Findings For the components of the Social Capital measures itself, some increased over the period, some stayed static, and some declined:       Trusting neighbours (+43%) Feeling safe in the neighbourhood (+14%) Not being careful what you say (+2%) Joining others to raise an issue (-2%) Attending community meetings (-6%) Being a member of a voluntary association or community group (-9%) There were minimal changes in the percentage of women from 2012 to 2017 that could be described as having “hi social capital”. Secondly, these changes in social capital over the three rounds were statistically significant: social capital rose from 2012 to 2014 and then dropped from 2014 to 2017. The interesting finding here is in the difference between “hi” and “lo” groups of women, with the former showing an increase in social capital by 2017, with the “lo” group showing a decline. It is not possible to see an obvious reason for this difference other than the difference being due to residence, with the “hi” group being mostly rural women. Although radio is far and away the most common source of information, there was a decline in accessing all forms of information. There was small increase in the number of women using the internet, but it was a very small percentage of women that had access to internet. Unfortunately it was not possible to compare use of social media over the period as the question was only asked in 2017. The decrease in the number of women using radio and television is not offset by a significant increase in the use of other media. Women’s perception of freedom – to say what one thinks or join any organisation one chooses – rose significantly from 2012 to 2014, but then dropped enormously from 2014 to 2017. The changes in the first period might be attributable to improved economic situation under the Inclusive Government and a largely peaceful, but surprising election. But the changes in the second period, 2014 to 2017, may either reflect the deteriorating economic situation or a growing disenchantment with politics in the light of the faction fighting in all political parties. One of the more disappointing findings was in the very low levels of engaging public officials, what we described as a measure of agency. Zimbabwean women do not engage duty bearers on our evidence. Whether this is a lack of agency as such or the perception that duty bearers are unapproachable or ineffective, or both factors are operating, is not clear. However, less than 10% of women report that they often engage duty bearers. This is extended by the responses of women on political participation. Virtually none of the women reported working for a candidate or political party, and the numbers that attended a political meeting dropped markedly between 2012 and 2017: from 59% attending a meeting or rally in 2012 to 43% in 2017. This all suggests that women are disengaging from politics, and this is worrying. Women continue to express support for democracy as the best model for the country, but this support is declining. Only 46% of women thought that Zimbabwe was a democracy in 2012, which declined to 35% in 2014, rising to only 38% in 2017. 4