*3.2. Research Hypothesis*

According to the research model, there are three independent variables, namely perceived usefulness (PU), perceived ease of use (PE), and ad attitude (AT), and four antecedent variables, namely social interaction (SI), informativeness (IR), intrusiveness (IN), and relevance (RE). In addition, there are two mediator variables called perceived usefulness (PU) and ad attitude (AT). These variables are supposed to impact the dependent variable named purchase intention (PI).

To reveal the relationships among these factors, we propose hypotheses to the research questions presented earlier, as listed in Table 1. The right column in Table 1 shows the corresponding hypotheses that are proposed to answer the research question. In the following subsections, we will explain each hypothesis.

**Figure 2.** The research model extended from Technology Acceptance Model (TAM) [23] (each rectangle represents one variable. The mediator variables are shaded grey. Purchase intention is the dependent variable, and the remaining five are independent variables).


**Table 1.** Research questions and corresponding hypotheses.

Sociability has been defined as 'the extent to which the computer-mediated communication environment facilitates the emergence of social space by allowing social affordance [33]. Social interaction can mitigate users' perceptions of ad avoidance and enhance the effectiveness of ads [19]. Interactive online media are an increasingly preferred format for users and advertisers. Particularly, health-related short-video ads are more social than traditional social media advertising, whereby audiences can, in a timely manner, view, share, and post barrages on these ads with anchors and peer audiences. Given the massively real-time interactivity distinguishing live video from traditional social

media, real-time interactivity may enhance the influence of social interactions or depress viewers' experience because of barrages filled with mobile screens. Therefore, we make the following hypothesis:

**Hypothesis 1 (H1)**. *Social interaction of health-related short-video ads positively a*ff*ects users' perceived usefulness (PU) of ads.*

Intrusiveness describes the extent to which the content is messy and irritating to users [34]. Intrusiveness in social media advertising will be negatively related to the perceived usefulness of ads [2,14,15]. When an inconsistent ad is posted with live videos, audiences may spend efforts on non-major tasks and cannot effectively deal with advertising information, resulting in the ineffectiveness of advertising. In addition, the small screen of a mobile phone may also make intrusiveness be a negative factor in live video ads. Moreover, intrusive ads may bring the negative emotion to users [35]. Hence, we make Hypothesis 2:

**Hypothesis 2a (H2a)**. *The intrusiveness of health-related short-video ads negatively a*ff*ects users' perceived usefulness (PU) of ads.*

**Hypothesis 2b (H2b)**. *The intrusiveness of health-related short-video ads negatively a*ff*ects users' perceived ease of use of ads.*

Informativeness represents the information that is helpful and resourceful [36]. It can inform users about product alternatives. Consumers tend to gain information more through unconditional, interpersonal information exchange [37]. Studies have found informativeness to be important in the formation of consumer attitudes to electronic commerce websites as well as to SNS advertising [10] and to social media advertising [13]. The informative content of a message positively influences the perceived value of online advertising [15]. Hence, we make Hypothesis 3:

**Hypothesis 3 (H3)**. *Informativeness of health-related short-video ads positively a*ff*ects users' perceived usefulness (PU) of ads.*

Relevance is used to capture users' general perception of similarity between short-video ads and live video content. According to the limited capacity model of attention [18], the total capacity allocated to process activities can be divided into the primary capacity for the most important task and the spare capacity for less important tasks. When an inconsistent ad is placed in the primary task field, audiences may reduce the likelihood of primary capability being allocated to processing the ad. Advertising relevance decreases ad skepticism and ad avoidance in e-mails, direct mails, telemarketing, and text messages [38]. Hence, we make Hypothesis 4:

**Hypothesis 4 (H4)**. *The relevance of health-related short-video ads positively a*ff*ects users' perceived usefulness (PU) of ads.*

Perceived usefulness (PU) refers to the degree to which individuals perceive the use of health-related short-video ads. If users think that short-video ads are useful, they are much likely to have a positive attitude on using them. Attracting and motivating short-video ads can also lead to purchase intention [2]. Hence, we make Hypotheses 5 and 6 as follows:

**Hypothesis 5 (H5)**. *Users' perceived usefulness of health-related short-video ads positively a*ff*ects their attitudes toward ads.*

**Hypothesis 6 (H6)**. *Users' perceived usefulness of health-related short-video ads positively a*ff*ects their purchase intentions*.

Perceived ease of use (PEOU) is defined as the mobile phone users' expectations about the e ffort required to use live video advertising messages. Short-video advertising requires some cognitive effort from users such as maneuvering through screens, reading the advertising, and making a quick evaluation of whether it is worthy (in terms of time or e ffort) to take further action. Therefore, the easier they perceive using short-video advertisements to be in general, the more likely they are to later engage in the action implied by the ad. Hence, we make Hypotheses 7 and 8:

**Hypothesis 7 (H7)**. *Users' perceived ease of use of health-related short-video ads positively a*ff*ects their perceived usefulness.*

**Hypothesis 8 (H8)**. *Users' perceived ease of use of health-related short-video ads positively a*ff*ects their attitudes toward ads.*

Attitude towards advertising is regarded as an important factor for advertising managemen<sup>t</sup> because consumers' attitude can influence purchase intention [39]. Previous research has indicated that purchase intention is a critical indicator of advertising e ffectiveness and may be a ffected by indicators such as attitude towards ads [2,12,36]. For example, Lin and Kim found that the attitude towards ads a ffected brand awareness and purchase intentions [2]. Tran found that purchase intention was positively a ffected by ad attitude on Facebook [40]. Whether consumers prefer a product or an advertisement depends on the informativeness of the advertisement, i.e., whether the advertisement can provide the opportunity and convenient ways for purchase. Therefore, the informational nature of advertisement a ffects the consumer's purchase intention. Based on these observations, we make the following hypothesis:

#### **Hypothesis 9 (H9)**. *Users' attitudes of health-related short-video ads positively a*ff*ect their purchase intentions.*

Note that we only set up the Hypotheses H1–H4 to measure the impact of the four constructs on PU but not on PEOU. This is mainly because the constructs are all about the information o ffered by advertisements, but not about the operations or behavior of advertising. On the other hand, the perceived ease of use (PEOU) refers to the ease of operations that users perceive, i.e., PEOU is operation related. Thus, in this paper, we only focus on the impact of the constructs on PU.

#### **4. Data Collection and Measurement**

### *4.1. Data Collection*

We mainly conducted questionnaires on users who have online shopping experiences due to mobile live ads, so that we could gain useful feedback from experienced users on short-video platforms. The data for this study were collected using an online survey in Douyu.com (a famous short-video platform in China) during January 2018, which lasted for three weeks. Three hundred and fifty questionnaires were sent to the viewers with anchors' help, and 289 questionnaires were received. Thirty-three questionnaires that contained inconsistent answers or incomplete information filling were removed from the data set.

Note that it is not possible to ask the Douyu platform to only present health-related short-video ads to the participants, because the platform does not o ffer the function of filtering health-related short-videos (actually, this could be a complicated task which needs e ffective algorithms). We o ffered the participants three weeks to return questionnaires, and all questions were restricted to health-related short-video ads (an example is given in Figure 1). Although we do not know what health-related videos the participants watched, we believe that the involved short-videos covers a wide range of health-related products. This is because the number of short videos on the Douyu platform is dramatically increasing day by day. In addition, di fferent users have various interests, so the platform will recommend di fferent short-video ads to each user.

Specifically, we asked three Douyu anchors for help, namely Prisoner, Valila, and Liu shaji, to distribute the questionnaires to their fans. The time span of the questionnaires was from January 3, 2018 to January 24, 2018. Consequently, we collected 256 valid questionnaires, which met the analytical requirements of the structural equation model.

In the data set, the proportions of males and females were 53.5% and 46.5%, respectively. There were 79.8% users that were under 30 years old. The duration of using live videos ranged from 0 to 2 years, and 58.6% of the users had experienced live videos for 1 to 2 years. Furthermore. 50.4% of the users were undergraduates and graduates (see Table 2). The attributes of the surveyed data were similar to those in the IResearch report (IResearch, The research report on the China mobile live video market, http://wreport.iresearch.cn/uploadfiles/reports/636161408839324949.pdf).


