ORANGE COUNTY ANNUAL SURVEY

METHODS

The Orange County Annual survey was directed by Mark Baldassare, a professor of social ecology at UC 
Irvine. For the survey, 1,085 adult orange County residents were interviewed by telephone Sept. 6 to 23. 
In Orange County, where more than 97 percent of households have telephones, this method of interview 
gives highly representative data.

Interviewing was conducted on weekend days and weekday nights, using a random sample of 4,500 listed and 
unlisted telephone numbers. These were generated by computer from a list of working blocks of telephone 
exchanges. The telephone sample was generated by Pijacki and Associates of Shoreham, N.Y. The field work 
was conducted at the Center for Survey Research by UCI's Public Policy Research Organization.

Of the telephone numbers called, 23 percent resulted in completed interviews and 15 percent were 
refusals. The completion rate for the survey (completions divided by
completions plus refusals) was 61 percent.

Other telephone outcomes included the following: 21 percent disconnected numbers; 3 percent computer or 
fax lines; 15 percent businesses and other non-Orange County households; 20 percent persistent no answers 
and l percent persistently unavailable respondents. Two percent were not completed because of language 
problems, including non-English speaking households, and hearing impairment.'

Within a-household, respondents were chosen for interview using the Troldahl-Carter method. This method 
randomly selects a household member from a grid that includes information on the number of adult 
household members and the number of adult men in the household. Up to six callbacks were attempted per 
telephone number.

Each interview included 90 questions and took an average of 20 minutes to complete. Most interviews 
ranged in length from 15 to 25 minutes.

The surveys were designed in three stages over several months. In the first stage, UCI undergraduate 
students conducted face-to-face, interviews on orange County topics with randomly selected adult 
residents. The second stage involved feedback on questions and topics from the annual survey's Steering 
Committee, Advisory Committee and colleagues. The final stage included pre-tests by students, followed by 
final revisions of the questions.

The interview began with questions about housing, moving preferences, consumer confidence and perceptions 
of life in orange County. These were followed by questions on growth, transportation and crime issues. A 
major section of the interview was then devoted to questions about air pollution and the Air Quality 
Management Plan. Later in the interview, we turned to the topics of charities. The conclusion of the 
survey was devoted to questions about work and commuting patterns, personal characteristics, household 
status and political attitudes.

The survey's validity was checked by comparing the sample's characteristics to available information on 
Orange County's population. We compared the 1989 survey results to the 1980 U.S. Census, previous annual 
surveys and other recent survey data. Age, income and other demographic features of our sample were 
comparable with those noted in other studies.

For data analyses, we statistically weighted the sample to represent the actual regional distribution of 
Orange County residents. The 1989 population estimates for north, west, central and south county regions 
were issued by the Demographic Research Unit, County of Orange.

Other efforts were made to correct for possible errors in the course of interviewing and data processing. 
Approximately 10 percent of the completed interviews were verified through callbacks. All questionnaires 
were checked by the interviewer supervisor immediately after completion. Finally, keypunched data were 
double-checked for all cases in the survey sample.

The sampling error for this survey is +/- 3 percent at the 95 percent confidence level. This means that 
95 times out of 100, the results will-be within 3 percentage points
of what they would be if all adults in Orange County were interviewed. The sampling error for any 
subgroup would be larger.

Sampling error is just one type of error to which surveys are subject. Results may also influenced by 
factors such as question wording, survey timing and other aspects of survey design.