Polling in Iraq takes extensive planning, coordinated effort -- and some luck. Not least is that every interviewer, all Iraqis, returned home safely from the field work for the latest ABC News poll, co-sponsored by USA Today, the BBC and ARD German TV.
That's no small feat. More than 100 people worked in the field to complete the survey, randomly selecting and interviewing 2,212 Iraqi adults in 458 locales across the country from Feb. 25 to March 5. Interviewers and supervisors kept journal entries of their experiences -- and while most were relatively uneventful, that wasn't always the case.
Several brought back harrowing tales of having witnessed some of the bombings, shootings and beatings that, as the survey shows, are widespread in Iraq (see main poll analysis).
In Diyala province, one interviewer reported, "Two bombs exploded in front of the public market and destroyed it, then dead bodies spread everywhere in front of us in a very awful way."
On March 5, also in Diyala, another wrote, "In front of me, an explosive went off under an American patrol and the other American patrol started random shooting, which almost got all of us killed hadn't it been for the mercy of God Almighty."
Several teams of interviewers were detained and questioned by Iraqi police. Others were harassed; one had his camera smashed by an "armed person." All nonetheless completed their assignments.
FIELD WORK -- Field work for the survey was managed by D3 Systems of Vienna, Va., and KA Research Ltd. of Istanbul, which began jointly polling in Iraq in summer 2004. D3, which specializes in polling in difficult conditions, co-managed the field work for ABC News polls (one of them with the BBC World Service) in Afghanistan in 2005 and 2006.
In Iraq, where it's prohibitively difficult and dangerous to send centralized teams to travel throughout the country, D3/KARL works with a dispersed staff of 190 interviewers on call to 19 district offices across Iraq. The system has held up in deteriorating conditions; D3/KARL has completed a total of more than 140,000 interviews in Iraq to date.
All polling staff in Iraq are Iraqis. For security purposes they're not told the identity of survey sponsors. Field teams include male and female interviewers accompanied by a supervisor; Shiite staff are assigned to Shiite neighborhoods, Sunnis Arabs to Sunni neighborhoods, Kurds to Kurdish areas. In mixed areas, many field workers carry dual forms of identification with separate Shiite- and Sunni-sounding names.
This poll included the participation of 150 field workers in all, including 103 interviewers and 47 supervisory or data processing staff.
LOCAL -- Supervisors' local contacts help: In Wasit province on Feb. 25, an interviewer reported, "during the field work the supervisor was taking pictures when suddenly we realized a police patrol in the region was following us. When they reached us, they started asking us about our work and whose side we are working for, then they took us to the police station.
"The supervisor is a member in the local council and has many good relations, so he was able to deal with the situation and solved the problem. He convinced them that our work is for the benefit of the public and aims to share opinions of the people."
In Maysan province on Feb. 27, said another, "before finishing my work the supervisor called me to say he will come to visit me to check my work. I waited for him for a long time, but he didn't come. I learned that a police patrol had arrested him after seeing him taking pictures. Because of the efforts of our field manager for this region, his telephone calls, and his personal relations with people in this governorate, the supervisor was released and returned back home safely."
SAMPLE -- The survey used an area probability sample based on 2005 population estimates by the Iraq Ministry of Planning. D3/KARL randomly assigned sampling points proportionate to the population of each of Iraq's 18 provinces, then proportionately to population size in each of the 102 districts within these provinces. Within districts, sampling points were assigned randomly among nearly 11,000 known settlements or neighborhoods. At each stage, the sample was stratified by urban or rural locale.
Maps (or a grid plan where reliable maps were unavailable) were used to select a random starting point for interviewers at each sampling point. Interviewers traveled to the starting point, randomly selected households (e.g., third house on the left), then randomly selected respondents within each household. (See separate methodological summary.)
In-person surveys of this type customarily interview 10 or more respondents per sampling point; 2,000 interviews would require 200 sampling points. Some use fewer points; a typical election exit poll in California (whose area and population are similar to Iraq's) may have 35 to 50 sampling points. To maximize this survey's geographical distribution, D3/KARL used 458 sampling points across Iraq, with about five interviews per point.
INTERVIEWS -- The questionnaire was prepared by ABC News in consultation with its media partners, translated into Arabic and Kurdish (using the Sorani dialect) by D3/KARL, then backchecked by translators working for ABC News. Interviews lasted about 30 minutes on average.
The survey had a cooperation rate of 62 percent -- far higher than usual in telephone surveys in the United States, albeit lower than ABC's previous surveys in Iraq, conducted in less violent times.
Photos from interviews show the range of homes visited -- from small villages to urban centers, from gated homes to shabby apartments.
"The situation of the neighborhood was good, as services are available, but there is no electricity line and there are only private electricity generators shared by people," one interviewer noted. "There is a shortage in fuel and it is very expensive, so people bring cooking and car fuel from the Iranian borders."
Said another, in Irbil, "I noted that the services in this region are average, but there is a shortage in electricity and it is not available before noon. Electricity comes for three hours only and there is no hygiene drainage [sewer] in the region."
Another, also in Irbil, sounded a common concern of farmers everywhere: "They complain that their harvests are being sold at prices that are too low because of the presence of vegetables and fruits from all neighboring countries of Iraq," the interviewer noted. Moreover, "People told me that most young people join the Peshmerga in order to get salaries and they don't work with their parents on the farm."
One of the more poignant interviewer entries -- emblematic of the struggles, conflict and ultimate tragedy of today's Iraq -- came from Mosul:
"When I went to this specific region for the survey I have noticed that there was a mourning ceremony, and I think that it was the last day for it. It was by coincidence that my work took me to this house."
"When I asked to do the questionnaire with one member of the family, they asked me about the reasons behind my visit and selection for their house. I told them that their house was selected according to the random rules of the survey and I told them that we have a project about the opinions of citizens concerning the general situations in the country."
"The process of the interview was very difficult with them, and I learned that their son had been an officer in the former Iraqi army and that he was killed by an armed militia. They don't know the reasons behind killing him, and they said that this is the destiny of former army officers because they receive threats and are killed."
Remarkably, even in mourning, the respondent in this family participated in the survey, along with 2,211 other randomly selected Iraqis. They wanted their stories told.
Sidebar: Methods Report
The following excerpt from D3/KARL's "Methods Report" summarizes the conditions encountered by the survey interviewers, nationally and in each of Iraq's 18 provinces:
A major U.S./Iraqi forces joint security operation in Baghdad.
Major terrorist attacks, assaults, assassinations, kidnappings, curfew, dead bodies found all across Iraq; and the continuing violence in all parts of Iraq (especially in Baghdad and mixed ethnic/religious provinces).
Basra: Intense fighting, suicide bombs.
Muthanna: Some fighting, suicide bombs.
Najaf: Intense fighting.
Anbar: Intense fighting, suicide bombs.
Dhi Qar: No major events.
Maysan: Some fighting.
Qaidisiyah: Intense fighting.
Wasit: Some fighting.
Babil: No major events.
Karbala: No major events.
Baghdad: Intense fighting, suicide bombs, major U.S. military operations, assaults on civilians, assaults on security forces.
Diyala: Intense fighting, suicide bombs, assaults on civilians, assaults on security forces, U.S. military operations.
Salahuddin: Suicide bombs.
Sulaimaniyah: No major events.
Tamim: Some fighting, suicide bombs.
Ninevah: Intense fighting, suicide bombs, assaults on U.S. forces, assassination attempts on politicians.
Irbil: No major events.
Dahuk: No major events.