Polls can tell you, with varying degrees of certitude, the state of play of individual midterm-election contests for the Senate and the House. But with the partial exception of the generic congressional ballot (which estimates House national popular-vote shares for both major parties), polls can’t give you the essential bottom line: how many seats Democrats and Republicans are likely to win or lose, and which party will control the two chambers. To fill that information gap, various outlets have devised “race ratings” to turn poll numbers (and other bits of information like past outcomes in a given state or district) into rough estimates of how each race seems to be developing. And some have gone the whole hog into election forecasts based on complex statistical models that provide probabilities of ultimate outcomes.
The best-known credible sources of race ratings for congressional contests are the Cook Political Report, Inside Elections, and Sabato’s Crystal Ball, all well-established and nonpartisan observers and number crunchers. They generally focus on identifying competitive House and Senate races and categorizing their likely direction. For example, Cook currently rates 37 House races as truly competitive, with 21 toss-ups, nine races “Leaning Democratic,” and seven races “Leaning Republican.” That sounds very even until you look more closely and see that of the 21 toss-ups, 15 involve seats currently held by Republicans, so they are more vulnerable to losses. Cook also lists 12 additional races as “Likely Democratic” and 20 as “Likely Republican,” which means they are on upset watch if a national partisan wave of significant strength em
Комментариев нет:
Отправить комментарий