2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

17
2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5

Transcript of 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

Page 1: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Survey MethodologySurvey Design

EPID 626

Lecture 5

Page 2: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Lecture Overview

• Discussion about survey study design issues

• Hands-on exercise with WIN data

Page 3: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Methods of data collection

• Mail survey

• Telephone survey

• Personal interview

• Computer-assisted interview

Page 4: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Considerations

• Available information

• Survey population skills and motivation

• Desired data

• Response rates

• Resources (money and time)

Page 5: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Mail survey• Available information?

– Need complete and accurate addresses

• Survey population skills and motivation?– Reading, writing, language skills– Knowledge of topic– Interest in subject– Limited opportunity to convince someone

to participate

Page 6: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Mail survey (2)

• Desired data?– No interviewer involved to clarify, prompt or

establish rapport– Respondents have more time to consider

responses– Can access hard-to-reach samples– Difficult to designate a respondent

Page 7: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Mail survey (3)

• Response rates– May be very low– May increase with telephone follow-up

• Resources– Relatively low cost– Requires minimal staff and facilities

Page 8: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Telephone survey

• Available information? – Need a good list of telephone numbers if

you’re not using RDD

• Survey population skills and motivation?– Skills needed are minimal– Language?– Limited opportunity to convince someone

to participate

Page 9: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Telephone survey (2)• Desired data?

– May be less appropriate for sensitive or personal information

– Can’t use visual aids– May miss those without phones, who screen

calls, who aren’t home etc.– Have advantages of interviewer

administration such as ability to clarify and probe

Page 10: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Telephone survey (3)

• Response rates?– Better than mail surveys– Usually not as good as personal interview– RDD especially is low

• Resources?– Requires less than does personal interview– May take less time than mail or personal

interview surveys

Page 11: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Personal Interview Survey• Available information?

– May need an address – May not (area probability sampling)

• Survey population skills and motivation?– Skills needed are minimal– Have an opportunity to convince people to

participate, establish rapport– Language?

Page 12: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Personal Interview Survey (2)• Desired data?

– Interviewer can explain, clarify, and probe– Can use visual aids– Rapport may encourage sharing of

sensitive information– Can get more information with a longer

survey– May be able to ask more open-ended

questions

Page 13: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Personal Interview Survey (3)

• Response rates?– Usually considered very good

• Resources– Usually very expensive– Need a well-trained staff – Need supervision– May take a long time

Page 14: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Computer assisted survey• More of a tool than a design really

– Can be used as a stand-alone self-administered survey

– Can be used in conjunction with telephone or personal interviews

– Can use in conjunction with video to simulate a personal interview

• If desired, you can match some characteristics, such as age, race, and gender of the interviewer with the subject (more on this later)

Page 15: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Computer assisted survey (2)

• Available information?– Depends on other elements of the study

design

• Survey population skills and motivation?– People who are not familiar with computers

may feel uncomfortable– Can adjust the language– Can read questions aloud

Page 16: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Computer assisted survey (3)

• Desired data? – Can follow complex question patterns or

questions tailored to previous answers– Can identify inconsistencies or missing

data before the interview is over– Can be useful in combination with

telephone survey

Page 17: 2/9/00 Survey Methodology Survey Design EPID 626 Lecture 5.

2/9/00

Computer assisted survey (3)

• Response rates? – Depends on other elements of the study

design

• Resources?– No data entry required– If it is self-administered, no interviewer

required