Manipulating Linguistic Complexity to Improve Child Language Treatment Outcomes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Manipulating Linguistic Complexity to Improve Child Language Treatment Outcomes (NCT03977701), a non-randomized na study. The primary outcome is Treatment Probe Accuracy Change.

2. Clinical Question

The study asks whether the complexity of a speech or language treatment target is associated with change in children's accuracy when producing that target during treatment. The registry describes English- and Spanish-speaking children aged 3 to 6 years with phonological disorder, specific language impairment, or both. The listed conditions are Language Development Disorders and Speech Sound Disorder.

The treatment targets vary by whether they involve singleton consonants or consonant clusters, and whether they occur in mono-morphemic or bi-morphemic contexts. The arm descriptions also distinguish children with PD, SLI, or PD-SLI. The interventions are described as phonological treatment and morpheme treatment, with particular target selections assigned to the study arms.

The primary question concerns change in accuracy for the treated consonant or consonant cluster, from the first treatment session to the final treatment session, over a period of up to 6 weeks. Accuracy is assessed in treatment-probe words selected to match the child's assigned experimental condition. The registry does not report the numerical change in accuracy for any arm, the size of differences between arms, or uncertainty estimates. It therefore records the question and measurement approach without establishing how the observed outcomes compare across target types.

Because allocation was non-randomized, the arms should not be interpreted as randomized, like-for-like treatment comparisons. Differences between arm outcomes, if reported elsewhere, could reflect factors beyond target complexity. The registry record does not report adjusted comparisons or analyses that address such potential differences.

3. Trial Design

This completed study used a non-randomized, parallel design with single masking. The registry does not identify who was masked, so the masking description cannot be interpreted more specifically. The listed enrollment is 41. The sponsor is San Diego State University. The study began on 2019-06-01, reached primary completion on 2021-12-31, and has completed status in the registry.

The arms cover several combinations of target structure and diagnostic grouping. For children with PD or PD-SLI, the descriptions include speech sound treatment on mono-morphemic singleton consonants and clusters, and treatment on singletons or clusters in bi-morphemic contexts. The SLI arms described concern singletons or clusters in bi-morphemic contexts. The registry identifies phonological treatment and morpheme treatment as behavioral interventions.

Non-randomized allocation means that assignment was not described as random. Random assignment is often used to make treatment groups more comparable at the start of a study, on average. Without it, differences in starting characteristics may be associated with both the assigned target and the outcome. The registry does not report the allocation process or baseline comparisons that could clarify this issue. The arms are therefore not a like-for-like control comparison, and observed arm contrasts should be read as descriptive unless further analytic details are available.

The primary outcome population is described as participants who completed the study protocol within each arm. The record does not give arm-specific enrollment or completion counts, and those counts cannot be inferred from the listed overall enrollment. It also does not report the rationale for the enrollment target, recruitment details, or reasons for non-completion. The registry records the design and intervention framework, but does not provide enough information to reconstruct participant flow or assignment procedures.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Of 41 participants recruited, 22 met eligibility requirements for participation. Two did not get assigned to study arms due to the COVID-19 pandemic, and one withdrew and opted for treatment elsewhere. Two did not complete protocol due to pandemic; corresponding Arms/Groups (PD-SLI) were discontinued. At pandemic onset, we limited enrollment to 2 Arms/Groups (Mono/Bi Cluster PD). Caregivers were NOT enrolled, were not participants in study, per protocol; no Arms/Groups for caregivers.

Only the first 6 of 10 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who completed study protocol within each arm. Reporting percentage of change in accuracy for treated target consonant or consonant cluster as presented in the verb stimuli.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: We transitioned to a telehealth approach due to the onset of the COVID-19 pandemic, and limited our project to specific arms of the study due to reduced feasibility and increased resources needed to complete the protocols. The participants in the Bi-morphemic Singleton PD-SLI and Bi-morphemic Cluster PD-SLI conditions discontinued due to the onset of the pandemic and we removed those Arms/Groups from the study going forward.

Only the first 6 of 11 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Treatment Probe Accuracy Change, measured over Change in percentage of consonant or consonant cluster accuracy (depending on Arm/Group) from first treatment session to final treatment session; up to 6 weeks..

The primary endpoint is Treatment Probe Accuracy Change, expressed as percent accuracy of target consonants. It measures the change in accuracy for the consonant or consonant cluster selected for treatment, comparing the first treatment session with the final treatment session. The timeframe is up to 6 weeks. The registry describes the reported outcome as a mean with standard deviation.

The probe uses words selected to match each child's experimental condition. A child is asked to say each word after seeing a corresponding picture and hearing a verbal prompt. The target sound is scored as correct or incorrect. Examples in the registry contrast singleton and cluster targets, as well as words in mono-morphemic and bi-morphemic contexts. The endpoint thus focuses on production accuracy for the treated stimuli rather than a general measure of language ability or performance on untreated words.

Change from the first to final session is used to describe whether accuracy on the selected treatment targets shifted during the treatment period. A mean provides a summary across participants in an arm, while a standard deviation describes variability around that mean. Neither summary alone reveals the range of individual outcomes or the reasons for differences among children.

The registry record does not report the mean change, standard deviation, or arm-specific endpoint results. It also does not report a comparative effect estimate or confidence interval. The definition and measurement procedure are available, but the magnitude and precision of the observed change are not.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean change represent?

For a continuous outcome, a mean summarizes the average value across the observations included in an analysis. In this study, the endpoint is a change in percent accuracy from the first treatment session to the final treatment session. A mean change would describe the average observed shift in probe accuracy among the participants represented in an arm's analysis. It would not imply that every child changed by that amount: individual responses may vary, and the registry's stated standard deviation is intended to convey the spread of observations around the mean.

General statistical interpretation: a mean can be useful because it compresses many observations into a single summary, but it can hide the distribution's shape and individual variation. A mean and standard deviation are not a full account of the data. The registry record does not provide the values needed to examine the actual distribution or calculate an uncertainty interval for the endpoint.

What does the probe's scoring imply?

The probe scores each target as correct or incorrect. A participant's accuracy percentage therefore summarizes correct productions relative to the tested targets. The study's endpoint then describes change in this percentage between the first and final treatment sessions. This is different from analyzing each individual response as a separate independent observation. Responses are grouped within children, and the observations from one child may be related to each other.

General statistical interpretation: when repeated measurements or multiple responses come from the same person, an analysis may need to account for within-person dependence. Treating related observations as independent can make uncertainty appear smaller than it is. The registry describes the outcome summary as a mean with standard deviation, but does not report whether the analysis used participant-level summaries, a model of individual responses, or another approach.

How should a comparison between arms be read?

An arm-level mean change is descriptive. A difference between arm means, if reported, would describe how the observed average changes differ. It would not automatically establish a causal effect of target complexity. Random allocation is not reported; the study allocation is non-randomized. If assignment relates to participant characteristics that also influence accuracy, those characteristics could contribute to an observed arm contrast.

General statistical interpretation: adjustment can address measured differences when the model and data support it, but adjustment cannot guarantee removal of bias from unmeasured or poorly measured factors. The registry does not report a comparative model, covariates, adjusted estimates, or a causal analysis. It also does not report the procedures used to assign participants to the described arms.

What do variability and confidence intervals tell readers?

The endpoint is specified as a mean with standard deviation. A standard deviation describes how dispersed individual values are around the mean; it is not itself a measure of how precisely the mean is estimated. A confidence interval, when calculated under a stated method, gives a range of values compatible with the data and model at the chosen confidence level. For a between-arm contrast, an interval would help show whether the estimate is precise or highly uncertain.

General statistical interpretation: precision depends on the amount and variability of information available and on the analysis method. A narrow interval indicates more precision than a broad interval, but neither proves that the estimate is unbiased. Since the registry does not report endpoint summaries, confidence intervals, or p-values, precision and statistical compatibility with a null difference cannot be evaluated from the record.

What is the role of planning assumptions?

Planning sample-size illustrations are not observed outcomes and are not observed power. The assumptions described for the planning illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The stated effect to detect is a fraction of the observed effect. For non-time-to-event outcomes, such illustrations are expressed as participants per arm.

General statistical interpretation: a planning calculation estimates the sample size associated with specified assumptions and a target effect. It does not show that a study achieved that power, nor does it establish that the assumed effect occurred. A sample-size illustration based on an observed effect is especially dependent on the stability of that estimate. The registry does not report the numerical endpoint effect or the numerical planning outputs in the record considered for this analysis. The planning assumptions therefore cannot supply missing outcome evidence.

Which analysis details are unavailable?

The registry reports zero analyses. It does not report a statistical analysis plan, the model used for the primary endpoint, whether an analysis was adjusted, how repeated responses were handled, or whether any interim analysis occurred. These details are important because they determine which observations contribute to an estimate and how uncertainty is calculated. They cannot be reconstructed from an endpoint definition alone.

General statistical interpretation: statistical methods should match the outcome structure, study design, and question. When the method is not reported, readers cannot verify model assumptions, reproduce estimates, or understand whether the analysis addresses the design's limitations. For this study, the non-randomized assignment and the binary scoring of individual probe responses are relevant features, but the registry record does not specify how the primary analysis handled them.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Mono-morphemic Cluster PD", "Bi-morphemic Cluster PD")
n   <- c(7, 10)
m   <- c(26, 23)
sp  <- c(36, 38)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Mono-morphemic Cluster PD  7  26.000  36.000  -7.294  59.294
CARM  Bi-morphemic Cluster PD  10  23.000  38.000  -4.184  50.184

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures, but provides no analysis entries. It identifies Treatment Probe Accuracy Change as the primary outcome and describes its measure and timeframe. The registry record does not report a multiplicity strategy, such as whether comparisons across target structures or diagnostic groups were prespecified, whether a subset of comparisons was primary, or whether any adjustment for multiple testing was planned.

Multiplicity is a general statistical concern when several outcomes or comparisons are examined. The more questions tested, the greater the chance that an apparently notable result could arise by chance if each test is interpreted separately. Methods such as specifying a primary comparison in advance or adjusting inference can address particular multiplicity concerns, but the appropriate approach depends on the study questions and analysis plan. The registry does not provide enough detail to determine what approach, if any, was used.

The endpoint population is described as participants who completed the study protocol within each arm. This wording identifies a completer-based population for the reported outcome description, but does not establish the exact analysis population used. The registry does not give arm-specific completion counts, missing outcome counts, reasons for missingness, or a method for handling missing measurements. It does not report whether all available observations were used or whether any imputation was performed.

General statistical interpretation: missing data can affect both precision and the representativeness of an estimate, particularly if the probability of missingness is related to outcome or participant characteristics. A completer summary may differ from a summary that includes all assigned participants. Without the analysis population and missing-data procedures, readers cannot assess the direction or scale of any such effect.

The registry also does not report a protocol-based analysis plan or interim analyses. These details are not reported in the registry record and should not be inferred from the listed outcome description or study status.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The adverse-event record lists zero serious adverse events, zero other adverse events, and zero deaths in every arm. These are registry counts of affected participants, not estimates of comparative safety. No serious adverse-event terms are listed.

The at-risk counts are not the same in every arm. The record gives serious-event and other-event risk counts of 7 for Mono-morphemic Cluster PD, 1 for Bi-morphemic Singleton PD-SLI, 10 for Bi-morphemic Cluster PD, and 1 for Bi-morphemic Cluster PD-SLI. The remaining listed arms show zero at risk in the adverse-event entries. A zero at-risk entry should not be read as proof that an arm had no participants; it indicates that the registry's safety entry does not give an at-risk denominator for that arm.

Zero reported affected counts mean that no affected participants are entered in these registry safety fields. They do not establish that an event was impossible or that the treatments were free of all harms. Interpretation depends on how safety events were collected and recorded, the period of observation, and the number of participants assessed for each type of event. Those details are not reported in the registry record.

Because the risk denominators vary and are absent for some arms, these entries do not support direct comparisons of event rates. They should be read as descriptive registry counts rather than evidence of equivalent safety across conditions.

13. What the Result Does, and Does Not, Mean

The registry describes a study intended to examine whether treatment target complexity relates to changes in children's accuracy on treated consonants or consonant clusters. Its endpoint is tied to treatment-probe words and compares performance from the first session to the final session over up to 6 weeks. The arm descriptions distinguish singleton and cluster targets, morpheme contexts, and diagnostic groupings.

What the registry does not establish is the outcome of that investigation. Numerical results for the primary endpoint are not reported in the registry record, and no analysis entries are recorded. There is no reported arm-level mean change, standard deviation, comparative estimate, confidence interval, or p-value. Consequently, the record cannot show whether accuracy changed or whether any target condition was associated with a larger change.

The non-randomized design also limits what an observed difference could mean. If results were available, differences between arms would need to be interpreted in light of possible pre-existing group differences and the unreported assignment process. The registry does not describe adjustment or other analytic steps that could address those concerns. A comparison of observed means alone would not establish that target complexity caused a difference.

The safety entries record zero serious adverse events, zero other adverse events, and zero deaths across the listed arms. These counts are descriptive and do not establish absence of risk, particularly because the at-risk denominators are not consistently reported. They also do not answer the primary treatment-accuracy question.

In general, an endpoint definition explains what an outcome is intended to capture, while results and uncertainty estimates are needed to assess what the study observed and how precisely it was estimated. The registry record supplies the former but not the latter. Any conclusion about comparative treatment outcomes, benefit, or lack of effect would go beyond the information recorded.

14. Why This Trial Matters Statistically

This study is statistically notable for asking a structured question about the complexity of treated language targets while measuring a specific behavioral outcome. The endpoint focuses on accuracy of selected consonants or consonant clusters in words that match each participant's assigned experimental condition. This makes the measurement closely tied to the treatment target rather than to a broad language score.

The design also illustrates why outcome definitions and study design need to be read together. The arms vary across target structures and diagnostic groups, while allocation is non-randomized. A set of detailed intervention descriptions does not by itself make arm comparisons equivalent to randomized comparisons. Interpretation of any observed difference would depend on how participants were assigned, how baseline differences were handled, and which comparisons were prespecified.

The endpoint's underlying responses are scored correct or incorrect, but the primary outcome is described as mean change in percent accuracy. This distinction matters methodologically: the summary is based on accuracy across probe items and measured across treatment sessions. Statistical analysis needs to account for the relevant unit of observation and for any dependence among responses from the same child. The registry does not report the analysis method, making it impossible to assess that aspect of the evidence.

Finally, the registry record demonstrates the difference between a completed study entry and a report of analyzable results. It documents the study's aim, design, intervention framework, endpoint, and safety counts, yet does not report the primary outcome values or analyses. For statistical readers, that gap is consequential: without estimates and uncertainty, the research question can be described, but its empirical answer cannot be evaluated from the registry record.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary limitation for interpreting this trial is that its numerical primary outcome results are not reported in the registry record. The endpoint is defined, but no arm-level mean changes, standard deviations, comparative estimates, confidence intervals, or p-values are available. The registry also lists no analyses, so the primary statistical method cannot be evaluated or reproduced.

The non-randomized allocation limits causal interpretation. The registry does not report how participants entered particular arms, whether assignment depended on participant characteristics, or whether baseline characteristics were compared. It does not report an adjusted analysis or the variables that such an analysis might include. As a result, potential differences between arms cannot be assessed from the registry information.

Participant flow is also unclear. Overall enrollment is listed as 41, but arm-specific enrollment, completion, and analysis counts are not reported. The primary outcome population is described as participants who completed the protocol within each arm; details about missing observations, reasons for non-completion, and methods for handling missing data are not reported in the registry record.

Other analysis details are unavailable, including the statistical model, any approach to repeated or clustered probe responses, multiplicity procedures, interim analyses, and the statistical plan. The masking description is single, but the identity of the masked party is not specified. The registry does not report the rationale for the enrollment target.

Safety interpretation is limited by variation in the reported at-risk counts and zero at-risk entries for some arms. The registry records no affected participants or deaths, but the observation and ascertainment details are not reported. These limitations mean that the record supports description of the study framework, not a full assessment of efficacy, precision, or comparative safety.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Manipulating Linguistic Complexity to Improve Child Language Treatment Outcomes — NCT03977701.
• Registry record last updated 2023-03-09; status last verified 2023-03.
• Results first posted 2023-03-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT03977701, titled Manipulating Linguistic Complexity to Improve Child Language Treatment Outcomes, is a completed, non-randomized parallel study sponsored by San Diego State University. Its listed enrollment is 41. The study began on 2019-06-01 and its primary completion date is 2021-12-31. The registry describes single masking but does not identify who was masked.

The study concerns children with language development and speech sound disorders, with the brief summary describing English- and Spanish-speaking children aged 3 to 6 years. Behavioral interventions include phonological treatment and morpheme treatment, with arm descriptions varying by consonant target, morpheme context, and diagnostic grouping.

The primary endpoint is change in treatment-probe accuracy from the first to final treatment session, up to 6 weeks. Numerical primary outcome results and statistical analyses are not reported in the registry record. Safety entries list zero serious adverse events, zero other adverse events, and zero deaths in each arm, with at-risk counts varying across arms.