Liberia Agriculture Census 2024 - Non Household Listing

The conduct of the 2024 Liberia Agriculture Census (LAC-2024) non-household component encountered several challenges and data limitations that must be recognized when using the microdata. These challenges reflect both the complexity of enumerating a diverse set of agricultural establishments outside of household structures and the logistical constraints inherent in implementing a nationwide census.

First, the construction of the non-household census frame posed difficulties. While the frame was based on administrative records from multiple agencies, including the Cooperative Development Agency (CDA), Ministry of Agriculture (MOA), National Bureau of Concessions (NBC), and Ministry of Internal Affairs (MIA), there were discrepancies between administrative lists and actual field conditions. Some holdings recorded in registers were found to be dormant, defunct, or had shifted into other forms of activity. Conversely, some holdings that were active in the field were missing from the official lists. Although a verification exercise was conducted to improve coverage, the potential for undercoverage or misclassification cannot be entirely ruled out.

Second, accessibility issues affected data collection. Many non-household holdings are difficult to access due to lack of physical addresses of holders and administrative bureaucracies. This sometimes led to delays in reaching holdings, shortened interview times, or reliance on proxies for information. In a few cases, enumerators encountered difficulties arranging interviews with key managers, particularly within concession companies or larger private farms, where authorization procedures were required. This affected the ability to obtain detailed responses on sensitive topics such as land size, land use, use of tools and equipment, and type of labor utilized. In fact, three of the agricultural concessions recognized by the NBC were not covered by the census despite several attempts by enumerators and census staff.

Third, while the census relied heavily on digital data collection tools (CAPI), connectivity constraints in rural areas occasionally delayed real-time data transfer and monitoring. Enumerators sometimes faced technical issues with tablets, particularly with battery life in areas lacking electricity. Although supervisors provided backup and troubleshooting support, these challenges highlight the need for stronger technological infrastructure to support fully digital operations.

Fourth, some respondents expressed reluctance to provide complete information, particularly on land size, land ownership arrangements, or labor practices. For example, questions on land size and labor hiring occasionally received partial or vague responses, reflecting both the sensitivity of the information and gaps in record-keeping among some institutions. Similarly, while the census intended to capture detailed data on equipment use and mechanization, many respondents could not provide exact numbers, leading to reliance on estimates.

Another challenge relates to the definition and thresholds applied for non-household holdings. For example, non-household livestock and poultry holdings were defined using specific size criteria (such as 25 heads of cattle or 500 chickens). Thus, smaller institutional farms were excluded despite contributing to agricultural production in significant ways at the local level. This limitation means that the census provides a picture of medium-to-large non-household holdings but may underrepresent smaller-scale institutional holdings.

Data were collected using Computer Assisted Personal Interview (CAPI) system. The CSpro application was used to develop the CAPI system. A total of 800 tablets was used for the household level interviews. In addition, some tablets were set aside for replacement of missing, damaged or malfunctioning tablets.
During the data collection, team supervisors routinely synchronized their enumerators tablets for transmission of collected data to the LISGIS central server. At the end of the data collection, all tablets used for collecting the household-level data were synchronized by the team supervisors in the field and later by data quality assurance officers at LISGIS central office. The synchronized data were downloaded in SPSS and Stata format, checked for completeness (by comparing mapping and enumeration data) and inconsistencies, and then prepared for editing, tabulation and analysis.
The census technical team developed a tabulation plan prior to the start of the survey, which helped to guide the data editing, tabulation and analysis processes.