Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
9 Cards in this Set
- Front
- Back
|
What are the features of attribute levels? |
Well thought through - not haphazard. Create combinations of alternatives. Includes: Choice (Dependent Variable) -> Train/Bus Attributes (Independent Variable) -> Fare, Time etc. |
|
|
What are the three stages of Survey Design? |
1) Specify the Model 2) Experimental Design 3) Questionnaire |
|
|
How do we 'specify the model'? |
Calculating the determinants to drive the alternatives. I.e. -> Is choice over time or seasonal? Is cost the main influence? Secondary data from groups can help develop hypothesis and understanding attitudes specific to the problem. i.e. Fernbelt Tunnel -> Sensitivity of travel time between Cardiff - Bristol. |
|
|
What goes into Experimental Design? |
Identifying the main attributes.
Value of Time = b time coach / cost. Unlike SI, must be supplied by appropriate levels. Requires background research such as interviews with operators, consumer groups. Maximum and Minimum Ranges required. |
|
|
Why is it important to consider the minimum and maximum levels of attributes? |
Wide range -> would provide more accurate parameters but may be unrealistic. Smaller would have the opposite effect. Must be willing to trade off between combinations of alternatives. But should be plausible and meaningful. Issue is the completeness Vs the size of choice situations. |
|
|
How do we reduce the size of attribute combinations? |
A full combination is called a "full factorial design". Would be perfect but way too large, costly and long to collect. So to reduce we collect a 'fractional factorial design': Reduces orthogonality, less interactions between attributes - requires min of five degrees of freedom. If sill too big....blocking can be used but not for every case. Subsets of choice situations. |
|
|
What is Orthogonality? |
Zero correlation between attributes, all independent from one another. Can be lost by reducing the size of a design if attributes are unevenly spaced, and missing responses because of non-response. |
|
|
What is an efficient design? |
Least standard error possible. Outcomes of survey for model parameters. |
|
|
In general, what is the difference in use between Revealed and Stated Preference? |
RP - Measures actual choice. SP - Measuring sensitivity for change. Increasing effort to combine both for Travel Demand Management. |
