Early assessments of polling accuracy in 2020 results point to new questions about methods and assumptions. Unlike in 2016, polls generally predicted the correct presidential winner, but there were plenty of surprises in the results. Pew Research notes that errors in 2020 favored Democrats, pointing to "a systematic cause or set of causes." These could include:
"Polling errors for Senate candidates were quite large," Sam Wang, a Princeton neuroscientist and polling analyst, writes. "Republican candidates had a considerable reservoir of support that was not apparent." Overall, he told Scientific American, "at some level we should stop expecting too much out of the polling data."