Evidence and evaluation
What we can say, and what we cannot.
Evidence based is used loosely in fatherhood programming. It is a description programs apply to themselves, not a status anybody confers. So here is our position stated precisely enough to be checked.
The short version. Legacy is evidence informed. It has not been evaluated against a comparison group. No study exists. We are not going to imply one does, and the study protocol is published before there are results.
Five frames
What it is built on, and where each one actually operates.
These were not applied to the curriculum afterwards. They were already running in it, written by men who had been in the room. A reviewer is not asking whether we read the literature. He is asking whether the thing has a mechanism and whether we can say what it is.
01 Attachment and the caregiving relationship
The claim. A child's security comes from a caregiver who is available, predictable and responsive. What builds it is not intensity, it is repetition.
Where it lives. Module 2 on being shelter rather than storm. Module 3 entirely, and 3.2 most directly. Module 6.2, where the argument against fear based discipline is that a frightened child conceals.
02 Social learning and modeling
The claim. Behavior is acquired by observation, especially of same sex caregivers, and especially under emotional load.
Where it lives. Module 1.2, which tells a man his child is not learning that he was stressed, they are learning what men do with feeling. Module 4. Module 6.1. Module 7.2.
03 Intergenerational transmission, and its interruption
The claim. Parenting patterns move across generations through both modeling and unresolved experience, and the transmission is interruptible rather than fixed.
Where it lives. Modules 1 and 4 as the spine of the whole course. The family map is a genogram in plain language. Interruptibility is the mechanism and the message at once, and it is the answer to the fatalism these men arrive with.
04 Father identity, and its salience
The claim. Fathering behavior follows from how central the identity is to the man and how clear he is about what it requires. Identity is more malleable than circumstance.
Where it lives. Module 1.1, which spends an entire lesson on the man before any lesson touches the child. Module 10 on shame. Module 12.2, which asks him to read his own week one answer and write forward a year.
05 Ecological systems
The claim. A child develops inside nested systems, and a parent's access to and standing within those systems shapes outcomes independently of anything happening in the home.
Where it lives. Module 9 in full. Module 2.2 on access. Module 5 on the co-parenting subsystem. Module 12 on the father's own support network.
Two mechanisms
The frames explain why the content is right. These explain why the format works.
Implementation intentions
A specific if then plan, made in advance and in the calm, is substantially more likely to be executed than a general intention. That is the entire design of the rage plan in 1.2, and of all twenty four weekly actions, which are physical, dated, small, and reported back. A plan made while angry is a wish is the participant facing statement of it.
A completed experiential learning cycle
Concrete experience, reflection, generalization, active experiment, repeat. Every lesson runs one full turn: teaching, the eight questions, the three reflections, one action in the world, reported back at the top of the next session. The eight minute opening block closes the loop, which is why the manual says never to cut it for time.
Boundaries
What Legacy does not claim.
Saying this plainly is worth more than a claim, because the first thing a skeptical reader does is look for the overreach.
- It is not a clinical intervention and it does not treat anything.
- It is not substance use treatment. Module 11 refers rather than treats.
- It is not a substitute for legal representation, and Module 9 says so on the page a man reads.
- It does not claim to change a co-parent's behavior, and Module 5 is explicit that it cannot.
- It has not been evaluated against a comparison group. No study exists yet, and we are not going to imply one does.
Target population
Both halves stated, because a program that only states one has stated nothing.
A man is eligible if all of these are true
- He is in a father role for at least one child aged 17 or younger. Biological, custodial, legal, adoptive, foster or stepfather, grandfather, non relative male, or male family member in that role.
- He is 15 or older.
- He can take part in the language of delivery, or a delivery in his language is available.
- He can attend the sessions in the configuration being run, or the make up path is open to him.
- Where the referral is court connected, it is documented.
And every one of these is a referral
- An acute safety concern for a child or a partner
- Legacy is not a batterers intervention program and does not substitute for one. Where there is a current, active safety concern, the safety protocol is followed and the man is referred to the appropriate program. Being enrolled in, or having completed, a batterers intervention program is not a bar to participation and is a population this program expects to serve. Men in both concurrently are supported, with coordination between providers where consent allows.
- An acute crisis that makes group participation unsafe
- Active psychosis, acute intoxication at sessions, a mental health emergency. This is a pause and a referral, never a bar. He is welcomed back.
- A court order that prohibits the contact the material asks for
- Verify before any assignment involving a child, and adapt the weekly action rather than excluding the man.
- He will not accept the group agreement
- Specifically confidentiality about the other men. One breach ends participation, and every man is told so before Lesson 1.
- Under 15
- The floor the platform currently sets. Where a younger father needs the material, it is a referral like any other.
Two exclusions this program deliberately does not make. A criminal record is not an exclusion. Being behind on child support is not an exclusion. Both would remove exactly the men this exists for. Nobody leaves without somewhere to go, and the hand off is recorded.
Measurement
Named instruments, and when each one is taken.
Four is the ceiling before a man in week one is filling in forms for forty minutes and deciding this is not for him. Attrition at intake is a bigger threat to an evaluation than any measurement gap.
| Domain | Instrument | When | Note |
|---|---|---|---|
| Father involvement | Inventory of Father Involvement, short form | Baseline, exit, 6 months, 12 months | Covers engagement, accessibility and responsibility rather than contact hours alone. |
| Parenting attitudes and discipline beliefs | AAPI-2 | Baseline, exit, 12 months | The most common instrument in this field, which makes results comparable to other programs. Licensed, with a per administration cost. |
| Co-parenting alliance | Parenting Alliance Measure | Baseline, exit, 6 months, 12 months | Administered only where a co-parent relationship exists. Where it does not, the reason is recorded. |
| Knowledge of curriculum content | Legacy program measure | Baseline, exit | Twenty four items, two per module, drawn from the learning objectives. Scored automatically on the platform, so no facilitator ever grades anything. |
| Father identity salience | Items inside the program measure | Baseline, exit, 12 months | The theoretical frame predicts this moves first, which is reason enough to capture it. |
| Program experience | Short exit survey | Exit | Not an outcome. Used for improvement. |
The comparison group
Designed before the first cohort, because it cannot be built after one.
A new cohort program has more referrals than seats. Where referrals exceed capacity, men are assigned to the next cohort by a documented rule rather than by facilitator preference, and both groups are measured at the same calendar points. Nobody is denied the program. Everybody gets it. The delay is real and would exist anyway.
Where demand does not exceed capacity, a matched comparison drawn from referral partners: eligible men referred in the same period, in the same circuit, who did not enroll. Matched on age, number of children, referral source and justice involvement. Weaker, and still substantially better than a single group pre and post, which is what the field is mostly built on.
Committed in advance. Instruments named, baseline, exit, six months and twelve months, and the findings published whichever direction they run.
Why this page exists