Chapter 5: Data Capture, Transmission and Management
Kenya
Kenya conducted several field tests before the pilot census to test the content and flow of the questionnaires as well as consistency checks and skip patterns. The pilot census was conducted exactly a year to the census date, thereafter, two pre-tests were undertaken.
-
The first pretest aimed to; test the census data collection, assess the suitability of the uploaded Enumeration Area (EA) maps, test the pre-enumeration household listing, and test the flow of the questionnaire with the newly added questions that were not tested during pilot census.
-
The second pre-test aimed to test the recommendations made from the first pre-test regarding data collection application, uploaded EA maps, pre-enumeration household listing system as well as test the supervisor module that had not been included in the pilot census.
Pilot census and testing activities were carried out in purposively selected areas to include; densely populated rural and urban areas, urban slums;,high class residential areas in urban setting, border counties with internet connectivity issues, outdoor sleepers, travelers, hotels and lodges, rough terrain (hilly areas), areas with arid and semi-arid land, plantation regions, sparsely populated rural areas, and nomads/pastoralists community.
Within the resources budgeted for the pilot census, two extra pretests had to be undertaken after the pilot census because it was not comprehensive enough to include testing of the supervisor module, additionally, the questionnaire was not yet final given that the NSO was still accepting to take on new questions.
-
The pilot census revealed several issues that informed the main census such as during the training, it was noted that time allocated for training on disability, labour, definition of concepts and map reading was not adequate, and this helped in revision of the training programme for the main census, among others.
-
This finding shows that the lessons from the pretests done before the pilot were not taken into consideration, for instance, the need to adjust the training timetable should have been seen and handled earlier during the pretests so that the pilot only identifies major issues. For example the failure rate of the tablets that was used to decide on the number of spare tablets to allocate per supervisory area and the emergence of fake job advertisements that informed the change in recruitment strategy for the main census.
Ethiopia
Ethiopia conducted three pilot censuses in November 2016, March 2017 and June 2018 and the main census enumeration did not take place in the 2020 round. The first and second pilots were conducted to test both enumeration methods - paper and tablet, while the third one was purely digital with backup paper questionnaires for emergencies such as tablet failure and power shortage. After the completion, the first pilot lessons were documented to inform subsequent census activities as outlined below.
-
Recruitment guidelines were not followed, leading to appointment of fieldworkers who were physically unfit, some were not familiar with digital technology, while others could not speak the local language, which negatively affected the pilot census operations. This, coupled with lack of or poor communication about field worker entitlements, led to discontent and disruption of the training.
-
The pilot census was done before a thorough review and finalisation of the data collection tools was completed, pretests should have been done earlier to address issues found such as inconsistencies in the translated questionnaires, inappropriate and non-functional skip patterns/rules in the data capture application.
-
The pilot census led to an overhaul in some of the IT equipment specifications indicating that there may have been limited prior research or non-adherence to recommendations. This lesson is drawn from the failure of power banks to charge the mobile devices due to low capacity, low battery time for the tablets, poor network coverage that necessitated modification of the specifications for the equipment hence a fresh budget.
-
Lack of preparedness on the part of the trainers and field operations office to ensure availability of appropriate training centres with LCD projectors to present video lessons affected the training in some centres and this can compromise on the data quality. This therefore necessitated the inclusion of specifications for an ideal training centre for a digital census as part of the training plan.
Namibia
Namibia conducted the Census Pilot Enumeration in September to October 2021 that provided important information on the adequacy of the field organisation, supervisory areas, training program, extent of respondent burden, the data processing plan and other important aspects of the census that would better inform the actual census. Prior to the census pilot implementation, tests were done, and the data generated was used to review the batch editing program and test the monitoring dashboard.
The sampling design was purposive and included; special populations such as the Nomadic people in Opuwo that are based in Kunene region and Tsumkwe, holiday homes in Walvis/Swakop (coastal areas of Namibia), and EAs in: commercial farms, resettlement farms, gated communities, communal settlements and urban (high class society, middle Class, low/informal).