I have done my own Malaysian Driver Behaviour Questionnaires (DBQ) which to me suits the Malaysian driving culture and speeding drivers scope as shown in the link below https://docs.google.com/forms/d/1Iy-39dPof4GrZPtXKl7g2yXLTvTR8Ck_qoqLu5EMb4w/viewform.
The questionnaires were modified from the past and existing questionnaires such as the ones from (1)https://www.duo.uio.no/bitstream/handle/10852/18410/Thexrolexofxpersonalityxandxattitudesxinxpredictingxriskyxdrivingxbehavior.pdf?sequence=1
(2) http://www.nhtsa.gov/DOT/NHTSA/NRD/Multimedia/PDFs/Crash%20Avoidance/Driver%20Distraction/810594.pdf
(3) http://www.nhtsa.gov/DOT/NHTSA/NRD/Multimedia/PDFs/Crash%20Avoidance/Driver%20Distraction/100CarMain.pdf and
(4)http://www.carbibles.com/docs/trl325.pdf.
The questionnaires were rather few and short since the respondents may get psychologically annoyed and exhausted in answering the DBQ and hence providing higher-biased response.
May I know how can I verify or validate the reliability or consistency of my DBQ?
May I know how can I analyse the DBQ responses received?
Dear Raymond,
Firstly consistency and reliability, from your description, seems to mean the same thing. You can perform this using Cronbach alpha. Scale validity, in your case, will be to test each construct in your questionnaire. For example, Driving Experience/Opinion On Driving Behaviour (thus testing the reliability of items included in this construct).
Validation is quite a different process and you first need to validate the questionnaire before you run reliability tests. Also, you might want to do a test-retest of your questionnaire to ensure temporal stability (the reproducibility of responses to the questionnaire).
Only after doing the above you can run reliability tests. Information on questionnaire validation is abundant. But your methods have to correct from the very outset.
Rabin
Thank you for the swift reply Rabin.
Regarding the procedures, can you kindly reassure me are they as follow?
Step 1: scale validity: test each construct of DBQ
Step 2: test retest ensure temporal stability. (Can i carry this out by firstly through online DBQ and secondly before real world test drive?)
Step 3: reliability/ consistency test via SPSS/PASW
Hi Raymond,
You got it correct, but little changes are required:
Step 1 (Scale Validity)
Testing each construct through factor analysis
See if these constructs are generated in the factor analysis (maybe these factors don't really exist, unless if you have adopted a validated scale)
If you have a validated scale, was it used in the Malaysia context, why isn't there a BM version of it (these things will definitely be asked by the reviewer later)?
Step 2 (Temporal Stability)
The answer to your question is no. You have to test it in the same environment. Hence you either do a test and retest online or in the real world. If you do it in the manner you suggest, the reviewer will say the measures are not compatible and you have a flawed research method.
Step 3 (Internal Consistency)
Once you have isolated valid factors in Step 1, you can go on and test the reliability of the various items in the construct.
Hope this helps.
Rabin
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification