Research and narrative for the work of building community.
We find out what your program is actually achieving — and turn it into evidence your board, your funders, and your council can act on.
Why Edict
Recreation is an essential public service. It deserves to be governed and funded with that understanding.
Those of us who work in community programs know what it takes to give people a place to belong: sustained attention, trusted relationships, and services people can count on. That work is part of how a city cares for its residents.
Research
Authority over these services does not require experience delivering them.
The people closest to the work must explain its value to the people who control its resources — often without access to the research and analytical support needed to make that case.
Edict Economics starts from that imbalance. We come from community work, and we believe the people doing it deserve a stronger voice in the decisions that shape it. So we put economic analysis and statistical modeling to work on your own records — to find what your program achieves, where it can improve, and what the evidence will support.
Narrative
Give people a reason to fund the work.
Councilmembers and board members hear more need than they can fund, and their task is discernment. A finding that takes ten minutes to explain rarely survives that room.
So we lead with the result: what changed, what it is worth, and how sure we are — in language a board member can repeat to the next person who asks. The studies below show what that looks like in practice.
Our Services
Three purposes, illustrated by actual work.
Each service answers a different question, and each ends with something you can put in front of a board, a funder, or your own staff — built from the records you already keep. Every finding below came from data an agency already held.
01
Program Analysis & Improvement Plans
“What could we do differently, and where should we focus?”
Find the stage where your program loses people, and aim the fix there. A focused review of one program, or one part of it: where participants drop out, who your outreach reaches, and which step costs you the most.
Considering similar work? Let's discuss your program.
Seattle Parks and Recreation — Career Explorations
Applicant pipeline and school reach review
The question
Where do applicants fall out of the process, and who is our outreach reaching?
What we found
404 applied. 98 finished.
The program keeps most of the people it places. It loses the most before placement. That tells management where to look first — and where better record-keeping would pay off, because the files do not say why those 189 were not placed.
For your program: any service with an application, referral, or intake step can be mapped the same way, so the next funding request points at a specific stage instead of a general need.
Show the full analysisShow less
Applicant pipeline, 2025–26 cycle
- 404 applied
- 215 offered or waitlisted
- 133 accepted
- 98 completed
The offered-or-waitlisted stage combines two decisions: 198 were offered a place and 17 were waitlisted only. Of the 198 offered, 133 accepted and 65 did not — the reasons for non-acceptance are not recorded. After acceptance, 35 records have no completion, and January–February accounts for 16 of them.
The two rates use different denominators on purpose. 24.3% (98 of 404) answers "what happens to an applicant." 73.7% (98 of 133 accepted) answers "what happens to someone who takes a place." They measure different things and should not be used interchangeably.
Who applies
Applicants can report more than one race or ethnicity, so these tags overlap: 39.1% African, 34.9% Black, 22.0% Asian or Pacific Islander, 13.1% Hispanic or Latino, 12.4% White. Somali, Spanish, Amharic, Tigrinya and Oromo are the largest named home languages. 7.9% request an accommodation, and only 10.9% have been through a previous Parks teen program — most applicants arrive new to the department.
School reach against enrollment
Five high schools send 67.6% of applications and produce 56.1% of completions. Measured against each school's own student body, Black and African applicants are over-represented at every one of the five:
| School | Applications | Gap |
|---|---|---|
| Franklin | 88 | +29.8 pts |
| Chief Sealth | 76 | +30.6 pts |
| Ingraham | 45 | +52.2 pts |
| Garfield | 32 | +48.1 pts |
| Cleveland STEM | 32 | +41.2 pts |
Enrollment benchmark: SPS 2024–25; school comparisons use one benchmark category per record. Hispanic share is higher only at Ingraham — the outreach gap is campus-specific, not general. These are differences in representation, not evidence of selection.
Completion by applicant attribute
Among those placed, 76.3% (58/76) of she/her applicants finish against 67.3% (35/52) of he/him; 76.5% (62/81) of non-English home language against 68.6% (35/51) English-only. Applicants reporting no transportation finish at 25% — but that is one completion out of four, so it is flagged as a support question, not a rate.
The records show where applicants stop, not why. Accepting a place does not prove attendance, and no completion recorded does not by itself mean a dropout. The next-cycle recommendations follow from that: broaden school referral routes, track each offer and response separately from the waitlist, record first attendance and check in when it stops, and keep every record linked so reasons are captured rather than assumed.
Produced in the course of employment with Seattle Parks and Recreation. Aggregate figures only; no applicant is identifiable, and denominators are shown wherever a rate is quoted.
02
Outcomes & Value Evaluation
“What is the program achieving, and what does the evidence support?”
Measure what your program produces — and, where the evidence supports it, what that is worth — so it can compete for funding on the same terms as the priorities around it. Built to hold up when a funder or committee asks, and to stay useful to the staff running the work.
Considering similar work? Let's discuss your program.
Seattle Parks and Recreation — Rec N' The Streets
Recreation as public-safety infrastructure
The question
Does bringing recreation into a park reduce violent crime?
What we found
Recreation rarely gets counted in a public-safety budget. This puts it there, with an estimate of what activating a park prevents and what that avoids in response costs — a benefit no attendance count would ever show. Underneath it: an estimated 0.49 fewer violent incidents per activated park-season, against a baseline of about 0.63.
For your program: if your work competes with enforcement or capital projects for the same dollars, this is the kind of evidence that lets it compete on equal terms.
Show the full analysisShow less
How the savings are calculated
| Effect per activated park-season | 0.49 incidents |
| Activations per year | ~176 |
| Violent incidents prevented annually | ~85 |
| Cost per violent incident | ~$15,000 |
| Annual cost avoided | ~$1.27 million |
The per-incident figure covers policing, courts and processing, drawn from the cost-of-crime literature. It is an estimate of costs avoided under these assumptions — not an audited saving, and not money recovered.
It also steadies the trend
A second specification tests whether RNTS is associated with how crime moves, not just how much there is. It finds a significant reduction in volatility: smaller increases when crime is rising, and larger declines when it is falling. For an operations manager that matters — a park that does not spike is easier to staff.
This is a short-run estimate covering the seasons programming is active, not a permanent effect. RNTS sites were not randomly assigned, and the roughly twenty park clusters are few enough that inference required a wild cluster bootstrap. The fiscal figure is a cost-avoidance estimate built on an assumed per-incident cost, not an audited budget line.
Prepared for Seattle Parks and Recreation's Outdoor Activation and Mobile Recreation program by Benjamin Scott, with methodological feedback from Dr. Plakias, Western Washington University, and research support from Edict Economics. Produced in the course of employment, not as a commissioned consulting engagement.
Seattle Parks and Recreation — RNTS Civic Engagement Study
Recreation infrastructure and civic turnout in Seattle
The question
Is nearby ballot access associated with greater turnout growth where recreation programming is present?
What we found
Both associations are statistically significant. The difference between them is not conclusive — but a dropbox associated with more than three times the turnout growth where recreation was active raises a question almost no one is asking about what parks do for civic life.
For your program: findings like this reposition recreation from amenity to civic infrastructure, a frame that opens different budgets and different allies.
Show the full analysisShow less
1,001 precincts, weighted by registered voters, with a ten-variable tract-level demographic control vector.
| Condition | Effect | p |
|---|---|---|
| Where recent RNTS activation is present | +3.33 pts | 0.016 |
| Where it is absent | +1.00 pts | 0.013 |
The gap between the two — the interaction term, +2.33 points — carries p = 0.103, above the conventional threshold. Each effect is individually significant, but the difference between them is not established by this evidence.
Two further results support reading the model seriously. The controls behave as the participation literature predicts: bachelor's-degree share positive, age 65-plus negative, median income negative, all at p < 0.001. And in a separate model, proximity to a permanent community center shows no independent turnout association at all (−0.33 pts, p = 0.726) — a null reported as a null.
The overlap group is small, and neither RNTS sites nor dropbox locations are randomly assigned. The defensible reading is that activation appears to complement formal voting access, while a durable facility's civic contribution is not separable from its neighborhood in a single-year cross-section.
Research note prepared for the RNTS Civic Engagement Study, on behalf of Seattle Parks and Recreation. Seattle precincts, 2023–2025, using King County Elections precinct results.
03
Policy Proposals & Strategic Plans
“What approach should we examine, what would it require, and how could it be implemented?”
A full treatment of a defined problem: what the problem is, what the evidence says, which approaches are available, what each would cost and who would bear it, and how the chosen one could be implemented. Benefit–cost analysis — weighing total costs against total benefits — is included where the question calls for it, not by default.
Considering similar work? Let's discuss your program.
Research collaboration with the Washington Department of Fish and Wildlife
Targeted habitat acquisition for western gray squirrels, Klickitat County
The question
Which land purchases could protect the most connected habitat within a fixed budget?
What we found
Three parcels. Half the habitat. Under $1 million.
The three-parcel package costs an estimated $925,788 and holds about 58 acres — 50.4% of the credited habitat in the narrower 15-parcel portfolio built from 2018–2020 survey records.
The analysis names specific acquisition candidates and shows how a small starting package could secure a large share of the modeled habitat before a larger commitment is made. All of it is existing habitat identified for protection, not new habitat created.
For your program: the same method — maximize what a fixed budget secures, then find the smallest package that keeps most of it — applies to any land, facility, or site-selection decision.
Show the full analysisShow less
Portfolios under each ceiling
| Scenario | Parcels | Credited acres |
|---|---|---|
| $18M, full historical record | 45 | 2,055.06 |
| $50M, full historical record | 124 | 3,306.38 |
| Either ceiling, 2018–2020 records only | 15 | 115.09 |
Ceilings are study scenarios, not verified appropriations. Each retains a 10% reserve. Credited acres are existing potential habitat — not newly created forest, and not confirmed occupancy.
Including 30 years of stewardship in the cost narrows each portfolio: 37 parcels and 1,799.75 credited acres under the $18M ceiling, and 102 parcels and 3,107.36 acres under $50M.
How the parcels were selected
County mapping was normalized into 21,341 unique parcels, with public, Tribal and conservation holdings screened out. Forest Service canopy and LANDFIRE vegetation layers on a shared 30-metre grid identify plausible primary habitat: 45–75% canopy, trees at least 15 metres, and one of five conifer systems. Habitat patches earn credit only where they connect edge-to-edge to an occurrence footprint through purchasable land — and because a parcel cannot be bought in pieces, the model pays for the whole parcel to secure the habitat inside it.
Mixed-integer optimization then maximizes credited connected habitat within the ceiling, and minimizes cost without giving any habitat back. Testing all 32,768 subsets of the 15-parcel portfolio shows five parcels retain 76.7% of its habitat at $2.64 million, and eleven retain 95.8% at $5.02 million.
What it costs, and how confident that is
Land prices come from 383 screened sale observations fitted to acreage, location, date and land-use category, with a planning formula of 1.02 × estimated land price + $15,000 + $25 per assessor acre. Spatially held-out validation returns a median absolute percentage error of about 67% — roughly half of held-out predictions miss the actual sale price by at least that much. It is a screening tool to order appraisals, not a substitute for them.
Acquisition changes who controls a forest; it does not make trees taller. The model credits zero newly created habitat, and its 20% avoided-loss illustration over 30 years is an assumption, not a measured parcel risk. The occurrence record is thin and dated: of 869 county features, 806 are from 1994–1996 and only 24 from 2018–2020, and those 24 are buffered survey transects rather than nest points. Screened habitat still requires field inspection for trunk diameter, nest use, mast production and actual crown connections.
Research collaboration with the Washington Department of Fish and Wildlife. The department assigned the original policy brief and used it in support of its own work; this expanded acquisition analysis builds on that brief. Conducted outside a consulting contract. Parcel identifiers are research outputs, not properties known to be for sale.
A clear scope. A price agreed in advance.
Every engagement begins with a free conversation about the decision in front of you. We review the records you hold, then agree the question, deliverables, timeline and fee in writing before work starts. If the evidence cannot answer the question, we say so then, and suggest what would.
The figures above are estimates for comparable work; they are not records of fees paid. Your quote reflects the question, the evidence available, the state of your records, and what we agree to deliver — well-organized data lowers the preparation cost.
Contact
Tell us what your program needs to understand.
Bring the question, not a research plan.
Benjamin Scott
Founder of Edict Economics
Benny came to economics as a community developer first: three years in community recreation with Seattle Parks and Recreation, including Rec N’ The Streets, alongside a bachelor’s degree in economics. He works on programs he understands from the inside, and brings the analysis to test and explain them.
Also building edictusa.com, a civic-data project under Edict LLC.
- contact@edicteconomics.com
- Phone
- 206-297-6883
- Location
- Washington State — serving public and community partners