I. Mail surveys
II. Telephone surveys
III. In-person interviews
IV. Focus group interviews
A. Mail surveys seek responses directly from respondents and so the quality of the instrument is extremely important. They are conducted in the "natural" setting of the respondent. They rely on probability sampling for representativeness. Large numbers of surveys are needed for making inferences to the population.
Mail surveys involve sending and receiving questionnaires through the mail. They are often used with two types of groups:
1) to obtain quantitative data on questions of general knowledge or opinion from large samples of a population; and
2) to obtain quantitative data on questions of specific knowledge from
small samples of experts.
| Aspect | Advantages | Disadvantages |
| Cost | Cheaper than in-person interviews;
No need to train interviewers; |
Repeated follow-ups increase costs |
| Time | Can be used repeatedly to track things over time; | Can be slower than other methods |
| Contact with respondent | R can decide the optimal time to complete;
Minimal inconvenience to R; |
It is not known who actually completed the survey;
Cursory or incomplete answers; Misunderstandings; |
| Interviewer Bias | No bias | |
| Quality and Type of Information | Can collect thoughtful, reflective answers from experts; | Limited to closed-ended questions for mass surveys;
Difficult to make completely unambiguous; Oversimplification may increase reliability but decrease validity; Expert information may be difficult to locate |
| Response rate | Good if sample is motivated | Poor response rate from mass surveys;
May need to re-contact respondents more than once; Surveys may be ignored due to budget cycles, holidays, personnel turnover, or lack of staff time |
| Suitability | Good for experts and highly motivated samples | Not appropriate for some target populations |
B. The two major concerns with mail surveys are response quality and response rate.
A sample of non-respondents should be contacted to see if they differ from respondents on any important characteristics which could affect the external validity of the study. Non-respondents tend to be poorer, have greater mobility, be less well educated, etc.
There is some indication of "survey fatigue" on the
part of the general public, which has been over-saturated by surveys, so
return rates in general may be falling rather than improving.
However, if there is a guarantee of anonymity, then
follow up must be done with the entire sample, since it cannot be known
which persons returned the survey and which did not. There may be a trade-off
between anonymity and response rate.
Another reason is the development of computer-based programs for telephone interviewing. Computer programs can now perform many tasks to make interviewing over the telephone easier, from picking a random sample of telephone numbers, to displaying the questionnaire on the computer screen, to recording the interviewer's entries into a data base that is ready for analysis as soon as the interview is finished.
Telephone surveys are often used to obtain quantitative
data on questions of general knowledge or opinion from large or widely
dispersed samples of a population.
| Aspect | Advantages | Disadvantages |
| Cost | Cheaper than in-person interviews | More expensive than mail surveys |
| Time | Faster than mail or in-person | |
| Contact with respondent | Allows close supervision of interviewers | People may mistake interviewers for tele-marketers;
Interviewers must be trained how to handle open-ended response categories; |
| Interviewer Bias | Minimal bias | |
| Quality and Type of Information | Data collection, coding, and data entry are immediate;
Reduces data coding and data entry errors; |
Need short questions and few response categories;
Lack of visual cues can lead to misunderstandings; Rank ordering or percentaging questions are too complex; May get many requests by respondents for repeats or clarifications |
| Response rate | Better than mail surveys; | Can distinguish between "unreachable" and "refused to participate";
California, Florida, and Texas residents are most likely to refuse to participate |
| Suitability | Covers a wide geographic area quickly;
Random digit dialing can reach unlisted numbers |
People without telephones are excluded;
Homes with more than one phone may be over-represented; People who answer the telephone are often different than those who do not; |
B. The two major concerns with telephone surveys
are response quality and response rate. A sample of non-respondents should
be contacted to see if they differ from respondents on any important characteristics
which could affect the external validity of the study.
-interviewer recruitment and training;
-scheduling the interviews;
-preparing the materials needed;
-organization and supervision of field work;
-follow-ups for missed interviews;
-translation services where needed;
-coding and transcribing interviews;
-analysis and interpretation of findings.
Interviews are usually of two main types: structured or unstructured.
A. Structured Interviews
1. All interviewers are trained to conduct each interview in exactly
the same way.
2. Interviewers are given a script to read to explain the purpose of
the study and to cover the respondent's rights.
3. Interviewers ask all questions with no change in wording or order
of presentation.
4. If probes are allowed, the probes are usually scripted beforehand
(e.g., "can you tell me a little more about...?").
5. Interviewers are trained how to handle incomplete or inappropriate
answers or misunderstandings.
6. Interviewers must record the respondent's answers verbatim; tape
recorders are sometimes used.
7. Interviewers are trained to deal with the interpersonal dynamics
of the interview situation.
B. Unstructured Interviews
Unstructured interviews are usually designed to collect qualitative information from a small-sized sample. They are also sometimes called intensive interviewing or in-depth interviewing. Unstructured or semi-structured interviewers use what is called an "interview guide" or "interview schedule."
These interviews assume that the respondents have had a particular experience, on which they can elaborate. The situation has been analyzed before the interview, and the researcher is seeking additional information. The interview guide specifies the topics on which information is sought. And the interview focuses on the respondent's subjective experiences. It allows the respondents to describe in detail the situation as it is meaningful to them, and allows the interviewer to freely probe and ask follow-up questions. It depends on the following:
Training: Interviewers are trained to develop good interpersonal relations, listening skills, and good interviewing techniques. Interviewers should make the experience pleasant and satisfying for the respondent. Interviewers should convey that the study is important and that the respondent's participation is valuable and worthwhile.
Commitment: Both parties must derive sufficient rewards from the interaction to continue the interview. It is important to build mutual rapport, to develop and nurture a relationship. On the one hand, the interviewer can give appreciation for the receipt of valuable information, and on the other hand the respondent can be gratified as the object of interest, having an audience.
Meaning: Shared meanings cannot always be assumed. It is important to probe, to ask for clarification, and to attempt to arrive at an equivalence of understanding between the interviewer and the respondent.
Flexibility: Some questions that don't elicit any response can be dropped or set aside for later in the interview, while those that prompt a response can be pursued at length. Any barriers to participation by the respondent should be addressed and, if possible, overcome.
Assimilation: when the interviewer can make send of the respondent's stories and can identify the relevant terms, values, and meanings, then the interviewer has been able to adopt the respondent's point of view.
Re-interviewing: new categories of analysis may emerge during the research project, so that former respondents can be re-interviewed about the new questions, if the interviewer established a good rapport with them in the first interview.
Access: Respondents are generally reached through a written contact, through a sponsor, or a person with access. The respondent must be told in advance how long the interview is likely to take, what topics will be addressed, and what incentives if any there are for participation. However, the researcher must not threaten or beg prospective respondents to participate!
Preparation: The interviewer must have completed some background research on the topic of the project, on the part the respondent can play in the project, and knowledge or assumptions that are commonly shared by members of the respondent's community.
Process: The interviewer should explain the
study, obtain informed consent, and get permission to take notes or to
record the interview. Notes should be taken on non-verbal as well as verbal
communication if that is important to the research.
| Aspect | Advantages | Disadvantages |
| Cost | More expensive than mail or telephone survey; especially if travel is required. | |
| Time | May be quicker than mail | Slower than telephone |
| Contact with respondent | Interviewer can build rapport with respondent;
Interview can last longer; Information can benefit the respondent; |
Requires experienced and knowledgeable interviewers;
Interviewer must limit outside interruptions |
| Interviewer Bias | Largest potential for interviewer bias | |
| Quality and Type of Information | Can obtain large amounts of data;
More complex and sensitive questions possible; Can use visual cues, charts, complex issues; Data likely to be correct; |
Less appropriate for obtaining factual information or widespread
opinions;
Interviewer must be able to tape record or take notes verbatim; Lack of standardization; |
| Response rate | Generally very good | Difficult to conduct back-to-back interviews due to interviewer fatigue |
| Suitability | Only method for some populations (homeless); requires no reading/writing ability for respondents; | Must be able to interview others with opposing or dissenting
views to get a complete picture;
Difficult to replicate |
A focus group is composed of individuals who are homogeneous on some important characteristic, such as marital status, income level, presence of children, political orientation, etc. The group is representative of some pre-defined segment of some population, rather than a random sample of all the segments of that population.
The steps in conducting a focus group interview are:
1. Decide on the purpose of the study, what information will be needed, and the questions that will be posed to the group.
2. Select a group of 3-12 people who are homogeneous on some characteristic.
4. After the interview, the facilitator and aides compare notes on the group interview and write up a preliminary report.
5. The transcripts of the notes and/or tapes are prepared for analysis.
6. Content analysis, often with specialized software, is conducted on the transcripts. Analysis looks for trends, patterns, key words and phrases, images, etc. It identifies the range and diversity of opinion, as well as areas of agreement and disagreement.
7. A final report is written.
| Aspect | Advantages | Disadvantages |
| Cost | More expensive than mail or telephone survey; | |
| Time | May be quicker than mail | Slower than telephone |
| Contact with respondent | Interview can last longer than mail or telephone; | Requires experienced and knowledgeable facilitator;
Difficult to schedule |
| Interviewer Bias | Group dynamics affect the results | |
| Quality and Type of Information | Can obtain large amounts of data;
More complex and sensitive questions possible |
Interviewer must be able to tape record or take notes verbatim;
Can use visual cues, charts, complex issues; Must have multiple focus groups to cover all sub-groups in the population |
| Response rate | Generally very good | |
| Suitability | Can be used with wide range of populations |