First point SWR are not requesting DOO to be used as general practice (and as I understand it, only on the metro services) and so (assuming that DOO is less safe than having a guard on board, which isn't as clear cut as the unions would like it to be) the risk to the travelling public would be mostly the same as most of the services would still have a guard on board. As an example it is safe to cross the road at a pelican crossing, however for the few times that the pelican crossing isn't working due to maintenance issues or power failure isn't going to result in a significantly higher accident rate.
As an example if we assume (all figures are arbitrary but are possible figures)
- 1 train in 100 being DOO (due to staff not being available for that journey, be that due to other delays, staff sickness, or whatever other reason that is) with an average load of 100 people
- 1 in 5 million rate of accidents for DOO passengers
- 1 in 6 million rate of accidents for passengers with a guard on board
Based on 1.7 billion passenger movements a year nationally that would equate to the difference if we moved to DOO vs guards on every train would mean the difference between 340 incidents and 283 incidents. However, if the proposed was brought in it would be the difference between 286 and 283 incidents (assuming 1/100th of all passengers being on DOO trains with all others being on trains with guards vs all trains having a guard)
However even the above is flawed as there isn't enough data to have a clear figure for rates of accidents, the last data I saw from TfL had it that the accident rate for trains on part of the London Overground network as higher than that for the rest of the network where DOO was being use (IIRC one rate was 1 in 8 million). However even then that TfL data could be wrong as just one indecent could then skew the figures one way or the other (for instance guards have an incident near the start of the time period having been safe for 10 million journeys previously whilst DOO has an indecent just before the start of the time period and then run for 8 million journeys before having another and then guards having another incident after 10 million journeys and then the time period ends. That results in 2 in a 12 million time period for guards whilst the DOO would be recorded as only having 1 in that 12 million range). Even then how do you determine if a guard on the train would be of assistance (for instance someone bashing into a closed door isn't going to be more or less injured if there is a guard to see them do it or not).
Then there's the complicating factor of station staff, if they are dispatching a train does their presence change the rate/risk to passengers and is that more/less than a guard's impact on the rate/risk. Likewise lighting levels, are trains run from well lit stations in urban London more/less likely to see an incident than a poorly lit station in the middle of a national park and is that difference more/less if it is DOO vs a train with a guard? Also train lengths can a guard dispatch a 12 coach train by looking from one end as one with cameras at intervals along its length, and how does that change if there is a bend in the platform? (basically even if there was a lot of data there's a lot of variables and has one of those other variables made the difference rather than if there is a guard or not?)
If it is all about passenger safety then where is the pressure from the unions for all guards first aid trained so as to deal with unwell members of public? As I understand it guards are not all first aid trained. Therefore, given that any member of the public could suffer a heart attack at any moment (which is a LOT more likely than even a near related to train dispatch/platform interface), a guard (if first aid trained) would be more likely to save someones life by performing CPR than stopping a rail related incident. In the event of a major crash, which is often highlighted as where DOO would cause multiple deaths, then having as many people as possible who are first aid trained could reduce the number of casualties. Likewise there could be a situation where a guard is on a near empty train and there are no members of the public with first aid training.
Another question, should all stations that have trains of over (say) 6 coaches in length have monitors to help guards and all stations that have more than 3 trains per hour have dispatch staff, would that make dispatch safer? (like DOO or guards I would suggest that the answer isn't as clear cut as some may like) However, assuming that monitors and dispatch staff would help then why aren't the unions calling for these to be standard methods of working as well to make the railways even safer?
The problem is that by striking over what to many people appears to be minor variation over what happens at the moment and when there are other safety issues that could see a bigger benefit to passenger safety is likely to result in passengers not being overly supportive of such a strike. It may even result in passengers calling for DOO as if they have to suffer the strikes anyway at least have the change so that they don't have to worry about future strikes and lack of train crew causing trains to be cancelled so much.
My personal view is that guards on trains are a good thing and should be kept, the above is just showing that the arguments put forward by the unions aren't as clear cut as they would like them to be.