15.10 Data collection Flashcards
1.Survey Research?
Variables are measured by asking respondents questions using standardized questionnaire or interview procedures.
Approach to asking questions is usually structured.
Answers are numerically coded (even nominal variables) for analytic purposes.
2.Survey Data Collection Modes?
Mail Self-administered in group settings Internet Telephone Personal (face-to-face)
Some Issues to 3.Consider in Choosing Data Collection Methods?
Cost Geographic location of respondents Speed of turnaround Length and complexity of questionnaire Sensitivity of topic
4.Total Survey Error?
Coverage error Sampling error Nonresponse error Item nonresponse errror Measurement error
5.Method of Reporting?
Paper and pencil
Technology assisted
Computer assisted
-CASI (computer-assisted self-interviewing)
-CATI (computer-assisted telephone interviewing)
-CAPI (computer-assisted personal interviewing)
- What Is an Acceptable Response Rate?
Response rate = # surveys completed/ total eligible sample
>=85% Excellent 70-85% Very good 60-70% Acceptable 50-60% Barely acceptable <50% Not acceptable
- Mail Surveys and Self-administrated Surveys?Advantages
Lower cost than telephone or personal Geographic dispersion okay Privacy for respondents Convenience for respondents No interviewer bias This method is reliable when closed questions are used, when the order in which questions are answered is unimportant, and when the questions and format are simple and straightforward.
- Mail Surveys
- Disadvantages?
Potential for coverage error Potential for nonresponse error Potential for item nonresponse error Length, complexity limits Open-ended questions problematic Selection factors (e.g., literacy) No potential of interviewer clarification/assessment
- Procedures/Practices for Ensuring Acceptable Response Rate?
Cover letter to respondent Postage on return envelope Guarantee of confidentiality or anonymity Reminders Incentives Length of questionnaire Clarity of instructions Aesthetics
- Example: Mail Survey?
School-based Substance Use Prevention Programs Study (SSUPPS)
Sample design: Stratified random sample of public and private schools
Respondents: Middle school teachers
Data collection: Mailed survey
- Example Data Collection Schedule for SSUPPS?
3/03/07 Lead letter sent to teachers 3/10/07 1st questionnaire sent by mail 3/17/07 Reminder postcard 3/31/07 2nd questionnaire sent by mail 4/14/07 Telephone prompt 4/28/07 3rd questionnaire sent by mail 5/12/07 Postcard w/calling hours
- Telephone Surveys
- Advantages?
Lower cost than personal interviews
Quality control is much easier than personal interviews
Speed
Can utilize special question formats in which certain questions are skipped, while such a format may be confusing for respondents completing a questionnaire
Cost is about 45 to 64 percent of personal interviews
In terms of sampling quality, the telephone survey mode is between personal interviews and mailed questionnaire
- Telephone Surveys
- Disadvantages?
Potential for coverage error
Potential for nonresponse error (but may be less than for mail surveys)
Reluctance of respondents to answer sensitive questions
Length/complexity limits
Open-ended questions problematic
Hard to establish trust and rapport with respondents than in personal interviews
- Personal Interviews (Face to Face)
- Advantages?
Higher response rate than mail or telephone (range from 85 percent in rural counties to 67 percent for large central cities in US)
Less potential for item nonresponse than mail or telephone
Appropriate for longer/more complex surveys (90 min for completing of some 400 questions)
May be appropriate for sensitive topics
Can use visual aids
Interviewer can clarify questions
Opportunity of unobtrusive observation
Can utilize special question formats in which certain questions are skipped
- Personal Interviews
- Disadvantages?
Cost (more than two times of cost in mailed questionnaire)
Quality control less easy than in telephone surveys (interviewers, data coders, and the researcher are geographically dispersed)
Difficulty locating respondents at home
Potential for interviewer bias