The poll field work was conducted by Prism based in Sarajevo, Bosnia and Herzegovina.  Oversight and analysis was provided by Jamie Fisfis of Centaur Research in San Francisco.

The field work was conducted October 30 - November 16, 2007.

The poll surveyed a nationwide random, multi-stage stratified representative sample.

Sample size: 1,550 Eligibility criteria: age 18 and older

The questionaire was developed by IRI and Prism Research in close consultation with political parties and civil society groups. Officials of parties submitted their questions after participating in a seminar about methodology of public opinion polls and surveys in Bosnia and Herzegovina.

For entire sample Bosnia and Herzegovina (N=1,550) the margin of error is ± three percent.  For each of sample of three ethnic majority areas (N=500) the margin of error is ± 4.5 percent.

The population under study is representative for the adult population of Bosnia and Herzegovina and included all regions.

Territory of Bosnia and Herzegovina was split into two entities plus District Brčko.

Territory of Federation Bosnia and Herzegovina was split into two large ethnic (Bosniak and Croat) majority areas. Republika Srpska was treated as area with Serb numeric majority.

Each of these ethnic majority areas were split further into 17 geographic regions, proportional to the size of population in each region to the total population in each ethnic majority area.

Minimum two and maximum three municipalites were randomly selected in each region – one per each category of municipality size (regional centers, middle and small size municipalities). Sample covered more than 40 municipalities in total.

In each municiaplity, urban and rural settlements were randomly selected from the list of all settlements. The sample covered almost 300 different settlements in more than 50 municipalities and 18 geographic regions. The sample was set to around 60-40 percent ratio in urban and rural areas. Maximum number of interviews per one sampling point was four.

Household addresses were randomly selected using random walk technique.

Household members were randomly selected using last birthday technique.

Results are weighted in two ways - Bosnia and Herzegovina nationwide and within each ethnic majority area.