Attendance Falls From 22 to 6: Fidelity or Access?
A 12-session evidence-based parenting program with a published fidelity checklist is being delivered in a rural agricultural county. Sessions run 6:00–8:00 p.m. Wednesdays for two hours at the elementary school, in Spanish, with one bilingual facilitator and an interpreter for a Mixteco-speaking subset. Enrollment was 22; attendance fell to 9 by session 4 and 6 by session 5. Parents tell the family advocate that the school feels 'like a place where we get in trouble,' that the start time collides with the second shift, and that the workbook has no pictures and asks them to write paragraphs. The developer's license counts a participant as a completer only at 10 of 12 sessions, and the funder's contract pays on completers.
1. Is this a fidelity failure, an engagement failure, or both? Lay out how you would tell the difference before changing anything.
Read the worked answer
The data pattern points to an access/engagement failure, not a content failure - but you must verify that before redesigning.
The distinction that organizes the whole analysis:
- Core components (what the program is): the parent–child interaction skills content, guided practice, home assignments, group process. These are what the evidence attaches to.
- Surface structure (how it is delivered): schedule, setting, language, literacy level, group size, facilitator ratio. These can usually move without breaking the evidence base.
Diagnostic steps this week:
- Exit interviews with the 16 who dropped. Ask what changed, not what they thought of the content. If they stopped because of shift schedules, transportation, or discomfort at the school, it is access. If they stopped because sessions felt irrelevant or they could not follow the material, it is content and possibly fidelity.
- Look at when they dropped. Attrition concentrated between sessions 3 and 4 often tracks the first home-practice assignment - a content or literacy barrier. Attrition from session 1 tracks logistics.
- Check what facilitators actually delivered. Supervisor observation against the checklist for sessions 2–4. If facilitators skipped practice activities for lack of time, you have a real fidelity problem layered under the attendance problem.
Reasoning: the most common error here is to start adapting content when the barrier is a 6:00 p.m. start time and a school building. Diagnose before you modify, because every modification has to be justified to the developer and the funder.
2. Propose up to five adaptations. For each, say whether it is fidelity-consistent or fidelity-threatening, and how you would document it.
Read the worked answer
| Adaptation | Fidelity status | Rationale |
|---|---|---|
| Move to a trusted community site (church hall, housing authority room) with a parent host | Consistent - surface structure | Removes the institutional threat; content unchanged |
| Offer a Saturday morning cohort, or split the 2-hour block with a meal break | Consistent, verify | Addresses second-shift conflict; check whether session length is a listed core element |
| Childcare, a meal, and gas cards or a van route | Consistent | Directly targets the stated attendance barriers |
| Redesign the workbook to low-literacy format with photos and drawing/verbal response options | Consistent if the home-practice behavior is preserved; threatening if the writing is the mechanism | Ask the developer; retain the skill practice, change the recording format |
| Add a second bilingual facilitator so the Mixteco-speaking subgroup is not routed through a third-party interpreter | Consistent | Interpreter-mediated delivery of a group-process program degrades the group process itself |
Documentation: use a structured adaptation log - what was changed, why, at what level (content / format / setting / provider), when, and by whose decision. Keep the developer's written approval attached to each content-level change.
Cut line: changing the number of sessions or deleting skills-practice content would be fidelity-threatening and would undercut the reason for choosing this program. Do not do those to fix an attendance problem.
3. The funder pays per completer and the developer counts completion at 10 of 12 sessions. What do you do about that contract, and what do you tell the developer?
Read the worked answer
Separate the three contracts - with the developer, with the funder, and with participants - and renegotiate the ones that create a perverse incentive.
- With the funder (act first, in writing, before the next cohort): request that payment be tied to enrollment plus documented engagement (e.g., sessions attended per enrolled family, or a graduated rate for 6–9 sessions) rather than to completers. Offer the data that makes the case: the 6:00 p.m. Wednesday slot is structurally incompatible with second-shift work in this county, so a completer-based rate punishes the provider for a design flaw, not for poor performance. Propose the revised metric as a contract amendment for year two and offer to report both measures in the interim.
- With the developer: request written guidance on (a) whether the adaptations above fall within licensed flexibility, and (b) whether a documented 'completer' definition can be adjusted for participants who miss sessions for documented shift work, e.g., a make-up session. Do not assume; get it in writing.
- Internally: never let the payment metric drive clinical decisions - do not keep families enrolled in a cohort that is failing them just to protect the completion rate, and do not deny a make-up session because the license is silent.
Reasoning: the funder's metric is a real constraint but it is negotiable; the developer's license is a real constraint and is usually narrower than people assume. Handling the funder conversation proactively, with dose data, is what keeps a defensible adaptation from looking like excuse-making.