TABLE 1 Comparison of Four Designs for Passenger Travel Surveys
TABLE 1 Comparison of Four Designs for Passenger Travel Surveys
| Single cross-sectional | Repeated cross-sectional | Longitudinal panel | Rotating panel | |
|---|---|---|---|---|
| Description | A single sample of households or individuals complete survey during a single period of time | Different samples of households or individuals complete survey for multiple periods of time | Same sample of households or individuals complete survey for multiple periods of time | Same sample of households or individuals complete survey for multiple, specified time periods. Sample gradually, occasionally replaced with new sample (cross between repeated cross-sectional and longitudinal designs) |
| Advantages | . Provides snapshot of behavior for given time period . Typically provides more representative sample of population of interest . Eliminates potential response bias due to respondent "conditioning" resulting from participating multiple times |
. Provides snapshot of behavior for given time periods . Typically provides more representative sample of population of interest . Eliminates potential response bias due to respondent "conditioning" resulting from participating multiple times . Allows for comparisons of population between field periods (assuming similar survey conditions) |
. Allows for analysis of change in behavior of same units due to changes in environment and other factors (cause and effect analysis) . Cost and resource efficiencies in subsequent waves from building off of previous interviews |
. Allows analysis of change in behavior of same units due to changes in environment and other factors (cause and effect analysis) . Allows for long-term analysis of population change (longer than the duration of a longitudinal study) . Cost and resource efficiencies from building off of previous interviews |
| Disadvantages | . Does not permit analysis of changes over time; cannot track trends . Operationally, higher costs often incurred to initiate a project; costs and resources not evenly distributed and maintained . Potential "telescoping" effects |
. Limited change can be implemented across enumerations to allow for population trend analysis . Does not allow for analyzing behavior and change among same sample units . If not continuously conducted or with small time lag between enumerations, cost and resources not evenly distributed and maintained . Potential "telescoping" effects |
. High respondent burden and panel attrition . Response bias due to respondent conditioning . High cost of respondent tracking (e.g., following "movers") . More complicated weighting and estimation . Potential "seam" effects |
Though often less than nonrotating panel, still subject to: . High respondent burden and panel attrition . Response bias due to respondent conditioning . High cost of respondent tracking (e.g., following "movers") . Even more complicated weighting and estimation . Potential "seam" effects |
| Examples | . The American Travel Survey (1995) | . National Household Travel Survey . American Community Survey . UK National Travel Survey . Sydney Household Travel Survey |
. Dutch National Mobility Panel . Puget Sound Transportation Panel |
. German Mobility Panel . Current Population Survey |