based account, based on the 1995 GHS time-use module. It found that more time was spent on unpaid than paid productive activity (see Table 1). Further, the value of unpaid work was estimated at between 40% and 120% of gross domestic product, depending on the method by which its monetary value was measured.

Table 1: Time-use of adults in Great Britain, 1995 clickHERE

ONS has now adopted a strategy to develop satellite accounts a lot further. This is a massive undertaking and will take many years. The time-use data will be an essential contribution on the inputs side. Pioneering work is also going on to piece together accounts on the outputs side, for example looking at numbers of children and adults cared for in different ways, at accommodation services, clothing produced and serviced, nutrition provided etc. This is going beyond what most other countries are attempting to do, and will significantly improve our understanding of the connections across socio-economic spheres of activity. ONS have established a steering group for this work involving external experts.

Other developments

A number of other developments have been taking place within government or commissioned by government, to provide new gender analyses of existing data, including for example, a range of work on women's individual incomes at specific points in time and over their life-times. Specific information gaps relevant to policy on women and gender issues (such as teenage pregnancy, and violence against women) have been researched. The Women's Unit continues to work with others to ensure that women's perspectives and gender dimensions are included in research across government, so far as appropriate. A report on lessons learned and good practice in the UK was being prepared for the UN Beijing Plus Five conference which took place in June 2000.

Data sources other than time-use have also been developed to improve the coverage of unpaid activities. Care questions have been regularly featuring on the GHS, but they are also now on the Family Resource Survey and a question will be included in the 2001 Census. Although the latter is a relatively crude question it will be particularly important in the context of the small area and demographic information provided by the Census.

Work to develop and harmonise mainstream statistical concepts has also had implications for gender statistics:

• The introduction of the 'household reference person' as opposed to the 'head of household' concept for all major government household surveys is a significant step forward in that it removes a gender bias that was positively misleading in a small proportion of cases;
• Similarly, the new Socio-Economic classification is more gender-neutral and makes more distinctions among types of work more commonly done by women; and
• Work is also going on to look at more sophisticated typologies of families and adult-child relationships within and between households. There has been some consideration of the range of definitions of household and family units (including benefit units, multi-family units within households, etc).

Where next?

This is an exciting time of change for official statistics, with much potential for developing statistics relating to gender issues in future. Following the 1999 White Paper, a new framework for statistical work in government will be launched in 2000. This will include more open and systematic mechanisms for consultation on issues such as the developing scope of official statistics and methodological issues, and improvements in joined-up dissemination of statistics, including through the Statistics Commission. The strategic approach to developing UK social statistics will continue. Following the work to harmonise and rationalise surveys of persons and households, a strategic approach is now being taken to longitudinal data sources, again working across government and with the academic and private sectors. Ways of releasing the potential offered by administrative sources while guaranteeing confidentiality of personal information are also being looked at. Developments both in Information Technology and the wider legal and constitutional context of course have great significance here.

There are a number of challenges and opportunities for gender-related statistics, including:

• to ensure the continuing development of appropriate concepts and methods data collection methods for statistics where there are still gaps, particularly in the 'private' domain;
• to ensure that gender issues are appropriately included in new areas of statistical activity (the Internet?); and
• that as analyses are developed e.g. in longitudinal and other models, and in new accounting systems, they are based on appropriate assumptions, consistent with wider research findings.

Effective engagement of users of statistics with those producing them is crucial if the data are to meet users' needs. The Gender Statistics User Group was established to facilitate this communication and it has already made some useful contributions to specific consultations, for example on work on women's individual incomes. To be effective, it should now build on the wide support demonstrated at the 1998 conference, and also build closer relationships with users and producers of statistics in other areas, and participate actively in the new consultation mechanisms with the Statistics Commission when they are established. It should continue to take a leading role in key areas, including thinking further about analysis of existing data, and conceptual issues, while working to mainstream gender across all areas of statistics. To the extent that it and producers of statistics are successful in this, it may eventually work itself out of a job. However, there is still a way to go in some areas.

NOTES

*Linda Murgatroyd is statistics Head of Profession at the Cabinet Office. The views expressed here are personal ones.

The Gender Statistics User Group can be contacted via Richard Laux at ONS, Social and Regional Division, 1 Drummond Gate, London SW1P 2QQ, email: richard.laux@ons.gov.uk

Further information on the UK Time-Use survey is available from: June Bowman, Socio-Economic Division, ONS, 1 Drummond Gate London SW1P 2QQ, email: june.bowman@ons.gov.uk

Sue Holloway is taking forward work on satellite accounts and can be contacted at Economic Assessment and Strategy Division, ONS, 1 Drummond Gate London SW1P 2QQ, email: sue.holloway@ons.gov.uk

The National Statistics website address is: http://statistics.gov.uk

The Women's Unit website is: www.women's-unit.gov.uk.

REFERENCES

Cabinet Office (2000), Pocketbook of Statistics on Women and Men.

Central Statistical Office (1995), Social Focus on Women, The Stationery Office.

Making Gender Count, Report of the Gender and Statistics conference, 1998 (available from Gender Statistics User Group)

Gershuny, J. and Smith, R. (1996), Report to the Central Statistical Office on the Development of a Simple Time-Use Schedule.

Koudra, M., Church, J. and Murgatroyd, L. (1996), 'Where have all the hours gone? Measuring Time-Use in the UK' in Statistical News.

McConaghy, M. (1998), 'The prevalence of informal caring. Characteristics of caring and those who are cared for', in Disability and Care: Questions and Needs considered: Proceedings of a Conference held on 15 June, ONS.

Murgatroyd, L. and Neuburger, H. (1997), 'A household satellite account for the UK', in Economic Trends, 527, October.

Noble, B., Dickson, M., Gershuny, J., and Fugeman, D. (forthcoming), 'Using Omnibus surveys to investigate travel', in Transport Trends 2000.

Office for National Statistics, (1998), Social Focus on Women and Men.

Office for National Statistics, (1998), A Brief Guide to Gender Statistics.

Sturgis, P. and Lynn P. (1998), The 1997 UK Pilot of the Eurostat Time-Use Survey, GSS Methodology Series No 11.