Why an external evaluator
Because a program that evaluates itself is a program nobody outside it has to believe.
SAFER's evaluation runs through OSU Extension. A school participating in SAFER contributes de-identified aggregate data to that evaluation.
What the evaluation asks
Four questions, and they are the ones SAFER most wants answered:
- Were the High Risk students already known to staff? This is the question that tests whether universal screening finds anyone a school would otherwise have missed. If the answer is "we already knew about all of them," SAFER has a problem.
- Did students connect to care, and were they still in it at 30 and 60 days?
- Did staff confidence change?
- What did it cost staff, in time and in strain?
What a school contributes, and what it does not
Contributes: counts, rates, distributions, barriers, timing, care continuation, staff survey responses, and the trust score. Aggregate, de-identified, nothing traceable to a student.
Does not contribute: anything from the care coordination file. Not names, not guardian details, not safety plan contents, not clinical notes. Evaluation happens on the de-identified side of the line described in Module 1.2, which is one more reason the line exists.
What is already built in
- Group pre and post surveys
- The student check-out, completed after every safety plan
- The planner debrief, completed after every safety plan
- A school debrief after every screening week
- School counsellor check-ins at 30 and 60 days
What evaluation is for
Not proving the program works and not satisfying a funder, although it does both.
It is for finding out where SAFER is wrong.
The same-day standard may not hold at scale. The two-hour promise may break in a building with one counsellor. The instruments may perform differently across populations than SAFER assumes. Those are answerable questions and SAFER would rather answer them than defend them.
# UNIT 11 – POSTVENTION
SAFER Universal Screening Toolkit. Two Modules. Brief by design.
This Unit is short on purpose. SAFER is not the postvention program.
