The Korea Broadcasters Association’s Joint Prediction Survey Committee (KEP) on June 11 formally apologised, saying errors occurred in some regional gender and age analysis data during coverage of the June 3 local election exit poll.

The exit poll for the ninth nationwide local elections was split across 16 provinces and major cities and carried out by three polling firms: Hankook Research, Korea Research and Ipsos Korea. For accurate forecasting, exit poll data from election day must be combined with forecast telephone survey data for early voters.

KEP’s own review found that in gender and age analysis for four areas handled by Hankook Research — Seoul, Daegu, Ulsan and North Chungcheong Province — early voter forecast data were omitted and only election-day exit poll results were reflected.

KEP said both surveys were combined normally in the final winner forecasts, but early voter forecast data were omitted from gender and age voter analysis data due to an apparent work error by Hankook Research, and that this caused misunderstanding among viewers. It added that early voter forecast data were properly reflected in areas handled by Korea Research and Ipsos Korea, and stressed that KEP did not modify the data with any specific intent but that it was a polling firm’s work error.

KEP said the three broadcasters had no way to verify at the data receipt stage whether results were produced as designed, while saying it was keenly aware of its management and oversight responsibility. KEP said it will fully reorganise its verification system and plans to take strict measures, including legal action over breach of contract, against the polling firm involved.