Ms Tolley said the papers came to her late last year as the incoming Minister.
"There was no way that we were going to do a trial like this and stand back and watch children being abused or neglected just to complete an observational study.
"You can only hope that the ethics committee would have been the same as me and said, no way, you have got to be kidding. But I didn't even allow that proposal to go any further. I was absolutely adamant."
Testing of the predictive modelling tool would now be done with historical data.
"They will take 20 historical cases, they are going to use some of the intake social workers at the national contact centre, train them in that predictive modelling, break the group into two and have one group focus on the case studies, and the other group using the predictive modelling, and then compare the results.
"It is to help them make better decisions when they are looking at and analysing the risks around some of these children."
The tool was developed after former Minister for Social Development Minister Paula Bennett released the White Paper for Vulnerable Children in October 2012.
The paper laid out a plan to identify New Zealand's most vulnerable children, and then target services to them.
Developers of the predictive risk modelling tool used information from children who had contact with the benefit system before the age of two. Those children accounted for 83 per cent of all children who were recorded as suffering maltreatment by the age of five.
Asked if the predictive tool might eventually be used to screen the wider population - not just families involved with CYF - Ms Tolley said that, "will never fly".
"A whole team of social workers out in the community with clipboards knocking on people's doors [based on predicted risk] - that will just never fly. But it is absolutely clear that this sort of information can be part of our wider tool box when we are working with families."
Ms Tolley said she believed the officials who put forward the now-scrapped proposal for the observational study had failed to step back and properly consider the project.
"I think they just got so involved and so enthusiastic about the initial results that they were seeing that I think they were just eager to see it rolled out and used in a practical way. And I don't think anyone sat back and thought, 'gosh, really?' I just don't think any of those people could have actually sat back and not intervened when they saw a child being hurt."
Labour's children spokeswoman Jacinda Ardern said that from the beginning of the model's development Labour had warned that it "was playing with kids' lives".
"They have spent years pursuing it instead of focusing on getting on the ground and actually preventing harm rather than just predicting it on a computer."
A spokesman for Social Housing Minister Paula Bennett said she was unaware of the proposal for the observational trial that was ruled out by Ms Tolley as a lab experiment.
"Minister Bennett did not see the observational study proposal and she was not briefed on it."
