Methodology for the CRG/BERCI poll
Following the release of our poll on October 25, 2016, there have been some questions regarding our methodology. In the interest of transparency, we have published the entire anonymized results here, along with the codebook here. A primer on how to use the codebook can be found here: http://xlsform.org and CRG/BERCI are happy to answer questions regarding methodology. You can also find the entire methodological note employed by BERCI here.
There have been three main complaints since the release of the data. First, that the sample was too small. In fact, our sample was extremely large by most standards, and the approach of conducting face-to-face interviews is much more accurate than phone or online polls. Most national polling in the United States, a population of 300 million, uses a sample between 800 and 2,000 and uses either online or phone polls. Recent opinion polls conducted in Nigeria and Kenya, for example, used sample sizes of 1,000 and 2,200, respectively.
Secondly, the government and some opposition politicians have said that the poll is biased because BERCI was co-founded by Olivier Kamitatu, currently an opposition politician who backs Moise Katumbi for the presidency. Kamitatu was indeed a founder of BERCI, but has not been officially associated with the organization for over 17 years. BERCI has conducted numerous polls for the government, as well as for the World Bank and international NGOs. We used an independent polling expert employed by CRG to conduct quality control and to monitor our enumerators in the field. The interviews were conducted on tablets, which took a time and place stamp for each interview, which were then uploaded to a cloud as soon as the enumerators entered into cell phone coverage. This allowed us to monitor for suspiciously high levels of support for particular candidates, for example.
Third, some critics complained that they felt the results do not reflect the reality of Congolese public opinion. I can only say that public opinion is something that we can only discover through these kinds of surveys, which we believe are more reliable than people’s intuitions or anecdotal evidence.
We strongly encourage other organizations to conduct their other polls and to publish their methodology and results.