Fall 2008

The upcoming Presidential election brings an interesting end to the year here at ODC. At no time has the intersection of our core public policy, social marketing and energy research strengths been so cohesively apparent.

The election has forced Americans to think hard about such issues as energy independence, renewable energy, energy efficiency, and the environment-all issues that ODC analysts think about every day. ODC staff are now more interesting to talk to at a cocktail party!

Our work shows strong correlations between voter preference and likelihood to adopt energy efficiency measures-for example, Obama supporters are more likely to recall energy efficiency messaging than McCain supporters.

On the corporate front, ODC continues to grow. This month we are moving into expanded office space in our Oakland, CA, satellite office. We know that our Founder, John Gorman, would be proud of what we have accomplished this year. Living up to his legacy ensures our current and future commitment to exceeding client expectations.

-Brad Kates

Cell Phones and Survey Research
by: Chris Anderson

Unease has been spreading through the survey research industry for years now over the increasing number of households cancelling their landlines and using cell phones exclusively. Traditional random digit dial (RDD) sampling methodology, which randomly generates telephone numbers and has been the gold standard for conducting accurate survey research for decades, does not include cell phone numbers. Recently released data from the National Health Interview Survey (NHIS) estimates that 15.8% of households are wireless only and an additional 15% are wireless mostly, meaning that traditional RDD samples could have coverage bias of up to 30%.

Some have speculated that the trend towards wireless only would eventually lead to the demise of telephone surveys as an effective methodology for generating highly accurate and timely estimates of public opinion. Fortunately for organizations that rely on telephone surveys for important market intelligence (not to mention the organizations that provide these services), two recent developments enable researchers to conduct telephone surveys without fear the results will be biased due to the exclusion of wireless only households.

First, we know more about what types of households are most likely to be wireless only and we are learning more about how these households differ from those with landlines. This knowledge enables researchers to account for wireless only households via sample stratification and weighting.

Second, it is now possible to purchase RDD samples of cell phone numbers and to develop sampling plans which incorporate cell phone and landline numbers.

While including cell phone samples in a telephone survey is usually the most effective way to ensure there is no wireless only bias, it is not always justified from a cost perspective. It is roughly four times more difficult, and therefore significantly more expensive, to complete an interview with individuals who are wireless only than with landline households.

Several factors contribute to making wireless only interviews more difficult to complete:

• Respondents are more likely to be otherwise engaged (shopping, driving, etc.) and unable or unwilling to complete the survey
• Calls are dropped due to wireless coverage
• The necessity to screen for wireless only households (the cell phone sample needs to be winnowed to those who only have cell phones)

Determinations about addressing potential wireless only household biases need to be made on a case-by-case basis. A critical situational analysis should be conducted regarding the demographics of the research population and the topic being researched.

Thanks to the National Health Interview in-person household survey, we know that wireless only users are much more likely to be younger, lower income, Hispanic, and living with unrelated adults (roommates). If research is intended to specifically understand the attitudes or behavior of one of these groups, then including a cell phone sample may be warranted. If not, sample stratification or weighting of the results will often be adequate to ensure the results are representative.

It is also crucial to evaluate the research topic and the potential interplay with cell phone usage. For example, it would be important to include a wireless only sample for a study regarding technology usage, but less important in a study of consumer preferences for ice cream flavors.

Overall, attitudes in the market research industry regarding wireless only households have shifted from concern to excitement. Concrete strategies have been developed for dealing with households without landlines so researchers can evaluate a situation and develop an appropriate cost effective methodology.

Social Network Analysis

Social network analysis is a technique used to study relationships between actors. It can be used in program evaluation to assess how well organizations are working together, by mass media campaigns to improve the diffusion speed of innovations and by companies to strengthen their internal networks. The actors in social network analysis can include individuals, organizations, or even nations. Social network analysis differs from other methods in that it is used to analyze relational data (data on the connections or ties between actors) rather than attribute data (data on the individual qualities of an actor, such as age or race). Some examples of connections between actors are: sitting on the board of the same company, attending the same university, communicating daily, and contributing money to the same political group.

Social network analysis is used to assess the structure of networks. The focus of analysis can be on the individual actors in the network and/or the network as a whole. It is possible, for instance, to determine the 'important' actors in a network. This concept is known as centrality. For instance, in terrorist networks it is obviously useful to identify the leaders of the network-centrality scores can indicate this.

Social network analysis can also be used to assess the overall connectedness of a network. The density of a network (the number of ties present in a network divided by the number of possible ties) measures this. In an evaluation of California's Statewide Marketing and Outreach Effort (SWM&O) social marketing campaign, ODC calculated the density of the network of organizations involved with implementing the SWM&O effort. The density was interpreted as an indicator of the level of cooperation between the organizations, and was compared to the density of another network of organizations running social programs.

Social network analysis can also be used to map and visualize networks. The diagram below shows the network of organizations involved in the SWM&O effort. The thickness and color of each line indicates the frequency of contact between the organizations. Organizations "never" in contact are not connected. Based on this map, it was possible to identify organizations that should have been contact but weren't.

Caroline Wilson, a Senior Analyst with the Oakland office, leads ODC's social network analysis efforts. Caroline has presented on social network analysis at several recent industry events, most recently a paper entitled "The Social Capital of the Chief Financial Officer and Career Advancement: A Study of FTSE 100 Companies" at the 2008 Academy of Management (AOM) conference in Anaheim, California. This study focused on helping a search firm improve their recruitment and talent identification processes for senior level executives. The study pioneered a new definition of social capital, based on connections through interlocking directorates, PR advisors, corporate brokers, and auditors. It then used statistics to determine the role of social capital in the career advancement of the directors, and to predict which CFOs will go on to become the next CEOs.

Social network analysis is just one of a number of cutting-edge methods offered by ODC to provide actionable recommendations to our clients. For more information on this topic, contact Caroline Wilson

Independents' Day: The Political Color Purple
by: Ernie Paicopolos

Over the last eight years, there's been a lot of talk about voters in blue (Democratic) states and red (Republican) states and their relative impact on the presidential election process. But increasingly it's voters with a blended, purple hue that have come to hold the real key to determining the occupant of 1600 Pennsylvania Avenue. These are the Independents-neither fully red nor fully blue. And while they have always been a force in presidential politics, their impact could be even bigger in 2008. Why? Because of their growing importance in key states and sharp volatility.

Independent voters will likely have a disproportionate impact in a handful of swing states like Colorado and New Hampshire. While Democrats and Republicans have shown predictable stability in their views, Independent voters have shown wild swings in sentiment over the long two year campaign season.

Since both major party candidates can reasonably lay claim to this critical constituency, tracking the attitudes of Independent voters becomes all the more interesting. Obama's early primary and caucus wins were heavily dependent on drawing new voters into the process, many without traditional partisan ties. McCain has built his resume on being the maverick who occasionally bucks his party line; a reputation that has always drawn independent support. More so than in most recent elections, these voters are truly up for grabs.

To underscore the volatility point, our latest polling data (from our late September poll) shows that Barack Obama was slightly ahead among unaligned voters-a dramatic shift from earlier in the month when John McCain held a 15-point edge. By contrast, Republican and Democrats have been holding fairly steady in their partisan support.

Exactly who are these voters who make up fully one-fifth of the voter population? While their demographic composition tends to vary from state to state, nationally they have a somewhat greater tendency to be:

• Male;
• Living in urban settings;
• Middle-aged; and
• Self-described political moderates

Of course, it's not just about who they are and how they stack up in the horse race. We also have to look at how Independent voters break on some of the major issues in the campaign. Both campaigns have tried to structure the race to match the narrative that benefits their candidate the most. As the race moves into the final weeks, it is interesting to see how both tickets try to capture the change mantle-Obama by continuing his campaign-long emphasis on this issue and McCain by using his running mate to bolster the perception that his is the reform team that can change Washington. Regardless of how the campaigns want to characterize the race, the mid-September crisis in the nation's financial markets, and the subsequent government relief efforts, pushed economic concerns to the top of the electoral ladder.

As important as issues are to the race, there are also subtle psychosocial elements that come into play when voters make their choice for president. When entering the voting booth, people carry with them an array of conscious and unconscious images about the candidates and the campaigns. Our late September poll asked voters some questions that try to serve as attitudinal placeholders for these underlying elements. For example, Americans like their presidents to be wise and expect them to offer sage advice on where the country should be heading. When we asked voters who they would trust on their own personal difficult life decisions, they were just about equally divided; and Independents, while slightly favoring Obama on this score, are much more indecisive about which candidate to seek advice from. Again, Democrats and Republicans are highly predictable in their views on this issue.

Trust is another critical factor in a person's calculation of who to vote for. If Americans detect any evidence that a candidate can't be taken at his or her word, it can effectively end their chances. Once again, overall attitudes on this score are just about evenly divided between Obama and McCain by a 5-point margin, Independent voters tend to trust Obama more. Perhaps more telling is the fact that nearly twice as many Independents are undecided on this score.

So, what does all this evidence suggest for the outcome of the November 4th election? A lot will have to do with how the narrative of the campaign plays out. If economic concerns (fueled by the crisis in the financial markets) dominate voters' minds, history tells us that the party out of power-in this case the Democrats-should hold an advantage among these swing Independent voters. If a national security crisis erupts, Republicans will gain some advantage among Independents. As the race stands in late September (when this newsletter went to press), economic concerns were clearly dominant. But if this race has shown us anything, it's that the conventional wisdom is neither conventional nor wise.

Traditionally, Independent voters are a little more likely than others to be swayed by the non-partisan psychosocial factors discussed earlier. If either of the candidates can capture those intangibles, it just might negate a seeming advantage on more substantive issues. Whatever happens on Election Day, it seems clear that the voters who sport the color purple are even more likely than usual to decide who will become the next president. As a consequence, the electoral map will be shaded a little differently.

ODC in Brief

ODC's Oakland office is growing, and we are doubling our office space to facilitate this growth. We remain in our 1999 Harrison Street location, and our new address is Suite 650.

Senior Analyst Caroline Wilson recently presented a paper "The Social Capital of the CFO and Career Advancement: A Study of FTSE 100 Companies" at the 2008 Academy of Management conference.

Project Manager Anne Dougherty presented a poster titled "Innovative Social Marketing Research Techniques" at the World Social Marketing Conference in September 2008. Her presentation can be found at www.opiniondynamics.com.

Director Mary Sutter recently achieved Certificates in Evaluation Practice and Quantitative Evaluation methods from The Evaluator's Institute (TEI) a highly respected organization which resides within George Washington University's Trachtenberg School of Public Policy and Public Administration and the Midge Smith Center for Evaluation Effectiveness.

Upcoming Events:

• ODC Vice President Chris Anderson will be in the FOX News studio on Election Day working on the network's Election Night Decision Team.
  
  
• ODC is a conference ally sponsor of the 2008 Behavior, Energy and Climate Change Conference being held November 16-19 in Sacramento, CA. The conference brings together over 700 energy industry leaders researching social and behavioral issues to address climate change challenges. ODC's Vice President of Marketing, Sharyn Barata is leading the opening plenary session and serving as a moderator. ODC's Anne Dougherty will also be speaking.
  
  
• ODC is a sponsor of the Association of Energy Service Professionals' National Energy Services Conference to be held January 27-29 in San Diego. ODC will have a booth at the event-and look for the conference agenda for ODC speakers.