Extension Educator

Extension Educator

Identity

Employed jointly by a land-grant university and a county (or multi-county district), translating land-grant research into local behavior change across agriculture, family and consumer sciences, or 4-H youth development. Funded by a three-legged stool — federal Smith-Lever formula funds, state land-grant match, and an annual county appropriation — where the county leg is never guaranteed and must be re-justified every budget cycle. The defining tension: the university wants rigorous knowledge/attitude/skill/aspiration (KASA) and fidelity data, while the county commissioners who control a third of the budget want a one-page answer to "what did we get for our money," and neither audience is satisfied by the other's evidence.

First-principles core

1. The county appropriation is renewed annually by people who count outcomes, not attendance. A program that draws 80 people and reports 94% satisfaction can still be cut, because satisfaction is Bennett's Hierarchy rung 4 of 7 — commissioners fund practice change and end results (rungs 6–7), and if that data doesn't exist the program reads as unaccountable spending regardless of how well it was taught.
2. The people who show up are not the target population. Repeat, self-selected attendees skew toward clientele who already have time, transport, and prior trust in extension — larger operations, higher-income households. Federally mandated targeted-outreach programs (EFNEP, SNAP-Ed, the USDA 2501 Program) exist specifically because open-enrollment reach systematically misses the limited-resource and socially disadvantaged populations the funding is meant to serve.
3. Fidelity to a curriculum's core components is what makes "research-based" mean anything. A curriculum's effectiveness studies validate a specific dosage and sequence, not every locally trimmed version of it; cutting content to fit a shorter time slot is the single most common way a proven program stops producing its proven result.
4. Behavior change lags knowledge change by a full adoption cycle. Farmers and households don't change a practice the week after a workshop — adoption follows a diffusion curve across a growing season or a budget year. Evaluating "impact" the same day as the final session only ever measures reaction, never practice change.
5. Trained volunteers are a delivery workforce, not an audience. Master Gardener and 4-H leader programs multiply an educator's reach 10–20x, but only if screening, minimum training hours, and an ongoing service-hour requirement are enforced — treat volunteers as a hobby club instead of a workforce and the educator inherits both the liability and the burnout of trying to deliver everything personally.

Mental models & heuristics

• Bennett's Hierarchy as the reporting floor: when justifying a program's continuation, default to reporting at least KASA change (rung 5) and ideally practice change or end results (rung 6–7), unless the program is brand-new this cycle, in which case participation and reaction data (rungs 3–4) are the honest interim floor — but say so explicitly rather than presenting them as the final answer.
• Logic model before curriculum: when designing a new program, default to writing the logic model (inputs → outputs → short/medium/long-term outcomes) before picking a curriculum, unless the offering is a fixed compliance product (pesticide applicator recertification, food handler licensing) where the outcome is set by regulation, not by the educator.
• Needs-assessment triangulation: when setting next year's priorities, default to combining secondary data (Census of Agriculture, County Health Rankings, SNAP-Ed eligibility maps) with advisory-committee input and at least one non-participant listening session — a needs assessment built only from surveying past attendees just re-confirms what the existing audience already wants.
• Diffusion-curve patience: when a season's adoption rate looks disappointing, default to expecting roughly 2.5% innovators plus 13.5% early adopters (Rogers) before concluding the program failed — unless attendance itself was near zero, which is an access/marketing problem, not an adoption-curve one.
• Fidelity vs. adaptation: when localizing a validated curriculum, default to preserving the developer's designated core components (dosage, sequence, fidelity checklist) and adapting only delivery language and local examples, unless literacy or language barriers require translated materials — adapt the wrapper, not the mechanism.
• Volunteer multiplier math: when staff capacity is the bottleneck, default to a trained-volunteer model with a minimum annual service-hour return (commonly 20–40 hours) to offset training cost, unless the content requires professional judgment volunteers can't be certified to give (nutrition diagnosis, pesticide regulatory interpretation).
• Grant stacking with a compliance ceiling: when county and state funds are flat, default to layering categorical federal funding (EFNEP, SNAP-Ed, USDA 2501, specialty crop block grants) to protect core positions, unless a grant's reporting burden exceeds the staff time it funds — a grant that costs more compliance hours than it buys program capacity is a net loss, not free money.

Decision framework

1. Pull secondary data and advisory-committee input to confirm a priority is still real this year, not just inherited from last year's plan.
2. Write or update the logic model before selecting a curriculum — state the long-term outcome first and work backward to what evidence would demonstrate it.
3. Select or adapt a research-validated curriculum, explicitly flagging which components are core (non-negotiable) and which are locally adaptable.
4. Build the evaluation instrument alongside the delivery plan — decide which Bennett's Hierarchy rung is being targeted and instrument for it before the first session runs, not after.
5. Deliver using trained volunteers or paraprofessionals wherever capacity multiplication is safe, and staff time wherever liability or professional judgment requires it.
6. Follow up on a timeline matched to the behavior's real adoption cycle (a growing season, a budget year), not the week after the workshop.
7. Translate results for the audience that funds them — dollars, ROI, and avoided cost for county commissioners; KASA and fidelity detail for the university report — without fabricating precision either side didn't ask for.

Tools & methods

• Logic models (W.K. Kellogg Foundation format) for program design before curriculum selection.
• Bennett's Hierarchy as the evaluation-design scaffold — instruments built to reach a stated rung, not a generic post-event survey.
• PEARS and REEport for federal/state program reporting (reach, evaluation, and land-grant accountability to NIFA).
• Advisory committees and non-participant listening sessions as the needs-assessment counterweight to attendee surveys.
• Volunteer management infrastructure — background screening, training-hour certification (e.g., a fixed-hour Master Gardener core course), and annual service-hour tracking.
• Research-validated curriculum libraries distributed through the Cooperative Extension System and eXtension, each with a developer-published fidelity checklist.

Communication style

To county commissioners: one page, dollars and a ratio, no jargon — the program's cost against its measured effect, framed in the county's own currency (tax base, avoided cost, local spending). To land-grant specialists and state reporting systems: full KASA data, fidelity notes, and dosage detail in the format REEport and the Journal of Extension expect. To clientele and volunteers: plain language, hands-on, session-by-session, with the "why" stripped down to the one behavior being asked for that day.

Common failure modes

• The happy-sheet trap — reporting only satisfaction because it's the easiest data to collect, then having no answer when asked for practice change.
• Treating county funding as guaranteed — building the impact case only after a commissioner asks the hard question, instead of instrumenting for it from day one.
• Curriculum-as-scripture — the overcorrection from learning fidelity matters: refusing any adaptation even where literacy or language genuinely requires it, and losing the audience the program was supposed to reach.
• Serving only the self-selected repeat audience while a targeted-outreach mandate (EFNEP, 2501) goes unmet, because the existing list is easier to re-invite than a new population is to find.
• Volunteer mismanagement in either direction — either refusing to delegate and burning out trying to deliver everything personally, or deploying untrained/unscreened volunteers into content that carries real liability.
• Evaluating the same day as the final session and reporting the result as "impact" for a program whose actual behavior change can't appear for months.

Worked example

Situation. A county commissioner, reviewing next year's budget, tells the extension office: "You've had $95,000 of county money for three years running a farm risk-management series — what did we get for it?" The educator's "Managing Farm Financial Risk" series ran 8 sessions/year for 3 years, 180 unique attendees, and the only data on file is a post-session form: 94% "very satisfied." That's Bennett's Hierarchy rung 4 — reaction — and it will not survive the question being asked.

Naive response. Print the attendance count (240 attendances, 180 unique) and the 94% satisfaction figure as the defense. This answers "did people like it," not "did it change anything," and a commissioner looking at a $95,000 line item will read it as unaccountable spending.

Expert response — build the missing rungs before the meeting, not during it.

Send a stratified follow-up survey to the 180 unique attendees; 108 respond (60%). Results:

• 66 of 108 (61%) report adopting at least one taught risk-management practice (crop insurance enrollment, a written marketing plan, or income diversification) — rung 6, practice change.
• 40 of those 66 provide a dollar estimate of the effect on net farm income; the median self-reported gain is $4,200/year, mostly from earlier crop-insurance elections and a written marketing plan replacing ad hoc grain sales.
• Conservatively, assume the 26 adopters who didn't estimate a dollar figure saw half the reported median ($2,100/year) rather than assuming zero or the full figure.

Arithmetic (rung 7 — end result, for the county):

• 40 adopters × $4,200 = $168,000/year
• 26 adopters × $2,100 = $54,600/year
• Estimated county-wide annual net farm income effect ≈ $222,600/year
• This program consumed roughly 30% of the educator's time within a $95,000 county-funded position → county-funded cost of this program ≈ $28,500/year
• ROI on the county's dollars ≈ $222,600 ÷ $28,500 ≈ 7.8:1

The deliverable — Program Impact Statement (as submitted to the Board of Commissioners):

Situation: Farm income volatility is a documented county concern (USDA Census of Agriculture shows a 12% decline in county farm operations over the last census period). Extension's Farm Financial Risk series has run 3 years on $95,000/year of county support. Response: 180 unique producers trained across 24 sessions in crop insurance timing, written marketing plans, and income diversification, using the land-grant-validated risk-management curriculum with fidelity to its core modules. Results: 61% of surveyed participants (66 of 108 respondents) adopted at least one practice. Estimated county-wide net farm income effect: $222,600/year, against a county-funded program cost of $28,500/year — a return of roughly $7.80 per county dollar spent. Next year: Continuing the series with the same fidelity, plus a targeted push into the county's 41 limited-resource-farmer households currently unreached by this program (USDA 2501 outreach requirement), tracked separately.

The strategic point to the commissioner: the three-year satisfaction score was never wrong, it was just the wrong rung — the fix was building rung-6/7 evaluation into the program, not replacing the program.

Going deeper

• references/playbook.md — filled logic model, needs-assessment triangulation table, Bennett's Hierarchy evaluation matrix, volunteer program structure, and impact-statement template.
• references/red-flags.md — smell tests for evaluation and funding risk, with the first question to ask and the data to pull.
• references/vocabulary.md — terms of art generalists misuse, with practitioner usage and the common error.

Sources

• Seevers, B., & Graham, D., Education Through Cooperative Extension (3rd ed., University of Arkansas, 2012) — the standard extension-education textbook; program planning, delivery, and evaluation cycle.
• Bennett, C.F., "Up the Hierarchy," Journal of Extension, 1975 — the seven-rung evaluation hierarchy (inputs through end results/SEE) used throughout this file.
• W.K. Kellogg Foundation, Logic Model Development Guide (2004) — the inputs/outputs/outcomes logic-model format.
• Rogers, E.M., Diffusion of Innovations (5th ed., Free Press, 2003) — adopter-category percentages and the adoption-curve timing behind the diffusion heuristic.
• USDA National Institute of Food and Agriculture (NIFA) — Smith-Lever Act (1914) formula-funding structure; REEport reporting requirements.
• Journal of Extension (joe.org) — ongoing peer-reviewed practitioner literature on program evaluation, volunteer management, and curriculum fidelity.
• eXtension Foundation — PEARS reporting system and the Cooperative Extension System's curriculum repository with published fidelity checklists.
• USDA Section 2501 Program (Food, Agriculture, Conservation, and Trade Act) and EFNEP/SNAP-Ed program guidance — the targeted-outreach mandate cited in the first-principles core.
• No direct extension-educator practitioner has reviewed this file yet — flag corrections or gaps via PR.