What Type of Survey is Best?

Carolyn Browne*
Principal
Carolyn Browne Associates

So, someone in your office says, "We should do a survey."

First, there should be clarity on the survey goals. A survey is a snapshot of public awareness levels and attitudes in the present moment.

Random Sample - Public Opinion/Attitude Survey

When an agency needs reliable, projectable data about the attitudes and opinions of its citizens, or a select group of its citizens, it is essential to conduct a valid, random sample survey. Telephone interview surveys are considerably more common than in-person interviews because they are far less expensive to conduct and tend to be widely accepted as an information-gathering tool. There is a margin of error, based upon the size of the sample (generally, a minimum sample of 200 is the industry standard for reliable data about any population segment). Overall, random sample telephone interview surveys provide reasonably accurate information about the population from which the sample is drawn.

While there is a statistical margin of error (the sample of 200 provides an error range of +/- 7% with a 95% confidence), this type of survey is the most democratic process there is, and the most reliable, for learning about the opinions of an entire community.

A random sample survey is not appropriate for educating people about an issue or trying to assess what people will do at some future point (i.e., "Will you vote for this bond issue?"). But, the results do provide a reasonably accurate portrait of the person's opinions in the present moment (i.e., a person's feelings or attitudes about the issues relating to the need for approving a bond). Questions asked in the past and present tense provide a reasonable degree of accuracy about a person's usage and habit patterns.

Self-Selected Survey - Newspapers, Mail, Internet and Written Questionnaires

When an agency has a political need to create a survey process that allows anyone who is interested to respond, it can do a self-selected process. A written survey can be distributed in public locations, such as the City Hall or Library, mailed directly, e-mailed or published in the city newsletter or the local newspaper.

When reporting data from a self-selected survey, it is important to begin with the understanding and the language, "Of those who chose to respond….." Most often, those who volunteer to respond to a self-selected survey have a strong opinion (frequently negative) about the issue being discussed.

A self-selected survey, however, can be an excellent public relations tool and a good way of giving information to the public. But, extreme caution must be exercised in drawing any conclusion about what the public, in general, thinks based upon the results from a survey when the respondents are volunteers.

*Carolyn Browne is principal of Carolyn Browne Associates, a successful community involvement and marketing research firm in the Seattle Area since 1977. Carolyn's firm has handled numerous community involvement projects, marketing research surveys, focus groups, public workshops and strategic planning meetings for a broad range of large and small municipalities and for several state agencies. She is an associate member of Association of Washington Cities and is a frequent presenter and workshop leader at their conferences. Carolyn can be contacted at: chbrowne@earthlink.net.