New Technologies in Survey Research

1
New Technologiesin Survey Research
Andrew E. Smith, Ph.D. UNH Survey
Center andrew.smith_at_unh.edu 603.862.2226
2
UNH Survey Center

• Collects primary data for UNH Faculty
• Works with UNH students on research projects
• Conducts Granite State Poll, a quarterly omnibus
  NH public opinion poll
• Media Election Polls
• CNN, FOX, WMUR, WCVB, Boston Globe, Philadelphia
  Inquirer
• State Local Governments
• Not for profits
• Business

3
Research Questions
Information Needs
Other Sources of Data?
Population
How Best to Interview?
How to Sample?
Specific Interview Topics
Actual Survey Questions
4
New Technologies, New Methodologies

• Better computers and telecommunications (lower
  costs) have led to development of new data
  collection methodologies
• Computer/Internet Advancements
• Web
• E-mail
• CAPI (Computer Assisted Personal Interviewing)
• CASI (Computer Assisted Self-Interviewing)
• Telecommiunications Advancements
• T-ACASI / TDE (Telephone automated
  computer-assisted self-interviewing, Touchtone
  Data Entry)
• IVR (Interactive Voice Response)
• New technologies may improve data quality (when
  used appropriately) but do NOT solve response
  rate problems

5
Web / E-mail Surveys

• Essentially advancements on self-administered /
  mail surveys
• Conducted in visual mode
• No interviewer

6
Web / E-mail Surveys

• Advantages
• Relatively low cost (sometimes free!)
• Zoomerang (www.zoomerang.com)
• Survey Monkey (www.surveymonkey.com)
• Inexpensive notification / follow-up
• Can include pictures, sound, movies, etc.
• Automatic data collection no need for data
  entry
• Quick turnaround
• Can use sophisticated branching
• Best with defined populations with e-mail
  addresses

7
Web / E-mail Surveys

• Disadvantages
• No interviewer ? no probing
• Difficult to collect complete open-end
  information
• No ability to sample general population ? no
  equivalent of RDD or even a telephone book
• Cannot generalize to general population
• Cannot guarantee confidentiality
• Tend to have low response rates (see Dillman,
  2000)
• Still need some programming / management

8
CAPI

• Computer Assisted Personal Interviewing
• Interviewer with a lap-top
• Reduction in cost of laptops make this feasible
• Many CATI software packages have CAPI
  capabilities
• Advantages
• Reduces data entry errors
• Can show materials to respondent
• Disadvantages
• Cost
• Maintenance
• Theft!

9
CASI

• Computer Assisted Self-Interviewing
• Respondent takes the survey on a computer with
  appropriate software
• Typically laptops are used
• Web surveys can be considered CASI
• Advantages
• May improve confidentiality of results reduce
  social desirability
• Can show materials to respondent
• Flexibility for respondent

10
T-ACASI / TDE

• Telephone Audio Computer Assisted
  Self-Interviewing / Touchtone Data Entry
• Mixture of standard telephone techniques and new
  technology
• Can use RDD
• Still use an interviewer
• Respondent enters response by touching a key on a
  touch-tone telephone
• Developed to increase confidentiality of
  responses, improve data accuracy
• However, recent research does not support this
  (Moscowitz 2004)

11
IVR

• Interactive Voice Technology
• Responses can be respondents voice or touch-tone
• Typically a computer is used as an interviewer
• Used more and more for market research and
  political polling. IVR did quite well in
  predicting 2004 presidential elections.
• Survey USA
• Rasmussen
• Performs better when questionnaires are short

12
IVR

• Advantages
• Low cost no interviewers to supervise or pay
• Can collect lots of information quickly
• Professional voices reading questions
• More honest answers (Tourangeau et al., 2002)
• Respondents more likely to report sexual
  behaviors (Turner et al., 1996) and illicit drug
  use (Gribble et al., 2000)

13
IVR

• Disadvantages
• Costly setup / equipment
• Typically lease time from providers
• Low response rates (Tourangeau et al., 2002)
• Increased break-offs (Tourangeau et al., 2002
  Cooley et al., 2000)

14
Conclusions

• Technological improvements in computers and
  telecommunication are reducing costs of data
  collection, but are not necessarily improving
  data quality.
• Need to choose the data collection methodology
  appropriate for the population, topic, and length
  of questionnaire.

15
andrew.smith_at_unh.edu603.862.2226