The Gallup Poll of Baghdad Sample Design
The strict, probability-based sample used by Gallup to conduct this survey projects with scientific accuracy to all adults (aged 18 and older) residing in urban areas within the governorate of Baghdad. All 1,178 interviews were conducted face-to-face, in the privacy of the respondent’s own home.
Interviewing was conducted during the period of Aug. 28 through Sept. 4, 2003. The cooperation rate exceeded 97%, that is, fewer than 3% of those we contacted refused to be interviewed. Average interview length was 70 minutes. For results based on this sample, one can say with 95% confidence that the maximum error attributable to sampling and other random effects is ±2.7%.
The governorate (
mohafatha) of Baghdad has a total population of 6,386,067 (2003 estimate) -- a quarter of the entire population of Iraq. It is comprised of administrative units known as
qadha.
Qadhas may contain both urban and rural districts; those
qadha closest to the city’s center, however, are comprised solely of urban districts. Because the objective was to survey the views of urban residents, rural districts lying within the outer boundaries of the Baghdad governorate were excluded from the survey’s sampling frame.
Each
qadha, in turn, consists of multiple neighborhoods, known as
nahiyas, which vary in population size. There are a total of 728 urban
nahiyas within the governorate of Baghdad. The adult population of each of these, (defined as those born in 1985 or earlier), was obtained by Gallup from a 2002 Central Statistical Office update. These data allowed Gallup to assign the appropriate population weight for each of the
nahiyas into which a sampling point (PSU) for our survey fell.
Each
nahiya is made up of blocks, or
majals.
Majals contain multiple, proximate housing units, often along the course of a given road or street (longer roads stretch across multiple
majals). Within our sampling frame, there are nearly 30,000 (29,779)
majals.
Selection of the sample was done at the
qadha level, with a total of 122 primary sampling units (PSUs) selected on a strict, probability-proportional-to-size basis. An average of 10 interviews, one per household, was conducted in each of these locations.
Once the number of PSUs to be allocated to a given
qadha was determined, these were then assigned -- again, on a probability-based basis -- to specific
nahiyas and
majals within that
qadha, with no more than one
majal selected within any given
nahiya.
For example, the
qadha of Rasafa was allocated, on the basis of its total population, 25 of the sample’s 122 PSUs. Rasafa consists of 210 urban
nahiyas, and 6,738
majals; our 25 PSUs for Rasafa were selected from among these 6,738
majals, not more than one per
nahiya, on a random probability basis.
Interviewers were given the all the relevant address details for each PSU. However, since the only available residential listings were based on a 1997 census, all residential listings had to be updated.
In addition, the available listings consisted solely of dwelling addresses, regardless of the number of independent families residing within a given housing unit. Therefore, we compiled separate listings identifying independent families, based on the criterion of their preparing or eating meals independently.
In those instances where more than ten independent families resided in a given, selected
majal, a random selection table was used to select which families would be interviewed.
Finally, within each selected household, the specific adult to be interviewed was selected by the Kish method. This research procedure, designed to ensure proper representation of all age groups and both genders in the sample, involves first recording the ages and sex of each of a selected household's adults on a grid. The respondent to be interviewed is then selected according to a prescribed systematic procedure.
