Summary of Pilot Implementation Findings

Between September and December 2020, IFES worked with Nigerian partner organization KDI to test five adapted gender-sensitive indicators as part of the NEVR initiative,¹¹ an existing KDI and IFES electoral violence monitoring project. Monitoring took place in six states holding by-elections and two states holding statewide local elections. One monitor also remained in Ondo to monitor any potential post-election violence and to test the common place monitoring indicator. A total of 42 monitors, including 23 women, were deployed across the nine states (Ondo, Bayelsa, Cross Rivers, Imo, Kogi, Lagos, Plateau, Gombe, and Abia). A full report on the pilot is included as Annex B.

Nigeria was chosen for the pilot in part due to IFES’ history supporting electoral violence monitoring throughout the country in collaboration with KDI. KDI works alongside local and international organizations in Nigeria to promote peaceful elections, active citizen engagement in democratic processes, and a sustainable economy. As well as having an existing working relationship with IFES, KDI was selected as the local implementing partner for this pilot due to its established network of local monitors across Nigeria; knowledge of local customs and ethnic groups in different parts of the country; and expertise in collecting and analyzing subnational data. Nigeria also fit the criteria developed for selecting a pilot country as it has experienced political, electoral and/or extremist violence; had upcoming pertinent political events (elections); had regional and cultural diversity across states; and had existing violence monitoring efforts.

In partnership with IFES, KDI adapted five indicators from the short list created for this framework into its existing electoral violence monitoring effort. The indicators needed to easily fit into existing electoral violence monitoring efforts but also be applicable or adaptable to other early warning systems with a broader monitoring focus beyond elections. The indicators were selected after conducting desk research on gender norms and women’s rights in the identified Nigerian states and holding consultations with KDI.

The five gender-sensitive indicators chosen were:

• Number of incidents of targeted violence and intimidation against voters, electoral officials, and party representatives – disaggregated by sex, victim, and perpetrator;
• Number of arrests of individuals active in political and electoral processes – disaggregated by sex and by the level of violence during the arrest;
• Number of campaign communications that utilize or refer to misogynistic, homophobic, or sexist references or propaganda;
• Percentage of individuals who are women present in designated common places; and
• Rate of gender-based violence, including sexual violence, leading up to and after the election.

As the pilot progressed, the focus further broadened beyond just monitoring electoral violence, in part due to the postponement of the elections and the temporary suspension of all campaigning.

The data collection strategy used for these indicators included a combination of a) filling out an incident reporting form that was updated to include additional gender, victim, and perpetrator disaggregates and pertinent pilot indicator information; b) weekly monitoring of social media accounts of candidates, whereby the local monitors developed a monitoring schedule to weekly check the Twitter and/or Facebook accounts of all candidates who had accounts as well as daily monitoring of local media outlets; c) identifying common places that are generally well attended by women and monitored them at the same time and day each week; and d) observing political events and campaign rallies to report on incidents and interview eyewitnesses. As per KDI’s existing verification methods, all reported incidents that were not witnessed directly by the local monitor or reported on in mainstream news outlets had to be verified by a second source. These secondary reports could include incidents reported through eyewitness accounts (but not from monitors) or the toll-free phone line set up as part of NEVR monitoring efforts. This second verification layer could include seeking a second eyewitness account, obtaining confirmation from local police stations, or visiting hospitals to confirm a victim of violence had been hospitalized, while retaining confidentiality without identifying the victim.

There is no value in collecting data without effective tools for managing it and quickly obtaining disaggregated values from new entries. To demonstrate that any organization – regardless of resources and capacity – could take up these efforts, it was important for the project to utilize cost-effective and user-friendly data management tools. The team created a spreadsheet-based database in Excel that can be hosted on a platform like Sharepoint or Google Sheets and permits simultaneous data entry and updating.

Given the remote collaboration between IFES and KDI, it was also beneficial to provide regular feedback on the quality of the data being collected and identify potential improvements in the data collection process (this was conducted on a weekly basis). Without this immediate feedback, adjustments would not have been made and the quality of the data would not have improved over the course of the pilot. As part of this effort to track data quality, a dashboard was set up to evaluate the weekly data being received across all five indicators. Data received for each of the indicators was scored weekly on a scale from one to five (with lower numbers reflecting weaker data quality). Despite the overall data collection challenges experienced during the pilot outlined below, this feedback system and data quality scoring process enabled the consistent improvement in quality of data being reported. The average data quality score improved from 1.5 to 4.3 out of five over the course of the monitoring period.

In addition to COVID-19 related restrictions on movement, there were other challenges that impacted the local contexts being monitored and the ability of monitors to safely and regularly collect data. These meant that some monitors were not able to systematically collect data on some of the indicators during this period, leading to less data than anticipated. Two of those challenges were: 1) deadly protests against police brutality in October, which led to curfews, lockdowns, and the postponement of several local elections, and 2) the small-scale by-elections occurring in six of the states, which meant that the elections were so localized that they did not generate either much mainstream media attention or interest from the local communities in which they were happening. Small-scale elections are accompanied by a limited number of campaign events held, which drew small numbers of people.

In addition, given that training for monitors was done quickly to ensure that there was enough time in the pre-electoral period to allow for data collection, some edits to the data collection methodology – that affected the data being collected – occurred during the data collection period. Where possible, IFES continued to work with KDI to improve the quality of the data that could be collected.

In the post-pilot survey, the top three challenges listed by monitors in terms of data collection were concerns for their safety; lack of consistent access to police reports; and too many changes in data collection/indicators during the monitoring period. These are important lessons that need to be applied to future efforts, and a summary of key learnings from the pilot to address these issues is included below.

Over the course of the pilot, there was also noticeable progress made in the collection of sex-disaggregated data. When reporting on violent incidents and arrests, monitors increasingly provided, where available, details on victims and perpetrators, including: the total number of victims; the number of female victims; the number of male victims; the number of victims whose gender is unknown; the total number of perpetrators; the number of female perpetrators; the number of male perpetrators; and the number of perpetrators whose gender is unknown. This sex disaggregation enabled IFES to have a greater understanding of the incidents being reported on and then analyze overall victim and perpetrator rates. IFES also worked with KDI to update its monitoring forms to disaggregate by type of perpetrator and type of victim to include perpetrator codes such as security force/police; member of the public; individual linked to a political party; partner/family violence perpetrator; sexual violence perpetrator; election worker; or unknown. The final victim codes included government/state actor; political party leader or supporter, candidate or candidate supporter; party agent; election observer/monitor; election worker; voter; protester; government/local authority property; gender-based violence victim; or member of the public. For future efforts, victim categories could be further expanded to include, where applicable, women activists, NGO workers, journalists, and prominent figures.

At the end of the pilot, 26 percent of local monitors indicated that data for gender-sensitive indicators was challenging to collect; 29 percent said it was both easy and challenging; and 45 percent thought it was easy. However, despite more than half of the monitors experiencing some challenges with gender-sensitive data collection, 88 percent of them agreed that gender-sensitive indicators are very helpful in understanding conflict in Nigeria, and 12 percent thought they were somewhat helpful. None of the monitors thought that gender-sensitive indicators were either somewhat unhelpful or not helpful at all. This finding highlights the understanding among early warning monitors of the need to continue improving the data collection strategies for gender-sensitive indicators and why these are important to integrate in early warning systems.

While the pilot highlighted some significant challenges in collecting data on gender-sensitive indicators, it also provided IFES with key lessons learned related to planning, training, and data collection, which could inform and improve future efforts:

Planning

• Allocating sufficient time before the start of the monitoring period to develop, test, and finetune the context-specific gender-sensitive indicators and their data collections strategies in consultation with local implementing partners, women’s organizations working on local conflict prevention and peacebuilding efforts, and monitors who will be collecting the data;
• Integrating indicators and data collection methodologies into reporting tools already being used by local monitors rather than creating entirely new tools, forms, or processes;
• Crafting exact definitions of each indicator to allow for consistent data collection and ensuring that monitors understand all the definitions; and
• Issuing identification to local monitors to use when seeking eyewitness interviews or speaking with local authorities, only in contexts where identification will make monitors safer rather than targets of violence.

Training

• Ensuring the security training and ongoing guidance provided to the monitors is sufficient to alleviate the safety concerns of local monitors;
• Allowing for at least two days for in-person training that draws on context-specific scenarios and examples for monitors to practice, if resources allow (instead of a daylong training session provided in a hybrid in-person and virtual environment due to COVID-19 travel restrictions, as was done in this pilot); and
• Developing an accompanying toolkit for monitors for reference after the training sessions have been completed.

Data Collection

• Establishing a well-publicized toll-free community phone line to collect information on incidents of violence against women, intimidation, threats, or attacks on women’s organizations and women in public roles, or instances of sexist, misogynistic, or homophobic hate speech and propaganda. Such a tool can allow for safe reporting of incidences of violence and increase the data collected, when resources allow. If the phone line is staffed with live operators, these operators should be trained on how to speak with survivors of violence and should be able to share resources with survivors, as requested. If the phone line is not staffed, the answering machine message could provide a phone number or website of a local organization to which survivors could go if needed;
• Providing regular feedback to local implementing partners that could then be immediately implemented and reflected in the following week’s dataset;
• Not relying on local police stations for arrest data and rates of reported gender-based violence. Instead, implementors should establish relationships with women’s shelters, women’s organizations, and/or humanitarian actors providing front-line services to survivors of sexual or domestic violence or working on other community-level gender equality initiatives. Implementors should ensure that the data collection strategy for arrests and gender-based violence in no way further jeopardizes the safety of the survivors or the organizations helping them;
• Seeking eyewitness accounts that can provide further details on who was involved in a particular incident, wherever possible, as media monitoring alone often does not provide sufficient sex-disaggregated data; and
• Developing, in consultation with local partners, a compendium of commonly used derogatory terms used in local contexts (e.g., a hate speech lexicon) as a way of providing more guidance on what to look for when monitoring for sexist, misogynistic, or homophobic hate speech and propaganda.

There would be significant benefit in retesting these five indicators across Nigerian states and conducting similar pilots testing these five indicators, or others included as part of this framework, in other countries and regions to expand on the lessons learned from this pilot.