Appendix F

Survey Methodology

Telephone Surveys

In order to assess public attitudes and perceptions of current LRSD issues and policies, the University Task Force enlisted the help of UALR's Institute of Government (IOG) to design and administer two telephone surveys of citizens in Little Rock. IOG and its predecessor campus units have 15 years of experience in field research of this kind. The initial telephone survey was designed to ascertain responses from adult citizens throughout Little Rock. The second telephone survey was designed to achieve responses from adult African-American Little Rock residents.

IOG's responsibilities included questionnaire development, administration of the survey, and primary statistical analysis. The questionnaires were developed by IOG staff in collaboration with the University Task Force on the LRSD. Some of the questions were adapted from the "Annual Gallup Poll of the Public's Attitudes Toward the Public Schools" published in the Phi Delta Kappan, a national education journal.

IOG conducted 800 telephone interviews between March 4, 1996 to April 4, 1996 with adult residents of Little Rock. (The survey is styled as "LRSD Telephone 1 Survey" in the narrative of this report.) The sampling design required that a total of 400 interviews be obtained from Little Rock wards one, two, and six, and an additional 400 interviews obtained from Wards three, four, five, and seven, collectively. Thus the survey was designed to capture a maximum number of African-American responses without screening and stratifying by race during the telephone interview.

Completion of the 800 telephone interviews yielded surveys from 564 white households and 211 African-American households. This number of African-American households, in the judgment of the University Task Force, did not provide a level of reliability with which to make conclusive research findings about African-American opinion. A second telephone survey was therefore organized to re-sample African-American opinion with the goal of completing at least 400 household surveys. (The second survey is styled as "LRSD Telephone 2 Survey" in the report narrative.)

The second survey of African-American households was conducted between April 26, 1996 and May 11, 1996. The sampling frame consisted of phone numbers from targeted census block groups, the total of which represented 50 percent of the African-American population of Little Rock. The tracts were identified by listing the census block groups of the city in descending order according to the percentage of African-American population in that block group. Ultimately, a phone number was selected if it was identified with a block group having at least 47 percent African-American population.

Sampling Error and Confidence Level

The samples of at least 400 households each for African-Americans and whites in the city provide a sample error of +/-4.9 percent with a 95 percent confidence level. These statistics can be very useful to a reader of these surveys. For example, if a survey showed that 80 percent of the sample indicated overall satisfaction with LRSD academic policy, we can be 95 percent confident that the actual figure for the entire population (if the LRSD population could be directly questioned rather than sampled) would be between 75.1 percent and 84.9 percent.

When survey findings from African-American and white households are combined (as in the LRSD Telephone 1 Survey) and reported as a single statistic, the sampling error is lowered to +3.5 percent with a 95 percent confidence level.

Mail Survey

The University Task Force on the LRSD sought additional opinions from the teachers of the school district. A mail questionnaire was developed by IOG and the University Task Force to obtain teacher viewpoints. The questionnaire included topics such as discipline, safety, parental involvement, desegregation policies, and academic achievement.

IOG was charged with mailing surveys to the teachers and principals of the Little Rock School District. The mailing included all teachers listed in the Little Rock School Directory. A cover letter encouraging their participation was included from the University Task Force on the LRSD. A follow-up postcard was mailed to every household on the initial mailing list. The postcard encouraged response from those who had not yet participated. The postcards were necessarily sent to all listings because no tracking was conducted of individual respondents. Tracking was not done to assure anonymity of respondents.

Completed questionnaires numbered 467. With approximately 1,660 questionnaires originally mailed to LRSD faculty, the completion rate was approximately 28 percent.

Demographic Characteristics of the Sample Surveys

Fundamental to the interpretation of opinion surveys is knowing the characteristics of persons in the sample and how well these persons collectively match the profile of the entire population, in this case all persons living in the LRSD. The two tables shown below summarize those characteristics that are believed to be most important in evaluating the representativeness of persons and teachers in the samples.

Table F-1 shows that African Americans and whites in the telephone sample were somewhat better educated and have slightly higher household incomes than comparative statistics reported by the 1990 Census. The telephone surveys were also heavily weighted with female respondents, a familiar problem in most telephone opinion surveys. Women apparently answer the telephone more frequently and are more inclined to participate in such surveys than are men.

Table F-2 shows that the sample of teachers closely reflects the gender and grade level assignment of the LRSD faculty. However, the mail sample was influenced by proportionately more responses from magnet school teachers and proportionately less by area school faculty. In addition, the mail survey of teachers did not capture a fully representative share of African-American faculty.
Table F-1
Table F-2

Table of Contents