Figure 1 displays the key statistic, percent approve, for every individual poll (except the one-nighters) conducted by the five organizations. During Bill Clinton's second term in office, approval ratings generally moved within a twenty-point percentage range. For the first year (1997), Clinton received ratings in the mid-50s to low 60s. In late January 1998, shortly after the Monica Lewinsky scandal broke, they jumped to around 70 percent, as the public rallied around the president. Clinton approval then fell back, but rose again to the high 60s and early 70s around the time the House of Representatives voted for impeachment in late December 1998 and the Senate acquitted him in February 1999. For the rest of Clinton's presidency, his ratings bounced around the 60 percent mark.

(Click for larger view of Figure 1.)

Approval ratings of George W. Bush during his first three years in office followed a very different pattern. Prior to September 2001, Bush approval moved within the 50-60 percent range. After the 9/11 terrorist attacks, it jumped to over 80 percent in all polls, affected by the largest "rally-round-the-flag" effect on record. As the rally faded gradually, Bush approval followed a steady downward path through early 2003, when it jumped upward in another rally effect from the Iraq war. It then resumed its downward trajectory until the end of 2003, when Saddam Hussein's capture produced yet another jump. With a few exceptions, Bush approval in polls taken from 2001 to 2003 ranged between 50 and 90 percent.

As Figure 1 makes clear, presidential approval ratings in recent years did not follow a straight line for very long. Such movement makes it very difficult for us to compare and contrast the results of polling conducted by the five organizations based on individual-level data. Since the five organizations tended to poll at different times and approval ratings were rarely static, timing was a powerful factor that tended to obscure any differences that might have resulted from variation in research methods and practices. Moreover, the sample size of a typical individual poll was no more than about a thousand, meaning that the difference between two approval ratings had to be more than six percentage points to be statistically significant.

Nonetheless, Figure 1 suggests that although the five polls followed the same general trend line, certain patterns occurred in the findings of particular polls that did not appear to be the result of random variation. In the last two years of Clinton's presidency, the FOX approval ratings appeared to be consistently higher than those of other organizations, and the Newsweek results appeared to be consistently lower. During the Bush years, ABC/Post approval ratings seemed consistently on the high side and the Newsweek numbers again consistently on the low side.

In order to reduce sample size limitations and minimize the impact of timing in comparing results, we combined individual polls to calculate quarterly and yearly averages in the percent approve, percent disapprove, and percent "don't know" for each of the five organizations. Table 1, which displays yearly average ratings for George W. Bush from 2001 to 2003, reduces the data for this period to fifteen sets of figures. These yearly averages confirm the impressions derived from the trend line. Overall, the approval ratings from the five polls were quite similar, but certain organizations displayed a tendency toward higher or lower ratings. The range in yearly averages for Bush approval across all five polls was only four to five points for these three years, but in two of the three years—2002 and 2003—ABC/Post's rating was highest and Newsweek's the lowest.

Table 1—George W. Bush Approval Ratings

Note: FOX News figures based on registered voters.

Oddly, 2001 was an exception, with Newsweek's average approval at the high end of the scale. Further examination, however, revealed this departure from the typical pattern as an artifact of the uneven distribution of Newsweek polls in the third quarter of 2001. Because Newsweek did no polls that quarter before 9/11, its third-quarter average was skewed by the extraordinary rise in Bush approval after 9/11 (85 percent). When the third-quarter data were removed from the calculation of yearly averages for 2001, the typical pattern reemerged, with ABC/Post having the highest average approval (68 percent) and Newsweek the lowest (65 percent).

These yearly averages for George W. Bush also showed at least as much variation in the percent disapprove as the percent approve, between five and six points for the three years. While ABC/Post ratings tended to be the highest or second highest, FOX, not Newsweek, tended to be lowest. The percent "don't know" also showed some consistency in the yearly averages, with ABC/Post tending to be lowest and FOX the highest.

For the four years of Bill Clinton's second term, the yearly averages had some gaps, and they did not show all the same patterns as the Bush data. As FOX did not begin asking the standard approval question until 1998, no data were available for 1997. With one exception, Newsweek did not ask the Clinton approval question on polls conducted during election year 2000, and so was not represented in the averages for that year.

Consistent with the patterns observed in yearly approval averages during the Bush years, Newsweek had the lowest approval rating in the three years it was active, as shown in Table 2. However, during the Clinton years it was FOX—not ABC/Post—that generally had the highest approval ratings. ABC/Post had the highest average disapproval during Clinton's second term. "Don't know" patterns were also somewhat different from the Bush years. ABC/Post's average ratings tended to be the lowest, but it was Newsweek and CBS—not FOX—that had the highest average "don't know."

Table 2—Bill Clinton Approval Ratings

Note: FOX News figures based on registered voters.