outset you said in 2012 he would win by 3.32, he won by 3.85. i wish i had more time. i don t. i m fascinated by your poll and like everybody else i ll keep watching it on a day-to-day basis. now columbia professor andrew gillman. you re the stats guy i find remarkable your analysis. tell everybody what you just did and with what result? we looked at above average of polls for various offices over the past several elections looking at polls during the several weeks before the election occurred. each poll has a margin of error so you can look at that but can you also look at how close the poll is to what actually happens in the election. so you can calculate an elm umpirical margin of error. it turns out the emperrical