CATI and CAPI Technology
The Office for Survey Research at Michigan State University's
Institute for Public Policy and Social Research primary
focus is on conducting surveys by telephone using its state-of-the-art
Computer-Assisted Telephone Interviewing (CATI) facility, which
houses 40 CATI workstations and interfaces with CAPI (Computer-Assisted
Personal Interview) use. This technology enhances interview capabilities,
as well as unobtrusive monitoring resources.
Telephone Survey Procedures
Interviewers work from scripted introductions, respondent selection
procedures, and questionnaires programmed into CATI and displayed
at their work stations. They read the questions displayed on their
monitors to respondents over the phone and directly record the answers
(whether pre-coded or verbatim) into the CATI computer data file.
System Software
OSR's CATI system uses CASES (Computer Assisted Survey Execution
System) software, developed at the University of California, Berkeley.
CASES is an advanced CATI software System that can be easily adapted
for CAPI (Computer Assisted Personal Interview) use. It also has
special program capabilities for DDE (Direct Data Entry) and SAQ
(Self-Administered Questionnaire) applications. The CASES system
is leased annually and includes full technical support of the U.C.
- Berkeley development team. Currently, OSR uses CASES 4.3 primarily.
CASES software facilitates fully programmable interviews. The software
integrates both closed-ended questions and open-ended questions.
For closed-ended items, the interview programmer can specify the
valid response codes for each item, such as:
<1> YES
<5> NO
<8> DON'T KNOW
<9> REFUSED/NO ANSWER
or the range of valid response codes for an item, such as:
<1-25> CIGARETTES PER DAY
<98> DON'T KNOW
<99> REFUSED/NO ANSWER
 Technology
Enhances Interview Capabilities
The software allows interviewers to record notes along with responses
to closed questions. To do so, the interviewer invokes the command
mode, indicates the wish to make a note with the command "n,"
and immediately begins typing the text of the respondent's comments
verbatim or records information about digressions, inquiries, probes,
or clarifications. Notes of great length (i.e., multiple pages)
can be entered if needed, although short notes are much more typical.
The software flags each item with notes attached for easy review
or checking for recoding purposes.
The software readily enables interviewers to immediately record
responses to open-ended questions, as well as "OTHER (SPECIFY)"
responses. The programmer simply defines the "other (specify)"
response in such a way that it automatically slips into text entry
mode and prompts the interviewer to record text. Any or all interviewer
notes and typed responses to open-ended questions
can be output and sorted by respondent or by question number for
review.
By default, the software moves directly from one item to the next
in sequence unless specific program commands direct the execution
path elsewhere. Different skip commands can be associated with separate
responses to the same questions. For example, the interview can
be directed to a separate battery of follow-up responses if the
respondent answers "<1> YES" to a question on smoking
cigarettes, and to an entirely different series of questions if
the respondent answers "<5> NO."
Commands can also be inserted between questions to direct the interview
to a particular battery of questions based on the combination of
responses to two or more previously answered questions. For example,
a special battery of questions on mammograms might be asked of only
female respondents over 35 years of age, while others might be directed
to respond to another series of questions. Such between-question
programming commands can also be used for checking response inconsistencies
between items and for skipping to follow-up questions to clarify
any inconsistencies discovered. Programming features minimize errors
since inappropriate questions are not asked and, as a result, appreciably
less editing is necessary after the interview.
The software supports both input items (i.e., those for which keystrokes
by the interviewer are recorded in the data set) and non-input items.
Non-input items can be variables created by the programmer that
store the result of a mathematical computation or logic check executed
during the interview or store management data such as the date or
time length of the interview.
CATI Components Support Detailed Data Retrieval
The typical CATI instrument has two major components, the section
containing the substantive questions of the interview, and the section
containing the case management items (i.e., the front end). The
front end section includes:
• Non-input items for recording the case status (e.g., whether
called or not, the number of call attempts, etc.)
• Calling history for the cases and the dispositions of call
attempts (e.g., dates and times of calls, outcomes such as no answers,
busies, not-in-service, respondent ineligible, call-back, refusal,
etc.)
• Scripted items that provide dialing information for interviewers,
introductory scripts, respondent selection protocols, callback calendars,
case reentry controls and scripts for callbacks, and scripts to
exit the interview and record interview evaluation notes.
• Program commands for measuring total elapsed time during
the interview, even for separate interview sessions.
• Auto-scheduling information to enable a dedicated PC to read
the callback information recorded in the data record of each case
and to build queues of appointments, which then are searched in
order of priority specified by the study manager to find any appointments
scheduled at the time an interviewer is requesting a case to call.
Appointment queues can be updated at whatever interval of time deemed
appropriate by the study manager but, typically, this is every five
minutes. The objective of the auto-scheduling program is to reduce
the need for paper call sheets since the front end program includes
an electronic call sheet (although the paper version may be used
as well) and to decrease the possibility of inadvertently missing
scheduled appointments.
CASES keeps track of calls to and appointments with targeted respondents,
prompts interviewers with survey questions, provides other important
information to interviewers and their supervisors, and records respondent
answers in a secure, backed-up data base.
Flexibility Is Key to Success
CATI is a remarkably flexible method of survey administration. A
great deal of latitude is available to tailor interviews to particular
respondents by building in skip patterns so that respondents are
only asked questions that are relevant to them. Furthermore, the
instrument can be programmed to recall information provided by a
client about a respondent, or to recall information gathered earlier
in the interview. It is also easy to introduce systematic or random
"variants" of question or answer categories to minimize
or to test for the effects of the order of questions or answers.
CASES also permits supervisors to unobtrusively monitor interviews
in progress and keep close track of the status of each case.
|
