Track Record

Rasmussen Reports measures the economic confidence of American Consumers on a daily basis. These daily releases are issued two-to-six weeks in advance of the more established measures of Consumer Confidence at the Conference Board, the University of Michigan, and ABC/Money Magazine. Time and time again, the data provided by Rasmussen Reports is confirmed by these other sources.

A recent and dramatic example of this was displayed in December 2003 and January 2004. The pivotal event in December was the capture of Saddam Hussein. The Conference Board’s data for December, however, was collected prior to that event. As a result, they reported a slight decline in confidence for that month. Rasmussen Reports showed a surge in confidence as early as December 15, the day after Hussein was captured. The Conference Board waited six weeks before issuing a report with data from after that big event. When they did, however, they confirmed the surge in Confidence first reported at RasmussenReports.com.

Jim Cramer, founder of TheStreet.com and co-host of CNBC’s Kudlow and Cramer says that Rasmussen is “uncannily accurate”.

While our focus in recent years has been on economic research and corporate business, Scott Rasmussen has a long history in political polling and has accurately projected hundreds of election results. Rasmussen Reports data is cited in the Almanac of American Politics and we issued a limited number of public polls during the 2001 and 2002 cycles. Those surveys have added to Scott’s solid track record of polling in statewide elections.

In 2002, during the final week of the campaign, we publicly released polling data on the Governor’s race in Wisconsin along with Senate races in South Dakota and Minnesota.

• In Wisconsin, we projected that Jim Doyle would be elected Governor by a 4-point margin and he won, 45% to 41%.

• The Minnesota race was complicated by the tragic death of Senator Paul Wellstone and his replacement as a candidate by former Vice President Walter Mondale. We projected a toss-up and Republican Norm Coleman won the race with just 51% of the vote.

• In South Dakota, the final Rasmussen Reports survey showed that Democrat Tim Johnson and Republican John Thune were separated by less than a tenth-of-a-point. The race was decided by an even smaller margin, just a few hundred votes statewide.

In 2001, Rasmussen Reports issued public polls in both of the races for Governor that year

• In Virginia, Rasmussen forecast that Democrat Mark Warner would win the election by 8 points. Warner won by 5. Of the four other public polls released before Election Day, Mason-Dixon showed Warner with a 6-point lead. The other three, including the Washington Post, indicated that Warner was leading by 10-13 points.

• In New Jersey, Rasmussen’s data was the closest of all the public polls. He showed Democrat Jim McGreevey leading by 13 percentage points and McGreevey ended up winning by 14 percentage points. One other firm projected a 9-point race, the New York Times found results similar to Rasmussen, and other polls showed a 16-22 point McGreevey victory.

In 2000, Scott Rasmussen released public polls in all 11 Governor’s races and was 11 for 11. He projected the actual candidate percentage within the 4.5 point margin of error for all 22 major party candidates. In fact, Rasmussen came within 3 points of the actual vote total 21 times out of 22 candidates. The lone exception was Utah, where he under-estimated the Democrat’s total by 4 percentage points. In that race, Rasmussen projected a 58-39 outcome for a 56-43 race.

In that last Presidential election year, Scott Rasmussen polled in Florida seven times between Labor Day and Election Day. Rasmussen’s data provided a remarkably accurate projection of what was to become the longest count in American political history. Six times out of the seven Florida polls, Rasmussen’s data showed that the race was dead even or within one or two points one way or the other. The only exception came the night after the first Presidential debate when George W. Bush received a bounce in all the polls. Even then, Rasmussen’s Florida numbers showed the race was well within the margin of error.

Of course, like all pollsters, Scott has missed the mark on occasion. When that happens, we simply examine the data, learn from it, and move on to the next election. Fortunately, this is a rare occurrence.

An entirely different way to consider the Rasmussen Reports track record is to look at an analysis offered by one of our competitors, Survey USA. Survey USA is a fine firm with more than a decade’s worth of public polling under their belt. They have compared their own results to other firms every single time that other firms issue polls in the same race. Survey USA reports that they have polled in the same race as Rasmussen 31 times. Using an analysis technique developed by Michael Traugott, former president of the American Association of Public Opinion Research, Survey USA found Rasmussen was more accurate 52% of the time. Survey USA was more accurate 39% of the time while results were comparable for the remaining surveys. In fact, Survey USA has gone head-to-head more than 25 times with just six other pollsters. Rasmussen is the only pollster with a winning record in these match-ups.

Our results are unparalleled in the survey industry.

Copyright © 2005 Rasmussen Reports