Writing IEP Goals from Language Sample Data: A 2026 Guide for School SLPs

The Individuals with Disabilities Education Act (34 CFR 300.320) is unambiguous about what an IEP goal must be: a measurable annual goal designed to meet the child’s needs that result from the child’s disability so they can be involved in and progress in the general curriculum. The word "measurable" is the load-bearing one. A goal that says "Marcus will improve expressive language" is not measurable and will not survive a serious due-process review. A goal that says "Marcus will increase his MLU in morphemes from a current baseline of 3.4 (SUGAR conversational sample, 50 utterances, 03/2026) to 4.6 (within one standard deviation of the SUGAR age-5 mean of 5.1) by the end of the IEP year, measured by quarterly 50-utterance samples scored with the MLU Calculator on conductscience.com" is the kind of goal a hearing officer reads twice and approves.

The difference between those two sentences is not vocabulary; it is data. Goals built on a defensible language sample baseline carry the four things every measurable goal needs: an objective baseline number, a target tied to a published normative reference, a clear measurement method that can be repeated across the year, and a mastery criterion specific enough that two clinicians scoring the same probe arrive at the same answer. Standardised tests provide the first three of those things only at evaluation; they do not generate the kind of repeatable, low-cost progress probes that the rest of the IEP year needs. Language samples do, and that is the structural reason every district SLP eventually ends up writing some version of LSA-anchored goals.

The other reason LSA-anchored goals matter is that they make the present-levels paragraph and the goal sentences refer to the same data. A team that walks into an IEP meeting holding a 50-utterance sample, an MLU value, an NDW count, a PGU percentage, and a Brown’s stage label does not have to guess what the child’s language profile looks like; the numbers are right there on one page. The same numbers then become the baselines for the goals, the targets become the next normative band the child is reaching for, and the year-end probes become the same scoring tools applied to a fresh sample. Everything traces back to a single data source. This pillar walks through every step of that pipeline and points at the specific free tool on this site that runs each step.

The pipeline from a scored transcript to a finished IEP goal is shorter than most clinicians expect: five steps, each of which has a specific free tool attached to it on this site. The five steps are (1) score the transcript and extract the metric values, (2) cross-reference the metrics against published age norms and decide which area the child is most behind in, (3) pick the goal area and look up the matching SMART template, (4) write the present-levels paragraph using the metric values as the supporting evidence, and (5) write the annual goal sentence with a baseline tied to the metric, a target tied to the normative band, and a measurement method that re-runs the same calculator on a fresh sample.

Step 1 is the calculator step. Take the cleaned, segmented transcript, paste it into the MLU Calculator for MLU-m and MLU-w, the Lexical Diversity Calculator for NDW, the DSS Calculator or IPSyn Calculator for syntactic complexity, and the PGU Calculator for grammaticality. Each tool returns a number in under five seconds. Write all of them on one row of a spreadsheet alongside the child’s name, age in years and months, and the date. This row is the structured baseline data the rest of the pipeline depends on, and it is the row a clinician will paste into the present-levels paragraph an hour later.

Step 2 is the normative cross-reference. Open the SUGAR Norms Lookup for the child’s age and the Brown’s Stages Lookup for the MLU value. The SUGAR Norms Lookup returns the published age-banded mean and standard deviation for MLU, NDW, and total parsable utterances; the Brown’s lookup returns the developmental stage label and the morphemes expected at that stage. A child whose MLU is more than 1.25 standard deviations below the age-band mean is in the clinical range for a morphology-focused goal. A child whose NDW is more than 1.25 standard deviations below the age-band mean is in the clinical range for a vocabulary-focused goal. A child whose PGU is below 80% in the school-age range is in the clinical range for a syntax/grammaticality goal. The numbers themselves are the trigger for which goal areas to write.

Step 3 is the goal-area match. Once you know which areas are clinically below band, open the matching IEP goal-area programmatic page on this site. There are eight: articulation, expressive language, receptive language, fluency, voice, pragmatics/social communication, AAC, and literacy. The expressive-language and receptive-language pages are the two that consume LSA data most directly; pragmatics and literacy also pull on transcript data in specific ways covered later in this article. Each page has five SMART goal templates already written in the format most US districts expect, plus a baseline-measurement protocol and a progress-monitoring cadence block. The templates are not finished goals — they are scaffolds calibrated to a typical baseline and target — but they are 80% of the work.

Step 4 is the present-levels paragraph. The structured baseline row from step 1 becomes the supporting evidence for one paragraph that names the assessment context, the sample length, the metric values, the published normative comparison, and the clinical interpretation. A typical paragraph reads: "On 3/14/2026, a 50-utterance conversational language sample was elicited from Marcus (age 5;3) using the SUGAR protocol. MLU in morphemes was 3.4 (SUGAR age-5 mean 5.1, SD 0.9), placing him at −1.9 SD. NDW was 84 (SUGAR age-5 mean 122, SD 18), placing him at −2.1 SD. Percent grammatical utterances was 71% (clinical range below 80% for school-age children). The pattern is consistent with a moderate expressive language delay with morphological and vocabulary components." Every number in that paragraph traces back to a specific tool on this site.

Step 5 is the goal sentence itself. Take the SMART template from the goal-area programmatic page, drop in the baseline from the calculator output, drop in the target from the normative band, and drop in the measurement method (rerun the same calculator on a fresh 50-utterance sample at the end of the IEP year). The IEP Goal Generator tool on this site automates this last step: it takes the baseline values from the calculators, the goal area, and the funnel parameters, and drafts a goal sentence in the format your district expects. The clinician’s job is to review and edit, not to type from scratch.

The most direct LSA-to-IEP-goal mapping is the one from MLU in morphemes to a morphology-focused expressive language goal. Brown (1973) described five stages of early grammatical development that map onto MLU bands: Stage I (1.0–1.99), Stage II (2.0–2.49), Stage III (2.5–2.99), Stage IV (3.0–3.74), and Stage V (3.75–4.49). Each stage is associated with the emergence of specific grammatical morphemes, in the well-known order documented across decades of normative data: present progressive -ing, prepositions in/on, regular plural -s, irregular past tense, possessive ’s, uncontractible copula, articles, regular past tense -ed, regular third-person singular -s, irregular third-person singular, uncontractible auxiliary, contractible copula, contractible auxiliary.

The clinical inference is straightforward. A 4-year-old with an MLU-m of 2.7 is at Stage III, which means present progressive -ing, prepositions in/on, and regular plural -s should be emerging or mastered. If the transcript shows that the child is producing Stage I and Stage II morphemes correctly but consistently omits the Stage III morphemes (no -ing on running verbs, no plural on multiple objects, no in/on used in prepositional phrases), the goal area is morphology and the specific targets are the missing Stage III morphemes. The annual goal sentence then targets emergence of those specific morphemes in obligatory contexts at a defined accuracy level by the end of the IEP year.

A typical SMART goal in this category reads: "Given a structured language activity, [Student] will use the regular plural -s morpheme correctly in 8 of 10 obligatory contexts across 3 consecutive sessions, as measured by clinician language samples and the MLU Calculator." The baseline number comes from the morpheme-tagged transcript output of the MLU Calculator (which returns morpheme counts in obligatory contexts, not just total morphemes). The target — 8 of 10 contexts across 3 consecutive sessions — is the standard SUGAR mastery criterion for an emerging morpheme. The measurement method is a fresh 50-utterance sample at the next probe, scored on the same calculator. Every component of the goal traces back to data the clinician already has on file from the initial assessment sample.

Number of Different Words (NDW) is the simplest defensible vocabulary metric in clinical LSA. The Lexical Diversity Calculator on this site returns NDW from a paste of the cleaned transcript in a single click and also returns the type-token ratio (TTR), the Moving-Average Type-Token Ratio (MATTR), and vocd-D. NDW is the headline number for vocabulary-focused IEP goals because it has the cleanest published normative bands across the SUGAR and Heilmann reference sets, and because it is the most intuitive metric for an IEP team to read in plain English ("the child used 84 different words in the sample").

A child whose NDW falls more than 1.25 standard deviations below the age-band mean from the SUGAR Norms Lookup is in the clinical range for a vocabulary-focused goal. The interpretation has to be paired with a structural check first — a child with extremely short utterances and a low MLU will mechanically have a low NDW for the same reason, and the right goal in that case is a morphology or sentence-length goal, not a vocabulary goal. The diagnostic question is: is the NDW low because the child used few words overall, or low because the child repeats the same words over and over? The TTR and MATTR values from the Lexical Diversity Calculator help distinguish these cases — a child with short utterances and low MATTR is producing repetitive language, and the goal should target lexical variety; a child with short utterances and a normal MATTR is producing limited but varied language, and the goal should target sentence length.

A typical NDW-driven SMART goal reads: "Given a structured narrative or conversational task, [Student] will produce a 50-utterance language sample containing at least 110 different words (current baseline 84, SUGAR age-5 mean 122) by the end of the IEP year, as measured by the Lexical Diversity Calculator on conductscience.com." The 110 target is the value at one standard deviation below the SUGAR age-5 mean, which is the conventional functional-range target for a year of vocabulary intervention. The measurement method is the same calculator on a fresh sample. The entire goal is defensible because every number in it points back to the SUGAR normative reference and the published Lexical Diversity Calculator output.

For school-age children, syntactic complexity and grammaticality are the two metric families that matter most for IEP goal writing because they are the metrics that change over an academic year of therapy in a measurable way. The DSS Calculator implements the Lee (1974) eight-category Developmental Sentence Score and returns a single composite number plus the per-category breakdown; the IPSyn Calculator implements the Scarborough (1990) 60-item Index of Productive Syntax and returns a 0–100 score; the PGU Calculator returns Percent Grammatical Utterances using the SUGAR rule. Together they cover the three core questions a school SLP asks about a school-age transcript: how complex is the syntax, how productive is the morphological inventory, and how often does the child produce ungrammatical utterances.

The DSS-driven goal targets syntactic complexity. A child whose DSS composite is more than 1.25 standard deviations below the published age-band mean is a candidate for a syntactic complexity goal targeting the lowest-scoring categories on the DSS breakdown. The target is movement of the composite into the within-functional-range band by the end of the IEP year, with the progress measurement being a rerun of the calculator on a fresh sample. The IPSyn-driven goal is similar in structure but with a 0–100 productivity scale: the baseline is the IPSyn score, the target is a defined increase tied to the normative band, and the measurement is a rerun.

The PGU-driven goal is the most common school-age grammaticality goal because it generalises across the entire transcript rather than targeting specific syntactic structures. A school-age child with a PGU below 80% is in the clinical range for a grammaticality goal; the typical annual target is 90% PGU on a fresh 50-utterance sample, which is the empirical floor for typically developing school-age children in the Pavelko & Owens 2017 normative data. The SMART goal sentence reads: "Given a 50-utterance conversational language sample, [Student] will produce grammatically complete utterances with at least 90% accuracy (current baseline 71%) by the end of the IEP year, as measured by the PGU Calculator on conductscience.com." The baseline, target, and measurement method are all anchored in published clinical data.

Not every language sample is a conversational sample, and not every IEP goal is a morphology, vocabulary, or grammaticality goal. The narrative samples elicited with story-retell or story-generation tasks (covered in detail in pillar #5 — "How to Conduct a Language Sample") feed into a different family of IEP goals targeting story grammar, narrative macrostructure, and pragmatic discourse skills. The Narrative Scoring Scheme Calculator implements the Heilmann (2010) seven-component scoring system (introduction, character development, mental states, referencing, conflict resolution, cohesion, conclusion) and returns a per-component score and a composite total. The Story Grammar Scorer returns a count of the classic story grammar elements (setting, initiating event, internal response, plan, attempt, consequence, reaction) per episode and a total episode count.

Narrative-driven goals are the right choice for school-age children whose conversational sampling looks unremarkable but whose teachers report difficulty with story comprehension, written narrative, or oral retelling tasks. The clinical question is no longer "how complex is the syntax?" but "is the child producing the macrostructure that listeners need to follow a story?". A typical narrative-driven goal reads: "Given a wordless picture book, [Student] will produce an oral narrative containing at least 5 of the 7 NSS components scored at proficient level (current baseline 3 of 7), as measured by the Narrative Scoring Scheme Calculator on conductscience.com." The baseline and target are both NSS scores from the same tool.

Pragmatic goals are the third family of LSA-driven goals and the one that most depends on transcript context outside the metric values themselves. The Conversation Turn Analyzer on this site returns a quantified speaker-balance ratio, average turn length per speaker, and topic maintenance counts from a turn-marked transcript. A child whose speaker-balance ratio is sharply skewed (the clinician produces 80% of the conversational turns) is in the clinical range for a turn-taking goal. A child whose topic maintenance count is consistently below threshold is in the clinical range for a topic-management goal. The pragmatic goals page on this site has five SMART templates calibrated to these patterns and drops them into the IEP Goal Generator alongside the morpheme, syntax, and lexical-diversity goals from the same transcript.

The four ASHA goal areas where LSA does not produce the headline metric — articulation, fluency, voice, and AAC — are the areas where the language sample is still useful as a contextual baseline rather than as the primary scoring tool. For an articulation goal, the language sample provides a connected-speech sample that can be scored for percent consonants correct (PCC) using the PCC Calculator on this site, which is the connected-speech complement to a single-word articulation screener. For a fluency goal, the language sample provides a connected-speech sample that can be scored for stuttering frequency using the Stuttering Frequency Calculator. For a voice goal, the language sample provides natural connected-speech audio that can be rated using the CAPE-V Voice Rating tool or the Voice Handicap Index Calculator. For an AAC goal, the language sample is a baseline of the child’s spoken output that anchors the AAC selection process and the device-trial measurement.

For articulation goals, the typical approach is to run the single-word articulation screener for the formal evaluation row and then use a 50-utterance connected-speech sample to compute PCC for the present-levels paragraph and the progress monitoring probes. A goal might read: "[Student] will produce target sounds /s/, /r/, /l/ in connected speech with at least 85% accuracy (current PCC for these sounds in connected speech is 62%), as measured by 50-utterance language samples scored with the PCC Calculator." The connected-speech metric is the more functional measurement and the more durable progress probe, even though the single-word screener is what runs at evaluation.

For fluency goals, the language sample is the only realistic source of an in-context disfluency count. The Stuttering Frequency Calculator on this site takes a marked-up transcript and returns the percentage of stuttered syllables by type (repetitions, prolongations, blocks). A typical goal might read: "[Student] will speak with no more than 5% stuttered syllables in a 50-utterance conversational sample (current baseline 14%), as measured by the Stuttering Frequency Calculator." For voice and AAC goals, the language sample plays a smaller role but still anchors the connected-speech baseline that the IEP team can refer to a year later when they re-record. In every one of these four areas, the language sample is not the primary metric but it is the metric that makes the present-levels paragraph specific enough to defend.

The single biggest practical advantage of LSA-anchored goals is that progress monitoring is structurally identical to baseline measurement: collect a fresh 50-utterance sample, run the same calculator, write down the new number. The cadence is constrained by clinician time more than by clinical judgement; the conventional school-based pattern is a baseline sample at the IEP date, a mid-year probe at the six-month mark, and a year-end probe one to two weeks before the next IEP meeting. Some districts collect quarterly samples on caseloads with a reasonable AI-assisted transcription budget, which is the cadence at which a clinician can spot a flat progress curve early enough to adjust the goal.

The mid-year probe is the most clinically valuable of the three because it is the one that drives goal-revision decisions. A child whose MLU-m has moved from 3.4 at baseline to 3.6 at the six-month mark is on a typical trajectory and the goal stays. A child whose MLU-m is still 3.4 at the six-month mark is on a flat trajectory, and the question for the IEP team is whether the goal target was unrealistic, whether the intervention dose is too low, or whether there is a structural reason the morphology has not generalised. None of those questions can be asked without the mid-year data point, and none of them can be answered without the same scoring tool that produced the baseline.

The year-end probe is the data the IEP team takes into the renewal meeting. A clean year-end probe row — MLU, NDW, PGU, plus the goal-area-specific metric — placed next to the baseline row from the previous year is the single most efficient way to communicate progress to a family in plain English. "Marcus’s MLU moved from 3.4 to 4.2 over the year, which is one Brown’s stage of progress and slightly above the typical annual rate of growth in the SUGAR data" is a sentence the family can understand without any clinical training, and it is a sentence the same calculator on this site can support with concrete numbers.

A clinician who anchors every IEP goal in language sample data has done most of the work of defending the goal before the question ever comes up. The four pillars of measurability — baseline, target, measurement method, mastery criterion — line up exactly with the questions a parent’s advocate or a due-process attorney will ask in a serious IEP review. "Where did the baseline number come from?" answers with "a 50-utterance language sample collected on this date and scored with the MLU Calculator on conductscience.com." "Where did the target number come from?" answers with "the SUGAR age-banded mean from Pavelko & Owens (2017), retrieved from the SUGAR Norms Lookup." "How will progress be measured?" answers with "a fresh 50-utterance sample collected and scored on the same calculator at the six-month and year-end probes." "How will mastery be determined?" answers with "an inter-rater reliable scoring rule from the published normative database."

The reason this matters in practice is that the alternative — a goal whose baseline is a clinician’s impression and whose target is a clinician’s judgement — does not survive a mediation or a due-process hearing in 2026. Hearing officers and parent advocates are increasingly familiar with the language of measurable goals, and a goal that says "Marcus will improve his expressive language" with no metric attached is the kind of goal that gets the school district held to a more aggressive intervention plan than the SLP intended. The defensible goal is the one that names the metric, the calculator, the normative reference, and the rerun cadence. The clinician’s job is to write that goal once at the IEP date and then run the calculator twice more during the year. The data does the rest of the defence on its own.

A second practical advantage of LSA-anchored goals in mediation is that the clinician can reproduce the entire pipeline in the meeting room with the parents present. Every tool on this site runs in a browser, takes a paste of the transcript, and returns a number in under a second. A clinician who pulls up the MLU Calculator on a laptop and pastes the child’s transcript while the parents watch has demystified the entire scoring process in 30 seconds. There is no proprietary software, no licence, and no installation step between the family and the data. The transparency itself is a defence; it is hard to argue that a number is wrong when the tool that produced it is open in the browser tab and the input is the family’s own child’s transcript.

The five most common ways an LSA-anchored IEP goal gets watered down into a vague goal are not clinical mistakes. They are documentation mistakes — small omissions that make the same number unrecoverable a year later when the rerun probe is collected. They come up at every district training and they are worth spelling out so that new clinicians do not have to discover them by writing a goal whose progress they cannot measure.

The workflow recommendation that comes out of this pillar is the same one that drives the rest of the LSA cluster on this site: collect the language sample, run the free browser-based calculators, cross-reference the published norms, and use the IEP Goal Generator to draft the goal sentences before editing them by hand. The workflow assumes a school SLP with a caseload of 20 to 50 students and a district that expects measurable goals at every IEP. It is calibrated for clinical realism, not for research-grade ideal.

The recommended pipeline, start to finish, is: (1) collect a 50-utterance language sample using the protocol from pillar #5, (2) clean and segment the transcript using the C-unit framework, (3) paste into the MLU Calculator, the Lexical Diversity Calculator, and the metric calculator matching the clinical question (DSS, IPSyn, or PGU), (4) cross-reference the metric values against the SUGAR Norms Lookup and the Brown’s Stages Lookup, (5) match the clinical pattern to one of the eight ASHA goal areas and open the matching programmatic page on this site, (6) write the present-levels paragraph using the metric values and the normative comparisons as the supporting evidence, (7) draft the annual goals using the IEP Goal Generator and edit them in the language your district expects, (8) run the same calculators on a fresh sample at the six-month probe and again at the year-end probe.

For clinicians whose caseload makes that pipeline tight, ConductSpeech is built specifically to automate steps 3 through 7 — audio in, scored transcript and draft IEP goals out — so the clinician’s role is to record, review, and edit. The honest framing is the same as in the rest of this cluster: ConductSpeech does not replace clinical judgement, it reallocates clinician time from the slow scoring and drafting steps to the fast interpretation and editing steps. For every caseload size, though, the workflow above is the frame: the specific automation level changes with caseload, but the metric, the normative reference, and the SMART goal structure stay the same.