When ScotRail train drivers make confidential reports on safety problems to Strathclyde University researchers, they can be sure that that confidence is respected.
The research findings are protected by burglar alarms, with the database kept on a computer that is not networked to prevent hacking. Information is taken off the computer each night and kept in a safe. And not only are names removed from reports, but places are also noted by a code known only to the two researchers, to avoid identifying drivers on the network's less busy routes.
"The data are sent back to management, but in such a way that no disciplinary action is possible," said John Davies, professor of psychology.
Strathclyde's Confidential Incident Reporting and Analysis System (CIRAS)is a two-year pilot project designed to gave a better picture of potential safety problems on the rail network. It is backed by a range of bodies, including Scotland's main train operating company, ScotRail, Railtrack, the Health and Safety Executive, and the rail unions.
Its main objective is to encourage drivers to report incidents, whether actual or "almost happened", which may not reach the existing standard reporting system, either because drivers fear they may incriminate themselves, or feel a report would be pointless.
Confidentiality is of key importance, Professor Davies said. In any official reporting system that may have disciplinary implications the tendency is to report only incidents which cannot be covered up, to blame technical failures or other people.
"Our system allows drivers to tell us about unsatisfactory situations or mistakes they've made. This means that in some instances the problem looks totally different, and it's often cheaper to solve," he said.
The Strathclyde team is also using the reports to help develop theories on what part human error plays in potential accidents. This is an increasingly important factor as technological advances reduce the chance of technical problems.
Drivers throughout the network can send in reports to CIRAS, and are generally happy to have a follow-up interview.
"They write about a particular incident, but during the interview may be five or six different issues come up," said research fellow Linda Wright.
The CIRAS database allows the researchers to spot groups of related incidents or trends which have the potential to lead to accidents. A liaison group, with representatives from ScotRail, Railtrack, the rail unions, the Health and Safety Executive and the university, then lobbies for appropriate action.
For example, certain lines are sanded because of problems with leaves, but the Strathclyde team may recommend that this be extended to other areas.
The Strathclyde team says that slips and lapses are more likely when an established routine is changed slightly than when a whole pattern changes.
The researchers produce a quarterly journal that goes to every driver in Scotland, outlining the incidents reported over the period, as well as what action has been taken.
"Even if we'd like something done but it's not practical at present, we want to be able to say that. It's important that drivers know their reports aren't being ignored," said Ms Wright.
"I think the fact that ScotRail are taking this project forward shows their dedication to safety. They're willing to have problems that normally don't come to light identified so that they can do something about them."