Approval of President Bush in 2005 and 2006

Updated through polls completed 10/22/2006

2005-2006 estimate of approval of President Bush.

The blue line is the estimated support for the President at each point in time. The gray dots are the individual polls which vary randomly around the estimated support. Each poll will differ from the common trend due to sampling variation, question wording or order among other ideosyncratic factors. The estimated support is based on a local regression which can flexibly fit the trend in the polls.

Approval of President Bush, 2001-present

Approval of President Bush. Gray points are polls, blue line is my estimate of approval trend, using a local regression. The spread of gray around the line indicates how much variation there is across polls taken at the same time.

The vertical grid marks quarters, with January 1 in a darker shade. Tic marks on the horizontal axis are months.

This graph is at 768 x 768 resolution. Some readers may need to click or double click on the graph to see it at full resolution.

Presidential Approval in Historical Perspective

Updated: 11/8/07























Presidential Approval since Roosevelt. The blue vertical lines indicate midterm and presidential election dates.

This graph shows all presidential approval readings by the Gallup organization since Roosevelt. Gallup's long series and consistent question wording provide a useful historical comparison across Presidents. The data are freely available at the Roper Center web site at the University of Connecticut here.

The data are graphed here in exactly the same perspective across all presidents. This means that trends will look the same regardless of whether a president served two terms, one term or only part of a term. Visual comparisons of trends will be valid because of this constant perspective.

For example, compare the decline prior to the 1992 elections in President George H. W. Bush's approval with those of President George W. Bush. The elder Bush is a clear example of "free fall", the sharpest and largest approval drop since President Nixon's in 1973-74. President George W. Bush's decline more closely resembles the long-term decline of Jimmy Carter's approval than it does the free fall of either the elder President Bush or President Nixon.

Because the vertical scales are also equal, it is easy to compare the level of presidential approval across Presidents.

A more complete explanation of how the almost universal failure to keep perspective constant distorts presentation of polling results can be found here.

Polling error distribution in CA propositions

Distribution of deviations between polls and final vote on seven California referendums, November 2005.

The several California referendums this week provide a good illustration of polling variability as well as a look at the average accuracy of the polls. Earlier I illustrated the distribution of the pre-election polls here. Now that we know the results, we can compare the outcomes with the estimates from the polls.

MysteryPollster has a comparison of the polls average accuracy across the referendums here. He focuses on the mean prediction error, while I focus on the “spread” of the distribution of errors. My concern is with the uncertainty we should attach to the polls as predictions of vote. The more spread there is, the greater our uncertainty. Because the prediction error also includes all non-sampling errors, these distributions are better measures of our practical uncertainty than is the sampling error alone.

It is worth admitting up front that proposition polling is unusually difficult from a technical perspective. MysteryPollster has addressed these difficulties here (and in related posts.)
I also address the issue raised here by “sennoma” to look at the change in the error distribution between early and late polling.

The figure above shows the distribution of errors, poll minus vote percentages, pooled across all the referendums. (I omit Proposition 80 which had few polls.) The dashed lines are for polls taken more than 10 days before the vote, while the solid lines are for polls taken in the last 10 days of the campaign. The reason to pool across polls and propositions is that we have no a priori reason to assume one poll is “better” than another and no reason to assume polling is easier or harder on any given proposition. Pooling allows us to maximize cases and assess our uncertainty prior to seeing any results.

The distribution of errors tightens for polls taken in the last 10 days. For “yes” responses the middle 90% of errors from more than 10 days prior to the election range from -16.5 to +17.1. For polls taken in the last 10 days the 90% range is -8.6 to +10.6. The median error shifts from +4.1 to +1.8, showing that the middle of the distribution comes closer to the actual outcome in later polling for the “yes” vote.

For the “no” responses, the 90% range is -20.1 to +6.5 for early polling, and -16.50 to +0.55 for the last 10 days. The median error is -12.5 for early and -8.5 for late polls, showing a systematic underestimate of the “no” vote.

Put another way, the 90% confidence interval for these distributions is 19.2% for yes votes and 17.1% for no votes. That range is quite large compared to the theoretical margin of error for sample sizes typical of these surveys, which are reported to range from 3-6% for a 95% margin of error. The empirical 90% range is +/- 9.6% for yes and +/-8.55% for no votes. This inflation of theoretical margins of error by a factor of 1.5 to 2.0 or more is typical of empirical estimates of survey accuracy as predictors of election outcomes. The inflation reflects sources of error in addition to those due to sampling, such as campaign dynamics, non-response, question wording effects and other sources of error. (And since I'm comparing 95% sampling margins to 90% empirical ranges, I'm being generous to the polls here.)

These results also demonstrate the problem of bias in the survey estimates. While the percentage of “yes” votes had a median error of only +1.5, the error for no votes was a rather large -8.5% median error, a quite large error. While smaller than for early polls, the median of the “no” distribution falls rather far from the actual outcome. One might chalk this up to “undecideds” who ultimately voted no, but this is a problem of modeling from the survey data to a prediction of the outcome--- something that few pollsters actually do in public polling. (MysteryPollster calculates the errors using the “Mosteller methods” that allocate undecideds either proportionately or equally. That is standard in the polling profession, but ignores the fact that pollsters rarely adopt either of these approaches in the published results. I may post a rant against this approach some other day, but for now will only say that if pollsters won't publish these estimates, we should just stick to what they do publish-- the percentages for yes and no, without allocating undecideds.)

The bottom line of the California proposition polling is that the variability amounted to saying the polls “knew” the outcome, in a range of some +/-9.6% for “yes” and +/- 8.55% for “no”. While the former easily covers the outcome, the latter only just barely covers the no vote outcome. And it raises the question of how much is it worth to have confidence in an outcome that can range over +/- 9 to 10%? In general, I'm pretty confident elections will fall between 40% and 60%. That's only slightly larger than the empirical uncertainty in these polls. If anyone would care to pay me for that prediction, I'm pretty sure I will be as accurate as these polls. (Apologies to my pollster colleagues. But don't worry, no one will pay me for my opinion!)

The next post below (I'm posting these in reverse order) looks at polling in the New Jersey, Virgina and New York City partisan elections. These are interesting to compare to the much harder to poll proposition issues I've focused on here. 

Polling error distribution in NJ, VA and NYC

Distribution of polling errors in New Jersey and Virginia Gubernatorial elections, and New York City Mayoral election, November 2005.

As with the California propositions, we can calculate the variability of the errors for the New Jersey and Virginia Gubernatorial and the New York City Mayoral election polls. Again, there is little a priori reason to think the poll performance will differ across these, so I pool over the three elections and all the polls.

The figure above shows the variability of the errors and the shifts in the location of the distributions between early and late polls. For polls taken more than 10 days before the election, the middle 90% of poll errors is -10.70 to -0.67 and -9.04 to -1.65 for polls taken in the last 10 days of the campaign. The medians are -6.65% and -5.19 for early and late polls, a modest improvement but still a substantial underestimate of the final vote.

For the Republican vote, the 90% interval is -7.78 to +4.22 for early polls and -8.53 to -1.86 for the final 10 days. The medians are -2.78 and -2.39 for early and late polls.

These results show a modest decrease in uncertainty in late polls, and small shifts of the median estimates towards the true outcome. However, the median underestimates the Democratic and Republican share of the vote, even in late polls. As with the California proposition polling, these underestimates reflect the failure to allocate the undecided voters in the polls. But this begs the question of how such votes SHOULD be allocated and by whom. We can arbitrarily choose either of the Mosteller suggestions mentioned above: proportional or equal allocation of the undecideds. But the more reasonable method of allocation of undecideds requires access to the raw individual level data, which is not available in general and never before the election.

On the face of it, one might estimate the predicted probability of voting and of voting for each candidate among the undecideds, using the decided voters for the estimation, and from this estimate the final vote prediction. Pollsters are apparently unwilling to do this, preferring to simply report poll percentages rather than make modeling decisions as well. (This seems a reasonable business choice, even if it frustrates election prognosticators.)

The bottom line here is that the empirical margin of error for late polls in these three races is +/- 3.7% for the Democrats and +/- 3.3% for Republicans. Unfortunately, neither of these ranges covers the final vote percentage, so the bias in estimating the voter percentage is larger than the margin of error. Again, without the raw data, we cannot calculate an optimal allocation of undecideds which would reduce the bias estimate. (One can also calculate the “spread” between the candidates, which estimates who is ahead. Those calculations make the polls look a bit better in picking winners, but with still substantial variability, which is my story here, rather than prediction per se.)

Polling for the 2005 elections

Today's elections have produced a good deal of polling and considerable uncertainty as to likely results. MysteryPollster has done several excellent posts on this here, here, here and here. MP stressed the difficulty of polling on California's propositions, and provides a thorough discussion of the implications for poll variability.

Here I want to show you the variability across polls in the propositions plus the NJ and VA governor's races. I'm going to do this a little differently than you usually see it. Here's why.

We can think of each poll as a sample of the likely outcome. The uncertainty arises from both sampling errors, campaign dynamics and non-sampling errors such as non-response, question wording and a host of other polling demons.

It is common to toss out early polling and focus only on the last week or even days of polling. I'm not going to do this. Rather I want to show the total distribution of results across the polls. This reflects the uncertainty due to campaign dynamics, as well as the other factors. The reason I think that's worth doing is that changes over the campaign are unpredictable. What has gone down may come back up in the end, or vice versa. We place a lot of faith in the last week of polling (not wholy misplaced) but in representing how uncertain we are about the likely results, I think it makes sense to consider the total variability in polling. This provides more polls as part of the evidence as well.

Of course where there are clear trends in the polls, this may make the uncertainty appear greater than would be the case only with late polls. That matters if we are primarily interested in predicting winners. But if we want to represent uncertainty, I think my approach is preferable.

So in the graphs below I present the distribution of polls for each California proposition and for the NJ and VA governor's races. Where there have been clear trends, I mention them. But you won't find the usual time series plot here. That isn't my point. Rather look at where polls have been highly variable--- we should be very uncertain there. Where they have not been, we should be less uncertain.

The wide variability in percentages and in spread should be appreciated by readers of polls (and perhaps by pollsters themselves). Sampling error is only part of the story of uncertainty. These plots give a better sense of how much uncertainty comes from all sources, not just sampling.

The conclusions from most of these plots: don't place big bets based on the polls. (But do note that polls DO seem to provide clear evidence in some cases. The exceptions count too!)

Data: See RealClearPolitics.Com for the numbers and for the trends.

California propositions 73 and 74 polling

Polling on props 73 and 74 show a wide spread with no clear leader for either proposition.

Prop 73, parental notification for minors seeking abortions, has produced widely divergent polling results with alternating polls finding substantial leads for both sides. The distribution of the spread shows a range of over +/- 15%, a huge range of uncertainty.

Prop 74, modifying tenure rules for public school teachers, has also produced substantial variation in results. Some of these are due to trends over the fall. Early polling produced a number of "yes"-leading results, but the most recent polling has found "no" leading by single digits. 

California propositions 75 and 76 polling

The polling on Prop 75, controlling the political use of union dues shows an interesting bimodal pattern. Much of the spread is concentrated around zero, a dead tie. But there is a second mode- centered around a 20% advantage for the "yes" side. Much, but not all, of that right tail comes from early polling, but even polls in the last week find results with both sides winning.

The proposal to limit state spending (and modify previous propositions regarding spending), Prop 76 seems likely to lose. This is a case where the simple "no" percentage has been polling well over 50%, rather than both options polling below 50%.

California proposition 77 polling

The redistricting issue might seem to be a juicy target for reform via proposition. It should be easy to paint the districting process as corrupt and that "simple honest effort" would make the process much more fair. And who would want to campaign in favor of gerrymandering? Yet the polling here is largely "No", though there is a little right-tail in the spread graph. But bet "no" on what might have seemed like a proposition with a real chance.

Polling also shows redistricting reform trailing in Ohio, faced with strong opposition from Republican officials (who stand to lose control over the current redistricting process.)

"Fair redistricting" is a lot harder than proponents suggest, so there is a lot of oversimplification of the issue. But I remain surprised that this issue hasn't been welcomed by voters.

California propositions 78 and 79 polling

Support for these drug discount propositions have produced polls that are both evenly divided and very spread out. One might speculate that "duelling propositions" such as these, which are common in California propositions, produce very large uncertainty among voters which is reflected in the polling. These typically produce competing propositions claiming to "solve" some problem but with each backed by different interest groups. Not only must voters discover the implications of the proposition, but they must also decide which of the alternative proposals might best accomplish their goals. Result: uncertainty.

There are also fewer polls available for these propositions than for the others, which further increases uncertainty. Bottom line on these two: who knows! 

NJ and VA Governor's race polls

Based on the polling, the Virginia governor's race looks impossible to predict. The New Jersey Governor's race polls have generally given both candidates less than 50% but virtually all show Corzine leading Forrester.

Other races to come. 

Cases mentioned in Alito coverage

Number of articles appearing in "Major Papers" discussing cases involving Judge Samuel Alito.


When discussing legal issues, news coverage has focused overwhelmingly on abortion cases in reporting on Judge Samuel Alito's nomination to the Supreme Court. In the week since Alito was nominated, 79 articles discussed Alito's position in relation to Roe v. Wade, with 70 articles discussing Alito's position in Planned Parenthood v. Casey. The next most mentioned Alito opinion appeared in only 10 articles. That case, United States v. Rybar, involved issues of federalism in regulating gun sales. The fourth most mentioned case, Bray v. Marriott Hotels, appeared 7 times and involved employment discrimination. No other case was mentioned in more than four articles.

During this period, 10/31-11/6, a total of 528 articles appeared in the Lexis/Nexis "Major papers" database. Of these, only 92 (17.4%) discussed specific cases in which Judge Alito was involved. An additional 50 (9.5%) articles mentioned Roe v. Wade but no other case involving Alito. The remaining 386 (73.1%) articles on the Alito nomination failed to mention any case by name. It is possible that some of these 386 articles discussed the substance of cases without naming the case. To the extent this is widely true, the estimates here will understate the prominence of cases in the coverage.

The disproportionate coverage of abortion related cases reflects the emphasis that both conservative and liberal groups have placed on the Roe decision as a factor in Supreme Court nominations. It probably also mirrors the public's limited awareness of issues before the Court with the exception of the most highly visible cases, of which Roe is the premier example.

Nonetheless, the reporting demonstrates a very narrow appreciation by reporters of the legal issues that are relevant to the Court and to the appointment process. While the public appetite for detailed legal analysis may be limited, the press is falling short in informing the public on the range of issues which define the differences between judicial conservatives and judicial liberals. Instead, the differences are reduced to abortion and only occassional mentions of other issues.


Data: The data are a result of a search of the Lexis/Nexis "Major Papers" database, covering 89 mostly US sources. The search terms used were distinctive parts of case names, rather than the full formal name. Thus "Casey" rather than "Planned Parenthood v. Casey". The list of searched cases was developed by examining all articles that appeared to cite a case (as determined by a search for " v. " or " v ". The list of cases is as follows, with the number of articles found in parentheses.

United States v. Allegheny Ludlam (1)
Baker v. Monroe Township (1)
Blackhawk v. Pennsylvania (1)
Chittister v. Department of Community and Economic Development (1)
Cruz v. Chesapeake Shipping (1)
D.R. v. Middle Bucks Vocational Technical School (1)
Doe v. Groody (1)
United States v. Igbonwa (1)
Pennsylvania Coal Association v. Bruce Babbitt (1)
Public Interest Research Group v. Magnesium Elektron (1)
Saxe v. State College Area School District (1)
Specter v. Garrett (1)
W. R. Grace v. EPA (1)
Zubi v. AT&T (2)
Child Evangelism Fellowship of NJ v. Stafford Township School (3)
Fraternal Order of Police v. Newark (3)
Elizabeth Blackwell Health Center v. Knoll (4)
Planned Parenthood v. Farmer (4)
ACLU v. Schundler (4)
Sheridan v. E. I. Du Pont de Nemours (4)
Bray v. Marriott Hotels (7)
United State v. Rybar (10)
Planned Parenthood v. Casey (70)
Roe v. Wade (79)

News coverage of Alito nomination, part I

Use of "Scalito" in news coverage of Alito nomination with multiple measurements.

Initial coverage of the nomination of Judge Samuel Alito to the Supreme Court relied heavily on the "Scalito" nickname to convey the Judge's conservatism and possible similarity to Justice Antonin Scalia. That usage declined rapidly after the second day of coverage, falling to well under 10% of all articles by Wednesday, November 2. In last week's post here, I wrote about this, based on searches of web news sources via Google, and Lexis/Nexis searches of US papers and major papers (as defined in the Lexis/Nexis database.) The current results are updated through searches of Monday, November 7, and confirm the rapid decline in the use of "Scalito" and the continued low level of usage.

In response to the earlier post Robert Chung raised an important question concerning how fast these databases are updated, and hence how reliable searches taken no more than a day after the source appeared might be. These data give us a nice chance to answer that question here.

In the graph above, I've plotted the results based on searches done on four different days: 11/2, 11/3, 11/5 and 11/7. For searches of Google News, the four searches produced nearly identical results, reflecting the very rapid updating of Google's data. In the figure, only the blue line (from 11/7) shows up, but that is because all four searches produced identical results, with the exception of the 11/3 search of news from 11/3, precisely when we would expect some lag in Google's indexing (or in web posting of articles, for that matter.) The implication is that Google searches of news within 24 hours of the date of publication appear to be quite stable.

For the Lexis/Nexis data, the delay in updating is apparent, though there is rapid convergence to stable search results as well. The Major papers database appears to update slightly faster than the US Papers database. In either case, within 48-72 hours of publication the database is quite stable. One reason for the US papers variability is that this database contains some non-daily publications, hence longer lags, while the Major papers data are all daily publications.

So the bottom line is that we should be careful about using searches without verifying them a couple of days after publication. On the other hand, the search sources are quick enough that, at least in this case, the conclusions we reach are unchanged.

For whatever reason, the Web news sources used "Scalito" in over 40% of their initial stories. Print used it in 20-30% of their Monday and Tuesday stories. After that, all sources dropped the use of the term to well under 10% of articles. Language Log has a nice discussion of this from a non-quantitative perspective here.

What did the press write about instead? Stay tuned.

Presidential Approval by Party Identification

Presidential approval by party identification in four post-indictment polls compared to results from January 2005.

Barry Burden posted a comment here asking if approval has begun to drop among Republicans. That seemed a good enough issue to deserve a post of its own. He points out that Republican support has been running over 80% which has provided a strong base that would support President Bush despite declines among independents and Democrats.

The answer is that support has declined among Republicans as well as Democrats.

In these latest four polls, the approval rate among Republicans is 77, 78, 78 and 76%. In a January 2005 Pew poll the Republican support level was 89% (and an AP poll from December 2004 put the rate at 91%.) At the same time the Democratic support rate was 17% among Democrats while it is now 11, 11, 9, and 11 in these late polls. So there has been an approximately 12% decline among Republicans and a 6% or so decline among Democrats. Among Independents, there is more variation in support: 31, 33, 24 and 28. This compares to 47% in the January Pew poll, a fall of some 17%. Thus the rate of decline has been strongest among Independents, then Republicans and least among Democrats (who had less room to drop.)

The Pew Center for the People and the Press has a very nice graph of this in this report from early October. Their graph is nice because the comparison is all between their own polls, avoiding the house effects that are present in my comparison here. The salient paragraph in their comparison is

The president continues to draw strong support from Republicans, 81% of whom approve of the job he is doing. But that number reflects an eight-point decline since January, with most of that drop occurring in late summer. Among independents, a plurality of 47% approved of Bush's performance in January; now just 34% do so. Approval among Democrats is now in the single digits (9%), down from 17% in January.

There has been some discussion concerning “how low can approval go”. MysteryPollster has a nice discussion here and here. Some have argued that there is a floor for President Bush due to his exceptionally strong support among Republicans. But that only holds if his base remains constant in their strong support. These data suggest that base has been eroding at an alarming rate. While the level is still high, it is not as stable as other commentary would suggest.

A related concern for the President is the percentage of Republicans in the public. There is evidence that this has been declining modestly over the course of the year. See this piece by Alan Abramowitz at the Cook Political Report. See figure 3 on page 4.

The declining support among Republicans and the modest decline in Republican identifiers suggests that while it is way too early to panic, there is a significant possibility for further erosion in support for the President. Those who think there is a hard floor of support may be overestimating partisan loyalties.

Presidential Approval since the Libby Indictment

Presidential Approval since January 1, 2005 with estimated trends and model fits.


President Bush's approval rating has fallen in the first six public polls taken after the indictment of Vice-President Cheney's Chief of Staff Lewis “Scooter” Libby on October 28. Based on a model of approval that takes into account the year-long decline in the President's approval rating, the effect of Hurricane Katrina and variation across polling organizations, I estimate the immediate decline since the Libby indictment as -1.59%. This compares with an estimate of -1.28% for the effect of Hurricane Katrina.

During the week that included the Libby indictment other events occurred which may have also contributed to the decline in approval. The press widely reported the 2000th U.S. military death in Iraq and the President accepted the withdrawal of his second Supreme Court nominee, Harriet Miers. Further, some of these polls interviewed respondents after the President nominated Judge Samuel Alito to the Supreme Court on October 31. It is not possible to separate out the effects of each of these individual events which all may contribute to the -1.59% estimate.

Since January 1, 2005, the President's approval has declined at a rate of -.031% per day. The immediate post-Libby indictment effect thus amounts to a 51 day drop in approval (-1.59/-.031). This compares to a drop equivalent to 41 days due to Hurricane Katrina and its aftermath.

While the decline of 1.59% is statistically significant, it is considerably smaller than the sudden drop of some 20% following the Iran-Contra revelation in the Reagan administration and the 10% drop in the Clinton administration during the Monica Lewinsky scandal. Both of those declines, however, occurred when approval was over 60% for both Presidents Reagan and Clinton. The estimated approval for President Bush as of October 27th was 40% in my model, suggesting that sharp declines may be somewhat less likely than for the previous presidents.

The recent polling results, along with previous polling, are available at PollingReport.com.

My model includes estimates of differences in approval across the various polling organizations. The estimated effects of Katrina, Libby and the long term trend are only slightly affected by ignoring these effects, though the inclusion of the “polling house” effects reduces the uncertainty of the estimated effects considerably.

One of these days I'll get around to posting on the house effects estimates. In the mean time, see Robert Chung's excellent pages on estimating these effects. His results and mine differ only modestly, though our estimation techniques are somewhat different. In my case, when looking only at the 2005 data, as here, I use a simple linear regression model with variables for date, post-Katrina and post-Libby plus indicator variables for each polling house. When modeling the entire Bush administration I use a different approach, closely related to the one Chung uses, though the approaches were developed independently of one another.

The model is based on 143 national polls by 17 different polling houses. There are 40 polls post-Katrina and 6 post-Libby. Three of these post-Libby polls are partially or completely overlapping at 39% approval in the figure. The six post-Libby polls included here are Associated Press-Ipsos poll Oct. 31-Nov. 2, 2005
ABC News/Washington Post Poll. Oct. 30-Nov. 2, 2005
Zogby America Poll. Oct. 29-Nov. 2, 2005
CBS News Poll. Oct. 30-Nov. 1, 2005
CNN/USA Today/Gallup Poll 10/28-30/05
ABC News/Washington Post Poll. 10/28-29/05

Initial News Coverage of Alito

Initial news coverage of the nomination of Judge Samuel Alito frequently repeated the nickname of "Scalito" to compare the nominee to Supreme Court Justice Antonin Scalia. Web based news sources were especially prone to use this nickname in articles appearing on Monday, October 31, the day President Bush announced the nomination of Judge Alito. Print papers made less, but still substantial, use of the term in their Tuesday coverage. In the small number of Monday print articles, "Scalito" was used about as much as in the much more extensive Tuesday coverage.

In the second day of coverage, however, news articles made much less use of this nickname. Web article use fell below 10% by Tuesday with print articles falling to under 7% on Wednesday, their second day of substantial coverage. Web use of the term fell to a mere 2.8% by Wednesday.

This illustrates the problem of initial interpretation in news coverage and raises the question of the impact such coverage has on public understanding of Judge Alito (or other newly visible figures.) As a memorable phrase, "Scalito" adds color to a news story while also conveying the writer's fundamental point: "(Alito) has been called "Scalito" for a judicial philosophy akin to that of Justice Antonin Scalia" (Steve Lash, The Houston Chronicle, May 26, 2000, the earliest published reference I could find.) To the extent this phrase is an accurate short hand for a complex argument it serves to inform the public in much the same way that saying Judge Alito is "conservative" or "right-leaning" or "a strict constructionist" might also do.

However, there is another element to this, and that is the tie to a figure (Scalia) who carries not only informational but emotional connotations. It is unlikely that many average citizens can cite in any detail at all the opinions expressed by Justice Scalia, but both liberals and conservatives may have a stronger emotional response to him as a symbol of judicial conservatism. Because this emotional response is quite available for politically aware citizens, even if the legal details are absent, then the "Scalito" nickname may serve to orient readers to Alito in ways that less personalized descriptions of his legal views may fail to do. If so, we might expect readers exposed to the "Scalito" reference to form stronger and more extreme views (positive or negative) of Judge Alito than they would had they only read about his opinions or his legal philosophy.

The news media relied substantially on the "Scalito" shorthand in their initial reporting but have moved away from it dramatically in subsequent reporting. This may reflect the pressure of deadlines. When the nomination was first announced, few reporters could have had the time or the background knowledge to produce an overview of Alito's judicial writtings and so the immediately available shorthand of "Scalito" was used to convey the general orientation of the Judge. With more time in an aditional news cycle the reporting turned to other (one might hope more detailed) reports of Alito's past rulings and philosophy and the use of the nickname was dropped. (Or it could just be pack journalism at work on day one of the story.)

None the less, the use of shorthand labels such as this may affect readers' first impressions of new public figures. In this case, the "Scalito" moniker would seem likely to both inform readers as to Judge Alito's philosophy but also to increase polarization of views because liberals might be expected to react more negatively to Alito when he is explicitly identified with Scalia, while conservatives might welcome this as a sign that the new nominee is what had been hoped for. Initial polling does in fact suggest that opinion towards Alito is a bit more polarized than for either Miers or Roberts, though it is impossible to know how much to attribute this to the "Scalito" nickname.

Subsequent coverage has dramatically reduced the use of the "Scalito" term. Whether the public learns more from this new and more detailed coverage, or is more shaped by initial impressions is an interesting question.

(Data for the figure are from a Google News search of web news sources using "Alito and Scalito" or "Alito and Not Scalito" for each of the three days. The U.S. and Major papers counts are from Lexis/Nexis searches using the same search criteria. The "US News Sources" include 374 publications, though many of these are not daily papers and would not have been available in the search done on November 3. The "Major Papers" include 89 daily publications, a modest number of which are outside the US.) 

Presidential Approval as of 11/1

Presidential approval polls (gray dots) with local regression trend (blue).

In the aggregation of 1014 polls taken by various public polling organizations during his administration, President Bush remains near the low point of his public approval rating. The blue trend line is an estimate of approval that pools across all survey organizations using a "lowess" fit. The gray dots represent each poll and their spread around the line gives a sense of how much variation there is across polls around the trend.

The trend line does not reflect public reaction to the CIA Leak indictment or to the withdrawal of the Harriet Miers nomination or her replacement by Judge Samuel Alito. As such, this is a good point to take a deep breath before plunging into what comes next-- the post-Katrina, post-indictment, post-Miers, post-Alito phase of the Bush administration.

Gallup and ABC/Washington Post Polls taken after Miers withdrew and after the "Scooter" Libby indictment showed insignificant changes from previous results. However, a good deal more polling will be needed to gauge the impact of these events. More on that as results come in.

Why does the graph go through 2008? When we judge the rate of change in approval or the range of approval, we need a constant visual perspective in the plot. All polling organizations I know of do not understand this. As a result their trends change shape as new data are added (and sometimes old data removed.) When looking at approval over an administration, it is important to maintain the same perspective. My graph above will grow to the right as time passes, but the trend shown will look exactly the same in future graphs as it does today. For a more extended example of how misleading other presentations of these data can be, see my paper here.

Second term presidential approval trends

Estimated linear trends through October of post-war second term presidential approval.

President Bush has started his second term with falling approval ratings. In this, he is not significantly different from Truman or Eisenhower in the rate of decline through October of their second terms. However, President Bush started his second term at a lower overall approval rating of any second term president since opinion poll data have been available.

To clarify the differences between administrations, I estimated the simple linear trend in approval through October of each second term. This deliberately ignores the details to focus on the broad trend and not the bumps and wiggles.

Four of the six presidents began their second term with a decline: Truman, Eisenhower and Bush with declines of -1.1, -1.7 and -1.3% per month respectively. Nixon's Watergate problems produced by far the most dramatic rate of decline, -4.4% per month.

Two presidents gained support in the first 9 months of their second terms: Reagan at a +0.38% per month rate, and Clinton at +0.34% per month.

From this perspective, President Bush's rate of decline is not at all exceptional. Where he does suffer is from a lower initial approval rating combined with the negative trend. Truman started his second term 5% above Bush's 52% estimated approval. Clinton started his second term insignificantly more popular than Bush, but his postive trend improved his standing over time.

(The estimates presented here are linear trend estimates from Gallup Poll data for the various presidents. The data are available free of charge at the Roper Center here. Because these are linear trend estimates, the actual percentage support at the beginning of the Bush second term (estimated at 52%) differs from any single poll taken near that time.)

Second Term Jinx, part II

First and second term approval since 1945. The vertical line marks the second inauguration.

Each president has ups and downs. The figure above plots the first and second term approval in the Gallup poll since Truman. The data are available free at the Roper Center here. Scroll to the bottom of the page for presidents prior to Bush. The Gallup Poll web site also has a Presidential Approval Center page here (subscription required) which has party breakdowns and the trend for President Bush. I'm sure they have the data for past presidents also, but I couldn't find them in a quick check.

As I showed in an earlier post here, the second term approval rate has been worse on average than the first for three of the five two term presidents in the post war period (not counting Bush who still has 39 months to go before we know the answer to this.) Of these, Nixon and Truman had uninterrupted declines. Eisenhower had a bad first two years of his second term, then rebounded only to decline somewhat towards the end of his term.

Reagan's second term had two years of bliss--- steady and high approval. Then the Iran-Contra scandal hit, and you can see the nearly instantaneous drop of some 20 points right after the 1986 midterm election. Reagan only slowly recovered from this, trending up until the very end of his term when he enjoyed a final lift in approval to his old levels.

Clinton, by comparison, enjoyed long term positive gains in approval. When the Lewinsky scandal struck he suffered only a 10% approval decline. Even impeachment had only a modest effect on his approval ratings, which remained near 60% in his last two years. The contrast with Nixon is, of course, striking. Nixon was driven from office by collapsing ratings and the threat of impeachment. Clinton was impeached, but never lost the support of the public, and of course was not convicted by the Senate.

Another aspect of the second term is its variability. Despite their second term scandals, Reagan and Clinton has less variation in their second term approval ratings than in their first. Reagan's standard deviation was 7.6 in the first term, 6.7 in the second. Clinton's were 5.6 and 3.9. Truman also had a smaller standard deviation in the second term, but that came from the highly erratic trace of his first term rather than any good news in the second term. Not surprisingly, Nixon's second term was far more variable than his first (standard deviations of 5.1 and 12.0 in the first and second respectively.) And Eisenhower was only a bit more variable in the second term, 5.1 and 5.8 in first and second.

So, if the second term is not generally worse than the first in mean or median approval, and the second term is not generally more variable than the first (2 of 5 are more variable), and if in the face of scandal some presidents can rebound (Reagan and Clinton), though others resign (Nixon) I think we should say that, at least when it comes to public opinion, there is precious little evidence to support commentary which presumes a systematically worse second term than first.

If President Bush suffers in the polls, or if he rebounds, it is a matter of the politics he makes and the circumstances he faces, not a systematic tendency for second terms to go badly. 

The Supreme Court will shift only slightly right

Distribution of Bayesian Ideal Point Estimates of Rehnquist Court Justices.


The nomination of Judge Samuel Alito to replace Justice Sandra Day O'Connor has dramatically changed the public debate over the nominee compared to that of either Harriet Miers or now Chief Justice John Roberts. Roberts high quality but lack of a paper trail allowed his hearings to avoid direct confrontation over his positions. Miers weak qualifications allowed judicial conservatives to dominate the debate, again avoiding a direct ideological and partisan confrontation. Judge Alito, in contrast has very high qualifications, which I think cannot seriously be questioned, plus a long and clear track record of judicial opinions which will be strongly attacked by liberals and defended by conservatives. The hearings will focus on the most extreme statements that can be found by liberals in his published opinions, a defense of his views by conservatives, and an interesting opportunity to see how this plays out with the mass public. (We visited this issue before here.) It is going to be fun. This is the judicial war that both extremes wanted to have.

But how much is it likely to matter for the alignment of the Court? Both sides will claim much but the evidence is not so much.

The key to any Court is the median voter, the decisive fifth vote for any decision. In the Rehnquist Court, that vote was O'Connor's. In the Roberts Court (with Alito replacing O'Connor) that fifth vote will be Kennedy's. But as the figure above shows, Kennedy is only slightly to the right of O'Connor. Even if we assume Alito is as conservative as Scalia, and that Roberts turns out to be to the right of Rehnquist, there remain only four votes for very conservative positions. Someone must convince Kennedy to join the four most conservative Justices. (And I continue to wait for the evidence on Robert's positions.)

This configuration raises an interesting problem for the Roberts-Alito additions to the Court. Kennedy was only slightly to O'Connor's right. But there will be a substantial gap between Kennedy and the group of Roberts-Alito-Thomas-Scalia. There will be a similarly large gap between Kennedy and the group to his left: Stevens-Bryer-Ginsburg-Souter. If Roberts and Alito turn out to be near Thomas and Scalia, Kennedy will actually be closer to the liberal wing. As such, a stronger conservative wing of the court could drive the swing vote to the left. The conservative Justices will have to make sure this doesn't happen, and in doing so will have to moderate their views enough to win Kennedy's support. The liberals will similarly have to moderate their views to win Kennedy over to their side.

Regardless of where the wings of the Court stand, Kennedy is the key vote. With O'Connor on the Court, it was possible to win a majority by gaining either Kennedy's or O'Connor's vote. The figure shows considerable overlap between their estimated locations. The liberal wing could win with either O'Connor's or Kennedy's votes, and sometimes both. The conservative wing required both Kennedy and O'Connor to win.

Now Kennedy stands alone in the middle of the Court. Neither wing can win without him. This gives him tremendous leverage to shape opinions to his satisfaction, and requires both wings to be far more deferential to his views.

The upshot is that conservatives are likely to be disappointed once more with the outcome of a new conservative appointment to the Court. Despite adding Roberts and Alito, the Court is likely to remain a center-right Court, and not a "right-right" Court. Not unless Kennedy changes his positions.

So sit back and watch the huge fight we are going to see over Alito, with claims that the fate of civilization as we know it hinges on his confirmation or defeat. But don't be shocked when the Court fails to shear sharply to the right after his confirmation.

If you want to look for the day of real change in the Court, wait for the next appointment after Alito's. And remember that President Bush had 39 months still to serve. The odds are good that he'll have at least one more appointment. If one of the liberals, or Kennedy, retires, that appointment will be the one that shifts the median in a big way.

(The ideal points presented here are my estimates based on the Justices voting on decisions during the Rehnquist Court. For much more excellent analysis of this see the work by Martin and Quinn here. Their work produce slightly different estimates, but the same qualitative conclusions.)

Second Term Jinx, Part I

First and second term presidential approval, 1945-2005.

There has been a predictable flurry of articles on the presidential second term "jinx". Some examples are The New York Times, The Christian Science Monitor, and Newsday to name only a few. This morning's Milwauke Journal Sentinal has a rare sane view.

The evidence for second term decline is relatively weak. Two of the five two-term post-war presidents have improved their second term standing over that of the first. (The final story is not yet written for Bush.) Truman, Ike and Nixon all are lower on average in the second term, though Ike's low isn't that low. Reagan has a slightly higher median in the second term than the first. Clinton breaks the mold with clearly higher second term ratings than first, and that despite being impeached!

But no matter what, five cases do not amount to an inevitable second term jinx. Truman had post-war economic adjustments and then Korea. Ike had a weak economy in the second term, Nixon had Watergate, Reagan Iran-Contra and Clinton Lewinsky. So scandals haven't been uncommon, but not universal either.

President Bush's current problems are much more like Truman 's and Eisenhower's problems-- driven by political decisions and policies started in the first term. With the Libby indictment this week, there will be a "scandal" as well, but it wasn't the scandal that drove his ratings down this year-- it was the failure of the social security proposal, the long term costs of the war, and then topping that off with gas prices, hurricanes and a less than brilliant nomination.

It is hard to argue that 3 out of 5 cases constitute a clear pattern of poor second terms. Rather it looks like a toss up.

There is another problem with the second term jinx argument: it ignores those presidents who had a first term "jinx" and never got to serve the second term. Recognizing this selection bias in second term jinxes wipes out any remaining argument for systematic decline in the second term.

We have five postwar presidents with 2nd terms (not counting Bush) three of whom do clearly worse in the second term. Compare that with the four ONE term postwar presidents who had such bad first terms they couldn't win reelection. (Johnson, and Ford are special cases, neither initially elected but both eligible for another term). So it looks to me like four presidents had "jinxes" in the first term, and three of five had "jinxes" in the second term. It isn't a jinx.

When you do bad in the first term, you don't get to have a second. When you do ok in the first, there is still a just about even chance (3/5) that the second term is worse than the first. I conclude there is no systematic difference between first and second terms, at least in presidential approval ratings.

I'll have more to say on this soon. 

Public Understanding of Miers Withdrawal

The Gallup/CNN/USAToday poll, taken 10/28, N=516, finds that response to the Miers withdrawal follows partisan and ideological lines, with liberals more pleased and conservatives more disappointed by the withdrawal. The CNN story on the poll is here. Gallup has a Frank Newport writeup of the poll here. (Free as of Saturday, but subscription required after that, I think.)

The interesting thing is that the most vociferous opposition to Miers came from the most conservative wing of the Republican party, yet among the public conservatives were more disappointed than pleased by the withdrawal. How come?

One simple partisan answer would be that liberals like seeing the President twist slowly in the wind, while conservatives are disappointed by anything that hurts the President. While plausible in general, this story line doesn't address the giddiness at the National Review web site on Thursday morning:

CORKS [Kathryn Jean Lopez]
I'm getting some of those e-mails too. But no champagne here. No gloating. Just relief. It was the right thing to do--withdraw--and it has been done. Super. Was this whole thing unnecessary? Yes. But I don't think that's our fault. The nominee made no sense. So it's over. Live and learn. Posted at 09:34 AM

RE: CORKS [Jonah Goldberg]
Kathryn -- That would be easier to buy if you weren't wearing a giant party hat and drinking out of your shoe and Chaka wasn't in the mailroom snockered out of his gourd xeroxing his butt. But, yes, you're right. Blessed are the peacemakers.
Posted at 09:37 AM

As any regular reader of conservative opinion can attest, the criticism from the "Intellectual Right" was overwhelming, and based on both Miers judicial philosophy (or lack of same) and limited qualifications compared to many other prominent conservative jurists and legal scholars. Among "Social Conservatives", criticism more commonly focused on whether Miers was a reliable conservative, or might show a moderate-to-liberal streak that would not provide the kinds of outcomes that are of most concern to social conservatives. So whether one is a reader of The Weekly Standard or Eagle Forum (or a daily listener to Rush, for that matter), the message was the same, a poor choice who should be withdrawn.

Pitted against these conservative messages were those coming most prominently from President Bush himself, and the White House, combined with those religious conservatives such as Dr. James Dobson who spoke in favor of Miers. Both the White House and its surrogates talked about Miers religious faith, evangelical views and the "near unanimous" opposition to abortion in her church.

So the messages from conservatives and Republicans were quite divided in their valence towards Miers' nomination. What about the liberals and Democrats?

Some prominent Democrats, most notably Senate Minority Leader Harry Reid, supported Miers but the vast majority remained quiet or expressed concerns. But none launched a full attack against her qualifications or judicial philosophy.

So what does this configuration of elite messages suggest about public opinion in response to Miers? The most important point is to appreciate how little attention the mass public pays to complex issues such as this that delight Washington and us political junkies. As a previous post here mentioned, the level of "don't know" response to Miers averaged over 30%. But even that does not mean that the 2/3 who did express an opinion about her nomination understood the context of the debate elite opinion makers were waging. The nature of the mixed signals that were coming from elites suggests that the public was even less able than usual to make sense of the Miers nomination and its issues.

UCLA political scientist John Zaller, author of "Nature and Origins of Mass Opinion", argues that citizens look to elite signals for considerations that support their positions. When the message is entirely one sided (think aftermath of 9/11) then the public moves more or less uniformly in the direction of elite consensus. When there is division, citizen look for cues as to who to believe. The most prominent of these are partisan cues. Here think of the run-up to the Iraq war. A unified Republican party sent unambiguous messages to Republicans and conservatives in support of the war (despite a considerable isolationism visible two years earlier in the 2000 campaign.) The Democrats sent a mixed message, with both supporters and opponents of the war. Mass opinion was overwhelmingly in support of the war as a result, with much more division among Democrats in the mass public than among Republican partisans.

So what happens when both parties are sending mixed signals, as in the Miers case? Conservatives and Republicans would be exposed to two conflicting messages-- both from usually trusted sources. Because most people pay limited attention to politics, the President's message would be expected to affect mass Republican opinion more, simply because even among Republicans only a small proportion are avid consumers of conservative sources such as George Will, The Weekly Standard, National Review, the Wall Street Journal or Rush. A secondary source of information was through the "mainstream" media coverage of the controversy, which Fox, CNN, MSNBC as well as the broadcast networks made prominent parts of their coverage. The upshot is that any conservative or Republican had to decide not only who to believe but also to sort out what the issues really were. The difficulty of this task was guaranteed to produce a split in conservative opinion, with the advantage going to the best known, most widely visible, and most familiar source-- President Bush. However, the counter message was also available and produced a significant split in opinion. The result, a split among conservatives of 34% pleased with Miers withdrawal and 44% disappointed. Such a small edge to the President's position is testimony to the strength of the conservative critique of Miers.

Among Democrats, the message was much more muted. Despite a few supporters, such as Reid, the overwhelming response was... silence. Democratic leaders stood on the sidelines and watched the Republican death match. As a result, Democrats failed to send strong signals that a Miers defeat (or withdrawal) was clearly a good thing (for their side), but they didn't send a strong signal in favor of her either. This left the mass Democratic partisans to fall back on their partisan perception of the debate, and overwhelming disapproval of President Bush, in order to judge the meaning of the Miers withdrawal. The result was Democrats and liberals were a bit more united in their response to the withdrawal than were Republicans and conservatives-- a rarity in itself. Liberals split 55% pleased to 25% disappointed, while conservatives split 34%-44%.

Roberts drew enough Democratic support that the partisan response to him was relatively muted, while Republicans were nearly unanimous in support. Republicans split 68-8% in favor of Roberts, while Democrats were evenly divided 31-33% in a Pew poll taken September 8-11, just before the Senate hearings.

Both of the Miers and Roberts figures would be expected to be more extreme with more typical homogeneity of partisan elite messages. If the next nominee provokes unified partisan messages, we should expect to see a much more divided response in mass opinion.

Miers v. Roberts

Support to opposition ratio. Roberts in red, Miers in Blue.


Harriet Miers withdrew her nomination for the Supreme Court this morning. President Bush "reluctantly accepted" her decision to withdraw.

While Miers drew strong opposition from the conservative legal movement, her support in the mass public never approached the levels of support for now Chief Justice John Roberts.

Polling on Supreme Court nominees is tricky because large portions of the public pay little attention and lack information to support their judgments. In Robert's case, the median "Don't Know" response was 27%, and the mean 29.5. Even at the end of his confirmation process, in September, the median rate of "don't know" responses was 27%. For Miers, who was a nominee for less time, the median DK response was 31.5% and a mean of 34.6%.

To avoid confounding "don't know" with opposition, I look at the ratio of support to opposition, removing those without an opinion. This lets us focus on the more attentive public who may have more meaningful opinions, and avoids depressing apparent support due to high levels of DK. For example, Roberts median support was only 50%, which alone suggests weak support. However, his opposition median was only 24%, with the rest not holding an opinion. This results in a 2-1 support to opposition ratio, and a healthy confirmation in the Senate. As the figure shows, there was little trend in the support ratio for Roberts, which mostly ranges between 1.9 and 2.6 throughout his confirmation process.

The Miers support ratio is strikingly lower. Her median ratio was only 1.2, with half of the polls putting this support between 1.07 and 1.24. Her highest support ratio was equivalent to Robert's lowest ratio. And in the last poll released before her withdrawal, Miers' support ratio had fallen below 1.0, to 0.98 in the Gallup/CNN/USAToday poll of 10/21-23.

The Miers nomination collapsed due to strong conservative elite opposition and lack of support in the Senate, rather than public opposition or outrage. But the public perceptions reflect the failure of the administration to build a convincing case for her qualifications.

(Data in the figure reflect all public polls asking some form of the question "Do you support or oppose John Roberts/Harriet Miers nomination to the Supreme Court". There is quite a bit of variation in wording, which I have not statistically modeled in the figure. Such modeling would probably reduce the variation a bit for both candidates but would certainly not affect the central conclusions.)

Rosa Parks, 1913-2005

This isn't the usual post here, but I cannot let Rosa Parks' passing go without saying something.

My viewpoint is of a white boy growing up in a small town outside of Montgomery. I missed the bus boycott (I was 1 when it happened), but it was still talked about when I was becoming aware of the civil rights movement as I started school in 1960. If that seems young for political memories, take a moment to put civil rights and school integration into the context of small town Alabama of the time. For whites this was an upheaval of the social order. The New York Times obituary describes that social order well:

On Montgomery buses, the first four rows were reserved for whites. The rear was for blacks, who made up more than 75 percent of the bus system's riders. Blacks could sit in the middle rows until those seats were needed by whites. Then the blacks had to move to seats in the rear, stand or, if there was no room, leave the bus. Even getting on the bus presented hurdles: If whites were already sitting in the front, blacks could board to pay the fare but then they had to disembark and re-enter through the rear door.

These mundane injustices were, of course, only the tip of the iceberg of racism that pervaded my hometown and state. My home county was 70% black, but had no blacks registered to vote in 1960. Education was, of course, segregated and anything but equal. All the businesses on the two blocks of "main street" were white owned and none employed blacks, at least not "out front" dealing with customers-- quite a few blacks worked in the back, doing hard chores. And the police and sheriff were friends of mine, but not all that interested in equal justice. And of course the churches were as rigidly segregated as the schools. The Klan was strong in my home county.

So when Rosa Parks didn't get up for a white man to sit, it was a radical action, one that my students today cannot really imagine, let alone understand as white southerners of my generation can. She challenged a social system that was an abomination and a mockery of the words of the Declaration and the Constitution (as amended.) And that act required a courage that surpasses my comprehension.

But that courage ceased to be rare after Rosa Parks' example. In Montgomery at first, with the boycott and then throughout Alabama and the south, African-Americans risked much and sometimes lost their lives in the pursuit of equal rights. The famous are remembered and revered, as they should be. But I also remember the un-famous who integrated the schools in my little town, who registered to vote when their names were noted by white officials and who spoke out when speaking was so terribly lonely. And continued to do so year after slow year of gradual progress. Their courage was at least as great as that of the famous.

What Rosa Parks did for black folks is obvious. From her small act grew a mighty movement. So what did Rosa Parks do for me, a white boy from Alabama, and all my kind? She saved our souls. The social order I grew up with was a mortal sin, damning the souls of all of us who supported and perpetuated and benefited from that evil. Thank God for Rosa Parks and all the brave souls who followed her. My kind may yet have a chance at forgiveness and salvation.

SurveyUSA 50 State Approval

SurveyUSA has new 50 state approval ratings that are quite interesting. The data at their site are here.

SurveyUSA uses automated polling that is open to question by most survey research standards. MysteryPollster.com has discussed the methodological issues here and here. I think the validity of their polling is an empirical question, and by and large their results appear to track well with conventional surveys. Perhaps I'll post on this later, but for now, let's take their results at face value. And you'll see below a very nice match between their results and national surveys using conventional methods.

My graph compares President Bush's approval in June (before either the Vacation with Cindy Sheehan or Katrina or Harriet) and his rating in October. (SurveyUSA took the October poll 10/14-16/2005, with 600 respondents in each state. The June survey was 6/10-13/2005, also with 600 interviews per state.)

What is interesting is the uniform drop in approval of President Bush across the states. The regression fit (the purple line in the figure) is OctApprove = -5.130 + 1.0086*JuneApprove. You can fiddle with adding differences between Bush states and Kerry states, but the difference is quite small. After trying several specifications, I think the simplest model with across the board decline is the best. (See Philip Klinkner's comments at PolySigh for a somewhat different view.) While there is an obvious relationship between states won by Bush and approval levels, the critical fact is that ALL states (except Oklahoma) show a shift downward, below the diagonal line in the figure, representing stable approval levels.)

So the conclusion is a little over a 5% decline in approval between June and October that crossed partisan state lines. Given the year-long decline of -.03% per day reported in my previous posts, the trend would account for -3.75% of this decline (125 days x -.03 per day), with the post-Katrina shock taking the remaining -1.38% change. This is very close to the estimates based on national polling reported here, of -1.3472 (after adding the latest Gallup poll into the data. It was -1.2% before the October Gallup data came in.)

So what is to be said? The President's problems are very widespread. This is not a "Blue State versus Red State" phenomenon. Instead, the President is suffering equally everywhere. The causes of the decline are widely reported, though poorly documented. We all cite gas prices, the war in Iraq, Katrina and so on. There is little empirical evidence for which of these (or other) variables are really driving approval ratings down. Regardless, this is a presidency in trouble across the board.

Of particular political relevance are the Red states in the figure that are below 50% approval in BOTH June and October. There are sixteen (16!) states that Bush carried in 2004 but in which fewer than half the adults approve of his handling of his job in either June or October. The midterm election is fast approaching. These are not encouraging signs for Republican candidates, or (especially) potential candidates. Don't count on the Democrats to produce a coherent and winning platform, but Republicans must fear self-inflicted wounds that criple them even against a poorly organized opposition. Should the Democrats put together a coherent critique AND positive proposals for change, this could be a devastating midterm for the hoped-for long term Republican Majority we heard so much about in January and February. If Clinton squandered his second term with Monica, Bush may also be losing his chance to create an enduring partisan legacy for his party and for the conservative movement. The "political capital" of January appears to have been spent, and then some.

Two views of Presidential Approval, Oct 13.

Presidential Approval, January-October 13, 2005.

Two views of President Bush's job approval ratings come from the recent flurry of new polls. The first focuses on recent events. I've been calling these "Post-Katrina" but perhaps should now call them "Post-Harriet".

This first graph shows the new results with a total of 26 post-Katrina polls (in red). The decline in the President's approval is clear visually, and the local trend line in red shows how sharply the trend has departed from the blue linear trend for PRE-Katrina polls. (Note the close fit of red and blue trends before August.)

The model, corrected for differences among polling houses, now estimates a Katrina effect of -1.2%. Prior to Katrina, approval was falling at a rate of -.03% per day throughout 2005. So the post-Katrina drop is equivalent to 40 days of pre-Katrina trend.

This will be the last estimate I give of the "Katrina" impact. I believe events have now moved on, so attributing any further change to Katrina, is probably a confounding of effects. The estimate of that effect has remained stable at between -1.1 and -1.3 for a while now, so I'm satisfied with this estimate.

The bottom line: the President's approval has fallen all year, declining about 1% every month since January. But since August we've seen a sharper drop. One important remaining question is whether the rate of decline has also increased, after we take account of the immediate Katrina effect. So far that can't be estimated because it is confounded with Katrina. But the next month should give us the data for an answer.

The next graph takes the long view of the entire Bush administration. Due to a change in Blogger that I can't seem to circumvent, this graph and comments appear in the next post, rather than right here, as I might wish. My HTML skills aren't what they might be.

Two views of Presidential Approval, Part 2

Presidential Approval, 2001-2005

The most crucial observation here is that over the course of President Bush's administration, his approval rating has never fallen as fast as it has since August. This is not "free fall" (the Elder President Bush fell faster in the fall of 1991), but the drop since August 1 is disturbing. The effects of Katrina are significant, but these are now absorbed in the rating. The current difficulties with the Miers nomination, the ongoing Iraq casualties and gas prices (even if they are moderating a bit) are now going to drive the approval ratings. So far the administration has failed to regain control of the agenda to arrest this decline.

And as bad news brings bad news, the Miers nomination has probably added to the President's problems. Not that all that many citizens are as glued to the coverage as we junkies are, but because the tone it sets for coverage of the President is likely to continue to reflect his troubles rather than his successes. Approval of the Iraqi constitution may help a little, but we will pass 2,000 U.S. troop deaths within the next 6 weeks, and that will almost certainly produce renewed focus on the cost of the Iraq war. If the Miers hearings are underway at about the same time, the President could face damaging news on two fronts. The President's attempt this week to reassert the positive case for our involvement in Iraq was generally well reviewed but almost immediately lost in a week of other controversey. His talk with the troops today probably isn't going to help matters.

Second terms often seem to depend on second string players in the White House. Whether it is the new players, the distractions of grand jury investigations, a loss of agenda control by the President (after a failed Social Security initiative), or uncontrollable events both natural and political that have taken their toll, the result is a seriously damaged presidency. Someone in the White House needs to take control and make better decisions or the 2006 elections may slip away--- even if Democrats are not, so far, producing a clear alternative message. The elections may be over 12 months away, but candidate recruitment is now, and rational Republican candidates must see a hard road while Democrats are encouraged by the current turmoil.