Survey research II

1
Survey research II

• Interviewing

2
In person surveys

• Instead of respondents reading questionnaires and
  recording their own responses,
• Interviewers ask the questions orally and record
  the answers
• Small scale study might use one interviewer
• Larger studies have a staff of interviewers

3
Interviewer

• More personal, higher response rates (80-85)
• Usually decreases no answers and dont knows
• Can clarify questions (control of clarifications)
• Make observations about surroundings, nonverbal,
  reactions to questions

4
Interviewers

• Use of probes
• Reliability of interviewers a potential
  problemwill different interviewers get different
  answers?
• Interviewers must be trained

5
Training interviewers

• Appearance depends on the target population
• Norm middle class standards
• Businesses business attire
• More casual for street people, incarcerated
  individuals
• May want to ask members of target population

6
Training

• Pleasant nonjudgmental demeanor, put participants
  at ease
• Be familiar with the questionnaire, understand it
• use of laminated responses sheets as a guide
• Some example interviews

7
Probes

• Silence, pencil poised
• How is that? In what ways? Anything else?
• Must be neutral, must not lead the interviewee
• Probes should be indicated next to the question

8
Conducting the study

• Training
• Interviewers should understand the questionnaire,
  what the study is about to the degree that they
  can properly ask the questions
• Practice sessions
• Conducting interviews and feedback

9
Supervision

• Verifying that interviews have been completed
• Random checks, verify with participants
• Reviewing questionnaires as completed to ensure
  that questions are understood

10
Computer assisted

• Computer-based self-administration
• Faxes, web
• Disk-based or CD at a computer

11
In person computer assisted

• Interviewers might read questions from a computer
  screen and then key in respondents answers
• Or, interviewers might hand computers over to
  subjects who then key in the responses themselves
  (might seem more confidential)
• studies indicate that subjects admit to more
  with this technique

12
Advantages of computers

• Data analysis
• Complex contingency questions can be automated
• Less monotonous, especially when combined with
  face to face
• Different languages

13
Disadvantages

• Costs
• More programming
• Computer problems

14
Specialized interviews

• Unstructured interviews, open ended
• Rather than very specific questions, a list of
  general questions or topics with each subject
• Directed conversation rather than a formal
  interview
• Interview guide more like a checklist

15
Focus groups

• 8-15 people
• Assembled to engage in a guided group discussion
• Used originally by market researchers
• Ask some open ended questions, specific questions
  or issues, somewhat controlled
• Example determining market acceptability of a
  new program

16
Focus groups

• Generalizability
• Selecting the right participants to be
  representative of a larger group
• Hopefully the members will bounce ideas off each
  other that will provide useful information for
  the researchers

17
Telephone

• Structured interviews
• Still need training
• Dress not an issue, nor demeanor
• Pleasant telephone voice
• Need to state purpose quickly, because of
  telemarketing (do not call lists)
• Structured introduction

18
Telephone

• 94 of household have telephones, which means
  that representative samples are likely to be
  obtained
• Exceptions very poor, transient, homeless,
  unlisted numbers (RDD)
• Will cell phones change telephone surveying?
• Potential language problems

19
Computer assisted telephone interviewing

• Computers can generate and dial phone numbers
• Interviewers wear headsets at computer stations
• Program dials numbers
• Interviewer reads questions from computer screen,
  records responses

20
Computers

• Software formats responses into a data file
• Nationwide opinion polls typically use such a
  technique
• Some simple polls can be automated so that an
  interviewer is not needed (i.e., how people will
  vote in an election
• There could be a privacy issue

21
Which method to use

• 1. Self-administered cheapest and fastest, but
  only works if groups can assemble and can read
  reasonably well (with offender populationsmight
  want to tape record or read the questions)

22
Four methods

• Interviews get more in-depth information, and are
  better for less educated populations who might
  not be able to manage written questionnaires
  (again particularly relevant to offender
  populations), because interviewers can read and
  explain the questions

23
Four methods

• Interview studies get good response rates
• However, face to face interview studies will be
  more expensive and time-consuming than other
  types of studies

24
Four methods

• Mailed surveys can reach large numbers of people
  relatively quickly, and the costs are relatively
  lowbut return rates are problematic
• More structured and more impersonal
• Best when the answers to closed ended questions
  are sought

25
Four methods

• Telephone surveys are best for short
  questionnaires when the results are needed
  quickly
• Also better than mailedif literacy is problematic
• Return rates better than mailed, not as good as
  face to face, but acceptable

26
Surveys in general

• People might not be truthful or might not know
  the truth
• Somewhat superficial, do not give us the social
  context of field research