Advertising avoidance is considered to be one of the biggest obstacles of advertising and it has been a focus for researchers to find out the determinants of ad avoidance. Ad avoidance is defined by Speck and Elliott [13] as “all actions by media users that differentially reduce their exposure to ad content”. Many early studies explored the audience avoidance behavior on television, radio, newspapers, magazines and other traditional media. Clancey [14] found that audience avoid ads by ignoring distracting ads which is cognitive avoidance, leaving the room which is physical avoidance, changing the channel which is mechanical avoidance. Speck and Elliott [13] found that in the traditional media, the factors affecting ads avoidance are demographic characteristics (gender, age, income), ads perception (advertising interruption), communication problems (search barrier) and so on.

In recent years, scholars have focused on the exploration about ads avoidance on the Internet [11] [14] - [16] . Many prior researches conceptualizes ad avoidance as a tripartite disposition composed of cognitive, affective, and behavioral components, namely cognitive avoidance, affective avoidance and behavioral avoidance [11] [12] [17] . Cognitive ad avoidance is the psychological defense mechanism that results in users intentionally ignoring an ad they are exposed to [17] . Cognitive avoidance is rooted in consumers’ beliefs about ads, which can lead to them deliberately ignoring an ad. Affective ad avoidance, on the other hand, involves negative feelings and the expression of emotional reactions toward an advertisement [18] [19] . Behavioral avoidance is “consumer avoidance actions other than lack of attendance” [11] . This article is also measured by these three dimensions of ads avoidance in the context of online behavioral advertising as follows: consumers deliberately ignore any ads (cognitive avoidance), hate online behavioral ads (affective avoidance), close or intercept online behavioral ads (behavioral avoidance).

Goal impediment has been considered as one of the most significant predictors of ad avoidance. The Internet is a goal-oriented medium where users are usually engaged in pursuing specific tasks that can be interrupted by exposure to unsolicited ads, which will make consumers have negative experiences and ads avoidance. Li et al. [20] found that the perception about advertising interruption will result in the cognitive avoidance and behavioral avoidance directly. Speck and Elliott [13] also found that, interrupted by advertising will directly affects consumers’ advertising avoidance behavior. Consumers may also occur negative emotions because the advertising content is not consistent with the webpage they are browsing [21] . Therefore, we propose the following hypotheses:

H1a: Goal Impediment is related to negative experience of online behavioral advertising positively.

H1b: Goal Impediment is related to avoidance of online behavioral advertising positively.

Privacy concern is defined as “the degree of consumers’ concern about potential privacy was invaded” [22] . With the rapid advances of information processing and communication technologies, most marketers collect and track personal information about specific consumer purchase histories and characteristics to identify the best prospects, customized advertising, and promotion strategies; implement highly targeted direct- marketing efforts; and establish reward and loyalty programs [23] - [25] . However, information privacy is a great concern to consumers [26] [27] . These marketing tools will make consumers to worry about their privacy has potentially exposed. Privacy concern has a negative impact to consumers’ buying behavior, trust and information control [28] . Whether the privacy is protected influences consumers’ trust in e-commerce a lot [2] . With the intensification of privacy concern, consumers are likely to have a negative experience and avoid online behavioral advertising such as don’t provide complete information, cancel or block ads, or make a very bad comments about the ads [29] . Therefore: we propose the second hypotheses:

H2a: Privacy concern is related to negative experience of online behavioral advertising positively.

H2b: Privacy concern is related to avoidance of online behavioral advertising positively.

Perceived Personalization refers to the process that is customizing solutions according to consumers’ information [30] ; it is the personal communication on the basis of individual preference [31] . The biggest difference between online behavioral advertising and traditional advertising is that OBA is customized; it is personalized. Prior research [32] has further asserted that consumer reactance to personalized messages can be determined by whether the perceived utility of the advertised products or services offsets the psychological costs of receiving inappropriate personal messages. Perceived utility (e.g., some rewards and benefits from the restricted freedom) significantly decreases consumer reactance against the loyalty program [32] . Studies [33] also found that personalized content is the most effective way to avoid the ads that are regarded as disturbing. When the ads are considered useful and valuable, consumers are inclined to have positive experience to ads and have a lower rate of ads avoidance. Therefore, we propose hypothesis 3:

H3a: Privacy concern is related to negative experience of online behavioral advertising negatively.

H3b: Privacy concern is related to avoidance of online behavioral advertising negatively.

Consumer advertising experience also has a significant impact on consumer attitudes and behavior [34] [35] . In previous studies [11] , prior negative experience is one of independent variables, which is from consumer’s learning perspective; the consumer prior experience will affect the type and manner of information processing. This article is from perspective of “perception, attitude, and practice”. When consumers are contacted with the online behavioral advertising, goal impediment, privacy concerns and perceived personalization will affect the consumer experience immediately the n affect consumers’ attitude and behavior. For online behavioral advertising, the negative experiences may include lack of usefulness, lack of incentive, and so on. Consumers will avoid the sources generating negative experience. Accordingly, we propose hypothesis 4:

H4: Negative experience is related to avoidance of online behavioral advertising positively.

The theoretical model of this paper is shown by Figure 1.

We collected the data through questionnaire. Questionnaire is divided into two parts. The first part is basic information, including gender, age, education, duration, and daily access. The second part is a measure of the model five variables including a total of 26 problem items. All measurement scales are from prior research and adjusted according to the research target. Goal impediment adopts three question items from Cho &

Cheon’s [11] research; privacy concern adopts the scale from Malhotra [36] , Shintaro et al. [37] ; perceived personalization is measured by the scale of Hyejin [38] and Srinivasan’ s research [39] negative experience adopts four question items from Seyedghorban [16] ; advertising avoidance is measured by scale from Cho & Cheon [11] .

All variables used five-point Likert scale to measure. Likert scale is used widely in the current study, which asks subjects to express their own views to a set of statement related to the topic. Using five-point Likert scales (1 = strongly disagree and5 = strongly agree), five latent constructs―advertising avoidance, goal impediment, perceived personalization, privacy concern, and negative experience―were measured in this study.

Questionnaires were issued and collected from May 18 to May 25, 2016. We collected questionnaires online and offline. There are 154 online questionnaires and 120 offline questionnaires. We chose the largest CBD of Guangzhou, Pearl River New City to distribute questionnaires.

There are valid 231 questionnaires excluding the non-standard questionnaires and the effective rate is 84.3%. In the recovered sample, men are 64.5% and women accounted for 35.5%; the population between the ages of 18-35 accounted for 88.9%. 83.1% of the sample population has a university degree or above; 33.3% of the population’s online time is 1 - 2 hours per day and 35.1% of the population daily online time is 3 - 4 hours.

The research used AMOS20.0 and SPSS17.0 software to test reliability and validity. As shown in Table 1, all variables’ Cronbach’s α values are greater than 0.7 and it has reached the requirements. It means that the scale of this paper has good reliability.

In this paper, the scales are mature scales referring to the prior research and modified according to the purpose of this study, which has good content validity. From Table 1, the factor loading of all variables are greater than 0.6; composite reliability (CR) values of each variable is greater than 0.8 and average variance extracted (AVE) is greater than 0.5, which indicates that the measurement scale has good convergent validity. From Table 2, the square root of AVE values of each variable is greater than the correlation coefficient with other variables, which means that scale has good discriminant validity.

The value of diagonal is the square root of AVE.

Amos20.0 was used to explore the relationship between the variables and establish model structure. This model, the absolute fitness X2 = 325.190, df = 292, GFI (0.906) is greater than 0.90; AGFI (0.887) is greater than 0.80; NFI (0.941) is greater than 0.90; RMR (0.060) is less than 0.08; RMSEA (0.022) is less than 0.08. The indicators reach an acceptable range, which means that the structural model has a good degree of adaptation.

Hypothesis test results are shown in Table 3. From the path coefficients and significance of theoretical model, it can be known that the assumptions are through testing except H3a. Goal impediment is related to negative experiences (P < 0.001) and advertising avoidance (P < 0.001) positively; perceived personalization is related to negative experience (P < 0.001) and advertising avoidance (P < 0.001) negatively; privacy concerns is related to advertising avoidance positively (P < 0.001), but there is no significant impact on the negative experience (P > 0.05); the negative experience is related to advertising avoidance positively (P < 0.001).

The study established a theoretical model of online behavioral advertising and enriched the theory in the field of advertising avoidance.

First, in the research of traditional advertising, television, newspapers, magazines and other mass media are difficult to customize content of advertising, so it does not require much thought about perceived personalization and privacy concern. However, as the characteristics of online behavioral advertising, it should not ignore the impact of perceived personalization and privacy concern. This study introduced the two new factors into advertising avoidance theoretical model. It’s an expansion of the research of advertising avoidance.

Secondly, the research introduced negative experience as a mediator of model. The negative experience is independent variable in the traditional research model, while the

research found that goal impediment and perceived personalization affect the consumer experience and further influence the advertising avoidance. The research found out the more complicated relationship among variables.

Behavioral advertising is popular in the Internet age, and it may become more precise in the future. What factors affect avoidance of online behavioral advertising? How do consumers view privacy and personalization? The paper explores these issues.

From the results, the perceived personalization is related to advertising avoidance negatively. When there is a high degree of personalization, consumers can more easily find the desired product, or to know the incentive timely, so they will think the ads are useful, attractive and incentive and therefore they will have lower degree of negative experience to the ads and have a lower tendency of advertising avoidance. This is consistent with previous studies, targeted advertising will increase the good feeling [40] . Maybe this because personalization will make consumers feel that they have contacted with the company. People are less likely to have negative experience to the company if they had contacted with it before [41] . For companies and businesses, this is an important revelation. More accurate analysis of consumer demand will let consumers reduce the negative experience and ads avoidance.

The correlation between privacy concerns and negative experience is not obvious, and its impact on advertising avoidance is lower than perceived personalization on ads advertising. Even so, companies or businesses should pay attention to consumers’ privacy. When collect consumer’s information, companies or businesses should promise the security of personal information and ensure that consumers' right to know and control the use of privacy. When the perceived utility from personalization is greater than the psychological loss from privacy concern, the avoidance will decrease.

There is insufficient in this research. First, the sample selection: as the younger age groups is the main force of the Internet activities, the sample selected mainly was composed by university students and young workers. Therefore future research can expand the sample. Second, behavioral targeting advertising is a relatively new kind of forms of advertising, which is closely linked with Internet. Then whether the form of the ads or the network of ads delivering influence the avoidance, can be explored in the future.