Gender analysis must become integral to development policy. But the first step is to count women in, says UNCSTD gender adviser Shirley Malcom.
Research, policy and development programmes have, for the most part, failed to take women into account. But without separate and gender-specific assessment of these programmes, we will not know what works best for both women and men.
There are plenty of examples of where 'counting women in' has led to good policy, programmes and outcomes. We need to build on them by 'mainstreaming' gender into science, technology and innovation (STI) policy a process that should begin with reliable baseline data.
Agricultural policies often do not support smallholder subsistence farmers who, in Africa, are mostly women; or they fail to provide them with scientific knowledge to improve the quality and yield of their food crops.
For example, African women farmers are estimated to produce 20 per cent more than men from the same access to land and inputs, with only one per cent of the land and seven per cent of extension services.  By how much could global agricultural production increase if women had the same access to support and services as men?
In post-genocide Rwanda, women have made major contributions to the economic recovery of the country by supporting their families and community development. This was a result of changes in policy that allowed land ownership by women, who are now heads of households in the country. It also followed from women seeking knowledge of growing coffee, available from agricultural extension workers.
In Namibia, a palm tree conservation project was failing until the rights of local women to manage the forest were reinstated. Palm trees began to die months after women were asked to cut production of baskets made from palm leaves, used to store and carry milk and water.An investigation eventually revealed that the trees started to die when their care was transferred from the women, but once they regained this responsibility the plantation revived. 
Examples abound in research too. The symptoms of heart disease and action of aspirin were assumed to be the same for men and women, until research done on both groups showed this was not the case.
Experience has taught us it is a mistake to make assumptions about sex or gender.
Embedding gender analysis
These examples of positive change are a good start, but they are not enough. Unless this practice of gender analysis becomes an integral part of policy and decision making, long-term, sustainable development will not be realised.
Too often, good policy depends on the work of individuals committed to gender equality, but changes in leadership or priorities can derail good intentions.
We need a systemic approach if a requirement is in place that includes a process for pre-decision assessment and post-decision monitoring of proposals, for research or development projects, the information needed for advocacy and action will be in place.
Baseline data, once available, can be used to help establish a case for gender-sensitive policymaking as well as to document its value.
Some may argue that this kind of data collection, analysis and application would be too costly. But we need to compare this with the cost of programmes that do not work, people who are not served and development that is not realised.
And while establishing an assessment system that can be used to support gender mainstreaming may add costs, maintaining it should not.
Surveying the situation
Recent initiatives by the UN Conference on Trade and Development (UNCSTD) to document the importance of applying a 'gender lens' to STI policymaking in development are important first steps. 
One such step is a current UNCSTD survey of member country policies, which will help to determine if the infrastructure for gender analysis even exists. This involves a review of a country's data systems sources and locations, for example, and access to computers to determine what types of gender data are being collected. And if they are being collected, can they be disaggregated by sex and other characteristics?
In the long run, we will need a process to determine what is being counted in member countries, what these data can tell us, and how this information affects the choices we make in STI.
But this is only the first step. At the end of the day, leadership will be needed to act on the results of the survey to invest in gender-sensitive policies and programmes, to understand and value women's roles in development and to exercise the political will to translate data into deeds.
Mainstreaming gender in STI will lead us to ask different questions and ask questions differently; to formulate more effective policies and programmes, and to seek solutions in development that are sensitive and responsive to the roles that women and men play in the lives of their communities and the future of their countries.
Shirley Malcom is head of the directorate for education and human resources programmes at the American Association for the Advancement of Science (AAAS). She is also co-chair of the gender advisory board at UNCSTD and the Gender InSITE campaign.
 World Development Report on Gender Equality and Development. (World Bank, 2012)
 Villalobos, GR, et al. Diversity Makes the Difference. World Conservation Union. (2004)
 UNCTAD. Applying a Gender Lens to Science, Technology and Innovation. (UNCTAD, 2011)