Statistical Methods, Standards and Guidelines 102
Tabulation Plan Preparation The tabulation plan is developed following user group workshops and thus reflects the information needs of the end users. Also the improvements on the tabulation plan will be made following the lessons learned from the previous Agriculture Sample Census. It takes into consideration the tabulations from previous censuses and surveys to allow trend analysis and comparisons.
Sample Design The National Master Sample (NMS) was developed from the 2012 Population and Housing Census. It was developed by the National Bureau of Statistics to serve as a national framework for the conduct of household based surveys in the country.
The sample design is a stratified two-stage sample, where the rural part of Tanzania is stratified into districts. The first stage (Primary Stage) units are the villages in the case of Tanzania Mainland and rural enumeration areas (EAs) for Zanzibar. In the first stage, villages/EAs are selected in each rural part of the district and in total an optimum sample size of villages and EAs is selected for Tanzania Mainland and Zanzibar respectively. In the second stage, agricultural households are selected in each of the selected village/EA. The villages/EAs in the first stage are selected with probability proportional to the number of households (PPS) in the village/EA. The cumulative total method is used to achieve the PPS selection of villages/EAs. In the second stage (Secondary Stage), agricultural households are selected using systematic random sampling procedure. A list of agricultural households is compiled from each of the selected village/EA and a systematic random sample is drawn.
Questionnaire Design and Other Census Instruments The questionnaires are designed following user meetings to ensure that the questions asked are in line with users’ data needs. Several features are incorporated into the design of the questionnaires to increase the accuracy of the data:
(i) Where feasible all variables are extensively coded to reduce post enumeration coding error; (ii) The definitions for each section are printed on the opposite page so that the enumerator could easily refer to the instructions whilst interviewing the farmer;
Statistical Methods, Standards and Guidelines 103
(iii) The responses to all questions are placed in boxes printed on the questionnaire, with one box per character. This feature makes it possible to use scanning and Intelligent Character Recognition (ICR) technologies for data entry; (iv) Skip patterns are used to reduce unnecessary and incorrect coding of sections which do not apply to the respondent; and (v) Each section is clearly numbered, which facilitate the use of skip patterns and provide a reference for data type coding for the programming of CSPro, SPSS and the dissemination applications.
Three other instruments used: (i) Village Listing Forms are used for listing of households in the village and from this list a systematic sample of agricultural households is selected; (ii) A Training Manual which is used by the trainers for the cascade/pyramid training of supervisors and enumerators; and (iii) Enumerator Instruction Manual which is used as reference material for enumerators.
Field Pre-testing of the Census Instruments The Questionnaire should be pre-tested in areas where different characteristics of farming and livestock keepers are found. This is conducted in order to test the wording, flow and relevance of the questions and to finalize crop lists, questionnaire coding and manuals. In addition to this, after pre-testing, several data collection methodologies can be finalized, namely, livestock numbers in pastoralist communities, cut flower production, mixed cropping, use of percentages in the questionnaire and finalizing skip patterns and documenting consistency checks.
Training of Trainers, Supervisors and Enumerators During training, cascade/pyramid training techniques are employed to maintain statistical standards. The top level of training is provided to national and regional supervisors. The trainers are members of the Planning Group from the National Bureau of Statistics and the sector Ministries of Agriculture. In each region, two training sessions are conducted for the district supervisors and enumerators. In addition to training them in field level Census methodology and definitions, emphasis is placed on training the enumerators and supervisors in consistency checking. Tests are given to the supervisors and enumerators and the best 50 percent of the
Statistical Methods, Standards and Guidelines 104
trainees are selected for the enumeration of the smallholder questionnaire and the community level questionnaire. The household listing exercise is carried out by all trained enumerators.
Information, Education and Communication (IEC) Campaign Radios, televisions, newspapers, leaflets, t-shirts and caps are used to publicize the Agriculture Sample Census. This helps in sensitizing the public for the field level activities. The t-shirts and caps are given to the field staff and the village chairpersons. The village chairpersons help to locate the selected households.
Data Collection The data collection methods used during the census is by interview and no physical measurements, e.g., crop cutting and field area measurement is taken. Field work is monitored by a hierarchical system of supervisors at the top of which is the Mobile Response Team followed by the Regional Supervisors and District Supervisors. The Mobile Response Team consists of Principal Supervisors who provide overall direction to the field operations and responded to queries arising outside the scope of the training exercise. The mobile response team consists of the Manager of Agriculture Statistics Department, the Desk Officer for the Census and senior officers. Decisions made on definitions and procedures are then communicated back to all enumerators via the Regional and District Supervisors.
Field Supervision and Consistency Checks
Enumerators are trained to probe the respondents until they are satisfied with the response given
before they record them in the questionnaire. The first check of the questionnaires is done by
enumerators in the field during enumeration. The second check is done by the district supervisors
followed by Regional and National Supervisors. Supervisory visits at all levels of supervision
focus on consistency checking of the questionnaires. Inconsistencies encountered are corrected,
and where necessary call-backs are made by the enumerator to obtain the correct information.
Further quality control checks are made through a major post enumeration checking exercise
where all questionnaires are checked for consistencies by supervisors in the district offices.
Data Processing and Analysis Data processing consists the following processes: (i) Data entry
Statistical Methods, Standards and Guidelines 105
(ii) Data structure formatting (iii) Batch validation (iv) Tabulation
Data Entry Scanning and ICR data capture technology for the small holder questionnaire are used. This not only increases the speed of data entry, it also increases the accuracy due to the reduction of keystroke errors. Interactive validation routines are incorporated into the ICR software to trap errors during the verification process.
Prior to scanning, all questionnaires undergo a manual cleaning exercise. This involves checking that the questionnaires have a full set of pages, correct identification and good handwriting.
CSPro is used for data entry of all Large Scale Farm and Community based questionnaires due to the relatively small number of questionnaires. It is also used to enter small holder questionnaires that are rejected by the ICR extraction application.
Data Structure Formatting A program is developed in visual basic to automatically alter the structure of the output from the scanning or extraction process in order to harmonise it with the manually entered data. The programmes automatically checks and change the number of digits for each variable, the record type code, the number of questionnaires in the village, the consistency of the Village ID Code and saves the data of one village in a file named after the village code.
Batch Validation
A batch validation program is developed in CSPro in order to identify inconsistencies within a
questionnaire. This is in addition to the interactive validation during the ICR extraction process.
The procedure varies from simple range checking within each variable to more complexes
checking between variables. After the long process of data cleaning, the tabulations are prepared
based on a pre-designed tabulation plan.
Statistical Methods, Standards and Guidelines 106
Tabulations
Statistical Package for Social Scientists (SPSS) is used to produce the Census tabulations and
Microsoft Excel is used to organize the tables and compute additional indicators. Excel is also
used to produce charts while ArcView and Freehand are used for the maps.
Analysis and Report Preparation The analysis in the reports focuses on regional comparisons, time series and national production estimates. Microsoft Excel is used to produce charts; ArcView and Freehand are used for maps, whereas Microsoft Word is used to compile the report.
Data Quality A great deal of emphasis is placed on data quality throughout the whole exercise from planning, questionnaire design, training, supervision, data entry, validation and cleaning/editing. As a result of this, it is believed that the census is highly accurate and representative of what is experienced at field level during the Census year. With very few exceptions, the variables in the questionnaire are within the norms for Tanzania and they follow expected time series trends when compared to historical data.
5.10 LARGE SCALE FARMS SURVEY The survey covers all Large Scale Farms in Tanzania Mainland. It is based on a complete enumeration. The survey is intended to generate information on agriculture that would serve the needs of a wide range of people and institutions down to regional level. These include policy makers at regional and national levels, rural development agencies, funding institutions, researchers, NGOs and farmers’ organizations.
5.10.1 Objectives and uses of the Large Scale farms statistics
The survey is normally conducted in order to:
(i)
Identify the structural changes in the size of farms, crop and livestock production,
livestock population etc;
(ii)
Provide trend analysis data on productivity, production and agricultural practices in
relation to policies and interventions promoted by the Ministry of Agriculture,
Livestock and Fisheries and other stakeholders;
Statistical Methods, Standards and Guidelines 107
(iii) Establish data that can be used to measure the impact of high level objectives of the Second Five Year Development Plan (2016/17 – 2020/21), other rural development programmes and projects; and (iv) Obtain data that will be used to address specific issues such as: food security, poverty, agro-processing, marketing, service delivery etc.
5.10.2 Methodology of data production The designed questionnaires that are used in the Agriculture Sample Census for large scale farms have been reviewed based on the census results and are posted to the Regional Statistical Managers who collect the data from all large scale farms. The reference period is the agriculture year. An agriculture year in Tanzania commences on 1st October and ends on 30th of September of the following year. A list of large scale farms is updated annually by the Regional Statistical Managers. The questionnaires are edited by the Regional Statistical Managers before they are posted to NBS headquarters for data entry and analysis.
Data entry The data is being entered manually using CSPro. Prior to data entry, data cleaning is done manually in order to maintain the quality of data. It includes questionnaire checking to see if it has a full set of pages, correct identifications and correctly filled-in information.
Tabulation Plan Tables are prepared based on a predesigned tabulation plan. Statistical Package for Social Scientists (SPSS), Census and Survey Program (CSPro) and STATA can be used to produce the tables while Microsoft Excel is used to organize tables.
Analysis
The analysis focuses on regional and trend comparisons, time series and national production
estimates.
5.11 AGRICULTURE ROUTINE DATA COLLECTION SYSTEM
Quarterly agricultural production and producer unit prices are normally collected by the Extension
Officers from the Villages.
Statistical Methods, Standards and Guidelines 108
5.11.1 Objectives and uses of the Agriculture Routine Data Agriculture Routine data is used as an input in the compilation of the national Gross Domestic Product.
5.11.2 Methodology of data production The data is collected at the village level by the extension officers who estimate production of various crops grown in the respective villages. They then send the data to the Districts who compile them to give District estimates of production and produce unit price of each crop. The Regional Statistical Managers collect the returns and send the data to the NBS head office.
Data Entry The Regional Statistical Managers enter the data in the e-excel spread sheet and send them to the NBS head office.
Data Cleaning The Regional Statistical Managers check the data for consistency before entering them in the spread sheet.
Analysis The Department of Agriculture does not produce any report on the data collected but compiles them to give Regional Production and unit price of each crop and submit to the National Accounts Department for the calculation of GDP.
5.12 ANNUAL AGRICULTURE SAMPLE SURVEY The Annual Agriculture Sample Survey (AASS) is conducted on annual basis. The main purpose of this survey is to provide more timely and accurate estimates of area and production of major crops and livestock. The quantity of the crops and numbers of animals are crucial information needed by many people involved in agriculture. Estimates derived from this survey provide basic information needed by farmers, agribusinesses and government policy makers to make decisions for both short and long term planning.