For speech-language pathologists
How should SLPs measure progress in speech sound therapy?
Published · Updated
The short answer
Measure progress mainly with probes of words the child hasn’t practiced, not just accuracy during drills. Useful measures include percent consonants correct (PCC) from a speech sample, a short list of untreated words probed every few sessions, an untreated control target, and a functional rating such as the Intelligibility in Context Scale. Session data shows how practice is going; probe data shows whether the skill is generalizing. ASHA asks for measurable goals and objective progress data in notes and reports.
Percent consonants correct (PCC)
PCC, introduced by Shriberg and Kwiatkowski (1982), is the percentage of consonants produced correctly in a continuous speech sample. They presented it as a severity metric with four levels (mild, mild-moderate, moderate-severe and severe), and as a way to gauge the effects of intervention programs (Shriberg & Kwiatkowski, 1982).
A 1997 paper extended it into a set of related metrics computed from a 5- to 10-minute conversational sample, reported reliability and standard error of measurement estimates for them, and discussed choosing among them for clinical and research needs (Shriberg et al., 1997).
ASHA’s Practice Portal lists PCC as one way to place severity on a continuum from a transcribed speech sample, while noting there’s no clear consensus on the best way to determine the severity of a speech sound disorder (ASHA).
- Elicit the sample the same way each time, such as the same retell or conversation topic, so samples are comparable.
- PCC takes transcription time, so it suits baseline, discharge and periodic checks better than every session.
Probes, generalization lists and control items
A probe samples the target in items the child hasn’t practiced. In a tutorial on speech motor chaining, Preston, Leece and Storto (2019) describe a list of 25 untreated one- and multisyllable words with the target in varied word positions, taking about 3 minutes. They suggest probing every 3 to 4 sessions in clinical practice (about every two sessions in their research), at the start of a session, and moving to 10 to 15 untrained sentences once word accuracy reaches about 80% (Preston et al., 2019).
Untreated words are also how trials measure generalization: in the 2026 C-RESULTS trial of children with residual /ɹ/ errors, generalization was measured on untreated words (Preston et al., 2026).
Control items are targets you probe but don’t treat, such as a second error sound you expect to stay stable. This borrows the logic of a multiple-baseline design: treatment starts on one behavior while baseline measurement continues on others, so change that appears where and when treatment starts is more convincing. The behaviors need to be independent, so choose control items that are unlikely to improve from work on the treated target (Byiers et al., 2012).
Session data versus probe data
Accuracy during practice reflects the cues, models, feedback and items being drilled. In their tutorial, Preston and colleagues note that practice data reflect acquisition but not generalization, so probe data should be the primary outcome (Preston et al., 2019).
Session data still matters. ASHA’s guidance on documentation in schools describes data collection that records students’ performance during intervention activities, and notes goals are often written as percentage accuracy or the number of trials completed successfully (ASHA). Counting trials also lets you report dose: see treatment intensity and dosage.
Olswang and Bain (1994) argued that decisions about whether treatment is working should, in general, rest on data, both quantitative and qualitative, within a data-based decision-making model, while leaving room for clinical intuition.
- Record the level (syllable, word, phrase, sentence, conversation) and the support given, so a rise in accuracy isn’t confused with a drop in cues.
- Keep probe conditions the same each time: the same list, the same instructions, and no feedback on probe items.
Intelligibility in everyday life
Accuracy measures don’t show how well a child is understood. ASHA describes intelligibility as a perceptual judgment of how much of a child’s spontaneous speech a listener understands, and notes that a parent-rated scale can show intelligibility across contexts and listeners (ASHA).
The Intelligibility in Context Scale (ICS) is a 7-item parent-report scale that rates how well a child is understood by parents, immediate family, extended family, friends, acquaintances, teachers and strangers, on a 5-point scale. In its first study of 120 preschoolers, it had high internal reliability (α = .93) and correlated moderately with PCC (r = .54) (McLeod, Harrison & McCormack, 2012).
Normative data from the parents of 803 Australian English-speaking children aged 4;0 to 5;5 gave a mean score of 4.4 out of 5. Scores weren’t affected by socioeconomic status or the number of languages spoken, and the authors recommend the ICS as a screening measure. In 184 children with speech sound disorders, its correlation with PCC was low (r = .24), so it captures something different from accuracy (McLeod, Crowe & Shahaeian, 2015).
These studies describe intelligibility and screening; we didn’t find studies validating the ICS as a measure of change during treatment, so interpret small changes cautiously.
Single-case design basics
- Baseline first. By convention, at least three baseline data points are needed to establish stability, and more are preferable; the baseline shouldn’t already be trending toward improvement (Byiers et al., 2012).
- A–B alone doesn’t show cause. A baseline-then-treatment comparison shows correlation but not necessarily causation, because other factors could explain the change (Byiers et al., 2012).
- Withdrawal designs fit learned skills poorly. Stopping treatment is unlikely to make a child unlearn a sound, so multiple-baseline designs suit behaviors that can’t be reversed (Byiers et al., 2012).
- Multiple probes save time. Multiple-probe designs replace continuous baseline measurement with intermittent probes, though they may not suit highly variable behaviors (Byiers et al., 2012).
- Read the graph. Visual analysis looks at changes in level, trend and variability, and how immediately they follow the start of treatment (Byiers et al., 2012).
In everyday practice, a treated target, an untreated probe list and a control item, graphed over time, give you much of this logic without a research protocol.
Documentation, and where home practice data fit
ASHA’s guidance on documentation in health care says progress notes typically include objective measures of progress toward functional goals and are written at intervals set by the payer or facility; for Medicare, at or before every 10th session. Treatment notes typically include subjective and objective data on progress, compared with prior sessions (ASHA).
In schools, IDEA requires regular progress reports, and the IEP must say when they’ll be provided. ASHA advises SLPs to include both qualitative and quantitative information about progress (ASHA).
Home practice data can add to the picture, with caution:
- It’s usually self-report. Most intervention studies involving parents give limited detail about what parents actually did (Sugden et al., 2016), so treat home logs as a record of practice, not a measure of skill.
- Home accuracy is practice data. Like session drills, it reflects acquisition rather than generalization (Preston et al., 2019). Your probes remain the outcome.
- Automated scores need checking. A systematic review found automated tools were typically less than 80% accurate on words containing mispronunciations (McKechnie et al., 2018). Spot-check any tool against your own judgment before it shapes goals.
- Dose is worth recording. Home programs show preliminary benefit when dosage is high and parents are trained directly (Tosh, Arnott & Scarinci, 2017), so the number of practice trials at home is useful information.
See tracking practice between sessions and getting more home practice.
Questions SLPs ask
What is percent consonants correct?
PCC is the percentage of consonants a child produces correctly in a continuous speech sample, introduced by Shriberg and Kwiatkowski (1982) as a severity metric. ASHA notes there’s no clear consensus on the best way to rate severity.
How often should I probe generalization?
One tutorial suggests probing untreated words every 3 to 4 sessions in clinical practice, at the start of a session. Research protocols often probe more often.
Why isn’t session accuracy enough?
Accuracy during practice reflects cues, feedback and drilled items. It shows acquisition, while probes of untreated words show whether the skill is generalizing.
What does the Intelligibility in Context Scale measure?
It’s a 7-item parent questionnaire rating how well a child is understood by different listeners, from parents to strangers, on a 5-point scale. It complements accuracy measures rather than replacing them.
Can home practice data count as progress data?
Treat it as a record of practice. Home logs are usually self-report and reflect practice rather than generalization, so base progress decisions on your own probes.
Sources
- Shriberg, L. D., & Kwiatkowski, J. (1982). Phonological disorders III: A procedure for assessing severity of involvement. Journal of Speech and Hearing Disorders, 47(3), 256–270.
- Shriberg, L. D., Austin, D., Lewis, B. A., McSweeny, J. L., & Wilson, D. L. (1997). The percentage of consonants correct (PCC) metric: Extensions and reliability data. Journal of Speech, Language, and Hearing Research, 40(4), 708–722.
- ASHA Practice Portal: Speech sound disorders (articulation and phonology).
- Preston, J. L., Leece, M. C., & Storto, J. (2019). Tutorial: Speech motor chaining treatment for school-age children with speech sound disorders. Language, Speech, and Hearing Services in Schools, 50(3), 343–355.
- Preston, J. L., Hitchcock, E. R., Leece, M. C., Benway, N. R., Hill, J., & McAllister, T. (2026). Rhotic generalization is more rapid in biofeedback than motor-based treatment for residual speech sound disorder: Secondary outcomes of a randomized controlled trial. Journal of Speech, Language, and Hearing Research, 69(4), 1362–1378.
- Byiers, B. J., Reichle, J., & Symons, F. J. (2012). Single-subject experimental design for evidence-based practice. American Journal of Speech-Language Pathology, 21(4), 397–414.
- ASHA Practice Portal: Documentation in schools.
- Olswang, L. B., & Bain, B. (1994). Data collection: Monitoring children’s treatment progress. American Journal of Speech-Language Pathology, 3(3), 55–66.
- McLeod, S., Harrison, L. J., & McCormack, J. (2012). The Intelligibility in Context Scale: Validity and reliability of a subjective rating measure. Journal of Speech, Language, and Hearing Research, 55(2), 648–656.
- McLeod, S., Crowe, K., & Shahaeian, A. (2015). Intelligibility in Context Scale: Normative and validation data for English-speaking preschoolers. Language, Speech, and Hearing Services in Schools, 46(3), 266–276.
- ASHA Practice Portal: Documentation in health care.
- Sugden, E., Baker, E., Munro, N., & Williams, A. L. (2016). Involvement of parents in intervention for childhood speech sound disorders: A review of the evidence. International Journal of Language & Communication Disorders, 51(6), 597–625.
- McKechnie, J., et al. (2018). Automated speech analysis tools for children’s speech production: A systematic literature review. International Journal of Speech-Language Pathology.
- Tosh, R., Arnott, W., & Scarinci, N. (2017). Parent-implemented home therapy programmes for speech and language: A systematic review. International Journal of Language & Communication Disorders, 52(3), 253–269.