Living Standards Measurement Survey 2002 (Wave 1 Panel)

Pilot Test

The questionnaire was field tested in February 2002 in Korca, Shkodra and Durres by the core INSTAT team and the local teams, with the assistance of international consultants and World Bank staff. The pilot testing covered all sections except food consumption (section 10), anthropometrical measurements (section 15), and Other Income (section 14), which had not been completed at the time.

The pilot was done by three teams, one per region, which tested the 12 modules that were ready for testing, and did additional testing on the more complicated modules. The south team tested the agricultural module, the north team the non-farm enterprise module, and the team in Durres the labor module.
The questionnaire was then updated following the results of the testing. The agriculture module had fairly major changes in structure and questions, and it was tested again after the changes were implemented.

Training

The enumerator training began on March 18 and finished on April 8, one week in Tirana and two weeks near Durres. Each training week was organized with five days of classroom training and one day of hands-on fieldwork practice. Both the classroom training and the fieldwork practice served also as pilot tests for the questionnaires, that were revised as the training proceeded. Approximately 100 people took part in the training. The training also covered logistics, the use of the GPS devices, and the plans for the fieldwork, including replacements, the numbering of PSU's and households, instructions on revisiting households. The supervisors received an additional 45 minute specific training session each day, plus and extra day of training in Tirana on April 12, shortly before the beginning of the survey work in the field on April 15. The training for the supervisors was organized in three parts: One on the community questionnaire, its design and how to administer it. The second part covered the plans for the field work, the use of the household tracking sheets, how to properly track the reasons for replacements, how to structure the visits to the PSU's, the food diary questionnaire, and the second visit to the households. The third part of the supervisor training was another session on using the GPS's, to reinforce what had been learned on the last two days of the enumerator training. At the end of the supervisors training, all the equipment was handed out to each supervisor and driver, and questionnaires and food booklets for use during the first week on the field were given out. Training for the data entry operators (DEO’s) ran from April 12 to April 16th. DEO’s were trained on the use of the custom data entry program developed in CS-Pro for this survey.

Fieldwork

By early February, the selection of PSU’s for the LSMS had been completed except for those in Tirana, i.e. for 375 of the 450 total PSU’s. At this stage a preliminary division of the PSU’s into team responsibilities was made. Eventually, the field staff was organized into 20 teams, four of which were in Tirana. Each team had one supervisor and, generally, three enumerators. Some remote districts had one enumerator in each area of the district (in Korca, for instance, there were four enumerators, each in a different area), while other districts that have a particularly large number of PSU's also employed a greater number of enumerators (in Fier, for instance, there were 4 full-time enumerators to complete the 33 PSU's in that district).

The supervisors were all chosen from permanent INSTAT staff, and almost all had experience in supervising surveys on the field. Supervisors from within the regional INSTAT offices were identified for the 16 teams outside Tirana, and each regional office then identified enumerators from those who had successfully worked on the 2001 census. The four Tirana teams each had a supervisor, and they in turn were supervised by a Tirana head supervisor. Each team also included a data entry operator and a driver. The supervisors were given the list of PSU’s to be completed , and then were free to schedule the visits to each household as they wished, with the only general constraint that rural PSU’s should be completed first, then the urban ones. For each PSU the supervisors were given a large envelope with the 8 questionnaires, a tracking sheet listing the households to be interviewed, with one reserve household listed on the tracking form, a listing sheet for the supervisor of the 12 selected families including the 4 reserve households, and a map of the corresponding census EA.

Each listing sheet for a selected EA included the following information:

1. Name of the district
2. Name of the commune
3. Name of the city
4. Identification of the EA based on the 2001 Census Frame
5. Cartographic address (building, entrance, apartment, nr. of family inside the apartment)
6. Postal address (building, entrance, apartment)
7. Code identifying a selected HU or a reserve unit

Each tracking form included the following information:

1. PSU number
2. Name of the commune
3. Identification of EA based on 2001 Census Frame
4. Names of the heads of households for the selected housing units
5. Space for explanation by the enumerator of the reason why it was not possible to complete any of the household interviews

In order to control more closely the number of reserve households that were used, the enumerators received only the name of the first reserve household on the tracking sheets. If the enumerator needed more than one replacement household, he had to return to the supervisor to explain why more than one replacement was needed. The enumerator had to write a full reason on the tracking sheet for the replacement – and these reasons were verified and kept in the central office in Tirana in a continuously updated Excel sheet. This process helped ensure the reason for using a reserve unit was documented and provided more control on their use.

The only exception to this practice was for remote mountain villages, where the access was difficult, and it was decided to provide all four replacement names to the enumerator before the first visit. Out-migration from these villages is fast, and even in the year since the census many dwellings had become vacant. In four PSU’s (one of them urban) more than the four reserve families initially selected turned out to be needed. In these cases, extra housing units were randomly chosen from the remaining housing units in the EA.

Identification of the selected households in the field was made by using the head of household names for all the rural PSU’s, because this was the only reliable method to identify the selected housing units when there were no street addresses. However, the enumerators were trained to understand that the name of the head of household was used to identify the proper housing unit, but it was the housing unit that was selected. For instance, if the listed household head was Arbur Rexhepi, but at the enumeration visit another family was found in the housing unit, as long as the enumerator could determine that this was the HU in which the family of Arbur Rexhepi had lived at the time of the census, it was clear that this was the proper HU, and the enumeration continued with the family in the HU at the time of the visit.

For the urban PSU's, since there are explicit addresses for the selected dwellings, it was initially deemed not necessary to use the names of the household heads for identification. However, as confusion and inconsistencies started to arise in urban areas at the beginning of the fieldwork, household head names were used in urban areas as well, to control the urban PSU's more carefully, and also make the selected HU very clear. Having the list of household head names in urban areas also removed the necessity of letting the enumerators choose which of several families to interview in a selected housing unit, because the family had already been selected and
could be identified by the name of the head.

The fieldwork started on April 15, with close supervision from the core team at INSTAT Tirana head office. By April 27, 13 of the 16 rural teams were visited by the supervising Tirana team- Kukes and Gjirokaster, were visited in the week of April 29. Tropoje was not visited directly as the head office INSTAT staff did not feel safe to visit this region, and instead the Tropoje supervisor went with the completed questionnaires to Tirana. The local staff from Tropoje did not however report security problems during the data collection The enumeration was completed in two visits – the first, to complete Part 1 of the questionnaire, and to explain the food booklet that was left with the household, and the second visit at least two weeks later to collect the food booklet and to complete Part 2 of the questionnaire. The core team also paid additional visits to those enumerators whose questionnaire seemed to contain a higher number of errors - but overall the enumerators seemed to be willing and diligent.

In Tirana some families accepted to have the main questionnaire administered, but did not want to complete the 14 day food diary. In some cases these were families with only elderly people who did not feel they could properly keep the diary. In these cases the enumerators were instructed to try to obtain assistance from neighbors who came into the households and filled out the food diaries for the respondents. In several cases with older persons in urban areas, the enumerators returned every few days to help the respondents keep the diaries up. Supervisors visited and discussed personally with the households all cases where the household had refused to complete a questionnaire with the enumerator.

The rural enumeration was finished by the third week of June and the urban enumeration was completed by first week of July 2002.

The supervisors completed most of the community and price questionnaires for the rural areas in parallel with the household enumeration, but worked on the urban community questionnaires and price questionnaires mostly after the urban enumeration of households was completed – in July. The community data collection was finalized in September.

The food diaries were sent to Tirana as soon as they were completed in the field. Data coders began coding the food booklets. The coding of all the food booklets was completed by a small staff in Tirana to guarantee consistency.

The monitoring of the entire fieldwork process was ensured by the core team in Tirana, maintaining constant contact with the supervisors by telephone, tracking the progress and keeping apprised of any problems, and traveling to the field as necessary. Household replacements were also constantly monitored as the fieldwork continued, to make sure that replacements be kept at a minimum. In Tirana, where the refusal rate was higher than elsewhere, one of the field supervisors or a core team member revisited with the supervisor many of the households who had refused.