Recent posts

October – December 2021

The quantitative fieldwork got underway again towards the end of October. Eshetu reassembled the field team in Arsi, consisting of 14 field assistants and 2 data entry clerks and a supervisor. Some of the field assistants had been involved in the pilot study and household census collection, while others were new to the team. Eshetu ran a training / refresher training for everyone, going over the data collection protocols, as well as some changes that had been made to the questionnaire and procedures since the pilot.

Carrying out the fieldwork during a global pandemic AND an emerging civil war in Ethiopia has been very challenging. Firstly, extra measures have had to be put in place to mitigate Covid risks both for staff and study participants. The team are equipped with masks and appropriate sanitation, however, even with these measures in place difficulties remain. Wearing a mask makes communication harder which is not ideal when interviewing people about sensitive topics, and the uncertainty about how the outbreak might progress in Ethiopia and what restrictions could be put in place, has made planning complex. Additionally the pandemic has prevented Mhairi or Janet from travelling to Ethiopia to assist with the fieldwork, and as a result Eshetu has handled the entire operation single-handedly on top of many other commitments.

Secondly, the Tigray conflict in Northern Ethiopia has worsened in the last couple of months, and in November 2021 the government declared a State of Emergency and imposed a 6 month curfew in the Oromia Region which surrounds Addis Ababa, and is where our fieldsite is located. As our field team are recruited locally in the Oromia region, they were already in the area and have been able to get to the villages to carry out fieldwork without violating the curfew. However, the situation remains uncertain, reminding us of the substantial difficulties involved in data collection on this scale and in this context.

Eshetu has done a remarkable job in keeping everything going despite these significant challenges.

September 2021

We have been analysing the household census data. One interesting outcome is that we have been able create a map which identifies the location of all ~2,000 households using the GPS data and QGIS software. Being able to visualise the data points on a map (in context with other landmarks, roads and topographical features) allows us to start to understand how individuals living in these households might interact. Some households are along the roads, or in small villages, while others are very isolated.

We have also been ironing out some of the challenges in data entry which we identified during the pilot study, and setting out protocol for data entry staff to follow in order to minimize errors. We have created a codebook for each of the nine kebeles in the study site from the household census data, which lists all the individuals in alphabetical order, with their individual ID numbers and other identifying features (age, education etc.). These codebooks will be used by the data enterers when keying in the social network data from the questionnaires. They will use these codebooks to match the individuals named by the respondents in the name generators questions to their individual ID.

August 2021

During this month we have made the difficult to decision to defer the quantitative fieldwork until October/November, once the rains have passed. Fieldwork was attempted, however, the conditions proved too challenging. Access was not an issue as the roads were passable, but once the enumerators were in the villages it was difficult to complete the questionnaires. Everyone was at home sheltering from the weather and there was nowhere to conduct the questionnaires with respondents in a private setting that was under cover.

July 2021

Following the census (household listing) which was carried out in May/June, the listing team in Ethiopia have completed data entry. The data entry was a lengthy process due to the number of individuals and the details captured for each household and individual, including GPS location. However, we now have a catalog of ~2,000 households and the ~5,500 inhabitants over 15 years of age, half male and half female.

All households and individuals have been assigned unique IDs, and this census catalog will be used throughout the next phase of fieldwork to match names generated through the questionnaire to build up the social network analysis.

The main fieldwork has been postponed by a week as it is seeding time in the area where the fieldwork will be conducted. During seeding time many people will be out in the fields, meaning that the enumerators would have difficulty in locating the individuals in the villages to interview.

Before fieldwork begins all the enumerators are taking part in a refresher training for a weekend. This will allow Eshetu to go over some areas of the questionnaire which have changed since the pilot study, as well as reminding the enumerators about important aspects of administering the questionnaire to keep the standards of the data collected high.

The delayed 2021 election was held this week, and the rains should not be too disruptive as the area in which the fieldwork is being done so we are optimistic that the fieldwork will be carried out as planned in the next month.

June 2021

Fieldwork for the main survey has started!

The first step has been to carry out a comprehensive census of the villages where the survey will be conducted. The team in Ethiopia, led by Eshetu, have spent the past three weeks out in the field completing the household listing for the three kebeles (villages) identified from our sampling exercise. There are just over 2,000 households within these three kebeles.

The household listing provides a complete list of all the people who usually live in the household who are over the age of 15 years. Additional data concerning their education, sex and relationship to the household head, as well as any specific role they may have in the community, is also collected. All individuals are assigned a household ID and a unique individual ID which will be used to match individuals named in the social network survey.

In addition to collecting data about the people who live in the household, the GPS location for each household was also recorded. This provides spatial data which can be used in the social network analysis, as proximity may impact on social networks.

The census data is being entered and once the team have a complete listing of all individuals then the next stage of the fieldwork can begin. We are finalising a few practical and budgetary details in the meanwhile, and anticipate that the main survey will start to be conducted in July.

Feels like we’re off to a great start!

April 2021

One of the key activities this month has been finalising the locations for the main fieldwork. Our main criteria is to carry out the survey in locations primarily populated by Arsi Oromo, as FGM/C is more prevalent among these populations than among other ethnic groups. Through local contacts several kebeles (districts) have been identified, and Eshetu has been talking with local government administrators to obtain population information for these kebeles at the zone and household level. Through this process we have identified an appropriate location, which consists of three neighboring zones which in total will give us the necessary sample size for our analysis. Although they are remote, we hope that travel time for the enumerators will be reduced as all households to be sampled are concentrated in a small geographic area. The necessary research permissions from local authorities have been obtained.

Other activities this month have included quite a lot of practical but time consuming matters such as sorting out bank transfers, obtaining permissions, and getting the refined questionnaires translated and ready for printing.

A recent covid surge in Ethiopia is a concern and this may affect the timing of the fieldwork.

March 2021

We have been reviewing the pilot data this month, which as been really interesting.

We have been able to come to a conclusion about the list experiment technique we will use in the main survey. This section of the questionnaire collects participants’ opinions about FGM/C for their daughters and daughters-in-law, directly and indirectly. We trialed a new method in the pilot study, which didn’t work as anticipated, therefore we will stick to the method used in previous studies. The pilot study was performed in an area close to the main highway where most people would have been exposed to anti-FGM/C campaigns, yet even in this area when asked directly a proportion of respondents said they would like their daughter/daughter-in-law to have FGC, and the figures indicate that their hidden support was higher. This is in line with previous studies (Gibson et al, 2018) , but is higher than anticipated in this area.

The results of the social network section of the questionnaire were very encouraging. We tested 8 different name generator questions in the pilot, and it was clear that the participants had understood the questions and answered them fully and freely. The network structures for advice, chatting, money lending and respect showed some different patterns. A range of people were named in response to each of the different questions (indicating that people were considering the specifics of each question and not just naming the same people repeatedly) and a sufficient number of people were named in response to each question which shows that the questions had been understood and administered correctly.

The rest of the pilot data also provided encouraging results about some new questions we had trialed, and based on all of the pilot data results we have been able to refine the questionnaire for the main fieldwork. This has involved cutting down the questionnaire length in order to achieve our sample size target within time and budget constraints.

Although the data entry process for the pilot turned out to be quite arduous, the process of setting up the data entry template and so on will have paved the way for a smoother data entry process for the main survey.

February 2021

The pilot data has been collected! This is a fantastic achievement by the team in Ethiopia.

As hoped the team were able to collect data from ~400 participants and the data is currently being entered and will be available for analysis in the next week or two.

23 staff were recruited and trained for the pilot data collection (2 listing staff, 16 enumerators (8 men and 8 women), 2 supervisors and 2 data editors). Our hope is that most of these same staff will be able to return and carry out the main study in the next month or two.

Before going out to the field the staff attended a training workshop during which they discussed the purpose of the questionnaire, went over the questionnaire in detail and were instructed on how to administer the different questions, as well as running a day of mock interviews and group discussions.

Once training was completed, the first step of the fieldwork was to list all households in the three zones identified for the pilot study, and record the relevant details of all household members aged 15 and over. This provided an entire listing of all householders to be interviewed, which could also be used to match individuals named in the social network name generator questions. Over the next few days the enumerators visited all the households to administer the questionnaires. They worked long days, travelling out by minibus to the zones early in the morning, and working in male/female pairs, they managed to administer up to 6 or 7 questionnaires each per day.

The data collection went smoothly. A few eventualities came up that we had not anticipated, which was useful to know for planning the full field study. For example, the enumerators found that many younger people were absent during the day (at school or work), and were only available at the end of the day. Several improvements for the questionnaire were also identified in some of the coding structures concerning relationships and occupations that can be amended in the final versions. The enumerators also found that the social network section of the questionnaire took a very long time to complete, so the number of name generator questions will be reduced for the final version.

All in all the pilot phase was a real success, and we look forward to seeing what the data will tell us.

January 2021

We have been working on refining the questionnaires to get them ready for translation into Afan Oromo. We want to trial several elements in the pilot study, and we’ve been discussing the best way to test them.

The social network section of the questionnaire will contain several ‘name generator’ questions, such as “If you needed to advice on health issues who would you ask?”. The participants will be asked several name generator questions (with no cap on the number of people they can name) and collectively their answers builds up a picture of the people in their social network. This sociocentric network information will be used to identify how positions, relationships and density of networks shape the social transmission of FGM/C.  A larger number of questions provide more nuanced network data, however, each question takes time to administer as several details of each person named must be recorded.

The pilot study provides an opportunity to test a variety of name generator questions, some of which we haven’t used before. We have spent time refining the precise wording of the name generator questions to trial in the pilot study, from our own experiences as well as from a review of the literature. This is to make sure they are as efficient as possible but still address our research questions. We will be interested to know if people name different people in response to the different questions we are testing, how many names are generated by each question, and whether the questions are phrased in a way that is understood and meaningful to the participants.

We also want to test a slightly different method for the list experiment / unmatched count section which we haven’t tested before. The list experiment is used to understand participants’ attitudes towards FGM/C for their daughters and daughters-in-law. Participants will be asked about several experiences that they might want/not want for their daughters and daughters-in-law, including FGM/C, and they are asked both indirectly and directly.

We will also be trialling some new questions in the general background section of the questionnaire concerning the participants’ opinions about the views of other people in their community towards FGM/C and IPV.

Our intention is to sample around 400 participants for the pilot study. Participants will be selected from three neighbouring ‘zones’ in a location that is convenient for the enumerators. A ‘zone’ is the smallest administrative unit in this area, equivalent to a village.