In a press release and a five-page letter to the scientific community, Peter Kim asserts that although Merck had to correct a statistical method, the 18-month conclusion is still valid based upon the data Merck had at the time.
Whereas your reported projection relied on a single poll question about people's second preferences, our method was to interview 8, 400 people across 18 different regions of Britain immediately after the May 1st general election.