Presentation is loading. Please wait.

Presentation is loading. Please wait.

Onderzoeksinstituut Workshop Valorisation of transport models for a sustainable transport Namur 19 December 2005 Travel behaviour survey in Flanders :

Similar presentations


Presentation on theme: "Onderzoeksinstituut Workshop Valorisation of transport models for a sustainable transport Namur 19 December 2005 Travel behaviour survey in Flanders :"— Presentation transcript:

1 onderzoeksinstituut Workshop Valorisation of transport models for a sustainable transport Namur 19 December 2005 Travel behaviour survey in Flanders : the latest developments Erik Nuyts Provinciale Hogeschool Limburg

2 onderzoeksinstituut Short History of OVG: Travel Behaviour Survey in Flanders 1993Haalbaarheidsstudie 1995OVG Vlaanderen 1997OVG Kust 1999OVG Antwerpen OVG Hasselt-Genk 2000OVG Vlaanderen 2 OVG Gent 2001(OVG Vlaams-Brabant) OVG Leuven, OVG Mechelen, OVG Aalst OVG Rand rond Brussel

3 onderzoeksinstituut Response Rates

4 onderzoeksinstituut Comparison OVG With Other Studies Time budget survey Mobel Socio-Economic Survey Neu Kontiv Design Dutch OVG Literature

5 onderzoeksinstituut Comparison OVG With Other Studies Mainly methodology Limited comparison of results, trying to recalculate OVG results with Mobel or TBE methodology

6 onderzoeksinstituut Main Results Suggestions for possible changes in methodology

7 onderzoeksinstituut Policy Decision: change wanted ? Do they prefer improvement or comparison between OVG’s

8 onderzoeksinstituut Main Results SAMPLE

9 onderzoeksinstituut Person or household level ? At present: household level household is a relevant entity often many respondents in 1 agreement  household less evident (divorce,..)  more difficult if recruiting by GSM

10 onderzoeksinstituut 1 household or group with 4 back-up households ? At present: 1 household cheaper  non-response results in more bias

11 onderzoeksinstituut Continuous survey or not ? 500 each year or 2500 every 5 year ? Smoother follow-up: every year a 5-year average  Never a full survey of 1 year  Every change in methodology lasts 5 years

12 onderzoeksinstituut Main Results CONTACT PROCEDURE

13 onderzoeksinstituut Incentive or not ? At present: no incentive  lower response rate incentive too expensive compared to expected increase of response rate

14 onderzoeksinstituut Postal contact: ask for phone/GSM At present: not asked Positive reactions according to NKD

15 onderzoeksinstituut First contact: closer to fill-in day At present: – phone call 8-10 days before fill-in day – if Ok, then send diaries Alternative: – send diaries (social pressure) – phone call 1-2 days before fill-in day

16 onderzoeksinstituut Diaries: keep paper en pencil At present: diaries Alternative 1: face-to-face if person level: item response   if household level : overall response   more expensive

17 onderzoeksinstituut Diaries: keep paper en pencil Alternative 2: CATI if person level: item response   if household level : overall response 

18 onderzoeksinstituut Main Results DIARY

19 onderzoeksinstituut 1 or 2 survey days ? At present: 2 survey days  2 nd day biased  too expensive when 1 day is used intra-individual differences

20 onderzoeksinstituut 1 or 2 survey days ? more trips for small groups more trips for multimodal models  not independent  decrease in response… … but not half of response

21 onderzoeksinstituut Proposition: 1 or 2 days If – Budget allowed to increase (= # trips) – Not interested in intra-individual behaviour – Increase of response rate is more important then 1 day survey

22 onderzoeksinstituut 1 or 2 types of questions on work/ school trips ? At present: 1 general question + detailed data in diary information on work / school trips for every respondent more information from trips done…  … but not independent

23 onderzoeksinstituut 1 or 2 types of questions on worktrips ?  Double work for half of respondents intra-individual information  increased burden, decreased response

24 onderzoeksinstituut Proposition: 1 or 2 types of questions on worktrips If – Budget allowed to increase (= # independent data) – Not interested in intra-individual behaviour – Increase of response rate is more important then no general question

25 onderzoeksinstituut Questions on attitudes/opinions or not ? At present: no more attitudes/opinions only facts decrease of answer to social willingness

26 onderzoeksinstituut Keep or skip hardly analysed questions ? At present: detailed questions concerning –Moped, motorcycle,… – type of working contract – full time / part time job – ….

27 onderzoeksinstituut Distance to bus stop, train station: search in external database At present: questions  answer very unreliable

28 onderzoeksinstituut Trips shorter than 100 m: no directives At present: directive: “do not write them down” more room for other trips  suggestion of unimportant trips  burden of decision for respondent

29 onderzoeksinstituut Coding/recoding trips

30 onderzoeksinstituut Coding/recoding trips : keep it Alternative: main transport mode, for details on public transport: survey afterwards (cfr Dutch OVG) Present Flemish OVG: full information multimodal trip  burden for respondent details still remembered

31 onderzoeksinstituut 3 kinds of ‘shopping’ ? at present: shopping smallest burden for respondent  hard to fit in multimodal models

32 onderzoeksinstituut Shopping: a purpose criterium ? Purpose: – weekly necessary grocery purchase – recreational shopping – exceptional purchase Frequency: – ± 1 or 2 per week – ± 1 per month – ± 1 per year

33 onderzoeksinstituut Shopping: see what fits in lay-out In present lay-out: increasing complexity is decreases response If lay-out is changed (price !) then motive shopping reconsidered

34 onderzoeksinstituut Main Results EDITING

35 onderzoeksinstituut Editing: stay conservative At present: replace empty fields only when very sure  Dutch OVG: ± every empty field filled in

36 onderzoeksinstituut Main Results WEIGHTING PROCEDURE

37 onderzoeksinstituut Eduction in weight: if reliable source At present: no education in weight  education most relevant socio- economic variable  no information available about population ? See if ArbeidskrachtenEnquête and Socio-Economic Survey fit

38 onderzoeksinstituut Calculation of weight At present: different calculation of weight in OVG, Mobel, TBE To do: – Try different calculations – Select the one with the smallest St. Dev the smallest range

39 onderzoeksinstituut Main Results NEW TECHNOLOGY

40 onderzoeksinstituut New technology ? At present: paper and pencil accepted by everyone  some prefer internet  no detail on road used

41 onderzoeksinstituut New technology: as an extra  Only internet, PDA,GPS… then bias  internet, PDA,GPS as an extra, then more expensive Wait results of SBO

42 onderzoeksinstituut QUESTIONS ? REMARKS ?


Download ppt "Onderzoeksinstituut Workshop Valorisation of transport models for a sustainable transport Namur 19 December 2005 Travel behaviour survey in Flanders :"

Similar presentations


Ads by Google