Sunday, January 21, 2007

Hillary Clinton, Favorable/Unfavorable, 1993-2007

(Click on graph for full resolution)

Hillary Clinton has served in the Senate for only six years but she has one of the longest polling records in American politics, going back over 13 years to the beginning of her time as First Lady in 1993. No other Presidential contender has this long a track record of public opinion readings. As a result we have an unparalleled record of the ups and downs of public response to Sen. Clinton.

There was considerably more variation in favorable/unfavorable ratings during her time in the West Wing than there has been since Sen. Clinton became an elected official in 2001. The White House years describe the downturn in favorability during first two years, which included the failure of her leadership of the health reform initiative. Sen. Clinton's favorability rebounded during and following the 1996 presidential election. The effect of the Lewinsky scandal was largely favorable to Senator Clinton, if not to President Clinton.

The decision to run for the U.S. Senate in 2000 produced a new downturn in favorable opinion (and upturn in unfavorable views). However, since she took office in 2001, Senator Clinton has enjoyed relatively stable favorable ratings of around 50%, while unfavorable ratings have averaged in the mid-40s. This sharp split is, of course, one of the more widely remarked aspects of Sen. Clinton's public image.

Most recently, there has been a modest upturn in favorable ratings, especially since the fall elections.

The figure above plots and estimates the trend in favorable/unfavorable ratings after removing the poll to poll variation, also known as "house effects". These are effects due to the standard practices of different polling organizations that introduce systematic differences in results. For example, some polls make it easier for respondents to say they "don't know" their feelings towards a political figure, while others produce lower rates of "don't know" response, pushing more survey respondents to pick a positive or negative response. In the plot above, I've used an iterative estimation procedure to remove the "house effects" and norm the results to those of the Gallup organization. This provides better comparability because it adjusts all polls to a common standard. However, it changes the observed poll results to fit the model. Some may wish to see the raw data. They are presented below.

The figure below plots the raw, unadjusted, data. The greater spread in the gray data points is due to the house effects. In particular, you can see the "don't know" response has considerably greater variation in the unadjusted data. The broad outlines of Senator Clinton's favorable and unfavorable ratings are similar to the adjusted ones. The late upturn in approval, however, is largely only visible when house effects have been removed. The raw data suggest less recent movement.

(Click on graph for full resolution)

The bottom line of these data is that Senator Clinton remains a public figure who has both a large following of supporters and one who suffers from a substantial antipathy. At the moment, the Senator enjoys more supporters than opponents. Whether than margin is enough to win a national contest or now is, of course, what Democrats (and perhaps the rest of the electorate) will have to decide in the coming months.