When Telephone Interviewing Is Most Appropriate

Four common methodologies are available for marketing research. The trick is knowing which one is the best one(s) to use for your particular research project. In this two-part series, MR Perspectives will first explore the most popular methodology � Computer Assisted Telephone Interviewing (CATI) � explaining what it is and when it is best used. In Part II, we will describe web, mail and interactive voice response methodologies and explain when they are best used.

With CATI research, the entire survey is programmed into an interactive computer-based script that looks and navigates much like a web survey. The technology supports both closed and open-ended questions, and all or parts of the survey may be recorded and saved as audio files, such as .WAV files. The surveys are dynamic and support complex skip logic and branching, thus preventing the need for multiple versions of a survey when there is a need to differ the questions by customer segment.

Once in the field, a computer dials the phone numbers. After a respondent is reached, a live interviewer conducts the interview by reading verbatim from the programmed survey script. The interviewer reads the question and possible responses to the respondent, marks the appropriate answer, and then moves to the next screen (one question per screen). If the correct person is not available, the interviewer may schedule a callback for a specific time in the future.�If the person to be called back is at a different phone number or has a different name, that information may be recorded in the sample file and used during the callback. The data set contains not only answers to each question, but also a complete log of what happened with each sample record (e.g., how many records were dialed, the exact status of each record, how many phone numbers were bad, how many people were screened out and why, and more).

It is BEST to use CATI, when�.

Now that you have a better defined problem and an identified set of objectives, the next step of the marketing research process is to specify the approach to reach your objectives. From deciding on the best methodology to funding the project:

  • The surveys are transaction-based, and thus time sensitive.�Typically, if asking questions regarding a specific experience or representative involved in a specific experience, the survey should be completed within four weeks of the transaction/experience.

  • There is a need to screen for the appropriate participants.�For example, finding the person who had direct contact with the Service Department. Here the interviewer asks for the person listed in the sample file.�Once that individual is on the phone, the interviewer asks if they had direct contact with the Service Department.�If so, the interview continues.�If not, the interviewer asks to speak with the person who did.

  • The sample does not contain the respondent�s name.�For instance, there is a business name and number for business policies, but no contact name.�In this case, the interviewer can explain whom he/she is looking to speak with.

  • There are multiple versions of the survey.�For example, there are differences in which questions are asked of whom and/or the survey is being fielding for multiple brands or product types. In this case, the script can read directly from the sample file and pipe in different text depending on the respondent�s brand or type.�It can also skip the respondent over questions that are not applicable.

  • The survey requires extensive quota cell management. For example, you may want to get specific number of completes for each brand.�Then within those brands, you may want a certain percentage to be for B2C policies and another for B2B policies.

  • The survey requires extensive explanation or clarification.�For example, a respondent may need clarification regarding some of the industry terms used in the survey.�

  • The survey contains follow-up or probing open-ended questions.�For example, the respondent may say that they were very dissatisfied with their overall service experience.�In this case, the next question may be to probe for specifics (e.g. �Why specifically did you say you were very dissatisfied with your service experience?�)�Often, the initial answer is fairly vague (e.g., because the service wasn�t good).�The interviewer, however, will probe to get specifics that are actionable within your organization.

  • You have accurate phone numbers for the majority of the intended respondents (note: cell phone numbers cannot be used by law).

  • You do not have enough sample for IVR and/or mail.�Since the response rate is higher for CATI, it will take less usable sample to achieve the desired number of completes.

  • The targeted respondents are busy people.�CATI can record scheduled callback times, allowing the interviewer to call the respondent back at a time/number convenient for them.

  • The subject of the survey is non-sensitive enough that the respondents are willing to discuss it with the interviewer (MAIL is best when it is a sensitive subject).

Next edition, Part II: Web, Mail and IVR Methodologies

All Content � Copyright 2004-2011. Polaris Marketing Research, Inc. All Rights Reserved.

Send inquiries to or call 1-888-816-8700.