Every serious observer and participant in the formulation of Democratic political strategy is familiar with the work of Democracy Corps – the polling and strategy organization founded by Stan Greenberg and James Carville. Democracy Corps is unique in making all its polling research public and providing a perspective that is firmly partisan and progressive but insistently objective in its data collection and interpretation.
Democracy Corps is an independent project run by the polling firm Greenberg, Quinlan, Rosner Research (GQRR). Democracy Corps defines its role as “providing research, strategic advice, and a public voice to the issues important to the American people. It acts as a resource for the unions, public interest organizations, party and congressional leaders who are working for a more responsive government.”
But, while Democracy Corps frequent memos are read by essentially every major political commentator and analyst, there are few if any articles or commentaries that examine the unique aspects of D-Corps methodology and the quite significant methodological advances they have recently introduced.
In order to better understand these topics, The Democratic Strategist interviewed Erica Seifert, the co-author of recent memos with Stan Greenberg and the chief coordinator of D-Corps day to day activities.
Q – What is your background and how did you come to work at Democracy Corps?
I actually have a PhD in political history and was an academic specializing in the recent history of political communication and candidate images before I arrived at Greenberg Quinlan Rosner. My academic training included substantial work in oral history and quantitative methods. Also, I like to think that, as an historian, I have a unique perspective–especially in terms of witnessing and analyzing trends in our research.
Q – What did you find when you got to GQRR and Democracy Corps?
What I found was that there was actually a wide range of backgrounds among the people who work here at the firm. There’s really no standard bio, there are social science people and quantitative people working together.
Q. Does GQRR have a formal methodology that it teaches new employees?
No, there is no one particular book or specific formal academic or technical methodology. The core methodology comes from the really vast internal knowledge-base that has been developed over the years and the extensive experience accumulated within the firm. The main training and orientation for new employees is a series of unique workshops that Al Quinlan, Jeremy Rosner, Stan Greenberg, and other in-house experts conduct.
Q – One thing that distinguishes D-Corps’ strategy Memos from other poll-based strategy analyses is the more extensive use of data from focus groups that is closely integrated with standard opinion polling. Many people – particularly Democrats – are often somewhat suspicious of focus groups because of their close connection with advertising rather than political science research. Is this sense of distrust at all valid?
Actually, the need for focus groups in political polling derives from the fact that opinion polls — and specifically the design of questions — is really quite subjective and subject to bias. We’ve all seen extreme examples of questions slanted to produce a particular answer but there’s a more subtle problem as well. We in D-Corps, for example, because we are progressives, could easily write questions that sound right to our ears but which fail to capture the linguistic and psychological framework that ordinary voters employ when they think about a particular political issue.
It’s for this reason that both Stan and Al insist on doing focus groups before writing poll questions. It’s simply too easy to write poll questions that sound right to you but not necessarily right to your audience.
As a result, we do focus groups to learn the language people actually use in normal conversation and how they spontaneously describe things. We use this information to improve our questions as well as to improve the messages we test. Focus groups allow us to hone the grammars and textures we use in our polls and messages.
Q – Do you videotape your focus groups?
Yes, for the same reason. It’s a very different thing to actually see the expressions and emotions people express during a discussion than it is to read a transcript or scan a statistical summary of responses.
Let me give you an example. For some time we have been honing what we call our “middle class message.” It is an attempt to capture what people tell us about how the modern economy “feels” – the very real pain of losing a way of life that defined past generations.
In one of our groups, when we presented this message to the group two of the women participating began to cry. They were touched by the message and said that no-one else seemed to understand how it feels to lose so much. How, for example, they had always counted on being able to contribute something to their children when they were retired and their children had grown but that they now find that they have nothing to give.
This kind of emotional reaction tells you that you are on the right track in a way that no amount of data from poll questions can match. It tells you that you are striking an emotional chord with the people you are trying to talk to. And this is vital. Emotionality is core of political campaigns – if you don’t connect emotionally with a person, that person could end up as a drop off voter.
Q. Apart from giving you a sense of emotional intensity, are there specific political insights that you can learn from focus groups that you can’t get from polling?
Most definitely. For example, when people say that they want “compromise” and “bipartisanship” many analysts misunderstand what they are saying. They are not saying that they want political leaders to split the difference on every issue and find some exact 50 percent middle point between extremes. The focus groups have shown us that voters mean something quite different. They see the partisan positions of both Democrats and Republicans as rigid, dogmatic and inflexible – as vacuous canned talking points rather than attempts at serious discussion. As a result, they don’t see any exact midpoint between two inflexible positions as representing a genuine solution. They see it as an entirely mechanical “split the difference” approach that doesn’t really grapple with the real choices and real issues that have to be faced.
When these people call for compromise or bipartisanship they really mean something closer to “independence” or “freedom from rigid dogmas.” They want politicians who do have real core principles and convictions but who are not tied to mechanical formulas and will honestly look at the facts without preconceptions and try to find practical solutions even if they lead sometimes in one direction and other times in the other.
In fact, they think of politicians who are tied to a party line as people who really have no genuine convictions of their own, who won’t listen to other people’s points of view and who won’t change their minds no matter what someone says. They think most politicians don’t really believe most of what they say but just parrot a party line.
This is a radically different image of what people have in mind when they say they favor “compromise” or “bipartisanship” then what some analysts, pundits, or casual observers suppose they are thinking but it comes out very clearly, again and again in our focus groups.
Q – In recent months Democracy Corp has been developing something they call the Economy Project. Can you explain what this is?
Essentially, the main goal of the Economy Project is to gain a deep and ground-level understanding of economic experiences. We also develop and test economic messages and narrative frameworks built around the economy. Democracy Corps first launched the Economy Project in 2009; the scope of the research has evolved since then. It is probably the most dynamic project we do–we started the year with dial meters during the State of the Union, followed by break-out focus groups immediately following the speech. We did in-depth interviews in March and April. Our big report came after we conducted focus groups and a national survey in May. We began the second phase of 2011 research with a large webtest in June and will be tracking economic metrics throughout the fall.
Q – What is one of the main things that this new approach has revealed
This year, anyway, I think our research has allowed us to understand the real economy separate from the economic crisis/recession. The jobs environment–which includes pervasive underemployment (far more important than unemployment) and an insecure and “smashed” middle class–goes beyond the recession. As we wrote in our memo, “‘The economy’ is not the recovery, but a set of powerful on-going realities: a middle class smashed and struggling, American jobs being lost, the country and people in debt, and the nexus of big money and power that leaves common people excluded.” This is not the recession–it is a result of a fundamental shift in the economy–politicians must come to terms with this in order to make this economy livable for the majority of Americans.
Q One important research method that the Economy Project adds to D-Corps’ traditional methods are in-depth interviews. Tell me about these.
These are one-hour one-on-one telephone interviews. The people conducting the interviews pre-call the subjects to let them know what general topics will be covered and to set a time for the interview. The interviews themselves are conducted according to a detailed script with a series of questions and branching follow-up questions depending on the subject’s responses.
Q – What are the advantages – and disadvantages — of this technique?
Well, the most significant advantage is that very often in focus groups one or two people tend to dominate the discussion and you get relatively little from the quiet ones. Yet these quiet ones are very often the disengaged or ambivalent voters you want most to understand. The interview technique gets you the feedback from these important voters. Another advantage is simply the depth and range of issues you can cover in in-depth interviews.
On the other hand, one advantage of focus groups is that they are more interactive than an interview that follows a fixed set of questions. In a focus group, the people conducting the research are sitting right there and can signal the moderator to veer off the planned agenda and pursue issues or topics that arise unexpectedly in the course of the discussion. In this respect, in-depth interviews and focus groups are complementary techniques, not alternatives to each other.
Q – The original questions in the Economy Project about personal difficulties with the economy were about spells of joblessness, reduced wages, loss of home or fringe benefits. You have now added some additional issues to this list. What are they and what new information do they provide.
We have now added questions about gas prices – for example, have you had to cancel a trip because of prices — and other aspects of the cost of living. This gives us more texture in our understanding of the effects of the economy on ordinary people.
One thing we discovered with these new metrics is that the national unemployment rate is not a particularly good way to measure the ways in which people actually feel “employed” or “unemployed” or secure in the workplace. We have also been able to see where the official cost of living index falls short–that is, where traditional ways of measuring changes in the cost of living do not really reflect what people see in their daily lives. What we’ve found is that items that are not tracked–for example, the cost of a college education–are actually central. Education costs have gone through the roof, and people perceive this as a major increase in the cost of living – one with important long-term consequences for their children and the future. But the cost of education (and the burden of the loans that increasingly finance it) is rarely considered in the traditional short-term “weekly shopping cart” conception of living costs.
Q – Another methodological innovation you’ve recently begun using more consistently are web surveys. How do these differ from traditional phone surveys and what new insights are they providing?
Well, first off, these are really a different kind of research instrument from phone surveys and are not directly comparable. They are based on large samples which we then weight by a whole series of demographic and psychographic factors to make them as reflective of the electorate as we can.
There are two major advantages to web surveys. First, you can get much larger sample sizes than from phone surveys. Our web surveys have samples of 2000 rather than the 800-1000 typical in phone surveys.
This is important because with only 1,000 respondents, many demographic or geographical subgroups have so few respondents that the data is statistically insignificant. With 2,000 respondents, on the other hand, the numbers in many of these subgroups reaches useful levels.
The other important benefit is that we can test different political messages in a quasi-experimental format, creating several matched groups of respondents and running two or three different variations of message, message order and other variables in a way you just can’t with phone surveys.
One of the most interesting things has emerged from some of our recent research is validation for something that Stan has been saying for a long time. In our phone surveys we have found that a message centered on the middle class tends to be the most popular with voters but we have also found that voters seem equally insistent that government corruption and inefficiency cannot be ignored.
With the web surveys, we were able to disentangle the separate effect of these two factors. We found that the middle class message was indeed the most powerful – but only when preceded by a strong message calling for reform. As Stan has been arguing for several years, tackling government reform seriously is a kind of “entrance fee” for being taken seriously by voters. The generally popular message about jobs or restoring the middle class dimply doesn’t get treated seriously if a candidate or platform does not first make a clear and meaningful commitment to government reform.