en-1749018255-Tanzania Data Quality Assurance Framework (TDQAF) for Official Statistics_June2024.pdf

Type: Document | Status: ready

35 Requirement Element to be assured Description 8.8 Risk analyses addressing the quality of important statistical products and processes are performed. Risk and quality management are closely coordinated (e.g. by institutional arrangements and regular meetings if responsibilities for these activities are placed differently) Quality assessment report should be presented and discussed at length to wider range of stakeholders Risks linked to core recommendations and principles of the NQAF (e.g. for lack of independence and confidentiality breaches) are analysed and measures taken if needed to improve compliance Risk analyses addressing the quality of different stages of the statistical production are conducted according to the GSBPM. Risk analyses addressing the quality of important statistical products such as population statistics and censuses, national accounts and Consumer Price Index (CPI) are performed (e.g. risk of poor accuracy expressed by errors, poor timeliness and lack of comparability)

36 3.2.6 Principle 9: Assuring adequacy of resources The financial, human, and technological resources available to NSOs and Statistical Units should be adequate both in magnitude and quality, and sufficient to meet the needs regarding the development, production and dissemination of statistics. This principle is explained by three requirements elaborated in Table 3.9.

37 Table 3.9 Principle 9: Assuring adequacy of resources Requirement Element to be assured Description 9.1 Financial, human and technological resources are sufficient to implement the statistical work and development program. A resource mobilization strategy such as a NSDS is in place. NSOs should have resource management and mobilization policies or strategies that guide on mobilizing and use of such resources. The evidence documentation includes Human Resource Planning (HRP) and Personal Emoluments (PE) The annual work plan is feasible given the available resources. Costs (staff costs and other costs) of each stage of the production processed are measured. 9.2 Planning and management principles are aimed at the optimal use of available resources. Information technology is employed to increase efficiency. Cost and benefit analysis should be done to optimize the use of resources for achieving greater results Standardization, integration and automatization of statistical production and dissemination are pursued to increase efficiency of operations and to save costs. 9.3 The statistical agencies’ use of resources is reviewed. Indicators on the use of human and financial resources are monitored centrally and regularly reported to management. There should be regular reviews of the budget and implementation of work plans, for example mid- term reviews The use of human resources is evaluated annually based on established guidelines and procedures. The evaluation covers allocation, performance and training needs of staff. Staff opinion/satisfaction surveys are conducted regularly.

38 3.3 Level C: Managing Statistical Processes International standards, guidelines and good practices are fully observed in the statistical processes used by the NSOs and Statistical Units to develop, produce and disseminate official statistics, while constantly striving for innovation. The credibility of the statistics is enhanced by a reputation for good management and efficiency. The relevant principles to be assured under this level focused on balancing between quality of statistics, cost and respondent burden. Managing statistical process has four principles: a) Assuring methodological soundness
b) Assuring cost-effectiveness
c) Assuring appropriate statistical procedures
d) Managing the respondent burden
3.3.1 Principle 10: Assuring methodological soundness In developing and compiling statistics, NSOs and Statistical Units should use sound statistical methodologies based on internationally agreed standards, guidelines or best practices and consistent with established scientific principles. Effective and efficient statistical procedures should be implemented throughout the statistical production chain. The following five quality requirements are recommended for assuring methodological soundness, see Table 3.10.

39 Table 3.10 Principle 10: Assuring methodological soundness Requirement Element to be assured Description 10.1 The methodologies applied by the statistical agencies are consistent with international standards, guidelines and good practices and are regularly reviewed and revised as needed. Organizational structures for the development and application of sound statistical methods are commensurate to the needs. NSOs should have manuals which guides on the processes of statistics production and these guidelines need to be harmonized with international standards. For example, statistical classifications (ISIC Rev. 4, SNA 2008, COICOP etc.) There are review and reporting processes in place that allow the management of the statistical agency to be assured that sound methodological approaches have been adopted and applied throughout the production process. The methodologies of surveys and the use of administrative data and other sources of data are evaluated periodically.
Sampling design is based on sound methodology Proper follow-up procedures are planned and implemented in the case of non-response. Statistical editing procedures and imputation methods are based on sound methodology. When statistical modelling is used in the statistical production process (e.g. for seasonal adjustment), the validity of model assumptions is carefully considered and the impact on final estimates evaluated. Statistical agencies review the methods used by external partners for the compilation of data and the production of statistics. 10.2 The statistical agencies recruit qualified staff and have regular programs to enhance their methodological skills. Staff of the statistical agency are recruited based on their academic background, qualifications and experience. Recruitment of statistical staff should consider professionalism and there must be a program to train these staff regularly. Supporting evidence includes the Scheme of Services; Training Needs Assessment and Training Plan Appropriate qualifications requirements are specified for all posts. Training and development programs are in place to ensure the staff acquires and continuously update their methodological knowledge. Staff skills are regularly updated so that staff is able to utilize new data sources and tools, and able to easily change positions.

40 Requirement Element to be assured Description Attendance of staff at relevant training courses and/or to national or international conferences is encouraged. 10.3 Statistical agencies are to choose the data source with regard to accuracy and reliability, timeliness, costs, the burden on respondents and other necessary considerations. The use of alternative sources of data, including existing surveys and census, administrative data, “big data” or other sources of data is constantly evaluated. Data collection should be planned only when, and for items which are, necessary at a certain time. Choice of data source should consider all data quality dimensions including cost- effectiveness Quality has to be assessed when using administrative data or other data sources. Ideally, when using administrative data, it should be assured that:
• The population is consistent with the statistical output requirements;
• The classifications are appropriate;
• The underlying concepts are appropriate; • The records are complete and up to date; • The geographical coverage is complete and the measurement units are appropriately defined/identified. When using other data sources (such as big data), the specific methodological challenges such as the ones linked to the statistical population and the veracity and volatility of such data have to be considered. 10.4 The registers and the frames for surveys are frequently evaluated and adjusted. A systematic approach is in place for updating the survey frames to ensure accurate coverage of the target population. NSOs must maintain and update sampling frames of households and establishments For all surveys the appropriate statistical population frames are updated regularly.

Information gathered during the conduct of surveys is used to assess and improve the quality of the frame, especially its coverage and the quality of the contact variables and the auxiliary information (variables used in the sampling design). 10.5 The statistical agencies cooperate with the scientific community to improve methods and promote Collaboration with the scientific community is in place, for example through conferences, workshops, task forces, and training/courses, to discuss relevant methodological and There should be a formal collaboration agreement between NSOs, R&D and the scientific community

41 Requirement Element to be assured Description innovation in the development, production and dissemination of statistics

technological developments, e.g. in exploiting new data sources. There are agreements with academic institutions on cooperation and the exchange of qualified personnel. Staff collaborate on methodological issues with colleagues at international level. Regular participation and presentations at relevant national and international conferences is encouraged for exchange of knowledge and experiences. National and international conferences, seminars, workshops, or similar events with the participation of the scientific community are organized by the statistical agencies.

42 3.3.2 Principle 11: Assuring cost-effectiveness Principle 11 emphasizes the responsibility of NSOs and Statistical Units to ensure the effective and efficient utilization of resources. It is imperative for NSOs and Statistical Units to demonstrate the extent to which objectives have been met, the attainment of results at a reasonable cost, and the alignment of outcomes with the primary purposes of the statistics. This principle is managed by six quality requirements which are elaborated in Table 3.11.

43 Table 3.11 Principle 11: Assuring cost-effectiveness Requirement Element to be assured Description
11.1 The costs of producing all individual statistics are measured and analysed, and mechanisms are in place to assure cost- effectiveness of statistical activities or processes. There is a system for registering cost and time used for all statistical products, and it should be possible to estimate time used on the main processes. Cost per unit individual should determine for each statistical activity and should be calculated for alternative methods of data collection. Concept notes and proposal may serve as evidence The costs of producing the statistics are well documented at each stage of the production process, and regularly reviewed and analysed across statistical products to assess the effectiveness of their production Cost–benefit analyses are carried out to determine the appropriate trade- offs in terms of data quality. The cost-effectiveness of every statistical survey is assessed. The need for each survey variable to be collected is justified. There is an ongoing review process that considers whether a particular program is still operating in the most cost-effective way to meet its stated requirements. Data collection instruments are designed to minimize coding and editing cost and time. 11.2 Procedures exist to assess and justify demands for new statistics against their cost. Demands for new statistics are regularly registered and assessed by statistical experts with respect to the proposed methodology and associated costs, and discussed by management, based on inputs from users and in cooperation with other stakeholders. User Satisfaction Survey, user feedbacks and user-producer forums can be used as a means of justifying demands for new data. Assessments and cost and benefit analysis reports need to be prepared Before contemplating a new data collection, there are mechanisms to review whether already available data sources can be utilized with minimal impact on their purpose and quality. When introducing new statistics, a cost-benefit analysis is conducted. 11.3 Procedures exist to assess the continuing There are regular discussions by management on the usefulness of all statistics, based also on inputs from users such as the results of user satisfaction surveys. NSOs must have a prioritization policy or guideline for statistics production.

44 Requirement Element to be assured Description
need for all statistics, to see if any can be discontinued to free up resources. The usage of different statistical products including statistical databases is monitored and assessed to evaluate their relevance. Questionnaires for periodic surveys are reviewed regularly before undertaking of survey Users and stakeholders are informed and consulted about the possible discontinuation of statistical outputs. 11.4 Modern information and communication technologies are applied to improve the performance of statistical processes. An appropriate IT strategy exists and is regularly reviewed and updated to improve effectiveness and efficiency of the statistical processes. NSOs have to ensure the use of ICT tools in planning, designing and implementing statistical activities at all stages. Primary statistical data collection are conducted using CAPI. Most of statistics are disseminated through websites, dashboards and social media

The IT architecture and hardware infrastructure are regularly reviewed and updated, and possibilities for innovation and modernization are identified.

Routine clerical operations and statistical processes (e.g. data capture, coding, data editing, data validation, data exchange) are automated where possible and are regularly reviewed.

Centralized IT and methodological units exist and provide possibilities for pooling of resources and investments. 11.5 Proactive efforts are made to improve the statistical potential of administrative data and other data sources. Statistical agencies provide input in the legislative process to obtain and maintain access to administrative and other data sources for statistical purposes, if needed.

NSOs should have strategic actions to improve administrative data and other non- traditional data sources. TSMP II prioritizes improvement of administrative data as among NSS focus Appropriate arrangements (e.g. service-level agreements or national legislation) with owners or holders of administrative data and other data collections are made and updated as needed, specifying the access to and flow of data and metadata and other relevant aspects An assessment of possible administrative data sources is carried out prior to launching any new survey. Data-linking and integration methods are pro-actively pursued while ensuring data security and privacy. Quality reports for administrative and other data used for official statistics are established by the responsible statistical agency in cooperation with the data owners or holders. 11.6 The statistical agencies define, promote and implement integrated and standardized The statistical agencies have developed strategies to move to a more integrated and standardized statistical production system within their organization. NSOs should have a guideline which document all procedures to be followed during statistics production. Evidence is shown through Manual for Strengthening and The statistical agencies promote, share and implement standardized