Here we specifically compare the data from white LGVs to the data from black cars. Black cars were chosen for the comparison because the number of observations in the dataset was almost the same as the number for white LGVs (although this nicety was hardly necessary as the results below barely change if all car colours are combined).
The mean proximity of overtaking black cars was 1.36 m (SD=0.37) whereas the mean proximity of overtaking white LGVs was 1.26 m (SD=0.35). In other words, white LGVs on average passed 10 cm closer to the experimenter’s bicycle than black cars across all the overtaking incidents recorded. This difference is statistically reliable, t(376)=2.69,p=.007 (two-tailed, performed on the square-roots of the overtaking proximities to correct for skew).
The data showed very similar distributions with a simple shift in means between the two groups of vehicles. Whether this shift is the result of personality factors, vehicle factors (systematic difficulty judging width), or social factors is a question for future research. Potential social factors could include a macho subculture, time pressures associated with driving for work, or the relative anonymity provided by a (usually unmarked) white LGV (for a discussion of anonymity and public behaviour see Festinger et al., 1952, and several subsequent decades of social psychological research).
This Observation is part of a much larger study being completed at the time of writing which systematically assessed several factors affecting how motorists overtake cyclists. This research was funded by the UK Engineering and Physical Sciences Research Council. The instrumented bicycle was constructed by Jeff Brewster.