SAMPLE DESIGN OF AGRICULTURAL CENSUS (AC) 2024-25
1. INTRODUCTION
Pakistan Bureau of Statistics (PBS) successfully conducted the Integrated Agricultural Census (IAC) by merging three censuses, i.e., Agricultural Census, Livestock Census, and Agriculture Machinery Census. Previously, three separate censuses on Agriculture, Livestock, and Machinery were conducted at different points in time using old sample designs that were distinct and independent of each other. Pakistan has extensive experience in carrying out separate censuses for Agriculture (2010, sample-based), Livestock (2006, sample-based), and Agricultural Machinery (2004, a mixture of complete counting and sample-based detailed interviewing).
The Integrated Agricultural Census, being a merger of all three types of censuses, was conducted in one go under the same logistic operation. Based on discussions between the PBS technical team and an FAO International sampling expert, the sample design of the IAC was developed in 2021. The sample design increases efficiency in targeting the three sub-populations of interest—crop producers, livestock holders, and machinery owners—in a sample-based census. The write-up of the Agricultural Census covers Introduction, Geographical coverage/universe/target population, and elements of the sample design.
2. SCOPE, COVERAGE AND UNIVERSE
2.1 SCOPE
The scope of the census includes the collection of data on the type, size, tenure, and parcels of farms; land utilization; irrigation; area under crops and orchards; number of fruit and non-fruit trees; use of manures, fertilizers, and plant protection measures; use of agricultural machinery; livestock population; loans and their sources; casual and permanent hired agricultural labour; household members by age, sex, qualification, and their contribution towards agriculture; type of residential structure; economic activities of household members; and their main source of income.
2.2 COVERAGE
The Agricultural Census covered the entire country, i.e., Khyber Pakhtunkhwa (KPK), Punjab (including Islamabad District), Sindh, Balochistan, Azad Jammu & Kashmir (AJ&K), and Gilgit-Baltistan (GB), which were covered as separate entities.
2.3 UNIVERSE
The universe of the census consists of all urban and rural areas of Pakistan. Results are representative at the overall district level.
3. SAMPLING FRAME
To prepare the sampling frame of the Agricultural Census (AC), the Mouza Census 2020 was conducted. Each mouza also contains Enumeration Blocks (EBs)/blocks; a mouza may consist of one or more blocks. After the latest Population & Housing Census 2023, the updated information was used in the frame of the Agricultural Census. Therefore, the mapping of the Mouza Census Frame 2020 and P&HC 2023 has been carried out. The latest information of EBs against each mouza of the Mouza Census Frame has been updated. The Mouza Census Frame 2020 was mapped with the updated sampling frame of the 7th Population and Housing Census 2023 for the 7th Agricultural Census 2024. The sampling frame consists of 45,753 mouzas (103,972 rural Enumeration Blocks / PSUs) and 39,985 Urban Enumeration Blocks / PSUs (with agricultural information).
4. SAMPLE SIZE ESTIMATION
Sample size estimation was carried out at the district level using a 15% Margin of Error at a 95% confidence interval. The previous sample size of Agricultural and Livestock Censuses in each district was considered, along with information on cultivated land and livestock concentration based on the Mouza Census 2020. Initially, a sample size of 11,054 mouzas/blocks was estimated; however, it was later enhanced with the addition of 53 mouzas/blocks/Gawala colonies.
5. STRATIFICATION
Various stratification strategies were adopted for the 7th Agricultural Census according to ground realities in Pakistan. First, two strata were identified as already mentioned in frames:
a) National Certainty Holdings (NCH): These are national-level large agricultural holdings or agricultural households (HHs) having at least 100 acres of agricultural land OR 50 cows/buffaloes OR 200 goats/sheep OR 25 camels. These HHs are enumerated on a 100% basis throughout the country.
b) Other Holdings: Comprising agriculture households with holdings below the limits of NCH. These HHs are enumerated through sample selection of Mouzas/blocks and then households.
c) Mouza Certainty Holdings (MCH): Within selected Mouzas/blocks, agriculture HHs with at least 20 acres of agricultural land OR 20 cows/buffaloes OR 50 goats/sheep OR 20 camels. These HHs are enumerated on a 100% basis within the selected Mouza/block.
For selection in stratum (b), further stratification was made across rural and urban areas independently.
6.1 STRATIFICATION IN RURAL AREA
In the sampling frame, rural Mouzas were classified into settled and unsettled areas. A two-way classification was developed based on agricultural information in each Mouza from the Mouza Census Frame 2020:
i. Cultivated land in the Mouza
ii. Livestock concentration in the Mouza
First way stratification:
• Z1: Mouzas without cultivated land
• Z2: Mouzas with cultivated land
Second way stratification (based on livestock concentration):
Four strata are created:
• L1: Mouzas having at least one type of animal in category 4
• L2: Mouzas with category 3 but not category 4 animals
• L3: Mouzas with category 2 but not categories 3 or 4
• L4: Mouzas without livestock
Proportional allocation is used for selecting Mouzas from each stratum.
6.2 STRATIFICATION IN URBAN AREA
PHC 2023 provided PBS with data on agricultural activities in urban areas. This block-level information was used to identify urban blocks with agricultural activities and for block selection in rural areas. Various strata were constituted:
• Stratum 1: 1-9 livestock-rearing households
• Stratum 2: 10+ livestock-rearing households
• Stratum 3: 1-9 households with self-agriculture
• Stratum 4: 10+ households with self-agriculture
• Stratum 5: Blocks without livestock (excluded)
7. ALLOCATION PLAN
A sample of rural Mouzas was proportionally allocated according to the number of Mouzas falling in strata Z1 and Z2, and within Z2, according to strata L1, L2, L3, and L4. In urban areas, stratum 5 was allocated zero blocks due to the absence of agricultural HHs, while the other strata (St-1, St-2, St-3, and St-4) were allocated sample blocks using power allocation with a 0.5 alpha quotient.
8. SAMPLE DESIGN
8.1 First Stage Selection of Rural Mouzas and Urban Blocks:
• Z1 Mouzas: Simple Random Sampling (SRS)
• Z2 Mouzas: Pareto sampling with cultivated area as a measure of size (PPS design)
• Urban Blocks: Simple Random Sampling (SRS); single-stage design
8.2 Second Stage Selection of Blocks from Big Rural Mouzas:
From large Mouzas with two or more blocks (identified through PHC 2023), a single block was selected using Multivariate Probability Proportional to Size (MPPS), with livestock and self-agriculture households as measures of size.
8.3 Third Stage Selection of Households (HHs):
Complete household listings were prepared for selected Mouzas/blocks. Categories of HHs identified:
i. NCHs
ii. MCHs
iii. HHs with agricultural land (with/without livestock/machinery)
iv. Livestock-only HHs (with/without machinery)
v. Machinery-only HHs
vi. Nomads and gypsies
HHs in categories i, ii, v, and vi were fully enumerated. For categories iii and iv, HHs were selected using Systematic Random Sampling fo