The National Statistics Quality Assurance Review 2000 of the NTS recommended it should be weighted to adjust for non-response bias. Particular emphasis was placed on this due to the fall in response rates since the early 1990s. Such declines have been observed over a wide range of surveys both in this country and abroad. The complexity of the NTS meant outside expertise was required to investigate weighting methods. DfT commissioned the National Centre for Social Research (NatCen) to develop a methodology to weight the NTS. The project involved several stages of work during which a weighting strategy for the NTS was developed. This report contains analyses of weighted data from the 2002 survey and compares it with unweighted data to gauge the impact of the adjustment. This report shows the effect of weighting on the distributions and absolute values of National Travel Survey (NTS) data, but not on trends. The implication for trends will not be clear until we have weighted data for other years. The aim is to publish data on trends (1995 - 2004) in March 2006. For 2002 NTS data, the key changes from unweighted to weighted data for non-response1 are as follows: • The proportion of men in their twenties increases by around 3 percentage points. The increase is around 2 points for women the same age. The proportion of people over 60 falls by around 3 percentage points. • The proportion of people in London and large urban areas increases by around 2 percentage points. • The proportion of full time employees increases by around 2 percentage points. • The proportion of households with one car falls by around 1 percentage point and households with three or more cars rises by around 1 point. • The proportion of households with older people falls and the proportion of larger households increases. • Despite the increase in household car ownership fewer people have the potential to drive. Cars per household increases but cars per adult falls. Driving licence holding falls in all age groups except 17-20 years olds. • There are fewer main drivers (down by around 1 percentage point) and more other and non drivers in households without a car. The fall in main drivers can be attributed to the increase in the proportion of London households, particularly those in the most deprived households. The households which contribute most to the fall in main drivers are households with adults of a working age with children. • There is a fall in average trips per person per year by around 1 per cent. • The overall fall in trip rates can be attributed to a fall in car trips, particularly car driver trips amongst men. On average, 14 fewer car trips are made by all persons. The fall in car trips is slightly offset by increases in public transport trips, where on average 5 more bus trips are made. • The fall in car driver trips is due to there being fewer drivers after weighting. This is true across all age groups but particularly so for people in their twenties. In this age group there is a shift from people who are main drivers to other drivers and to households without a car. This accounts for some of the increase in trips by public transport. • The fall in car passenger trips is in part due to 17-20 year olds personal car access where there is a shift from non drivers in households with a car to households without a car. This also accounts for some of the increase in public transport trips. • Average trip rates fall by 3.4 per cent amongst 17-20 year old men, 2.1 per cent for men in their twenties and by 3.1 per cent for women in their twenties, and around 1.5 per cent among people in their thirties and forties. The fall among 17-20 year old men is due to there being fewer non drivers in households with a car and more in households without a car. The fall in trips for people in their twenties, thirties and forties is due to there being fewer main drivers. • The rise in public transport trips (up 7 trips) mainly occurs on London buses (up 3 trips) and London underground (up 2 trips). • There is an increase in average commuting trips by 5 per cent. This is due to the proportion of full time workers increasing by 2 percentage points. • There are fewer shopping and non work related trips which are mainly made by car due to the fall in main and other drivers. For 2002 NTS data, the key changes from unweighted to weighted data for non-response and drop off in the number of trips recorded in the travel diary week are as follows: • Average number of trips and distance travelled both increase by 7 per cent. Weighting for non-response reduced trips by 1 per cent so adjusting for drop off increases trips by 8 per cent. This change occurs consistently across various subgroups of the sample. • Walking trips rise by 2 per cent, which is below the overall increase. This is due to most of these trips being short walks which are only collected on the seventh day of the travel diary week so are not weighted for drop-off. Further work will be done to establish if an under recording factor is also needed for short walks. • London bus trips increase by 5 trips per person per year (39 per cent) and London underground trips by 3 trips on average (47 per cent). The large increase in London buses and underground mainly occurs amongst persons without access to a car. • Bicycle trips increase more amongst men (by 4 trips) than women (by 1 trip). This is likely to be due to the increase in the proportion of men under thirty who are more likely to cycle. • The number of long distance journeys increases by a third. This is a result of the weighting process adjusting for drop-off in the reported number of LDJ occurring in the three weeks prior to the travel diary week. (Author/publisher)
Samenvatting