Devon: Reading Progress Data That Do Not Add Up

Plot it before reacting. Baseline 42; expected slope 1.25/week; after eight weeks the aim line sits near 52. Devon is at 46, a slope of about 0.5/week. Eight points is enough to draw a trend, and the trend is below the aim line - but that is a statement about the plan as implemented, not about the child's capacity.

Check implementation before changing programs. Fidelity was documented in 4 of 12 sessions. Twenty minutes three times weekly in a group of five is a modest dose; absence of documented fidelity means I cannot distinguish an ineffective program from an un-implemented one. Changing programs now would discard the only variable I have not yet measured.

Also check measure–goal alignment. Oral reading fluency is a proxy. Devon's relative strength is word reading and his weakness is comprehension (74) with listening comprehension at 79 - so comprehension is not purely a decoding artifact. A fluency-only goal with a fluency-only measure may be a poor match to the IEP's actual priority, and slope on ORF can plateau for students with autism for reasons unrelated to reading growth.

Decision rule I would put in writing: four consecutive data points below the aim line with fidelity at or above 85% triggers an instructional change - increase dose or group size, add modeling, or change program. Four consecutive points above triggers a goal raise. Right now the trigger is fidelity, not program.

Plan: fix fidelity first (paraprofessional training, coaching, daily log), reduce group to three or add a fourth session, keep the program four more weeks, then reconvene with a documented decision. Include Devon's own perspective and the parent's concerns in that meeting rather than presenting a verdict.

Treat the number as a hypothesis, not a fact. A standard score of 74 is roughly 1.7 standard deviations below the mean, but its meaning depends entirely on conditions of administration and on what the subtest actually demands.

Validity questions

• Conditions: unfamiliar examiner, a 45-minute session, sensory environment (fluorescent hum, transitions), anxiety, and fatigue can each depress performance for a student with autism.
• Task demands: does the comprehension measure require narrative retell of stories with social content, timed responding, or open-ended inference? Any of these can understate comprehension for a student whose receptive language is stronger than his expressive organization.
• Language load versus reading load: listening comprehension at 79 sets a ceiling. Comprehension cannot reasonably exceed oral language by much, so 74 may be consistent with his profile rather than an anomaly.
• Measurement error: use the confidence interval and the standard error of the difference before claiming two scores differ meaningfully. A few points is noise.
• Cross-check: classroom retell, question-answering with graphic support, and curriculum-based measures of the actual content being taught.

Interpretive stance. Converging low scores across reading and listening comprehension point at language comprehension, not at decoding. That reframes the intervention target and the IEP goal - but a single battery should not drive it. I would gather two more sources of evidence, present the pattern to the team, and explicitly tell the family what the scores can and cannot tell us. On the private evaluation request: acknowledge it, route it through the district's process for independent evaluations, do not stop services in the meantime, and treat any outside report as one more data source to be considered rather than as a threat or a mandate.

Depersonalize the data. The fastest way to lose the intervention is to make fidelity feedback feel like an evaluation of the paraprofessional. I would frame the question publicly as "was the plan delivered as written, and did it get a fair test?" - a systems question, not a competence question.

Next two weeks, concretely:

• Observe two sessions and score fidelity alongside the paraprofessional using the existing checklist, coaching in the moment rather than afterward from memory.
• Simplify the log to a one-page daily form: minutes delivered, students present, fidelity yes/no per component, plus the weekly probe score.
• Provide the scripted-program training that was apparently never completed, or identify that it was and document it.
• Keep the reading teacher in the loop with a shared graph so that the program decision is made against a trend line, not a feeling.
• Confirm the probe administration is standardized - same passage set, same timing, same directions. Inconsistent probing can manufacture a flat slope.

Protect the student's interests throughout. Devon should not experience two weeks of adult tension. If the intervention time is disrupted, that is a service-delivery problem the team must document and address.

Model the decision discipline I want the team to use: if, after two weeks of verified fidelity, the trend is still below the aim line, I will be the first to recommend a change - and I will have the data to show the family why the change is warranted rather than arbitrary.