Data Collection Notes
LISTING, PRETEST, TRAINING, AND FIELDWORK
Listing
A listing operation was carried out in the selected clusters starting on March 27, 2010. For this purpose, training was conducted for 88 listers and cartographers from CSO who had been recruited from all regions to do the listing of households and delineation of EAs. A manual that described the listing procedure was prepared and translated into Dari, as a guideline. The training was conducted using classroom demonstrations and field practice, with support from ICF Macro and MoPH. The listing was performed by teams composed of one lister and one mapper. Five core team members were also assigned to perform quality checks. At the end of the listing operations, out of the 751 clusters selected throughout the eight geographic regions and three zones, 719 were listed (with 2 of the listed clusters not surveyed).
Pretest
Prior to the start of the fieldwork, the questionnaires were pretested in both local languages, Dari and Pashto, to make sure that the questions were clear and could be understood by the respondents. A training of trainers was conducted by ICF Macro staff at MoPH for 35 participants over a two-week period from February 1-14, 2010. These trainees were recruited to conduct the pilot survey and to serve as trainers for the main training. The pilot survey was conducted in the two local languages from February 16-20, 2010, in three selected sites. The areas selected for the pretest were Estalif, Karte Parwan, and Karte Se from Kabul province which constituted rural, urban, and semi-urban households. Based on the findings of the pretest, the Household, Woman’s, and Verbal Autopsy Questionnaires were further refined in the two local languages.
Training
The training of interviewers, editors, supervisors, quality control staff, and reserves was conducted from March 23, 2010, to April 17, 2010. The Dari and Pashto questionnaires were used during the training, and both versions were simultaneously checked against the English questionnaires to ensure accurate translation. In addition to classroom training, trainees did several days of field practice to gain more experience on interviewing in the two local languages and in fieldwork logistics. In order to maintain uniform survey procedures, five manuals adapted from the core DHS manuals relating to different aspects of the survey were prepared in English and translated into Dari. The Interviewer’s Manual discussed the objectives of the AMS, interviewing techniques, field procedures, general procedures for completing the questionnaires, and included a detailed discussion of the Household and Woman’s Questionnaires. The Verbal Autopsy Manual discussed the process of conducting the verbal autopsy interviews and completing the verbal autopsy forms. The Supervisor’s and Editor’s Manual contained instructions on organizing and supervising fieldwork, maintaining and monitoring field control sheets, and general rules for editing completed questionnaires in the field. Trainers were given the Training Guidelines for DHS Surveys Manual, which described the administrative and logistical aspects of training and data quality checks using the field check tables. The Household Listing Manual described the mapping and household listing procedures used in DHS surveys. A total of 174 trainees were trained in three classrooms. The trainees were selected from 367 initial applicants through interviews and written tests in Kabul, Hirat, Mazar-e-Sharif, Kunduz, and Jalalabad between March 10, 2010, and March 14, 2010. An additional 60 persons were nominated from the MoPH, provincial and other government officials. In each class, the training was conducted by one senior staff member of the MoPH who had been trained during the training of trainers. In addition, six other participants trained during the pretest helped the staff of MoPH conduct the training. Two guest speakers from MoPH were invited to discuss family planning and reproductive and child health in Afghanistan. The Safety and Security office of USAID provided a security briefing for all trainees and a specialized briefing for the supervisors and editors. Training of field staff was rigorous. The 3-week fulltime training consisted of instruction in general interviewing techniques and field procedures for the survey, a detailed review of the questionnaires, mock interviews between participants using the Dari and Pasto questionnaires, and practice interviews in the field. After the training on how to complete the Household, Woman’s, and Verbal Autopsy Questionnaires was completed, all trainees were given written and oral tests to gauge their understanding of the AMS questionnaires and interviewing techniques. On the basis of the exam scores and overall performance in the classroom and during field practice, 140 trainees were selected to participate in the main fieldwork. From the group, 4 of the best trainees were selected as regional coordinators to monitor data quality, 25 of the best male trainees were selected as supervisors, and 19 of the best female interviewers were identified as field editors. The remaining 92 trainees were selected to be interviewers. The trainees not selected to participate in the fieldwork were kept as reserves in case of attrition. In order to minimize non-response, interviewers were clearly instructed to respect the cultural sensitivities of the country, with male interviewers conducting interviews with male respondents and female interviewers completing interviews with female respondents. The majority of respondents (97 percent) to the Household Questionnaire were male. Only women were administered the Woman’s Questionnaire. One or more key informants who were present during the death of a person were respondents to the Verbal Autopsy Questionnaires and were either male or female or both. In cases where respondents to the verbal autopsy were male and female, both the male and female interviewers worked jointly in administering the Verbal Autopsy Questionnaires and recording the responses. In line with the local culture, a female field staff was always accompanied by a maharam, a close male relative, who was either a member of the team or just a companion to a female member of the team.
After completing the interviewers’ training, supervisors and editors were trained for three additional days on how to supervise the fieldwork and edit questionnaires in the field in order to maintain data quality. The main duty of the field editor was to examine the completed questionnaires in the field and ensure that they were correctly filled out. An additional duty was to observe ongoing interviews in the field and verify the accuracy of the method of asking questions, recording answers, following skip instructions, and identifying eligible respondents. Supervisors were team leaders who ensured that the teams were in the correct cluster, who contacted the head of the village to inform them of the purpose of the survey and obtain their cooperation, and who ensured the general safety and well being of team members. Supervisors also maintained regular contact with the head office in Kabul, reported on the progress of fieldwork and alerted the head office of problems encountered in the field. The supervisors were also trained by a senior staff person from Roshan Clinic in Kabul on basic first aid, and they were provided certificates of completion. In addition, all fieldstaff were given a security briefing by the Safety and Security Liaison Office of USAID.
Apart from the main training, four additional training sessions were conducted for 28 males and 22 females in Kabul in the Pashto and Dari languages for fieldwork in the difficult and insecure provinces for which no trainee initially selected and trained had volunteered to go: 8 persons were trained from May 2-9, 2010 to work in Hirat and Badghis; 7 persons were trained from May 2-17 to work in Ghazni; 8 persons were trained from May 8-17 to work in Kandahar; and, 27 people were trained from August 20- 31 to work in the provinces of Uruzgan, Paktika and in Ghazni to cater for attrition of the fieldstaff who had been trained earlier.
Fieldwork
Data collection began on April 20, 2010 with 25 teams initially deployed to complete the 100 clusters in Kabul province. This strategy helped ensure close supervision of all teams during the early phase of the survey before they left for provinces outside of Kabul province. Most teams were made up of two female interviewers, two male interviewers, a male supervisor, and a female field editor. In total, 32 teams involving 102 males and 82 females worked over a period of 8 months to complete fieldwork in all the provinces. Fieldwork was completed on December 31, 2010. Fieldwork supervision was coordinated by IIHMR, ICF Macro, and the MoPH; data quality was monitored in all the provinces through continuous supervision by four regional coordinators. Regular communication was also maintained through cell phones. Additionally, close contact between IIHMR and the field teams was maintained through field visits by senior staff, members of the steering committee, and ICF Macro. Field visits by IIHMR and ICF Macro staff were not possible in the southern, south eastern and south western provinces of the country due to the fluid security situation.
Several important procedures were put into place to ensure the quality of the data collected. The four regional coordinators reported directly to the survey manager and were pre-assigned to the regions they would closely monitor and the teams in each region. Regional coordinators spent most of the eight months of fieldwork in the field moving among their assigned teams. They facilitated the movement of teams to their assigned EAs, ensured that teams spent at least three days in each cluster to minimize nonresponse, checked completed questionnaires before they were sent to the central office, and addressed field concerns faced by teams. Coordinators carried out spot checks to ensure that teams were in the selected clusters, that no clusters were replaced, and that no households were replaced during the interview process. They monitored the work of supervisors and field editors to ensure that household heads were properly informed about the purpose of the survey, that no household member was coerced into completing an interview, that male interviewers interviewed male respondents, that female interviewers interviewed female respondents, that questionnaires were administered to each eligible respondent, and that all interviews were conducted in private.
A considerable amount of time was spent during training to discuss interviewing techniques and how to establish privacy during the interview process. Interviewers were strictly trained to discontinue an interview if privacy could not be established and to resume only if privacy was possible. Interviewers were also trained on how to physically position themselves right next to the respondent to minimize intrusion by another person, male or female, during the interview process. Supervisors were trained on how to find the assigned households and resolve any problems with assigned households. They were also trained to carry out spot checks to make sure that interviewers visited the assigned households and followed up with eligible respondents. Field editors were instructed to observe at least one interview per day to ensure that survey protocol was adhered to, to rotate the observation of interviews among each team member and edit all questionnaires while the team was in the cluster so that inconsistencies could be resolved prior to the team moving to the next cluster. The cover page of every questionnaire was signed by both the field editor and the supervisor to attest that they had checked and verified the content of the questionnaires.