EXPERIENCE AND METHODOLOGY-ST. CLOUD TAX SURVEY
[last revised 5/17/2000]
I.
EXPERIENCE
A. Historical Background of the SCSU Survey
The
SCSU Survey is an ongoing survey research arm of the Social Science Research
Institute, College of Social Sciences, St. Cloud State University. Dr. Stephen Frank began the Survey in 1980
conducting several omnibus surveys a year, one of central Minnesota adults and
one of SCSU students in conjunction with his Political Science classes. Throughout
the year, the SCSU Survey conducts contract surveys. Past clients have included the Great River Regional Library, city
of St. Cloud, Allina Health Systems and the St. Cloud Times.
The
omnibus surveys are now done once or twice a year and have primarily a
statewide focus. Dr. Steven Wagner
joined The Survey in 1997 and serves as co-director along with Dr. Frank. Dr. Michelle Kukoleca Hammes joined the
Survey in the summer of 1999 and serves as co-director with Drs. Frank and
Wagner. Dr. Steve Frank, SCSU Professor
of Political Science, Dr. Steven Wagner, SCSU Associate Professor of Public
Administration and Dr. Michelle Kukoleca Hammes, SCSU Assistant Professor of
Political Science.
About
six upper level students serve as senior supervising student directors for the
Survey. These individuals review
questionnaires, field test questionnaires and supervise interviewing. The College of Social Science provides
technical support. Faculty from
departments within the College of Social Sciences and other applicable Colleges
occasionally are used to provide expertise to particular projects.
The
Survey conducts a variety of client-based surveys every year. Recent clients include Allina Health System,
Minnesota Planning, Minnesota State Lottery Commission, St. Cloud Area United
Way, Southwest State University, Great River Regional Library, St. Cloud Times,
National Parks and Conservation Association, and local governments and school
districts.
B.
Project Directors
Primary project personnel consist of the three
faculty directors of The SCSU Survey.
They are:
Dr. Stephen I. Frank
Dr. Steven C. Wagner
Dr. Michelle K. Hammes
All three directors are members of the American and
Midwest Association of Public Opinion Research. Below are brief professional vitas for Drs. Frank, Wager and
Kukoleca Hammes.
1. Stephen
I. Frank
Dr. Frank holds a Doctor of
Philosophy in Political Science from Washington State University. Dr. Frank teaches courses in American
politics, public opinion and research methods at St. Cloud State
University. Dr. Frank started the SCSU
Survey in 1980 and has played a major role in the development, administration
and analysis of over 150 telephone surveys for local and state governments,
school districts and a variety of nonprofit agencies. Dr. Frank has completed extensive postgraduate work in survey
research at the University of Michigan.
Dr. Frank recently co-authored with Dr. Wagner and published by Harcourt
College, “We Shocked the World!” A Case
Study of Jesse Ventura’s Election as Governor of Minnesota.
2.
Steven C. Wagner
Dr. Wagner holds a Doctor of
Philosophy in Political Science and a Master of Public Administration from
Northern Illinois University. Dr.
Wagner earned his Bachelor of Science in political science from Illinois State
University. Dr. Wagner teaches courses
in American politics and public and nonprofit management at St. Cloud State
University. Dr. Wagner joined the SCSU
Survey in 1997. Prior to SCSU, Dr.
Wagner taught in Kansas where he engaged in community-based survey research and
prior to that was employed by the U.S. General Accounting Office as a staff
researcher. While teaching in Kansas,
Dr. Wagner published two articles relevant to K-12 education, one on the “Kids
Voting in Kansas” initiative and the second on proposed federal changes in
school funding. Both articles appeared
in the Kansas School Board Journal.
3.
Michelle K. Hammes
Dr. Kukoleca Hammes holds a Doctor of Philosophy in
Political Science and a Masters in Political Science from the State University
of New York at Binghamton. Dr. Kukoleca
Hammes earned her Bachelor of Arts in Political Science in Political Science
from Niagara University. Prior to
SCSU, Dr. Kukoleca Hammes taught at The University of Pittsburgh Bradford,
Pennsylvania campus. Dr. Kukoleca
Hammes is a comparativist with an area focus on North America and Western
Europe. Her substantive focus is
representative governmental institutions.
She teaches courses in American Government, Introduction to Ideas and
Institutions, Western European Politics, and a Capstone in Political Science at
St. Cloud State University. Dr.
Kukoleca Hammes has recently joined the survey team and will be using her extensive
graduate school training in political methodology to aid in questionnaire
construction and results analysis.
B.
Relevant Research Experience
The SCSU Survey has a substantial history of relevant
research experience. As noted above,
The SCSU Survey has conducted telephone-based public opinion survey research
for almost 20 years. Today, The SCSU
Survey specializes in state and local telephone survey research. Throughout the past 20 years, The SCSU
Survey has conducted many studies similar to this study of the City of St.
Cloud. In the fall of 1998, the SCSU
Survey completed a study of residents in the greater St. Cloud area to
determine use of the local public library and the likelihood they would support
a tax to construct a new public library facility. In December 1998, the Survey additionally conducted a study of
residents in the greater St. Cloud area to determine how they might use a
multiple purpose events center, where they would prefer it constructed and
their willingness to pay additional sales tax to support the construction of the
facility. We found that about 59% of
area residents would support an addition to their sales tax for the
library. In regards to the proposed
Events Center, we found that only 42% supported a tax for the center. The taxpayers have now voted on a designated
tax to fund construction of a new library, acquisition of parklands, and the
construction of an Events Center. The
referendum (November 1999) failed by a vote of 53% to 47%. In addition, the SCSU Survey has conducted
surveys for the St. Cloud school district on issues pertaining to services and
bond referendum.
The
SCSU Survey maintains a computerized assisted telephone interview (CATI)
system, which not only allows efficient execution of a survey but also allows
daily analysis of completed interviews and almost immediate downloading of data
once interviewing is finished. For
surveys that require random selection of respondents, the SCSU Survey purchases
samples from Survey Sampling, Inc. of Fairfield, Connecticut.
In
addition to the SCSU Survey directors/principal investigators, approximately 20
SCSU students are employed to conduct the actual interviews. Approximately six senior students, who
assist the three SCSU Survey faculty directors/principal investigators, pretest
questionnaires and one or two of the directors supervise the actual
interviewing.
Mr.
Justin Wedeking served as senior supervising student director of the Computer
Assisted Telephone Interview (CATI) lab and Mr. Joe LeDuc provides technical
support for the computer network.
II. METHODOLOGY
About
20 SCSU students conducted the actual interviews. All callers were screened and received about five hours of
training. All calls were made from the
SCSU Computer Assisted Telephone Interviewing Survey Research Laboratory (CATI Lab). The survey software is developed by Sawtooth
Software. Mr. Justin Wedeking and Drs.
Frank, Wagner and Kukoleca Hammes monitored the callers.
Several
steps were taken to ensure that the telephone sample of St. Cloud residents was
representative of the city’s population.
The sample was drawn proportional to the population of the city of St.
Cloud. The sample was drawn by zip
code. Since the zip code for St. Cloud
extend out of St. Cloud’s civil boundaries, we screened out respondents who
indicated they do not live in St. Cloud [See Table 1]. The sample was developed by Survey Sampling,
Inc., of Fairfield, Connecticut, one of the largest and most prestigious
sampling firms in the country.
The
respondents matched known characteristics of the city. Survey Sampling, Inc. used a random digit dialed (RDD)
sample. A RDD sample ensures telephone
numbers not listed in telephone directories because they are new or unlisted
are included in the sample. A RDD
sample is a computer-generated list of telephone numbers. We find this a better way to sample than to
draw from a known list of numbers, either a local telephone book or voter
registration lists.
We
have found Survey Sampling, Inc. a particularly efficient sample production
company. They generate a sample that is
of very high quality because they:
n
construct a comprehensive database of all telephone working blocks
which actually represent residential telephones;
n
obtain, update and cross check working block information from the local
(U.S. West) telephone company;
n
confirm the estimated number of residential telephones with each
working block, excluding sparsely populated working blocks (industry standard
is to exclude those blocks with less than three known working residential
telephones out of the 100 possible numbers);
n
assign working blocks known to contain residential telephones to
geographic areas bases on zip code and most recent updates of census data;
n
mark each working block for demographic targeting;
n
check each RDD number against a list of known business telephone
numbers and generate new numbers as necessary; and,
n
arrange the ending sample in a random order to eliminate potential
calling order bias.
The
SCSU Survey has strict guidelines and procedures for sample management during
data collection. To assist in this
effort, the SCSU Survey uses a state-of-the-art Computer Assisted Telephone
Interview (CATI) system - Sawtooth Software’s Ci3 WinCATI. The benefits of Ci3 WinCATI are many.
1.
All interview stations are networked for complete, ongoing sample
management.
2.
Data is updated immediately, ensuring maximum data integrity and
allowing clients to get progress reports anytime. Data is reviewed for quality and consistency.
3.
Answers are entered directly into the computer. Keypunching is eliminated, thus decreasing
human error. Data analysis can start
immediately.
4.
The computer handles call record keeping automatically, allowing
interviewers and supervisors to focus on the interviewing task.
5.
Callbacks are handled by the computer and made on a schedule. We call each number ten times. Interrupted surveys are easily
completed. Persons who are willing to
be interviewed can do so when it is convenient to them, improving the quality
of their responses.
6.
Calls are made at various times during the week (Monday through
Thursday, 4:30 to 9:30) and on weekends (Sunday, 3:00 to 9:00) to maximize
contacts and ensure equal opportunities to respond among various demographic
groups.
7.
CATI maintains full and detailed records, including the number of
attempts made to each number and the disposition of each attempt.
In
order to reach hard-to-get respondents, each number was called up to twelve
times over different days and times and appointments made as necessary to
interview the designated respondent at her/his convenience. Most calls were made after 4:30 pm weekdays
and 4:30 pm on Sunday. Calls were made
between Monday, April 24 and Wednesday, May 3.
The
sample consists of 612 completed interviews.
In samples of 612 interviews the sample error due to sampling and other
random effects is approximately plus or minus 4.5% at the 95% level of
confidence. This means that if one were
to have drawn 20 samples of the city and administered the same instrument it
would be expected that the overall findings would be greater or lesser than
4.5% only one time in 20.
However,
in all sample surveys there are other possible sources of error for which
precise estimates cannot be calculated.
These include interviewer error, respondent misinterpretation, and
analysis errors. When analysis is made
of sub-samples such as gender the sample error may be larger.
The
cooperation rate of the survey is 70%.
This is several percentage points above the average for professional
marketing firms. Cooperation rate means
that once an eligible individual was reached almost nine of ten of the
respondents agreed to participate in the survey. Cooperation rate is calculated by adding the number of completed
calls to the number of refusals, and dividing the number of completed calls by
the sum of completed and refusal calls.
The disposition of all calls made during the course of the survey is
included below.
The
total survey consists of 25 variables.
Additional
material on the survey’s methodology and findings are available by contacting
Dr. Stephen Frank, Dr. Steven Wagner or Dr. Michelle Kukoleca Hammes at St.
Cloud State University.
|
Table
1: Call
Breakdown |
|
|
Description |
Record |
|
Completed Calls[1] |
854 |
|
Busy |
51 |
|
Not Working Numbers |
646 |
|
Not St. Cloud Resident[2] |
179 |
|
Not Minnesota Resident |
2 |
|
Callbacks[3] |
36 |
|
Refusals[4] |
362 |
|
Answering Machine[5] |
826 |
|
Business Phone |
347 |
|
No Answers[6] |
589 |
|
Fax/Modem |
159 |
|
Call Blocking |
1 |
|
Language or hearing problem, other |
29 |
|
Sick or In Hospital |
6 |
|
Out of Town/ Not Eligible[7] |
25 |
|
Cabin Phone |
1 |
|
Partial Completion |
1 |
|
See Director- Not Eligible |
20 |
Total
Attempted Dialings
|
4134 |
Total
Completed Surveys
|
612 |
