Quantitative baseline report Going for Gold: Tanzania & Ghana (revised version)
In the Tanzanian ASGM communities, women earned on average USD 56 per month, only 25% felt women always have the same earning opportunities as men, and mine workers had very little protection: on average 0.65 items of protective gear, 58% of men and 37% of women had worked with mercury in the past month, and 1-3% had received any job training.
Overview
Quantitative baseline for an impact measurement of the Going for Gold programme, which aims to improve women's economic empowerment and health in communities in and around artisanal and small-scale gold mines (ASGM) in Geita District, Tanzania, and Western Region, Ghana. Data were collected in February-March 2017 through a survey of vulnerable community women and a survey of male and female mine workers in all programme communities and mines. The study covers income, credit and savings, women's agency and leadership, household decision-making and intimate partner violence, mining practices and safety, and SRHR and health knowledge. It also reflects on how the survey questions performed, for use in later rounds.
Key findings
- In Tanzania, 68% of community women worked to earn an income, earning on average USD 55.82 per month (USD 108 for the 4% in formal employment); in Ghana 82% worked, earning USD 82 per month. Women with education above junior high earned significantly more.
- Only 25% of Tanzanian community women said women always have the same opportunities as men to earn money, and 3% rated their own earning opportunities as high; married women were less likely to work than single or divorced women.
- 77% of Tanzanian community women reported no access to credit; about 18-23% of respondent groups could get a loan within 24 hours, about 10-13% had taken a loan in the past 3 years, and only 3.5-4% of mine workers and 11.3% of community women had a bank or savings-group account.
- 55% of Tanzanian community women belonged to at least one group (mostly religious groups, 82%), but 44% of group members said they had little or no influence on group decisions; 20% reported ever being slapped or having something thrown at them by a partner, and 9% reported forced sex, which the report says may be underreported.
- In Tanzanian mines, women earned less per day than men in most positions (e.g. mining/extraction USD 2.31 vs 3.82), only 33.7% of women and 40.9% of men were satisfied with their income, and most men (89%) and women (77%) worked in extraction or ore transport.
- Mine safety in Tanzania was poor: 27% used any protective gear (average 0.65 items), 58% of men and 37% of women had worked with mercury in the past month, 22% said they or a colleague had an accident at work (mostly pit or wall collapses), 8% reported children working in their mine, and only 1% of women and 3% of men had received job training in the past year.
- SRHR knowledge was relatively high in Tanzania (72% of women could name at least 3 contraceptives; 85% said they influence the choice of contraceptive), but 39-44% said it is not acceptable in their community for a woman to decide when to have sex or to refuse sex; average satisfaction with health services was 5.8 out of 10.
In figures
27 % of mine workers
58 % of mine workers
37 % of mine workers
22 % of mine workers
8 % of mine workers
Data table
| Item | % of mine workers |
|---|---|
| Used any protective gear | 27 |
| Men who worked with mercury (past month) | 58 |
| Women who worked with mercury (past month) | 37 |
| Self or colleague had an accident | 22 |
| Report children working in their mine | 8 |
Recommendations
- Survey design (for later rounds): change savings-amount questions to multiple-choice ranges, and add financial literacy questions, because many respondents refused to state savings or confused savings with credit.
- Survey design: ask how often workers are paid and how income fluctuates, and ask for the highest and lowest real income in the previous year instead of 'good month' and 'bad month' estimates.
- Survey design: clarify what counts as 'training', add a 'my position does not require this' option for protective gear, and ask which chemicals respondents personally use.
- Survey design: rephrase intimate partner violence questions as community-level statements, and add Gender Equitable Men (GEM) Scale questions on gender attitudes.
- Survey design: use more pre-coded answers and 'not applicable' options; avoid changing questions unnecessarily between rounds, as this adds measurement error.
Lessons learned
- Some validated or borrowed questions (e.g. OPHI household decision-making and group-participation items) did not fit respondents' economic reality and produced uninformative answers.
- Direct questions on sensitive topics (savings amounts, intimate partner violence) led to refusals or likely underreporting.
Methodology
Two structured surveys, a community member survey of vulnerable women (849 in Tanzania, 804 in Ghana) and a mine worker survey (712 in Tanzania: 374 women, 338 men; 588 in Ghana), covering all programme communities and mines; data collected by CDC Consult in Feb-Mar 2017 and analysed by ICE using WEI-inspired indexes, ANOVA and regression.
Limitations
- Actual sampling per mine diverged significantly from the targets.
- In Tanzania, the number of women interviewed per community was set proportional to community size because of a miscommunication, instead of an equal number per community.
- Some communities and mines were removed from the Ghana dataset after the survey because they will not be included in the programme.
- Several questions (migration, payment method, savings and credit, training, IPV, group decision-making) were poorly understood or likely underreported.
- Positions held by mine workers differed between countries, and it is unclear whether this reflects reality or sampling bias.