**3. Results**

### *3.1. Visual Analysis*

The initial step for data analysis for this study was a visual analysis of the data plotted as a figure composed of the scores/ratings from the outcome measures (Daily Caregiver Survey and the Sense Sleep App). The data were evaluated observing changes in level, slope, and variability in data points across each phase for both participants' subjective and objective measures. Figures 1–6 represent the caregiver survey (sleep onset latency, sleep duration, number of night wakings, and morning mood) and the data from the Sleep sense app (sleep score, sleep duration, and deep sleep).

**Figure 1.** John's caregiver reported sleep onset latency and sleep duration [reported sleep onset latency is measured 0–60 min and sleep duration is measured 0–10 h (OBS = Observation)].

**Figure 2.** John's caregiver reported number of wakings and morning mood [reported number of wakings is measured 1–5 times and morning mood ratings are 1–5(OBS = Observation)].

**Figure 3.** John's Sleep Sense sleep score sleep duration and deep sleep duration [sleep score is measured between 1 and 10, total sleep hours are measured in 0–12 h, and deep sleep is measured in 0–8 h (OBS = Observation)].

**Figure 4.** Katie's caregiver reported sleep onset latency and sleep duration [reported sleep onset latency is measured in minutes and sleep duration is measured in hours (OBS = Observation)].

**Figure 5.** Katie's caregiver reported number of wakings and morning mood [reported number of wakings is measured 1–5 times and morning mood ratings are 1–5 (OBS = Observation)].

**Figure 6.** Katie's Sleep Sense sleep score sleep duration and deep sleep duration (sleep score is measured between 1 and 10, total sleep hours are measured in 0–12 h, and deep sleep is measured in 0–8 h).

John's caregiver reported a time to fall asleep during the baseline phase but was reduced to just below 40 min during the intervention and withdrawal phases (see Figure 1). The caregiver's perception of John's total hours of sleep did not see a significant change in level or slope across the phases of the study (see Figure 1). The number of night waking times reported by the demonstrated a downward trend across the baseline and intervention phases and leveled off during the withdrawal phases (see Figure 2). Finally, the caregiver's perception of John's morning mood began relatively level across the baseline and intervention phases but became improved towards the end of the intervention and withdrawal phases (see Figure 2). Visual analysis of the Sleep sense app for John yielded limited changes in slope, level, or variability (see Figure 3).

Analysis of Katie's data through visual analysis yielded notable changes in the caregiver's reporting of time to fall asleep (see Figure 4) during the intervention phase and return to higher levels of approximately 50 min in the withdrawal phase. John's caregiver noted no significant changes in the total hours of John's sleep across all phases. Katie's caregiver reported a slight increase over the baseline phase in the number of night wakings but a small drop and between 1 and 2 times across the intervention phase, returning to higher levels during the withdrawal phase (see Figure 5). Evaluating Katie's caregiver's perception of Katie's morning mood saw a trend on poorer ratings during the baseline phase with a significant improvement in morning mood ratings during the intervention phase with a return to poorer behavioral ratings during the withdrawal phase. Visual analysis of the Sleep sense app for Katie also yielded little changes in slope, level, and variability (see Figure 6).

### *3.2. Quantitative Analysis*

The percentage of non-overlapping data (PND) statistic was calculated to assess treatment effectiveness. Calculations were conducted using Microsoft Excel. Scruggs and Mastropieri [33] provide evaluative criteria for implementing this frequently used analysis method for single-case research. The index of treatment effectiveness is based on the percentage of non-overlapping data using the following criteria: PND ≥ 90% = Very Effective, PND 70–90% = Effective, PND 50–70% = Questionable effectiveness, and PND < 50% = Ineffective. When applying these methods in the current study, the only factor categorized as Effective was for John the morning mood category for Katie the categories time to fall asleep, and the number of night wakings (see Tables 1 and 2).


**Table 1.** John and Katie percentage of non-overlapping data (PND) analysis of daily caregiver survey.

PND ≥ 90% = Very Effective, PND 70–90% = Effective, PND 50–70% = Questionable Effectiveness, and PND < 50% =Ineffective.


**Table 2.** John and Katie PND analysis Sense Sleep App.

PND ≥ 90% = Very Effective, PND 70–90% = Effective, PND 50–70% = Questionable Effectiveness, and PND < 50% = Ineffective.
