That's an interesting and quite specific bit of analysis there in that it says for over 25's and in collisions with vehicles. It is also from 2012 and it says wholly or partly to blame, so that would mean that there was some amount of blame on the cyclist in some of the cases. To give an extreme example of a lorry turning left and killing a cyclist this could be recorded as both 'failed to look properly', therefore the lorry driver is held partly to blame and would be included in those stats. In fact, when you look at the graph for all ages the drivers and the cyclists are about even.... so my statement of cyclists often being at fault would be supported by that. Also bear in mind that those stats don't include cases like the one I linked to as no vehicle was involved.
The law says that all injury collisions must be reported to the Police and as is the case more and more these end up being self reported by the injured party, they do it online and there is very little, if any, police involvement. For those that do have an officer attend then it is that reporting officers' opinion that is recorded as the contributory factors. In the modern police service, about 10-15% of new recruits don't have a driving licence yet these are the ones who will be reporting collisions and giving their opinion on the contributory factors. Do you think that in either of these scenarios the contributory factors are going to be very accurate?
In the case of a fatal, or extremely serious injury, collision there is a very detailed investigation and a report from a Forensic Collision Investigator who is an expert. This investigation produces data that can be deemed as accurate, so any contributory factors recorded for these collisions can be relied upon. This is why it is the only data I use when looking at collisions and trying to prevent them from occurring again.