How We Conduct Our Public Opinion Research Programs

Our public opinion polls are not focused on the “horse race” issues of which candidate or party is leading.  Nor are our public opinion polls designed to produce a specific outcome.  Rather, we see opinion research as an integral component of IRI’s broader programs – for example, helping partners to better understand public sentiment, develop plans to contact citizens and prepare strategies for campaigns.  Much time is spent on analyzing cross tabulations from the poll so our partners better understand how issues and challenges are affected by demographic factors.

As part of our broader mission to advance democracy around the world, we work to strengthen the institutional capacity for public opinion research in the countries where we work.  We bring together statistical experts and political pollsters to share best practices in public opinion survey sample selection, frame design, variance, weighting, focus group research and political party research design.

Our public opinion research helps strengthen political and civic institutions, such as political parties and nongovernmental organizations.  In some cases, the information we gather and the identity of our partners is especially sensitive.  As a result, we may not publicly our findings.  We follow industry standards and make available the following information on our surveys:

  • When security permits, the name of the company that conducted the survey and the name of the pollster who oversaw the project.
  • A definition of the population under study.
  • A description of the sampling frame used to identify the population under study.
  • A description of the sample design, giving a clear indication of the method by which the respondents were selected by the researcher, or whether the respondents were entirely self-selected.
  • Sample sizes and, where appropriate, eligibility criteria, screening procedures, and response rates.
  • A discussion of the precision of the findings, including estimates of sampling error, and a description of any weighting or estimating procedures used.
  • Which results are based on parts of the sample, rather than on the total sample, and the size of such parts.
  • Method, location and dates of data collection.