Descriptive research surveys can be conducted in a variety of ways using many combinations of people and electronics and conducted in a variety of locations. Table 4.1 displays a few of the types of survey delivery methods possible from such combinations. We will first discuss the more “tradi- tional” methods of conducting surveys by telephone, mail, or in person, then discuss some of the newer ways these methods and the computer and fax technologies are being used in survey research.
Telephone Interviewing
Telephone interviewing is usually employed when the study design requires speedy collection of information from a large, geographically dispersed popu- lation that would be too costly to do in person; when eligibility is difficult (ne- cessitating many contacts for a completed interview); when the questionnaire is relatively short; or when face-to-face contact is not necessary.
From the Home
In this case, interviewing is conducted directly from the interviewer’s home. The interviewer has all the materials that will be needed to complete
TABLE 4.1. Survey Delivery Methods
Human/Electronic Involvement
Location
of Interaction Telephone Mail Personal
Computer or Fax
Home WATS or IVR Self-adminis-
tered or pan- els
Door to door Internet or disk by mail
Work WATS or IVR Self-adminis-
tered or lock box Executive interview Internet or fax
Malls — — Mall inter-
cept Computer assisted Central Location Research Facility — — On-site Computer assisted
an assignment in his or her home and use the specified period (per instruc- tions) on the telephone contacting respondents and conducting interviews.
From Central Location Phone Banks
A research firm may conduct interviewing from its facilities equipped with wide area telephone service (WATS). Central location interviewing has many advantages. It allows for constant monitoring and supervision. If an interviewer has a problem or question, it can be handled on the spot. Mis- takes can be corrected; respondents can be called back immediately if the supervisor finds that the interviewer’s information is not complete. Inter- views can be monitored to ensure correct interviewing procedures and tech- niques, and that quotas are being filled. Since telephone interviewing is the most commonly used method of data collection, a detailed discussion on the specific procedures of telephone interviewing is located in Chapter 8.
Most market research companies now use some form of computer-assisted telephone interviewing (CATI) system. The survey instrument or question- naire is programmed into a computer, and the “pages” of the questionnaire appear on the terminal screen in front of the interviewer. The interviewer can input the results either through the keyboard or in some cases by using an electronic light pen that can touch the appropriate response on the screen. The CATI System also provides for automatic skip patterns, immediate data input, and in some cases automatic dialing. This system promotes flexibility and simplification. In addition, increased speed and improved accuracy of the data are accomplished. Various programs of random-digit dialing can be attached to the CATI system. The computer-generated random-digit num- bers can eliminate the sampling bias of unlisted numbers.
Predictive dialing is another important feature that makes telephone in- terviewing more effective. With predictive dialing, the technology does the dialing, recognizes a voice response, and switches to the caller so fast that the interviewer virtually hears the entire “hello.” The system predicts how many dials to make to keep the interviewing staff busy without spending over half of their time just making contact. The system automatically redials no-answer and busy numbers at specified intervals. Some predictive dialing systems can be set to identify answering machines automatically and re- schedule those calls as well. This results in more efficient calling and higher penetration of the sample list.
Increasingly, research firms are turning to Interactive Voice Response (IVR) systems to conduct surveys by phone. In this approach respondents dial a toll-free number and respond to recorded questions either with a lim- ited set of voice responses indicated (e.g., “Please answer by saying home, work, or both.”), or by using the designated buttons on a touch-tone phone. Respondents may be members of a major panel (e.g., Market Facts, Na-
tional Family Opinion, or Home Testing Institute), or just responding to a one-time survey. The use of IVR, or Completely Automated Telephone Surveys (CATS), has been used to get quality data very quickly and has the advantage of letting respondents call anytime to answer the survey.
The advantage of having an interviewer ask questions of a respondent us- ing a WATS line is, of course, the ability to ask open-ended questions where probing for response is needed and answer respondent questions or provide clarification when necessary. These needs may sometimes be greater than the needs for the speed and cost savings of the IVR approach.
Telephone Communication
As compared to personal or mail interviewing, telephone interviews have the advantage of speed and relative economy. Interviews of this type are also easily validated and the personal interaction of a qualified inter- viewer maintains a relatively high degree of control. The proper sequencing of question response also can be maintained. While not as flexible and pro- ductive as a personal interview, a well-designed questionnaire administered by a skilled interviewer can gather relatively comprehensive information. The interviewer also is in a position to probe at appropriate times and follow appropriate skip patterns. Weaknesses of the telephone interview are its in- ability to be as long or as detailed as personal interviews and its inability to show the respondent any products, pictures, or lists. Specific advantages and disadvantages of telephone interviewing are listed in Table 4.2.
TABLE 4.2. Advantages and Disadvantages of Telephone Interviewing
Advantages Disadvantages
1. Provides ease of making a contact
(callback) with a respondent. 1. Exhibits cannot be shown unless amail/phone or phone/mail/phone methodology is used.
2. Administration and close super- vision is provided by professional staff at a central telephone center.
2. Question limitations such as limited scales.
3. Drawing representative samples of persons with telephones is relatively easy.
3. Long interviews are difficult to administer by telephone. 4. Sequential disclosure, and question-
naire flexibility with skip patterns, refer backs, etc., is possible. 5. Speed in gathering information. 6. Relatively low cost.
7. Simplicity at conducting a pretest. 8. Access to hard-to-reach people.
Mail Surveys
Mail surveys are commonly used in either ad hoc (one time) or mail panel research. In ad hoc research projects a sample of respondents is cho- sen to be representative of the population of interest and a questionnaire is sent by mail. The mail questionnaire may be preceded by and/or followed by a telephone call intended to encourage participation. The combination of telephone and mail is an attempt to reduce the high rate of nonresponse that plagues ad hoc, and to a lesser degree panel, mail surveys.
When mail is used in surveying people at work a lockbox approach is sometimes used to avoid secretaries or assistants completing a survey in- tended for an executive. Here an accompanying letter tells the respondent the purpose of the survey and that the box is a gift, but the combination will be provided at the time of a phone interview. The respondent opens the box at a scheduled interview time, completes the interview and, in some in- stances, returns the survey by mail.
Mail surveys are sometimes dropped off at homes or work sites with the intent of having the respondent complete the questionnaire and return it by mail. Such an approach has been shown to increase response rates and has the advantage of allowing the person dropping off the questionnaire to ex- plain the purpose of the interview.
Mail questionnaires allow for wide distribution. Also, the lack of inter- viewer/respondent interaction can give the feeling of anonymity that can encourage accurate response to relatively sensitive questions. In theory, the respondent has time to check records or even confer with someone else to make sure reported information is accurate. Disadvantages include a low re- sponse rate that results in nonresponse error. Control of the sample is mini- mal. Knowing the difference in results between those who participated and those who did not is not possible. Too often participants are those who are either more interested in the subject area or who have more free time to fill out questionnaires. Also, a mail survey is slow, less flexible, and does not allow for probing. Control is lost with a mail questionnaire and sequencing is difficult. Furthermore, researchers never actually know who completed the questionnaire. Many of these problems can be overcome through the use of mail surveys conducted by consumer panel organizations (see Chap- ter 2). Table 4.3 lists the advantages and disadvantages of mail surveys.
Personal Interviewing
This face-to-face method is employed when the survey may be too long to conduct over the telephone or there might be material to show the respon- dent. Personal interviews are effective when the interviewer is placing a product in a home or office; the respondent is tasting a product; or the sam- ple necessitates contacting homes in a specific manner, such as every fourth
home or going to every home until an interview is conducted, then skipping a specified number of homes before attempting the next contact as well as other applications.
Personal interviews allow for more in-depth probing on various issues and are the most productive, accurate, comprehensive, controlled, and ver- satile types of communication. There is ample opportunity for a well- trained interviewer to probe and interpret body language, facial expression, and other nuances during the interaction. Rapport can be developed that would put the interviewee at ease and gain his or her cooperation. The inter- viewer can explain any misunderstanding the respondent might have and keep the respondent on track and in sequence in responding to the question- naire. In spite of the advantages of greater depth and productivity, the per- sonal interview does take more time and money to administer. The advan- tages and disadvantages of personal interviewing are listed in Table 4.4.
There are several varieties of personal interviews. The most common are:
Door-to-Door Interviewing
In door-to-door interviewing, the client or field-service director will as- sign an interviewing location and often the exact address at which to begin. The interviewer then takes a specific direction in order to contact respon-
TABLE 4.3. Advantages and Disadvantages of Mail Surveys
Advantages Disadvantages
1. Cost-effective.
2. Efficiency in reaching large samples.
3. Access to hard-to-reach peo- ple.
4. Self-administered/no inter- viewer bias.
5. Limited use of exhibits is possible.
1. Low return rate.
2. Nonreturn bias; those who return a mail questionnaire may not represent the sample as a whole.
3. No control of who fills out the question- naire.
4. No ability for sequential disclosure of information.
5. Slow response. 6. Hard to pretest. 7. Question limitations.
dents according to the sample pattern that is specified to be followed. The interviewer will ask the questions and record the respondent’s answers ei- ther during the interview or immediately afterward.
Central Location/Mall Interviewing
The use of a central location or mall area is effective in instances where respondents will taste a product, look at a package design, view a commer- cial, or listen to a recording or view a videotape and report their impres- sions. Shopping malls are frequently used because of their high volume of traffic. Other central locations such as a church meeting hall, hotel, or other facility might be used to conduct a large number of personal interviews. When this method of personal interviewing is used, the respondents are generally recruited in advance. Shopping malls may be used as locations for prearranged interviews or involve “intercepts” where shoppers are recruited in the mall for the interviews.
Vendor/Dealer/Executive/Professional Interviewing
Interviewing of executives at their places of business is the business-to- business marketing research equivalent to consumer door-to-door inter- viewing. Locating people of interest to the researcher by title or name can be expensive and time-consuming if an accurate list is not available from a list vendor. Once identified, respondents must be contacted, an appointment scheduled for the interview, and then the survey can be administered in per- son at the workplace. Highly trained interviewers who are appropriately
TABLE 4.4. Advantages and Disadvantages of Personal Interviewing
Advantages Disadvantages
1. Exhibits such as ads, packages, pictures can be shown.
1. High cost 2. Flexibility and versatility. 2. Interviewer bias. 3. Speed can be accomplished with
multiple interviews being completed in multiple locations.
3. Administration/execution prob- lems.
4. Sampling can be done in a very repre- sentative way.
5. Observation of the respondent allows for viewing behavior as well as asking questions.
dressed and knowledgeable about the topic are much more likely to be suc- cessful in obtaining response, so this type of interviewing can be costly.
See Figure 4.5 for the strengths and weaknesses of the “traditional” means of surveying respondents.
Computer or Fax Survey
It is not uncommon for the computer to be used in telephone (CATI) in- terviewing, personal interviews conducted at shopping malls, or at a central location. See Table 4.6 for a summary of the various ways the computer may be used in survey research. See Table 4.7 for benefits achieved through the various computer-assisted survey methods.
Internet Research
Earlier in this chapter we discussed the use of the Internet for doing ex- ploratory research in the form of Internet focus groups. We will now discuss the use of the Internet for conducting survey research.7(Chapter 8 will dis- cuss a firm that specializes in Internet survey research.) Some of the newer, Internet and facsimile based approaches will be discussed here. Several rea- sons have been suggested for explaining the rise in the use of the Internet for conducting surveys. First, the questionnaire can be created, distributed to respondents, and electronically sent to the researcher very quickly. Since data are electronically delivered, statistical analysis software can be pro-
FIGURE 4.5. Strengths and Weaknesses of Methods of Data Collection
Interviewing Questionnaire
Personal Telephone
Self-
Administered Mail Back
Response Rate Moderate
to High
Moderate to High
High Low to High
Timing Slow Fast Moderate Slow
Cost/Coverage High Moderate Low Low
Interviewer Impact (bias)
High Moderate Low Low
Ability to handle complex questions
High Low Moderate Moderate
Control High High Low Low
Length Long Short Long Long
grammed to analyze and generate charts and graphs summarizing the data automatically. Second, Internet surveys are less expensive than using inter- viewers or printing, mailing, and tabulating mail surveys. Third, it is possi- ble to create panels of respondents on the Internet and longitudinally track attitudes, preferences, behaviors, and perceptions over time. A fourth rea- son is that other methods are not as cost effective as is the Internet for asking just a few questions. Another reason is the ability to reach large numbers of people globally very quickly and at low cost. Finally, Internet surveys can be made to look aesthetically pleasing and can, via Netscape or Internet Ex- plorer or similar browsers, add audio and video to the questionnaire.
Internet surveys can either be e-mail or Web-based approaches. E-mail surveys are simple to compose and send, but are limited to simple text (i.e., flat text format), allow for limited visual presentations (e.g., no photo or video based stimuli) and interactive capabilities, and can not permit com- plex skip patterns.
Web surveys, in comparison, are in HTML format and offer much more flexibility to the researcher, providing opportunity for presentation of com-
TABLE 4.6. Methods of Computerized Data Collection Available to Marketing Researchers
Survey Methods Characteristics
Computer-assisted personal interviewing
A method in which the researcher reads ques- tions to the respondent off a computer screen, and keys the respondent’s answers directly into the computer.
Computer-assisted self-inter- viewing
A member of the research team intercepts and directs willing respondents to nearby comput- ers. Each respondent reads questions and keys his or her answers directly into a com- puter.
Computer-assisted telephone interviewing
Interviewers telephone potential respondents, ask questions of respondents from a computer screen, and key answers directly into a com- puter.
Fully automated telephone interviewing
An automated voice asks questions over the telephone, and respondents use their touch- tone telephones to enter their replies. Electronic mail survey Researchers send e-mail surveys to potential
respondents who use electronic mail. Respon- dents key in their answers and send an e-mail reply.
Source: Adapted from V. Kumar, David A. Aaker, George S. Day, Essentials of Marketing Research (New York: John Wiley and Sons), 1999, p. 258.
COMPUTER-