Alan Wisler

The presence of outlying data points can have a significant impact on statistical modeling and significance testing. In the specific context of one-sample t-tests, prior studies have shown (primarily through simulations) that outliers make it more likely for t-tests to fail to reject the null hypothesis. In this study, we investigate the opposite scenario: when an outlier can cause the rejection of the null hypothesis. While it may seem intuitive that outliers aligned with the direction of an effect strengthen that effect, prior studies have shown that this is not always the case. Towards this end, we introduce mathematical bounds on how large outliers can be while still increasing the t-statistic in a given sample. These bounds are validated and supported using Monte-Carlo simulations and a survey of available data sets. From these results, we find that although it is not impossible for outliers to cause significant results in paired or one-sample t-tests, it can only occur under rather narrow circumstances. Specifically, it requires a concordant outlier, a minimal sample size of ( n ≥ 15 ), and a sufficiently small effect size ( μ ^ / σ ^ ≤ 1 / 2 ). Based on these findings, we argue that the risk of isolated outliers causing type I errors is low in many practical situations, especially when sample sizes are small.

Purpose: According to theoretical models, perceptual learning is modulated by the level of acoustic predictably in the speech signal. This study investigated whether the variable speech of children is learnable in a perceptual learning paradigm and examined whether speech variability, as quantified by the acoustic spatiotemporal index (STI), predicted perceptual learning outcomes. Method: Speech samples were elicited from 40 typically developing children talkers (ages 3–8 years) and presented to 410 adult listeners in a structured perceptual learning experiment. Listeners transcribed phrases before (pretest) and after (posttest) a lexically guided familiarization phase. Intelligibility improvements following familiarization were analyzed using paired t tests, and regression analyses examined whether acoustic STI values predicted perceptual learning outcomes. Results: Listeners demonstrated significant perceptual learning of children's speech, with an average intelligibility gain of 6.4% following familiarization ( p < .001). Regression analyses revealed a significant interaction between STI and pretest intelligibility: Lower STI values (indicating more predictable speech) were associated with greater intelligibility gains but only when pretest intelligibility was sufficiently low. Conclusions: This study provides the first empirical evidence that children's speech is learnable in a structured perceptual learning paradigm. Although the acoustic STI affords some insight into predictability, it does not fully account for variability in perceptual learning. Future work should explore measures that capture segmental predictability to refine models of speech learnability.

Purpose: The spatiotemporal index (STI) is widely used to measure variability across repeated speech stimuli. One of the most well-established findings is that STI values increase when speakers are cued to slow their articulatory rate. However, the mechanism behind this increase—and how it reflects changes in motor control—remains unclear. Recent evidence suggests STI differences may stem from template motion patterns as well as changes in motor stability. This study examines how variations in template motion patterns across speaking rates influence STI values. Method: Kinematic data from 20 participants at three speaking rates (fast, habitual, slow) were used to estimate template motion patterns. These patterns were then used to simulate rate-specific template effects on STI values while holding motor stability constant. Results: When motor stability is held constant, simulations show that STI values increase during slow speech ( p = .003) and decrease during fast speech ( p < .001), driven solely by variations in template motion patterns. In slow speech, these template effects increased the STI by 1.754, accounting for a sizable portion (31.9%) of the increased STI values observed in participants' real speech data. In fast speech, template effects lowered the STI by 2.085. Conclusions: These findings provide strong evidence that template motion patterns contribute to the elevated STI values observed in slow speech. Since these effects act in the opposite direction for fast speech, they may also be masking potential increases in motor instability when speaking quickly. While the results do not dispute that reduced motor stability contributes to increased STI values in slow speech, they suggest its role has likely been overestimated in prior research. Supplemental Material: https://doi.org/10.23641/asha.30625145

Purpose: When using the spatiotemporal index (STI) to measure variability across repetitions of the same stimulus, researchers will typically screen and remove productions that contain errors or disfluencies. However, this screening process is highly subjective, reduces the amount of data available, and may generate samples that are less representative of true speech difficulties. In this study, we quantify the degree to which the STI is skewed by the inclusion of highly deviating productions and whether alternative calculations could better facilitate their inclusion. Method: First, we conducted a controlled simulation to quantify how highly deviating productions skew STI values. The traditional STI calculation was compared to three robust alternative measures proposed to reduce the influence of outlying productions. Next, using audio recordings from typically developing (TD) children and children with childhood apraxia of speech (CAS), we investigated how effectively each STI measure differentiated the two groups. Results: Simulation findings demonstrated that the STI can be heavily skewed (more than doubling in value) by the inclusion of a single outlying production. In contrast, the robust alternative measures were all able to incorporate multiple outlying productions before their value was significantly altered. The proposed best-5 STI measure produced larger group differences between TD children and children with CAS compared to the traditional STI in both “Mom pets the puppy” and “Buy Bobby a puppy” stimuli. Conclusions: The STI is highly sensitive to outlying productions and requires careful consideration of the repetitions included in its calculation. However, conservative approaches to data removal may be problematic when studying populations that are prone to fluency errors. In these scenarios, more robust alternatives to the STI, such as the best-5 STI measure, may provide a more practical approach to measuring speech variability. Supplemental Material: https://doi.org/10.23641/asha.27973236

Purpose: When faced with challenging communicative situations, people with dysarthria are commonly advised to rephrase their message, using common words and keeping sentences short and manageable. However, it remains unclear whether relevant clinical populations can implement these changes on demand. The goals of this study were to (a) identify lexical changes that occur when speakers are prompted to rephrase sentences and (b) examine how rephrasing messages affects acoustic measures of speech production and listener perceptual ratings. Method: Speech samples were collected from 11 speakers with Parkinson's disease and 11 age-matched adults. In a baseline condition, speakers read 29 sentences from the Natural Stories Corpus. Following this, speakers received verbal instructions on how to rephrase statements and were prompted to make each sentence easier to understand (rephrasing condition). Transcripts from both conditions were analyzed to measure lexical diversity and sophistication. Sentence-level measures of average sound pressure level, pitch variability, speech rate, pause duration, and listener ratings of ease of understanding were also investigated. Results: Both speaker groups were able to significantly reduce the lexical diversity and sophistication of sentences following a brief educational protocol. However, rephrased speech stimuli were not judged as easier to understand. Acoustically, rephrased stimuli were associated with greater pause durations and a reduction in average speech intensity. Conclusions: Speakers, both with and without a diagnosis of Parkinson's disease, can simplify their language when prompted to do so. However, it is possible that the cognitive demands associated with rephrasing language could interfere with the clarity of speech production.

Purpose: Automatic measurements of fundamental frequency ( F 0) typically contain tracking errors that can be challenging to accurately correct. This study assessed to what degree these errors change F 0 summary statistics in speakers with Parkinson's disease (PD) and neurotypical adults. In addition, we include a case study examining how the removal of tracking errors influenced our ability to predict a perceptual outcome measure, speech expressiveness, associated with dysarthria and PD. Several different statistical approaches for characterizing F 0 variability were used to demonstrate the influence of tracking errors. Method: Eight speakers with PD and eight neurotypical speakers were recorded reading The Caterpillar passage. F 0 measurements were extracted in Praat and tracking errors were manually identified. The effect of tracking errors on F 0 mean and standard deviation was statistically analyzed. Twenty listeners rated speech expressiveness across 80 sentences. The relationship between listener ratings and F 0 variability was examined using different statistical approaches for characterizing F 0 variability (with and without tracking errors). Results: Measurements of F 0 standard deviation, but not F 0 mean, were significantly affected by tracking errors. Relationships between measurements of F 0 variability and expressiveness were strengthened when tracking errors were removed from data analysis. Conclusions: Tracking errors significantly alter F 0 standard deviation values for both speakers with PD and neurotypical adults. Case study evidence also suggests that tracking errors can reduce the strength of relationships between F 0 variability and perceptual outcome measures, such as speech expressiveness.

Purpose: Prior research introduced quantifiable effects of three methodological parameters (number of repetitions, stimulus length, and parsing error) on the spatiotemporal index (STI) using simulated data. Critically, these parameters often vary across studies. In this study, we validate these effects, which were previously only demonstrated via simulation, using children's speech data. Method: Kinematic data were collected from 30 typically developing children and 15 children with developmental language disorder, all spanning the ages of 6–8 years. All children repeated the sentence “buy Bobby a puppy” multiple times. Using these data, experiments were designed to mirror the previous simulated experiments as closely as possible to assess the effects of analytic decisions on the STI. Experiment 1 manipulated number of repetitions, Experiment 2 manipulated stimulus length (or the number of movement units in the target phrase), and Experiment 3 manipulated precision of parsing of the articulatory trajectories. Results: The findings of all three experiments closely mirror those of the prior simulation. Experiment 1 showed consistent underestimation of STI values from smaller repetition counts consistent with the theoretical model for all three participant groups. Experiment 2 found speech segments containing fewer movements yield lower STI values than longer ones. Finally, Experiment 3 showed even small parsing errors are found to significantly increase measured STI values. Conclusions: The results of this study are consistent with the findings of prior simulations in showing that the number of repetitions, length of stimuli, and amount of parsing error can all strongly influence the STI independent of behavioral factors. These results further confirm the importance of closely considering the design of experiments, which employ the STI.

Nonparametric estimation of information divergence functionals between two probability densities is an important problem in machine learning. Several estimators exist that guarantee the parametric rate of mean squared error (MSE) of O(1/N) under various assumptions on the smoothness and boundary of the underlying densities, with N being the number of samples. In particular, previous work on ensemble estimation theory derived ensemble estimators of divergence functionals that achieve the parametric rate without requiring knowledge of the densities' support set and are simple to implement. However, these and most other methods all assume some level of differentiability of the divergence functional. This excludes important divergence functionals such as the total variation distance and the Bayes error rate. Here, we show empirically that the ensemble estimation approach for smooth functionals can be applied to less smooth functionals and obtain good convergence rates, suggesting a gap in current theory.

Purpose: The purpose of this study was to investigate acoustic spatiotemporal variability in children with childhood apraxia of speech (CAS) compared to children with typical development (TD). Increased spatiotemporal variability in children with CAS relative to nonapraxic peers has been documented in multiple kinematic studies. To date, few studies have investigated spatiotemporal variability in CAS using an acoustic signal. Method: Data were drawn from 10 children with CAS and 10 children with TD, ranging in age from 5;0 to 11;2 (years;months), who participated in an online study. All children with CAS had a verified diagnosis and showed at least five CAS features across at least two tasks, independently confirmed by two speech-language pathologists with expertise in CAS. Children repeated the sentences “Buy Bobby a puppy” (BBAP) and “Mom pets the puppy” (MPP) 10 times each. The acoustic spatiotemporal index (STI), calculated from the amplitude envelope, was used to investigate acoustic spatiotemporal variability. Using a regression analysis, we analyzed group differences in STI values while controlling for age and gender. Results: Children with CAS did not produce enough usable tokens of BBAP to analyze. MPP revealed significantly higher acoustic STI values in children with CAS compared to children with TD. No significant effects were observed for age or gender. Acoustic data collected online without the use of specialized equipment yielded high-quality audio data from which amplitude envelope tracing could be reliably accomplished. Conclusions: This study adds to a growing body of empirical data indicating that children with CAS produce speech with more variable spatiotemporal control compared to children with TD. In addition to producing more variable speech, children with CAS were less consistently able to form productions free of phonetic errors. These findings also demonstrate the feasibility of using remote data collection to investigate acoustic spatiotemporal variability, which allows for the recruitment of larger samples of low-incidence populations. Supplemental Material: https://doi.org/10.23641/asha.26814883