The world is becoming increasingly data-driven. There has been an exponential growth in data production and collection in the last decade. Public and private institutions are generating volumes of data – with expansion in speed, variety, and complexity. Data based decision making has taken prominence and high-impact policy and business decisions are made using large-scale data. One of the major reasons for this uptake in data usage is its cost-effective and revenue-generating outcome. According to a report by McKinsey, data-driven companies are 23 times more likely to acquire customers, and 19 times more likely to be profitable.
The data being collected and used unfortunately represents only half the truth by excluding facts of half the humanity i.e., women.

Data on women are less documented and available, thus representing a partial picture of the women’s and men’s lives and the gaps that exist between them. This gender data divide is facilitating bias in product/service development and delivery by public and private institutions. In an era of big data, machines are fed with historical data which are biased given male overrepresentation across all areas of life, thus, leading to products and services created primarily for men, who are considered as default humans. The gender data divide occurs because researchers do not gather data from women and the historical data reflects societal gendered roles. For example, a language-categorization artificial intelligence software recognized computer engineers with men and homemakers with women.

In the groundbreaking book, Invisible Women: Exposing Data Bias in a World Designed for Men, which won the Financial Times and McKinsey Business Book of the Year Award in 2019, author Caroline Criado Perez mentions that women constitute around half of the globe but the consumer market, medical research, technology, government policy, and workplace culture have targeted and accommodated the average male, often leading to disastrous consequences. One of the examples stated by the author is that women are 47 per cent more likely to be injured in a car crash than men. Moreover, they are also 17 per cent more prone to die.

The major reason for this is since the 1950s, car crash test dummies have been modelled on a 76 kilograms male. Therefore, the design of seatbelts in cars for safety support men while women face the consequences of this lack of data for them in product development.

The data divide has become visible in recent years to researchers and organizations and efforts are being taken to address it. For example, technology companies like Amazon, Microsoft are removing algorithm bias for men from their software and artificial intelligence platforms, the Government of Uganda has formed a strategy for the development of gender statistics, to name a few. The realization that products, services, policies are not efficient for men and women with less representative data, has led to a concentration on the concept of gender data.

Gender Data entails data collected and presented by sex as a primary and overall classification. It reflects gender issues and is predicated on concepts and definitions that adequately reflect the diversity of women and men and capture all aspects of their lives. It is developed by data collection methods that consider stereotypes and social and cultural factors that may induce gender bias in the data.

Importance of Gender Data and Identified Gaps

To ensure development in the efforts to reduce the gender gap and to empower women and girls and curtail the negative impact of COVID-19 which has disproportionately impacted women, there is a requirement for more and better gender data. Moreover, gender data can assist in designing products/services/policies and delivery mechanisms that are gender-inclusive rather than biased. It will ensure robust evidence generation on use and impact of product/service/policies on the population.

Gender data can assist to understand the various issues and policies that impact women differently or solely. It is to be considered while informing policymaking and in the assessment of policy impact. Gender data is also important to monitor the progress of women and girls in society.

Gender-inclusive data collection mechanism can yield greater gross domestic product, revenues and generate other non-financial positive outcomes such as employee retention, and operational replicability and scalability.

Gender Data gaps exist given the lack of prioritization, data collection and reporting biases and lack of safety for girls and women to report their experiences. There are less than 13 per cent of countries that have dedicated budgets for gender statistics, less than a quarter of gender-related Sustainable Development Goal (SDG) indicators have reliable monitoring mechanisms, and more than 100 low and middle-income countries lack adequate civil registration and vital statistics systems.

The gaps exist given gender biases ingrained in the concepts, definitions and classifications used, in the way questions are asked, in how the samples are drawn for population surveys and in how data are collected. These biases impact the accuracy and reliability of collected data. For example, labour force participation surveys often ask and consider only primary economic activity for the respondents. This leaves out women’s contributions in unpaid care and domestic care. In Uganda, the incorporation of secondary activities as paid work increased female labour force participation to 87 per cent from 78 per cent. There are some international statistical standards and classifications that can assist to remove/reduce some biases and improve the international comparability of the data.

Investing in Gender Data: Collection, Dissemination and Decision Making

Understanding the need for Gender data, efforts should be made to increase investments in gender statistics through the provision of cost estimates for building core gender data systems and paving the way for improving the gender data funding ecosystem to produce data to measure gender gaps and ensure gender equality.

There is a deep requirement for mainstreaming gender into national statistical strategies. Investment within the national statistical system can improve data coverage, quality, and timeliness about women and men. Investments are also to be made to evaluate the impact of any disaster, natural or human-caused, to understand its impact on women and men differently to develop better disaster management strategies for recovery and rehabilitation, on a case-to-case basis.

Core gender data system should span data surveys across population and housing census, household health and wellbeing, income and expenditure survey including topics on poverty, civil registration, distribution of income etc., labour force survey, time use survey and agriculture census. Above this, the system should also possess a repository of historical data on women across all fields of science, arts, politics, to name a few. There should be an enabling ecosystem for administrative data to record gender-disaggregated data on regular operations of the government or any institution to understand the uptake and related challenges.

The true value of gender data can only be explored if it is used in dissemination and decision making. The countries/institutions should also lay out plans for linking data to policy. The leadership should be enabled to analyze and use gender data. Countries have faced consequences from a lack of production of gender data and a lack of effective publishing practices of such data that can assist citizens and analysts to engage in policy conversations.

In India, though National Statistical Office (NSO) releases data on employment, labour force and active participation periodically through the periodic labour force surveys, it does not account for or collect data on women's unpaid labour. There are time-use surveys that highlight the above issue collecting data on how each household member spends time, but it is less frequent. The last time use survey was conducted in 1998-99 for six states. Data periodicity needs investment to enable assessment over time.

The country also needs robust data on access to safe working conditions in the formal and informal workplace and childcare facilities to understand women’s employment conditions and make informed decisions to improve female labour force participation.
If we look at some of the skill schemes data for employability and entrepreneurship, it provides gender-disaggregated data on training sessions, beneficiaries and in some cases amount of credit accessed. It does not include gender-disaggregated data on types of training provided and attended which would help in understanding skill preferences and divergences for women to be used in designing schemes that would generate economic opportunities for women.

In the case of agriculture census, there is no gender-disaggregated data for titleholders of land within the household. Land ownership is integral to the definition of a farmer in India and since most women do not possess the land in their name they work, they remain excluded.

There is a lack of data on women’s participation in politics across all levels. The Election Commission of India and Ministry of Panchayati Raj does provide gender-disaggregated data on women participation in the lower house of parliament and rural self-government i.e., panchayats but disaggregated data on central and state leaders i.e., Prime Minister, President, Chief Minister, Governors, ministerial portfolios, and urban local self-government i.e., municipal corporations, Zila Parishad is unavailable. 

The above are some of the examples which need gender inclusiveness, in collection and use. Therefore, investments to generate Gender data in existing data systems – surveys and administrative can help in creating a better ecosystem for data-based decision making in India. Data is an important public good. Extensive and high-quality data can assist a multitude of stakeholders to create best in class products/services/policies and capitalize on the potential of women and men alike.

This article has been authored by Guriya.