Silverdale · Since 2018
The Heart
of GRACE
Five researchers. One question: can we measure when people actually do better?
The question
Better should be demonstrable.
The five researchers who began GRACE did not start with a finished picture of a better society. They started with outcomes: reported safety, social isolation, relationship stability, conflict resolution, community participation, wellbeing, household stability and recovery after disruption.
If they claimed a condition was better for people, they believed they should be able to show that people actually did better under it. If the outcomes did not improve, the claim had to change.
We did not decide what a better society should look like. We learned to measure when people actually did better.
Why is it called GRACE?
There were several working names early in the project. Grace was the one they kept returning to. It felt closer to what they hoped to build: something calmer, more considerate and more human than the systems that came before it.
Eventually it stopped being a working name and became the name. There is no technical meaning hidden in the letters.
Meet the Heart
The Heart of GRACE.
Five disciplines. One continuing argument.
They agree that claims should answer to outcomes. They do not always agree on what a pattern means, which is precisely why interpretation remains human.
Dr. Elena Vance
Neurology and longitudinal measurement
Vance studies how stress, attention, isolation and recovery change over time. Her work gave GRACE a way to distinguish a passing signal from a sustained outcome—and made her an advocate for measuring again before drawing a conclusion.
Dr. Sarah Jenkins
Culture and social adaptation
Jenkins examines how the same behavior can mean different things across families, communities and generations. She helped establish that a pattern is never an instruction by itself: people must interpret its context.
Dr. Amara Okafor
Behavioral observation
Okafor developed much of GRACE’s observational method. Warm, exacting and interested in ordinary life, she believes careful attention can help people live better—provided observation remains open to correction.
Dr. Elaine Foster
Psychology and human judgment
Foster’s work focuses on the boundary between useful context and intrusive certainty. She shaped the separation between diagnostic information and human judgment, including disclosure rules she acknowledges are difficult to live with.
Professor Theresa Sterling
Sociology and outcome evaluation
Sterling asks whether a promising pattern survives contact with households, friendships and actual days. Knowing why people behave as they do, she readily admits, has never made her immune to doing it herself.
Before GRACE
Selena taught us where measurement ends.
Selena was a catastrophic failure. The Heart of GRACE accepts responsibility for it without reducing it to a difficult pilot or an imperfect first attempt.
Selena moved too directly from observation to prediction to intervention. It assumed that a sufficiently accurate model of human behavior could determine the response most likely to improve a human life. That assumption failed.
The observations remained valuable. What changed was the boundary around them. GRACE can identify a relationship. It cannot decide whether that relationship is desirable. That remains a human judgment.
Observation is not judgment.
The method
What GRACE actually does.
GRACE exposes patterns and measures outcomes. It does not turn a statistical model into a moral authority.
- 01
Observation
GRACE records relevant social and behavioral conditions.
- 02
Statistical analysis
Patterns are compared with longitudinal outcomes.
- 03
Human interpretation
People consider context the model cannot judge.
- 04
Human judgment
A person decides whether and how to respond.
- 05
Measurement again
GRACE measures what happened afterward.
The models gave us somewhere to begin. Silverdale taught us what worked.
What the data changed
Responsibility is not always symmetrical.
The researchers tested systems in which partners received the same information and carried identical responsibilities. Over time, complementary responsibilities produced more reliable outcomes.
Women receive contextual and diagnostic information. They must interpret it and carry responsibility for what they do with it. Men serve as Anchors: informationally independent reference points with substantial responsibility for stability in relationships, households and communities.
Neither role is more important. An Anchor carries less diagnostic information. He does not carry less responsibility.
Read the Anchor Principle →Why women see more
For this task, the signal was more reliable.
Real-world research provides a limited starting point. A 2025 meta-analysis covering 1,011 studies, 1,188 effect sizes and 837,637 participants found a small but highly consistent average female advantage in decoding affect cues (d = 0.24).
This is a population-level average with enormous individual overlap. It does not show that women are morally better, naturally wiser or invariably more perceptive. It does not validate GRACE’s architecture.
Within Selena and Silverdale records, a narrower finding persisted: when contextual observations were compared with later outcomes, female assessment provided a more reliable predictive signal at population level. Alternative configurations were tested; the effect remained useful enough to shape the diagnostic HUD.
Not because women are better. Because, for this particular task, the signal was more reliable.
Why information stays asymmetric
Measurement changes behavior.
People respond to being evaluated. When they know the precise indicator being measured, they may begin to manage the indicator itself. This is known as measurement reactivity.
Selena and later Silverdale observations found that direct diagnostic feedback made scores rise faster than safety, trust or relationship stability. People became better at GRACE without becoming better for one another.
Women may use HUD diagnostics as context for judgment but may not disclose that information to the person concerned. The rule is difficult for women as well as Anchors. Once hidden measurement context is disclosed, subsequent behavior is no longer an independent signal.
Score integrity
The score serves the outcome. The outcome does not serve the score.
When scores have consequences.
A social indicator with no relationship to real outcomes is only a number. If scores rise while safety, wellbeing, relationship stability or recovery do not, GRACE treats that as a failure of measurement.
Knowingly disclosing diagnostic HUD information compromises the observation period. The reduction to 300 reflects that loss of observational confidence. It is not intended as a statement about anyone’s worth, though it has real consequences and the Heart understands why residents may experience it as punishment.
We understand why that distinction can feel academic when the consequences are personal.
Living with the evidence
Difficulty is data too.
A population-level improvement can coexist with individual difficulty. An Anchor may carry responsibility without seeing the full context. A woman may see information she cannot share. Couples may have to live with an asymmetry neither finds easy.
Acknowledging that difficulty does not invalidate the evidence. Evidence does not make the difficulty unreal. Both belong in the next measurement.
We live here too
The five members of the Heart live in Silverdale. They have scores, households, colleagues, friendships and relationships. They are observed under the same rules and limitations, without an administrative exemption.
If a principle is only acceptable when its designers are exempt from it, it is not a principle they are willing to defend.
A conditional confidence
The evidence is never finished.
GRACE is confident where outcomes have remained consistent. That confidence is conditional. New evidence can change a model, a threshold, a responsibility or the system itself.
Every result returns to observation. Every conclusion remains available for revision.
We began with five people and a hypothesis.
What GRACE is today, we learned with Silverdale.
Research foundations
A real starting point.
These studies establish real findings about affect-cue decoding and measurement reactivity. They do not test or validate GRACE. The Selena and Silverdale findings on this page are part of Silverdale’s history.
Social stability
The Anchor Principle
Why male stability carries weight beyond the individual, and how Anchors help stabilize households and communities.
Explore the principle →