Reading Lab · Dossier 15

Iraq as the CPA saw it

Of all the press releases’ references to governorates, Anbar got 5%; of all reporters’ questions’ references, 36%, 7.5 times as much. The CPA talked about Baghdad and the quieter south; reporters asked about Fallujah and Ramadi. The three Kurdish governorates are almost absent from both.

397press releases
2,897briefing questions
124weekly governorate updates to Bremer

What the CPA said, and what it was asked

Right: each governorate’s share of press releases’ references to governorates · left: its share of reporters’ questions’ references · darker means a larger share · hover a governorate for figures.

Where reporters asked more than the CPA said

A governorate’s share of question references divided by its share of release references · above 1: asked about more than spoken of.

Questions over time

Five maps for five periods: questions naming each governorate.

1

Established

Anbar: 24 press releases and 232 questions. Its share of questions’ references to governorates is 36% against 5% in releases. Next: Diyala (3.6×) and Salah ad-Din (2.0×).

2

Established

The Kurdish governorates: Dohuk 6 releases and 0 questions, Erbil 17 releases and 6 questions, Sulaymaniyah 8 releases and 0 questions. Largely outside the CPA’s day-to-day administration, and outside its talk too.

3

Established

Coordinators’ weekly updates to Bremer: Kirkuk 14, Diyala 13, Muthanna 10, Anbar 10, Ninawa 9, Basra 9 (in his memo log).

Bremer’s ear →

4

Open question

Did the CPA choose what to say about the governorates, or did reporters follow the violence? Probably both, and the figures cannot separate the two.

Method, confidence and limits

Press releases: every release in section 05 between April 2003 and June 2004 counts for a governorate if it names it or its main city (e.g. Fallujah and Ramadi for Anbar, Tikrit and Samarra for Salah ad-Din, Kut for Wasit). Questions: “Q” turns in briefing transcripts, by the same rule. Coordinators’ updates: “Weekly GC Update -- …” titles in Bremer’s memo log.

The comparison: most questions name no place, so we do not compare raw shares of releases and questions; we compare how each source’s references are distributed among governorates (each source sums to 100%).

Limits: “Baghdad” appears in release datelines (“Baghdad, Iraq”), so its share is inflated. Naming a governorate does not mean the document is about it. The archive holds no population figures, so we do not measure per capita.

The briefing room → ← Reading Lab