How is the Energy Survey conducted?
The U-M Energy Survey is conducted by telephone as a quarterly rider (add-on) to the University of Michigan’s long-running Surveys of Consumers. A representative random sample of 500 U.S. households in the lower 48 states are interviewed over the phone by consumer survey professionals at the university’s Survey Research Center in Ann Arbor. The interviews are conducted over one-month periods, every month for the general Surveys of Consumers and every third month (January, April, July and October) for the Energy Survey. See our Methodology page for further details.
What questions do you ask?
The Energy Survey asks a series of 18 questions, which are listed in our Questionnaire.
Why do you use a four-level scale for many of the questions?
The scale we use for a number of our questions, including those that ask how worried consumers are about various aspects of energy, is a variant of what is known as a Likert Scale, named after its developer, Rensis Likert. Professor Likert co-founded the University of Michigan’s Survey Research Center (which carries out our survey) and was the first director of the university’s Institute for Social Research (ISR; see the ISR timeline for historical background). Likert scales are widely used in opinion research to measure the extent to which respondents agree or disagree with a statement in a survey.
When defining a scale to probe consumers’ degrees of concern, we first needed to decide whether to use a bipolar (“black and white”) or unipolar (“shades of grey”) approach. Our cognitive interviews found that a bipolar scale confused respondents, who had difficulty seeing the choices of, say, their degree of concern about energy affordability, in such black-and-white terms. We therefore adopted a unipolar scale to assess consumer views along a continuum of concern.
A standard unipolar Likert scale involves five gradations, with a midpoint such as “neither agree nor disagree.” However, for assessing many aspects of energy-related concern, a middle option (such as “neither affordable nor unaffordable”) is not very meaningful. We therefore adopted a balanced four-point scale without a neutral midpoint.
What are the margins of error?
For a single survey, as conducted quarterly with a sample size of 500 household, the margin of error is 4 to 5 percentage points for most questions. This sampling error is for two standard deviations, i.e., for a 95% confidence interval. As the sample size increases, the sampling error declines. Therefore, for results based on a year (four quarters) of survey data — which provides a sample size of 2000 households — the margin of error is cut in half.
The margin of error depends on the percentage of respondents giving a given answer, and so also varies with the percentage response. For a detailed looks at how the sampling error varies depending on the percentage response and sample size, see this table [PDF].
Why don’t consumers seem to be more concerned about energy reliability?
That’s a good question. When our results first started coming in and reliability lagged both affordability and impact on the environment as areas of consumers’ energy-related concern, we were a bit surprised. Most energy professionals rate reliability as a great concern, and when we previewed our results with electric utility experts, they were also surprised by the relatively low degree of concern that consumers reported in our survey. Utility companies and policy makers who deal with energy certainly hear about it from consumers when there’s a power failure. And of course, they give a high priority to ensuring reliable service. As a result, we actually have a very reliable energy supply system in this country.
For most consumers most of the time, they can count on their power being on and can count on fuel being available at a gasoline station when they need it. We do see a relatively higher degree of concern about reliability from lower income consumers as well as from renters and those in the lower bracket of home value. So their experience of reliability is not as good as that of consumers in the higher income or home value brackets. Nevertheless, we surmise that the relatively low average level of concern about reliability is due to the fact that, by and large, Americans enjoy a highly reliable energy system.
Why don’t you ask about energy security?
When we conducted the focus groups leading up to the design of the U-M Energy Survey, we found that most people did not understand what was meant by energy security without quite a lot of explanation. Professionally, when we use the term energy security, we are often referring to risk of energy supply disruption for geopolitical reasons. The term is commonly used in the context of concerns about oil imports from unstable regions or hostile countries. But even among national security and energy experts, the term takes on different meanings in different contexts.
For many of the individuals we interviewed during focus groups, when we brought up energy security, they thought we referring to reliability (that is, whether or not they could count on their power staying on). On their own, focus group respondents did not relate the term “energy security” to the meaning it has for energy policy experts or others who closely follow energy issues. So we decided it was best to not ask about this issue, which would have required a longer (and potentially leading) question. See our post on how many consumers are confused about topics that are well understood by energy industry and policy professionals.