An on-line pilot for a SP survey

1
An on-line pilot for a SP survey

• Dr Eric CORNELIS
• FUNDP-GRT

2
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

3
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

4
Context

• Building a new RER around Brussels

5
Context(2)

• Impact of parkings on mode choice
• Inventorying actual parking places in the
  neighbourhood of the stations
• Surveying on how commuters actually travel
• Surveying on how they would change their mode
  according to modified parking characteristics
• gtNeed for a SP survey

6
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

7
Why on line ?

• Postal or phone protocols not very suitable for
  SP surveys
• Face to face OK but
• Resources (time budget) consuming
• Possible to interview commuters in trains
• BUT how surveying commuters using car ?
• Sampling from which base population ?
• Internet survey

? Recruiting
8
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

9
Advertisement

• Mobility managers from big companies in Brussels
• Free newspaper (METRO)
• Flyers in station parkings
• E-mails to university alumni
• Rem OK since more than 500 complete answers
  in one month

10
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

11
Technical issues

• PHP MySQL
• Advantages
• Dynamically built web pages
• Automatic control and checking
• Automatic storage of answers into DB
• Retrieving data in other tools (e.g. Excel)
• Free solution

12
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

13
Design

• Belgium gt 2 languages
• 2 parts
• RP part (actual behaviour)
• SP exercise varying attributes on base of RP
  data
• 8 pairs of scenarii
• Filters
• Home or workplace outside RER area
• Commuting with other modes than
• Car
• Bike train
• Car train

14
RP part

• Socio economical characteristics
• (gender, age)
• Commuting mode
• Location (home, workplace)
• Travel time (departure and arrival)
• Availability of driving license and car(s)
  (including company car)
• Reimbursement of travel costs

15
RP part (2)

• If commuting by car
• Knowledge according potential departure station
  (and parking opportunities)
• Parking at workplace (distance, time, cost, type)
• Knowledge according potential arrival station
• If commuting by car train
• Parking at station (distance, time, cost, type,
  comfort, security)

16
SP part

• If actually commuting by car
• Commuting by car
• Commuting by car train
• If parking AT working place, CAR alternative

VARIABLES CHANGES
Travel time 0 15 30 from actual travel time
Oil costs 0 25 50
Parking costs 2 4 6 per day
Probability of finding a parking place at destination 100 (every day) 70 (almost every day) 40 (two days a week) 10 (rarely)
17
SP part(2)

• If parking NEAR working place, CAR alternative

VARIABLES CHANGES
Travel time 0 10 20 from actual travel time
Oil costs 0 15 30
Parking costs idem 10 15 per day
Probability of finding a parking place 100 (every day) 80 (almost every day) 50 (half time) 20 (sometimes)
Security/ comfort level medium/good/very good
Distance 50m 100m 300m 500m
18
SP part (3)

• CAR TRAIN alternative

VARIABLES CHANGES
Travel time -30 -20 -0 from actual travel time
Travel cost Fixed travel ticket (according departure and arrival stations)
Parking cost According station type  18,60/18j ou 0
Probability of finding a parking place 100 (every day) 80 (almost every day) 50 (half time) 20 (sometimes)
Security/comfort level medium/good/very good
Distance 50m 100m 300m 500m
19
SP part (4)

• If actually commuting by car train
• Car train
• Car train
• 2 alternatives with different attributes for
  parking

VARIABLES CHANGES
Parking cost -30 -15 -0
Probability of finding a parking place 100 (every day) 80 (almost every day) 50 (half time) 20 (sometimes)
Security/comfort level medium/good/very good
Distance 50m 100m 300m 500m
20
SP part (5)

• If actually commuting by bike train
• Bike train
• Car train
• 2 scenarii different by parking characteristics

VARIABLES CHANGES
Parking cost -30 -15 -0
Probability of finding a parking place 100 (every day) 80 (almost every day) 50 (half time) 20 (sometimes)
Security/comfort level medium/good/very good
Distance 50m 100m 300m 500m
21
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

22
Some results from RP part

• Bias ?

23
Some results from RP part(2)
24
Some results from RP part(3)
25
Some results from RP part(4)
26
Some results from RP part(5)
27
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

28
Demo
29
Outline

• Context
• Why on-line surveying ?
• How advertising respondents ?
• Technical issues
• Design of the survey
• Some results from the RP part
• Demo
• Advantages Limits

30
Advantages

• No resources consuming
• No re-encoding (gt less errors)
• Automatic direct checking

31
Limits

• Which base population (for weighting) ?
• Problem for people unfamiliar with internet
• Perhaps too much parameters in SP
• Changes ranges ?