A door-to-door survey was conducted across Lebanon to obtain responses to the questionnaire.

The questionnaire that was used in this survey contains direct questions on a range of issues; it was divided into three parts that are as follows:

1. The introduction explained who is conducting the research and that respondents will be kept anonymous.

2. Direct questions included the following types :
   • Top of mind questions.
   • Prompted and non-prompted questions.
   • Closed-ended questions.
   • Open-ended questions.
   • Five-rank-scale questions.
   • Evaluation and leveling questions.
     
3. The third part covered the profile of the respondents, which was used in the analysis and the data cross-tabs.
   

• Respondents belonged to the range of social classes, A (upper class), B (middle class), and C (lower class), reflecting the various social classes in Lebanon;
• All were aged 18 years and older;
• The sample was distributed proportionally over the various selected areas to include the range of confessional groups.
  

A sample size of 2,400 respondents was used, with respondents selected from all categories as described under “Respondent Profile.”

The estimated margin of error in this survey for the TOP-LINE results is ±2.7 percent.

Distribution of clusters per region:       The sample was distributed in the various areas by clusters according to the following:   Sex Size of the sample Male 1,051 Female 1,349     Area Size of the sample Numbers of clusters Beirut 248 26 Mount Lebanon 826 82 North 480 48 South 506 50 Bekaa 340 34 Total 2,400 240       Age Size of the sample 18-39 1,205 40-50 965 60++ 230     Income Size of the sample <$1000 1,812 $1001-$2000 503 >$2001 85

Map of the sample distribution by Muhafaza (Governorate):

Statistics Lebanon Ltd. has developed an extensive database of clusters, and households covering all of the Lebanese territory. The steps taken in the preparation of this database make it reliable for drawing representative samples for this survey research.

Our sampling system is based on a complete listing of all localities, small villages, medium villages, small towns, large towns, and cities, obtained from door-to-door visits covering each and every household in Lebanon. It is also based on very detailed maps of the urban and rural areas comprising every city and large town, small town and/or neighborhood.

Statistics Lebanon has divided the urban and rural areas into clusters based on updated street maps, adjusted by field observation to update the maps where relevant and applicable. A random sample of all the clusters in the urban and rural areas was taken to produce by field inspection usable cluster maps with details of buildings, apartments and other residential dwellings. Statistics Lebanon also has a listing of non-residential land use of these clusters.

In the smaller townships and villages Statistics Lebanon drew a large sample of localities stratified by apparent size representing all dwellings in these agglomerations and divided the sample area by field inspection into clusters for which detailed maps where produced. Statistics Lebanon now has listings of all dwellings, buildings, apartments and other non-residential uses indicated on the cluster map, and they are therefore usable for selecting households at random.

The structure of the sample that was used is proportional to population numbers as shown in the survey conducted by the Department of Statistics (Lebanese official statistics office) in 2004.

The database contains a list of 770,000 households, out of the 841,000 households available in Lebanon; they are all pin-pointed on aerial images and 1:5000 scale maps, and usable to draw cluster based samples.

Geographic Information System (GIS) technology was used at-large to assist in the sample draw.

The methodology that achieved these objectives is as follows:

Interviewers were supplied with a questionnaire that covered the survey objectives listed above. They interviewed respondents using the door-to-door method in the respondents' own homes.

No amendments were made on the questionnaire after implementing a pilot test over a sample of 20 respondents. Sixty-three trained interviewer form thirteen teams to carry out the fieldwork; the team leader, who supervised the fieldwork, directed each group of four or five. An area coordinator controlled the team leaders' fieldwork in their specified geographic zone. Teams used detailed maps and an advanced GIS that led them to the location of the interviewees.

The fieldwork started on July 3, 2008, and finished on July 12, 2008, during a period of 9 working days.

All field teams were equipped with radio transceivers through which an operations center inside Statistics Lebanon was in direct contact with the field teams and their team leaders.

During and after the fieldwork, a team of office clerks within Statistics Lebanon carried out the record keeping and data entry. Eight to 10 employees led by a senior computer programmer completed this work and controlled the data entry procedures.

The software for this project was specifically designed to prohibit the entry of wrong data and was equipped with a data entry control system. An identical image of the questionnaire was created using the Access database. All controls and logical rules were pre-defined to prevent data entry errors. The control rules defined in the database minimized data entry errors (for example, by only permit two values (1 or 2) for a variable and rejecting any other value from being entered for that variable). Moreover, Statistics Lebanon conducted re-entry of 10 percent of the data to ensure the quality of the work.

Upon the accomplishment of the data entry, a senior programmer cross-tabbed the results and a senior statistician analyzed the results statistically.

As mentioned previously, all of the data were entered into a programmed version of Microsoft Access, which enabled data entry controls prohibiting the entry of wrong data into the database.

For the statistical data analysis Statistics Lebanon used SPSS software.

The team did the following:

• processed the data
• cleaned the data
• coded the data
• produced a data file with SPSS software
  

For data entry purposes, Statistics Lebanon used Microsoft Access, which enabled the inclusion of data entry controls that prohibit the entry of wrong data.

For the data analysis, Statistics Lebanon used SPSS.

For the production of maps, start-up points and illustrated samples, Statistics Lebanon used the following:

• Arc View 3.2
• Global Mapper
  

Quality control was conducted throughout the survey process.

Four techniques were used:

• Re-interviewing:
Re-interviewing was done by re-interviewers. Ten to 20 percent of the questionnaires were back-checked in person. Random spot checking was also applied during the fieldwork activity.

• Telephone re-interviews:
Ten percent of the completed questionnaires in which a telephone number was mentioned were checked by the area supervisor, covering key questions and any part of the questionnaire where there is a concern about quality.

• Accompanying interviewers: Ten to 15 percent
Office staff did the selection of the interviews which were checked systematically.

• Review all the questionnaires:
Reviewing the questionnaires was done by the data entry clerks on a daily basis. Any questionnaire with mistakes was cancelled and the interviewers were asked to redo the collection. (The total number of cancelled questionnaires was three.)