Workshop B7: Surveying

1
Workshop B7Surveying Hard to Reach Groups

• 13 participants
• Austria 1
• Denmark 1
• France 3
• South Africa 1
• Switzerland 1
• United Kingdom 1
• Unites States 5
• Resource paper Riandey and Quaglia (Surveying
  hard to reach groups)
• Contributed paper Contrino et al (A
  reexamination of methods in the US National
  Household Travel Surveys)
• Contributed paper Cowham et al (Prioritising
  street improvements for respondents with
  disabilities qualitative and quantitative
  research)

2
Questions discussed

1. Who are hard to reach groups in travel surveys,
   and how does this differ by type of survey (e.g.
   telephonic vs. mail-back vs. home interview,
   etc.) and by context?
2. What bias is introduced by lack of coverage of
   hard-to-reach groups? E.g. how does their
   absence/under-representation affect our estimates
   and forecasts of travel, or public policies?
3. Looking forward, how do we foresee this
   problemgetting less, getting worse? Why?
4. What measures can be taken to ensure that hard
   to reach groups are included adequately in
   sample frames and selections?
5. What measures can be taken to ensure that
   selected hard to reach respondents are
   successfully recruited and surveyed?
6. What forms of survey instruments are appropriate
   to reach hard to reach groups, and how should
   these instruments be developed and tested?
7. If we could completely re-design survey practice,
   what kinds of things would we include to
   integrate respondents from hard-to-reach groups
   into the survey?

3
1. Who are hard to reach groups in travel
surveys, and how does this differ by type of
survey (e.g. telephonic vs. mail-back vs. home
interview, etc.) and by context?

• Persistent non-responders
• Foreign language
• Illiterates
• Disengaged
• High/low income
• Adolescents
• Fearful
• Gate-kept individuals
• Physically disabled
• Mentally disabled

• Non-coverage in sample frames
• Illegal immigrants
• Transients
• Squatters
• Homeless
• Sub-letters
• Group quarters
• Inaccessible workers

4
2. What bias is introduced by lack of coverage of
hard-to-reach groups? E.g. how does their
absence/under-representation affect our estimates
and forecasts of travel, or public policies?

• Bias depends on
• Survey method (e.g. CATI vs. PAPI vs. CASI vs.
  CAWI)
• Context
• Significance varies by survey purpose
• Aggregate travel patterns
• Special group surveys
• Research gap size of HTR groups

5
3. Looking forward, how do we foresee this
problemgetting less, getting worse? Why?

• Drivers of change
• Technology
• Migration
• Demography
• Economy
• Legal barriers/benefits

6
4. What measures can be taken to ensure that
hard to reach groups are included adequately in
sample frames and selections?

• Multi-frame sampling
• Land line phone
• Mobile phone
• Address frame
• Association/organization membership lists
• Etc.
• Identification and recruiting
• General population surveys frame completeness
• Specific population surveys cost of screening
  to identify individuals, or access panels

7
5. What measures can be taken to ensure that
selected hard to reach respondents are
successfully recruited and surveyed?

• Toolkit of recruitment techniques
• Advertisements
• Association registers/lists
• Advance letters
• Endorsements
• Incentives/gifts
• Language specialists

8
6. What forms of survey instruments are
appropriate to reach hard to reach groups, and
how should these instruments be developed and
tested?

• Multi-modal with special components tailored for
  HTRGs
• But are results compatible with main survey?
• Non-response follow up surveys of two types
• Short last-ditch version to get essential
  travel-related information
• Short questionnaire for demographics/reasons for
  non-response
• Mode depends on HTRG
• Dont wait until end
• Design of instrument depends on purpose

9
7. If we could completely re-design survey
practice, what kinds of things would we include
to integrate respondents from hard-to-reach
groups into the survey?

• Ideal device to encourage HTRGs to participate
• Inexpensive
• Attractive/useful as a give-away after data
  collection finished
• Customizable for HTG group
• Non-intrusive
• Secure
• Passive fully automated data collection
• Multiple sensing capabilities not just GPS
• Close to real time data transmission w/feedback

10
Quantitative survey visual material (Cowham et
al)
STREET FURNITURE BENCHES EXAMPLE CHOICE CARD