التاريخ: 2008-05-19
Mission Report to the Census Monitoring and
Observation Committee (MOC) of the Sudan
Uzair Suhaimi
BPS- Statistics Indonesia
Mission Period: 23-30 April 2008
This report reflects briefly my view on what and how the 2008 Sudan Census has gone so far based on my observation during my one-week visit to the country. I came to Sudan at 23th April 2008 or at the 2nd day of field enumeration of the census taking, and left for Jakarta at 30th April 2008 or in midway the field enumeration period. By now the phase of field enumeration would have been completed (unless there was an extraordinary problem had occurred in the last week, something of which that I have not been informed). The census is now entering the phase of data processing, a crucial phase in a census undertaking, followed by the phase of launching the census results, a phase that seems to me also crucial if not most crucial for the 2008 population census as it is placed within its political context. I believe these two phases have been now of the concern of MOC, CBS and other relevant stakeholders. This report also highlights my views on some issues related to data processing and the time schedule for the whole activities of the census taking.
What and how the census has gone so far?
As far I observed, the census in general has gone so far quite smoothly. One of the reasons would be that the preparatory phase of the census has been well-managed as indicated by the promulgation of the census decree by the presidency at 11th February 2008, the involvement of a UN body (e.g., UNFPA) from very beginning, the establishment of the technical working group (TWK) and the census monitoring and observation (MOC) unit. The Presidential decree has obviously anchored the census with a strong political and legal basis. Meanwhile, the involvement of UNFPA and the establishment of TWK and MOC have shown a well-managed situation during the preparatory phase in terms of institutional capacity and resources mobilization.
Such a well-managed situation is more remarkable if the census is viewed within the context of political complexity that has been faced recently by the Sudanese. There was one-week delay of the census taking from that has been previously scheduled due to the surprising announcement of the Government of South Sudan to withdraw from participating in the census shows clearly the political complexity. On the other hand, the ability to resolve the problem within two days shows clearly the strength of institutional capacity of the central government. However, similar unprecedented political movements may occur in the coming weeks that may be harmful for further census processes. I have no clear idea what to do to anticipate such potential problems but it seems to me that maintaining what so-called command and control center and strengthening its capacity would be advisable. Recommendation: maintain the existing command and control center for the census operation (C4O) and strengthen its capacity to anticipate unwanted problems may occur in the coming weeks.
With regards to field enumeration process I would like to comment that the process in the first week had gone fairly well. I understand that some minor security accidents had occurred in that period but I understand also that they had resolved quite smoothly. In this case I shared similar view with Ms. Nancy Stiegler of the University of the Western Cape who told me in a short conversation the census has gone well so far: “… better that I had expected”, she said. I understand that many factors had contributed to the success of the filed enumeration. Nonetheless, in my view the share of MOC in a number of ways was notably remarkable.
My judgment about the success has been confirmed by my observation on the actual enumeration processes of the census in three counties (mahaliyyah) of Khartoum, Omdurman and Bahri at 28th April 2008. During the observation I had was impressed by high dedication of the census field workers, by the interviewers’ understanding of conceptual components of the questionnaires, their capability in applying appropriate technique in interview, and their skill in filling in the questionnaire appropriately, and in coordinating capacity of the census officers in the field work.
As I observed during the visit, the speed of field enumeration had been consistent within the scheduled duration for enumeration. From a brief conversation with an interviewer in Omdurman I was informed that until that day (e.g., the 7th day of the enumeration period) she had completed enumeration for about 50 percent of the total households (=340 households) in her enumeration area (EA). Another interviewer in Bahri reported to me that her EA covered 168 households and so far 155 of them have been enumerated.
Data Processing: How Prepared?
As scheduled, the process of field enumeration should be completed by now and the census has now entered the phase of data processing. I assume transportation of the questionnaires for storage in warehouses is now in progress and even has been ready for scanning and for further data processing. The question is, how has this census phase been prepared so far.
At technical level I have reviewed some documents indicating certain level of the preparedness of data processing. I have read, for example, a document describing the planning of work flow in the CBS computer center. The plan has identified quite comprehensively and in detailed the types of activities that would be the main responsibility of the center including, including:
• receiving questionnaires (Qs),
• the movements of Qs between warehouses and scanning operation (here, warehouse and scanning sites were assumed to be located in different building but as I recalled MOC had recommended scanning and further data processing are to be done in the same roof with warehouse),
• scanning of the Qs in the prescribed, validation and key correction of errors,
• editing data files, imputation and errors reports and manual of editing if required (surely is),
• storage and merging of clean data files, and
• tabulation and analysis of census data.
Another document I have reviewed is the validation rules for data capture. I was impressed by the comprehensiveness of its contents and by the clarity of validation rules that have been documented appropriately. I believe that all validation rules have been made consistent with the instruction rules in the questionnaires and in the manual for the interviwers as well. (I have no access to the manual.) I believe that some sort of office manual editing is going to be made before the questionnaires are ready for scanning (say, pre-scanning editing), to minimize the rejection rate of data capture by scanning machines. I also believe that a document showing the rules of data cleaning will be prepared.
There is a minor suggestion I would like to make concerning technical aspects of data processing about the importance key-in data entry on a sample basis (say 10%). The objective of key-in data entry here is to ensure that there is no mistake made by the interviewers in marking appropriate boxes as he or she intended to. It is possible, for example, that an interviewers coded correctly a respondent’s age (Q4), say 24, but marked unintentionally in the boxes 2 and 5. It also not impossible, as another example, an interviewer coded correctly county of usual residence of a respondent (Q9), say 17 and but marked codes 1 and 6. I shared similar view on this issue with Dr. Pali
These are all my observations about some technical aspects of data processing. Let me turn now to another aspect; that is, managerial aspects of data processing.
Regarding data processing at managerial level I have some observation. The question I had in mind is that: Are we ready for entering data processing stage by now? To my understanding, the answer to the question is yes only if all necessary preparation has settled by now. This means, among others, decisions on logistics issues (warehouses, data processing centers, etc.) have been firmly made, well-trained staff for data capture by scanning has been available, sufficient number of scanner machines and computer units needed for further data processing have been already installed, and well-tested software for data capture and further process has been ready for use. I have no opportunity to access necessary information to answer the question firmly. Nonetheless, from a few MOC reports I understand there are many issues related to data processing have not yet decided event though they need to decide soon as scheduled.
I have not been informed but believe that some activities related to data processing has been done because they are independent from filed enumeration and are to be completed before April 15th as scheduled. These activities are recruitment of data processing staff, specifications for editing/programming/tabulation, training of trainers and training of coding and editing staff. No information I can access about the status of training of scanning staff (scheduled 15-30 April 2008, parallel with scheduled field enumeration). However, given the schedule of scanning of short questionnaire (scheduled 14th May-14th July 2008) the training should have been started by now.
Information about the schedule as mentioned above come from the annex of a report prepared by Mr. Pali to MOC. I am not sure whether it is official or not but it is all I can access about the census schedule in detailed. To my mind, a detailed schedule that covers all major activities of a census taking, including clear statement about the status of each activity, is very an important monitoring tool.
The need for updating the census schedule
My quick evaluation on the existing schedule (that I can access) suggests that it needs to be updated. The delay of the commencing date of field enumeration by one week is a good reason for the update because it obviously influences the schedule of other sequence activities like scanning questionnaire. This in turns influences the schedule for tabulation and finally the schedule for the launching.
Another reason for the revision is that the existing schedule in a number of cases seems to less realistic or even less logical. As an illustration, in the existing schedule for scanning the long form should have completed at 31st July 2008 and the schedule for tabulation should have been started at 28th July 2008. How can we tabulate data that are not yet be scanned completely? (We obviously need both short and long forms to be scanned completely before tabulation for the whole population can be made, even for basic tabulations, given the fact that each of short and long forms was not used for complete enumeration.) In addition, between scanning and tabulation there are a number of activities to be complete for quality assurance that obviously requires sufficient time to allocate. Recommendation: Review the existing schedule for the whole census activities after field enumeration period based on the most recent status of preparation activities of data processing and by taking into account the logical sequences of activities, reasonable assessment of duration of time needed to complete each major activities. Use the updated schedule as a powerful tool for monitoring and on-going planning.
Acknowledgement
The opportunity to visit Sudan and to monitor the 5th Population and Housing Census of Sudan is a rare and even historical to me. I deeply appreciate and thank to Dr. Rusman Heriawan (the DG of BPS-Statistics Indonesia) and Dr. Abdelbagi Gailani (Chairman of MOC) for making that opportunity possible for me. I also appreciate and thanks to to J. Bill Musoke of UNFPA and to the Indonesian Embassy in Kahrtoum for all facilitation and supports provided generously to me during my stay in Khartoum. My thanks also go to all staff in MOC offices, notably Mr. Ismail (MOC Rapporteur) and Mr. Muhammd and Mr. Mabiker, for their hospitality and assistance given to me.
|
