An Introduction to CATI (Computer Assisted Telephone Interviewing) by moti

VIEWS: 0 PAGES: 17

									An Introduction to CATI (Computer
Assisted Telephone Interviewing)
Penn State Survey Research Center
Amanda Maull
Telephone Interviewing
   Some Benefits:

       Direct contact with respondent

       Less expensive than face-to-face

       Effective randomized sampling from most populations
        using RDD (Random Digit Dialing)

       More complex survey instruments
Telephone Interviewing
   Down-sides

       Interviewer Effects
         Sensitive Topics, Social Desirability Effects



       Still expensive compared to Mail and Web

       Can’t reach some populations
         Cell phone only households, transient groups
Sampling Methods for Telephone Studies

RDD (random digit dialing)
RDD studies are those which use a sample of telephone
numbers that have been randomly generated

Based on Exchanges and Working Blocks
           814-238-XXXX


Increases coverage by including unlisted and recently-installed
    numbers
Sampling Methods for Telephone Studies

List Samples
Similar to RDD but numbers are run through a predictive module
    that tries to associate secondary data with the number
    (driver’s license records, marketing reply cards, etc.).

These samples can save a great deal of time when specific
   demographics are needed

However, the sample is to some degree ―less random.‖
Sampling Methods for Telephone Studies

Listed
We start out with a list of respondents so we know who we want
    to interview before we even call.

These respondents have often been screened or contacted prior to
   the interview.

Common example: Secondary waves of data collection for a
  longitudinal (or panel) study.
Telephone Interviewing

The Telephone Interviewer plays a dual role as both recruiter and
   data collector.

   Unstructured vs. Structured

Important protocols structure both roles.

Interviewers are the key link between respondent and data.
Respondent Recruitment
The sampling method will determine the protocol for respondent
   recruitment.

RDD & List Studies
 Phone numbers are dialed randomly and then respondents are
  randomly selected from the household members.

   Different methods are used for randomly selecting the
    respondent.
     most recent birthday method
     computer selection method
Respondent Recruitment: RDD & List
     ESSENTIAL - Once a respondent is selected – he/she is
      the ONLY person in the household that we can interview.

     Respondent must be currently living in the household at
      least 50% of the time.

     Cannot conduct interviews on business, institution or
      secondary lines (unless specified by study design).

     Cannot call forwarded numbers, must only conduct
      interview at number provided by the sample.
Respondent Recruitment: Listed
Interviewers must interview the person whose name is listed and
    only that person.

They may call different numbers to contact them.

If the respondent no longer lives at previous number, interviewers
     will try to obtain current contact information from whoever
     answers the phone.
Interviewing Guidelines
   To obtain accurate information, all interviewers must collect
    data in the same, uniform manner from all respondents. Each
    interviewer should conduct the interview in the same fashion
    as all other interviewers

   The Main Rules: Stay neutral, follow the script, and record
    answers accurately.
Stay Neutral
   Responding in any way to answers given by respondents has
    the potential to affect their answers to future questions.
    Respondents generally want you to like them. If they see you
    approve or disapprove of an answer, they may answer
    differently to future questions.

       Never disagree or argue with the respondent
       Avoid ―fine‖ or ―good‖ after a response
       Always keep personal opinions to yourself
       Never release personal information
Be courteous and professional at all times
   The purpose of the interview is to obtain accurate information
    from the respondent. In order to obtain this information, you
    should avoid creating the impression that the interview is a
    test or cross-examination.

       There are no wrong answers.
       You have the right to ask the respondent for information,
        just as the respondent has the right to refuse to answer.
       Politely deal with difficult respondent.
Follow the Script
   Ask every question in the order presented. Some
    respondents will offer you information that answers future
    questions; it is still necessary to ask all the questions.

   Read each question verbatim. Each question is carefully
    designed and pretested to fulfill its exact purpose—to obtain
    specific information and to be understood by the maximum
    number of respondents.
Other General Guidelines
   Record all responses accurately

   Deal with inconsistencies politely

   Probe for answers to responses such as ―Don’t Know‖ or
    ―Refuse‖

   Thank the respondent periodically
Quality Control for CATI interviews
   Interviewers are monitored in real time

   Completed interviews are verified by calling back
    respondents and asking them about their interview
    experience.
Telephone Interviewing Positions
   The SRC often has telephone interviewing positions available
   Paid $8.15/per hour
   Flexible schedules: evening & weekend hours
   Requirements: good communication and computer skills,
    ability to work 12 hours per week and some weekends.

   Visit our website for more information:
    http://www.ssri.psu.edu/survey/employment.htm

   Or email: telsup@ssri.psu.edu

								
To top