What Ian Walker's study failed to do was measure the amount of traffic, let alone the oncoming traffic. As he did different parts of his data collection on different days and on different roads, this key measure which often affects how far a car is able to move out going past a cyclist is a very obvious parameter to take into account. Ian's data set has times of passes in it so it might be possible to crate a proxy for traffic density but I haven't attempted this. I think it's more valid to conclude this study has no credence than to draw any conclusion from it.