**3. The Conceptual Framework of Trust Perception-Based Purchasing Behavior**

This study analyzes the international purchasing behavior by Thai individuals. The structure of perception-based behavior is shown in Figure 1.

**Figure 1.** Conceptual framework and hypothesis.

The four hypotheses in this study are as follows:

**Hypothesis 1.** *Thai subjects have low satisfaction with YouTube advertising.*

**Hypothesis 2.** *Male subjects have lower satisfaction with YouTube advertising than Female subjects.*

**Hypothesis 3.** *The behavioral trend of Thai subjects in response to YouTube advertising is that of avoidance rather than confrontation.*

**Hypothesis 4.** *YouTube advertising has a significant e*ff*ect on behavioral trends.*

### **4. Research Methodology and Behavioral Modality Establishment**

This research used a quantitative approach through the collection of data from Thai subjects using a questionnaire and sample population are presented on Table 1.


**Table 1.** Sample Thai population.

#### *4.1. Sample Selection*

There are roughly 960,906 individuals living in Chiang Mai, Thailand [13]. Therefore, we determined the sample size for a precision level defined by a 95% confidence level and a degree of accuracy of 0.05. Thus [14],

$$\begin{array}{rcl} \text{Size} &=& \frac{X^2 \text{NP } (1-P)}{d^2 \ (N-1) + X^2 P \ (1-P)} \\\\ \text{Size} &=& \frac{3.84 \times 1.728,242 \times 0.50 \times (1-0.50)}{0.05^2 \ (1.728,242-1) + 3.84 \times 0.50 \times (1-0.50)} \\\\ &\text{Size} = 384, \end{array} \tag{1}$$

where *X*<sup>2</sup> is the Chi-Square value for df (degrees of freedom) = 1 for the desired confidence level of 95%, i.e., *X*<sup>2</sup> = 3.84; N is the population size; P is the population proportion (defined as 0.50); and d is the degree of accuracy (expressed as a proportion), these are shown on Table 1.

According to the above formula and the desired confidence level and accuracy, the sample Thai population size for this study was 384 Thai subjects. For easier analysis, we added 16 more subjects. Therefore, there were 400 subjects in the sample.

### *4.2. Random Sampling*

We used probability sampling by multi-stage sampling. The steps were:

Step 1: From Thai sample (400 subjects, these data are presented as Table 2), we conducted non-probability sampling by using quota sampling with respect to the age of Thai subjects.

**Table 2.** Sample population of Thai subjects.


Step 2: We used non-probability sampling, accident sampling, snowball sampling, and convenience sampling to distribute the questionnaire until the number of completed questionnaires was that of the calculated sample size.

### *4.3. The Study Instrument*

The questionnaire was written in two languages, English and Thai, and consisted of three parts.



**Table 3.** Importance scale.

#### *4.4. Data Analysis and Statistics for Analyzing Data*

Data analysis for analyzing data and classification of satisfaction are presented here as Tables 4 and 5. We used the standardized satisfaction of subjects by following the formula below [15]:

$$\text{Class Interval} = \frac{\text{Upper Class Limit} - \text{Lower Class Limit}}{\text{Amount of Class}} = \frac{5 - 1}{5} = 0.80. \tag{2}$$


**Table 4.** Data Analysis Statistics for Analyzing Data


**Table 5.** Average range of points for classification of satisfaction.

*4.5. Perception-Based Behavior toward Purchasing: Reliability Analysis*

This study analyzed the purchaser's perception-based behavior and the reliability of purchasing perception for relevant purchases in the future. Reliability is defined as the probability that an element (that is, a component, subsystem, or full system) will accomplish its assigned task within a specified time, which is designated by the interval t = [0, tM] [16]. Reliability is closely related to the following four factors: (1) probability value; (2) predetermined function; (3) predetermined life; and (4) prescribed environment. The probability function of reliability allocation is defined in the next subsection.

#### *4.6. Exponential Distribution*

Hazard rate:

$$h(\mathbf{x}) = (f(\mathbf{x})) (\mathbb{R}(\mathbf{x})),\tag{3}$$

where *f(x)* is the probability density function of exponential distribution,

$$f \text{ (x)} = \lambda \text{ } e^{-\lambda x} \text{ , x \ge 0} \tag{4}$$

where λ is the failure rate.

The mean time between failures (MTTF) is calculated by the following calculations.

Let *X* be a random variable that indicates the expiration time. Then, the probability of the product failing at a specific time *x* is

$$P(X \le \mathbf{x}) = F(\mathbf{x}), \mathbf{x} \ge 0,\tag{5}$$

where *F*(*x*) is the failure distribution function.

If the product still functions as intended at time *x*, then

$$R(\mathbf{x}) = P(X > \mathbf{x}) = 1 - F(\mathbf{x}).\tag{6}$$

#### *4.7. Weibull Distribution Probability Density Function*

The probability density function of a Weibull is

$$f(\mathbf{x}) = \frac{\beta}{\lambda} \left(\frac{\mathbf{x}}{\lambda}\right)^{\beta - 1} \exp\left[-\left(\frac{\mathbf{x}}{\lambda}\right)^{\beta}\right], \mathbf{x} \ge 0 \tag{7}$$

The cumulative distribution function is:

$$F(\mathbf{x}) = 1 - \exp\left[-\left(\frac{\mathbf{x}}{\lambda}\right)^{\beta}\right], \mathbf{x} \ge 0\tag{8}$$

Reliability is:

$$R(\mathbf{x}) = 1 - F(\mathbf{x}) = \exp\left[-\left(\frac{\mathbf{x}}{\lambda}\right)^{\beta}\right], \mathbf{x} \ge 0\tag{9}$$

The average time to failure is:

$$\text{MTTF} = \lambda \Gamma \left( 1 + \frac{1}{\beta} \right) \tag{10}$$

*Symmetry* **2019**, *11*, 989

The failure rate function is:

$$h(\mathbf{x}) = \frac{\beta}{\lambda} \left(\frac{\mathbf{x}}{\lambda}\right)^{\beta - 1}, \mathbf{x} \ge 0 \tag{11}$$

where the conditions are as follows: when β < 1, the failure rate decreases with time (early stage); when β = 1, the failure rate is constant (opportunity period); when β > 1, the failure rate increases with time (loss period).

Reliability defines the reliability of a product or system. This study had two key computations: the internal series calculation and the Internet of Things (IOT) system. The *RA* is the A-system reliability, *RB* is the B-system reliability, *RC* is the C-system reliability, and *RD* is the D-system reliability.

$$R\_S = (R\_A)(R\_B)(R\_C)(R\_D) \tag{12}$$

In the internal parallel calculation, the internal components of the system are connected in series, and the Internet of Things (IoT) system is also connected in series. The parallel equation is as follows:

$$R\_P = 1 - (1 - R\_A)(1 - R\_B)(1 - R\_C)(1 - R\_D) \tag{13}$$

The above scenario and calculations were applied in this experiment by using 400 subjects from Thailand, as shown in Table 6.


**Table 6.** Thai sample population.

From the Thai sample population (400 subjects), we conducted non-probability sampling by using quota sampling with respect to the age of Thai subjects.

The purpose of this study was to determine the reliability of using perception-based behavior to predict an individual's final decision on their willingness to purchase the product. The results are shown in Table 7. The reliability of the predicted purchasing behavior (i.e., whether the individual will purchase) was low when it was based on a single use of YouTube. With multiple uses of YouTube, the predicted purchasing behavior was highly reliable. Overall, the reliability of determining purchasing behavior on the basis of YouTube use was greater than 90%.


**Table 7.** Statistical analysis of reliability on purchaser and abandonment purchase behavior.

#### **5. Summary of Findings**

The data from the sample population are summarized in Table 8.


**Table 8.** Demographic characteristics of sample population.

The results indicate that most of the Thai subjects were students with a bachelor's degree, and the age range of the majority was 21–30 years old.

#### **6. Hypothesis Testing**

#### *6.1. Hypothesis 1: Thai Subjects Have Low Satisfaction with YouTube Advertising*

Table 9 reports the satisfaction of the subjects with the six types of YouTube advertising. The results reveal that Thai subjects had a moderate satisfaction with mastheads, display ads (banners), and TrueView in-search ads. Overlay-in-video ads and TrueView in-stream ads were scored as low satisfaction, and Thai subjects had very low satisfaction with non-skippable in-stream ads.


**Table 9.** Satisfaction with types of YouTube advertising.

We next used the ranges of average points to classify satisfaction, as specified in Table 5. The results reveal that Thai subjects have low satisfaction with YouTube advertising, with a mean score of 2.54.

#### *6.2. Hypothesis 2: Male Subjects Have Lower Satisfaction with YouTube Advertising than Female Subjects*

Table 10 compares the mean scores given by male and female subjects. These results demonstrate that the difference in satisfaction with YouTube advertising between genders was not significant at the 0.05 level, which means that males and females had the same opinion toward YouTube advertising.


**Table 10.** Satisfaction with YouTube advertising according to the gender of Thai subjects (independent samples *t*-test).

The mean satisfaction score given by male subjects was 2.48, which was lower than that given by female subjects, whose mean was 2.60. In other words, male samples have lower satisfaction with YouTube advertising than female samples.

#### *6.3. Hypothesis 3: The Behavioral Trend of Thai Subjects in Response to YouTube Advertising Is that of Avoidance Rather Than Confrontation*

The results show that most Thai subjects engaged in Avoidance behavior toward YouTube advertising. This option was selected 564 times, which was 76.2% of the total. The most common avoidance behavior was waiting for five seconds and then skipping the ad. This behavior was selected 328 times, which is 44.3% of the total.

Confrontation behavior was reported 176 times (23.8% of the total). The most common confrontation behavior was leaving the YouTube advertisement open, which was selected 124 times, which was 16.8% of the total.

This implies that the behavioral trend of Thai subjects was more avoidance than confrontation in response to YouTube advertising. These data are presented here as Table 11.

**Table 11.** The number and percentage of behaviors toward YouTube advertising reported by Thai subjects.

