"Surveys are the most widely used technique in the behavioral sciences for the collection of data. They are a means of gathering information that describes the NATURE and EXTENT of a specified set of data ranging from physical counts and frequencies to attitudes and opinions. This information, in turn, can be used to:

• answer questions that have been raised,
• solve problems that have been posed or observed,
• assess needs and set goals,
• determine whether or not specific objectives have been met,
• establish baselines against which future comparisons can be made,
• analyze trends across time, and generally,
• describe what exists, in what amount, and in what context."

Surveys should be:

• Systematic carefully planned and executed to ensure appropriate coverage and sound, efficient data collection
• Representative  closely reflecting the population of all possible cases or occurrences, either by including everyone or everything, or by using scientific sampling procedures
• Objective ensure that the data are as observable and explicit as possible
• Quantifiable  yield data that can be expressed in numerical terms

Except for surveys based on a search of records, "surveys are dependent on direct communication with persons having characteristics, behaviors, attitudes, and relevant information appropriate for a specific investigation. This makes them REACTIVE in nature; that is, they directly involve the respondent in the assessment process by eliciting a REACTION. Although direct interactions are often the most cost-effective, efficient, and credible means of collecting data, because the respondents are usually in the best position to speak for themselves, reactive measures run many risks of generating misleading information.

Among these risks are the following:

• Surveys only tap respondents who are accessible and cooperative.
• Surveys often make the respondent feel special or unnatural and, thus, produce responses that are artificial or slanted.
• Surveys arouse RESPONSE SETS such as acquiescence or a proneness to agree with positive answers or questions.
• Surveys are vulnerable to over-rater or under-rater bias ? the tendency for some respondents to give consistently high or low ratings.
• In the case of interviews, biased reactions can be elicited because of characteristics the interviewer or respondent, or the combination, that elicit an unduly favorable or unfavorable pattern of responses.

1. Define the purpose and scope of the survey clearly and in explicit terms. Avoid "fishing".
2. Avoid using an existing survey if it was designed for a different purpose or population. The questions may be used as a starting point to help you but your survey should be specific to YOUR question.
3. When you design interviews and surveys, it is often helpful to sit down with a group of people similar to the group you intend to survey and find out WITH them how to phrase questions that they understand and that will get you the information you are looking for.
4. Field test your survey or interview to spot unclear or unnecessary questions and to figure out the best method for quantifying your results.
5. As often as possible, use structured rather than open-ended questions to support uniformity of results and ease of analysis.
6. Do not ask questions out of idle curiosity. Only ask the questions you need to help you answer your question. Otherwise, your respondents will get tired, annoyed, etc.
7. Avoid loaded or biased questions and be careful of biased sampling.
8. Keep your questionnaire or interview as brief, clear, and simple as possible.
9. Be clear about the characteristics of the respondents that must be collected at the time the survey is administered (gender, age, ethnicity, occupation, religion, employment, education, demographic factors, etc.). Each survey or interview needs to be accompanied by this type of information.

These sources of data differ from those in the other survey types because they are nonreactive; that is, they do not involve a direct response from people. Although this circumstance can be an advantage to the objectivity of the information, there are other problems with how and why records are dept that can offset the possible objectivity.
 
 
 

ADVANTAGES	DISADVANTAGES
Records are non-reactive.	May involve confidential restrictions.
They are inexpensive.	Are often incomplete, inaccurate, and out-of-date.
Records often allow historical comparisons and trend analysis.	Changing rules for keeping records often makes year-to-year comparisons invalid.
If records are accurate and up-to-date, they provide an excellent baseline for comparison.	Can be misleading unless a knowledgeable person can explain how the records were compiled.
	Purpose of records usually is unrelated to the purpose of a particular survey.
	Factual data only (no input on values or attitudes) are present.

 
 
 

These are the most commonly used survey method and often the most misleading and uninformative UNLESS the disadvantages listed below are acknowledged and accounted for. Questionnaires sent through email would have similar advantages and disadvantages.
 
 
 

ADVANTAGES	DISADVANTAGES
Are inexpensive.	Low response rate can occur
Are wide-ranging	No assurance that the questions were understood.
Can be well designed, simple, and clear.	No assurance that the addressee actually was the one who answered.
Are self-administering.
Can be made anonymous.

 
 
 

Note: Questionnaires should be carefully field tested to eliminate unclear or ambiguous or biased items and to improve the format. The return rates can be improved if you include stamped, addressed envelopes, follow-up reminders, advance contacts, etc.
 
 

These are widely used in place of face-to-face interviews. In general, they seem to get the same results as face-to-face interviews.
 
 
 

ADVANTAGES	DISADVANTAGES
Less costly than face-to-face interviews	Unlisted numbers
Can be conducted daytime or evenings	Not everyone has a phone so sample may be biased
Allows for unlimited callbacks	Can be viewed as intrusive into home privacy and can be confused with a disguised sales pitch
Respondent is at ease in own home and tends to be more honest	Rules out many face-to-face advantages, including visual impressions
Extended geographical coverage

 
 
 
 
 

Although an interview is usually a one-to-one relationship, for many purposes interviewing in groups is appropriate. Not only does it save time, but if the behavior you are trying to understand takes place in group interaction settings, the group interview will give you a better picture of this phenomenon.
 
 
 

ADVANTAGES	DISADVANTAGES
More efficient and economical than one-to-one interviews	May intimidate and suppress individual differences
Results reflect group behavior and consensus	Fosters conformity
Reveals group interaction patterns	Intensifies group loyalties and can rigidly polarize opinions
As with brainstorming, can stimulate others’ ideas and responses	Is vulnerable to manipulation by an influential and skillful group member

 
 
 

This is the conventional method of collecting data face-to-face. It not only can be helpful as the principal method, but also can be especially useful to explore a problem area about which insufficient information exists. For example, you could use an interview as a pilot study on which to base a more extensive questionnaire.
 
 
 
 

ADVANTAGES	DISADVANTAGES
Is personalized	May be expensive and is time-consuming
Permits in-depth, free responses	May intimidate or annoy respondents with a racial, ethnic, or SES background different from the interviewer
Allows impressions of respondent’s gestures, tone of voice, home environment, etc.	Is open to overt manipulations or the subtle biases of the interviewer
	Is vulnerable to personality conflicts
	Requires skilled and trained interviewers to get optimum results
	May be difficult to summarize findings

Taken From:  Isaac, S and Michael, W.  (1995).  Handbook in Research and Evaluation.  3rd Edition. CA:  EdiTS.  p. 136-140.