Washington Post Revisits Its Own Data (9/30)

In the aftermath of yesterday's two main votes in the U.S. Senate Finance Committee against including various types of public option in the reform bill working its way through Congress, the Washington Post today went back and conducted new analyses on its own poll (with ABC News) that initially gathered data from September 10-12.

Of the Finance Committee's 23 members, 13 voted against both public-option amendments, eight voted for both amendments, and two split their votes (against one amendment and for the other). In their new analyses, the Post focused on the states represented by the first two groups of senators.

As part of a national survey, the pollster may ask respondents what state they live in (which the Post/ABC apparently did). With total sample sizes typically around 1,000 (it was 1,007 for the September ABC/Post poll), the number of respondents in any one particular state would almost certainly be too small for statistical analysis.

However, in a burst of creative thinking, the Post realized that aggregating the residents of states represented by the 13 strong-opposition senators, and doing the same with residents of states represented by the eight strong-proponent senators, could yield two fairly sizable subgroups in the data (although these exact subsample sizes did not appear to be reported).

I have created the following table to summarize how the residents of the two sets of states (those represented by opponent and proponent senators) came down on two key questions discussed by the Post (you make click on the table to enlarge it).


The Post's conclusion, which I think is consistent with the support percentages shown in the table, is as follows (extraneous symbols edited out):

Among those living in states represented by the 13 Senate Finance Committee members opposing both amendments, a majority... gave a thumbs down to the health reform proposals being developed by the Congress and Obama administration. But when asked about a package that excluded a public option, the results flipped and most stood in favor of the reform effort.

The opposite was true among those living in states represented by the 8 members of the committee who voted for both amendments, as support in those states held steady regardless of the inclusion of a public option.

The Post's re-analysis speaks both to the idea of ("small d") democratic representation and the seeming power of the public option's inclusion or exclusion in a final bill to turn around opinion in the "opponent-senator" states. As much as I admire the creativity of the approach, however, I feel it has at least one potential flaw.

While it may be true that in the aggregate the residents of the 13 "opponent-senator" states are negatively inclined toward a public option (although this is an inference from the rise in support for a hypothetical bill with the public option removed), it does not follow that each and every opponent-senator state opposes a public option. (I'm not saying that the Post asserted the latter, just that some readers could draw that conclusion on their own.)

In fact, as I summarize here, when many of the opponent-senator states are looked at individually, their residents actually support the public option by majority or plurality.

Kaiser Foundation's September Poll (9/29)

The Kaiser Family Foundation is out with its September poll, which I think is interesting to look at for a few reasons. First, Kaiser has been polling every month from June 2009 forward and also conducted surveys in February and April of this year (six polls total during 2009), providing useful benchmarks for current trends. Second, this latest poll was in the field from September 11-18, making it a good gauge of the public's short-term reaction to President Obama's big health care speech on September 9.

As seen in the following chart (which you can click to enlarge) in which I attempted to distill the key findings of this year's Kaiser polling, pro-reform sentiment currently appears to be at -- or near -- peak levels for the year, depending on which items one looks at.


For example, the percentage of respondents currently saying "the quality of your own health care" would be better under a reformed system was the highest it has been all year, whereas belief that the health care of "you and your family" would be better off was only one point shy of an annual high.

On the one hand, a skeptic might contend that support levels would be artificially high in roughly the week following the Obama speech (and South Carolina Republican Joe Wilson's "You Lie" outburst). On the other hand, however, some of the figures to which the current ones are being compared were obtained in February and March, during the initial "honeymoon" months of Obama's administration.

Some items still draw low support, such as the two pertaining to waiting times (one asking about personal situations and the other focused on the country at large), where only around 20% of respondents say things will change for the better. On these items, however, there are large percentages (31-38%) expecting things "would [...] stay about the same."

One last thing to point out is that Kaiser did some split-sample experiments to test the effect of certain wordings. As I highlighted in the above chart, inclusion or exclusion of the phrase "similar to Medicare" did not have much effect on support for a public option.

Politico on Interpretive Difficulties

The Politico takes a look at difficulties (real or apparent) in interpreting public opinion polls on health care reform.

Poll of (D-Leaning) Swing Districts (9/25)

A new survey of 1,200 respondents total "in ninety-one swing [congressional] districts (Blue Dog, Frontline, and Rural Caucus House districts)," taken from September 11-17, has now been released. The poll was sponsored by the organization Health Care for America Now (HCAN) and conducted by the firm of Anzalone-Liszt, which works for Democratic candidates (hat tip: Daily Kos). In general, the poll contains results that seem favorable toward President Obama's and the Democrats' health care reform proposals. Some details of the poll warrant particular scrutiny, however.

As a first step in analyzing a poll, one should know the larger population to whom the sample is meant to generalize. The term "Blue Dog" refers to relatively conservative Democrats, so I initially wondered if the sample consisted heavily of persons represented by Democrats in Congress. The term "Frontline," as I just learned via some web searching, also applies exclusively to Democrats. However, the Rural Caucus appears to include both Democratic and Republican members of the House. Thus, it appears safe to say that the sample in the HCAN/Anzalone-Liszt survey consists substantially of individuals who vote Democratic, albeit in many cases for candidates in the party's center-right ideological location.

Second, the poll appears to contain two types of items: ones that directly ask respondents if they approve or not of a given policy proposal, and others that preface the approval-disapproval question with a paragraph-length description of what's covered under the policy. Responses to all types of survey items can be affected by question wording, but some critics seem to really have a problem with the preface/description type of item.

Some of the key results are described in the official poll summary, as follows:

Although initial support for Obama’s healthcare reform plan is under 50% (44% Support / 49% Oppose), voters express strong support for its individual components and a majority favors it after hearing a detailed description (53% Favor / 41% Oppose)...

To say that this poll's overall support for the Democratic plan seems lukewarm, at best, may be missing the point, however. Some conservative Democrats have been threatening to vote against a reform bill that contains a public option. What the poll might convey to the Blue Dogs therefore is that, while bill is not necessarily wildly popular among their constituents, neither is it wildly unpopular.

U.S. Regional Differences in Health Care Reform Attitudes (9/23)

The political fortunes of the Republican party have been heading south in recent years, both literally and figuratively. A Daily Kos/Research 2000 national poll (September 14-17) found that, whereas 50% of respondents from the South gave the GOP a favorable rating, the percentages doing so from the Northeast, Midwest, and West regions were minuscule, ranging from 7-14%. These findings have sparked some commentaries from the blogosphere (here and here; hat tip).

One of the commentators, Joshua Tucker, offered the following analogy: "Quite seriously, if I saw this type of regional distribution of support for a political party in a country like Slovakia, I would assume the party represented an ethnic minority."

Tucker concluded with this suggestion: "...I wonder if we’ve hit the point where the mainstream media ought to be reporting support for the president, congress, political parties, etc. not in terms of the country as a whole, but rather by providing two numbers: support in the South and support in the rest of the country excluding the South?"

In today's entry, I provide such a breakdown for support of President Obama's and the Democrats' health care reform proposals. Before presenting the results, I want to warn that the information is limited, due to many pollsters' either not providing cross-tabulated percentages (i.e., demographics by support) at all in their free online documentation, or not doing so by region.

Thus, some of the polls are older than I'd like them to be, plus the question gauging support is not the same from poll to poll (i.e., sometimes regional breakdowns were available for general support, sometimes for public-option support, etc.). I encountered similar difficulties when I reported on age and gender breakdowns in support. Here are the numbers on region (as always, you may click on the graphic to enlarge it)...


Despite the different areas of focus in the questions used in this analysis (i.e., general support for reform, the public option, Obama's handling of the issue), the trends were pretty similar across polling outfits. As seen further in the graphs, Southerners were pretty consistently the least likely regional group to respond favorably to Obama and the Democrats' reform plans. However, the dip in support from the South (relative to other regions) is not always that large, particularly compared to the Midwest and the West. Thus, the South does not appear to be as different from the rest of the country on health care attitudes as it was on favorability toward the Republican party. The nature of the questions and dates in the field for these polls are as follows.

Economist/YouGov (September 13-15). “Overall, given what you know about them, do you support or oppose the proposed changes to the health care system being developed by Congress and the Obama administration?”

Daily Kos/Research 2000 (August 31-September 3). “Do you favor or oppose creating a government-administered health insurance option that anyone can purchase to compete with private insurance plans?”

Marist (August 3-6). “Do you approve or disapprove of how President Barack Obama is handling health care?”

Gallup (August 31-September 2). “Would you advise your member of Congress to vote for or against a healthcare reform bill when they return to Washington in September, or do you not have an opinion?”

I am also monitoring state-specific polling numbers.

Cost on Linkage of Public Opinion to Senators' and Representatives' Votes

Jay Cost at Real Clear Politics analyzes how public opinion on health care reform might translate into U.S. House and Senate members' votes.

One Week Post Obama Speech (9/18)

Observers of political survey research are no doubt familiar with the concept of a "bounce," the lift in polling numbers a presidential nominee tends to get immediately after his (or someday her) party's national convention. As seen in the two postings immediately below the present one, poll analysts have been examining the immediate "bounce" from President Obama's health care speech on September 9 with regard to support for his and congressional Democrats' reform plans.

Bounces tend to dissipate, however, so it's a good idea to look at polling numbers after additional days have gone by and the "buzz" over a speech has begun to die down. A week post-speech may not seem like sufficient time for assessing the persistence of any gains in support following Obama's address to Congress. However, with the hour-to-hour news cycles in the political blogosphere and on cable news channels, a week arguably is pretty long.

Another issue is whether any observed shift in aggregate opinion is broad-based, as opposed to being confined to partisans of one type or another. The graph below (which you can click on to enlarge) thus examines these two elements of post-speech public opinion on health care reform: what do things look like with a week's passage of time, and how do the trends break down by party? The traditional "red/blue" color scheme for self-identified Republicans and Democrats, respectively, is used, with Independents in purple, and overall (full-sample) trends in black. Each poll (Economist/YouGov [pre, post], FOX News, Pew Research Center) is identified by its initial letter. Finally, each dot placement represents that poll's final day in the field (e.g., the FOX "post" poll was taken September 15-16).


Two trends are apparent in the above graphs. First, there appears to be a small, but consistent upward bump in support for Obama and the Democrats' reform plans. Second, contrary to some media commentary, small increases in support were observed all partisan subgroups, not just the President's Democratic base.

A few qualifiers are in order:

1. Just because measured support for the Obama/Democratic reform plans appears to be higher after the speech than before, this doesn't mean that we can definitively say the speech caused the rise. Other news developments or social/political dynamics potentially could have caused the shift, although I can't think of any other, more salient explanation than the speech itself.

2. The polls depicted in the graph above are, of course, not the only ones available. For example, Rasmussen's daily tracking polls (based on two-day rolling averages) have shown support for the Obama/Democratic reform plans rising from 44% immediately before the speech to 51% roughly three days afterwards, but then falling back down to around 44% in the following days.