Exploring factors influencing Internet users’ adoption of Internet television in Taiwan
First Monday

Exploring factors influencing Internet users' adoption of Internet television in Taiwan by Kenneth C.C. Yang and Yowei Kang



Abstract
This study examined how demographics, Internet use motivation, and beliefs about Internet television influenced Internet users’ intentions to adopt Internet television in Taiwan. The belief factors of users in programming quality, government regulation, and media impact contributed significantly to predicting an intention to adopt Internet television. Results from hierarchical regression also demonstrated that gender and Internet use motivations were predictive as well.

Contents

Introduction
Literature review
Methodology
Discussion
Conclusion

 


 

Introduction

The rapid development of Internet and its applications has led to many new broadcasting services. One of the emerging applications is to stream video programs via the Internet (Lin, 2004). These applications are generally labeled as webcasting (Lin, 2004), or Internet television (Gerbarg and Noam, 2004; Katz, 2004). Gerbarg and Noam (2004) pointed out that Internet television is a product of digital convergence in telecommunications, the Internet, television, and computer applications. Technically, Internet television has been developed from the capability of the Internet to distribute full–motion video to users (Katz, 2004). However, the arrival of Internet television has been envisioned to affect perceptions of how television should be defined (Katz, 2004). In other words, Internet television will blur boundaries between telecommunications, broadcasting, and information technology.

Despite the popularity of the term “Internet television” in trade publications, there is no common definition for this new kind of broadcasting (Noll, 2004). As Gerbarg and Noam (2004) have observed, the term has been used interchangeably with other similar terms, such as Web TV, enhanced TV, personal TV, interactive TV, or IPTV. Gerberg and Noam (2004) provided the most elaborate definition of Internet television:

At the lower end of complexity, it is merely a narrowband two–way Internet–style individualized (“asynchronous”) channel that accompanies regular one–way “synchronous” broadband broadcast TV or cable. This Internet channel can provide information in conjunction with broadcast programs, such as details on news and sports, or enable transactions (including e–commerce) in responses to TV advertisements. This is known as ‘enhanced TV.’ At the other end of complexity is a fully asynchronous two–way TV, with each user receiving and transmitting individualized TV programs, including direct interaction in the program plot line. In between is one–way broadband with a narrowband return channel that can be used to select video programs on demand (VOD). [1].

Technologically speaking, the evolution of Internet television depends on several essential technological factors. Certain technological advances have led to the following developments: 1) increased ability to process user feedback; 2) a rapid increase in effective distribution capacity; and, 3) increased storage and processing power controlled by viewers [2].

There have been several Internet television experiments in the United States. They include Web sites such as ifilm.com, atomfilms.com, icebox.com, and etertaindom.com that deliver video contents to Internet users (Waterman, 2004). In Taiwan, Web sites operated by television stations or commercial Internet television sites (e.g., http://www.webs-tv.net/url/website.asp) provide either free or subscription access to a variety of Internet programming. In spite of many optimistic views about the potential impacts of Internet television, Owen [2] was pessimistic because of inadequate bandwidth and the Web’s unsuitable architecture for broadcasting applications. Past studies that examined Internet television have mainly taken an economic and technological approach [3]. Consumers’ needs and uses of new technologies are rarely studied [4].

The success of Internet television as a potential revenue source for broadcasting industry depends on technological, regulatory, and consumer factors. Waterman (2004) identified five economic characteristics of the Internet that will influence Internet television business models and contents: 1) lower delivery costs and reduced capacity constraints; 2) more efficient interactivity; 3) more efficient advertising and sponsorship; 4) more efficient direct pricing and produce bundling; and, 5) lower costs of copying and sharing. In terms of technology, Einhorn (2004) pointed out the following new capabilities provided by digital and Internet technology: 1) time–shifting; 2) space–shifting; 3) personalization; 4) screening; 5) transforming; 6) multimedia; 7) morphing; 8) archiving; 9) repackaging; 10) hyperlinking; and, 11) user communities and chat rooms.

Consumers’ needs often decide whether a new technology will become successful (Carey, 2004). Carey (2004) explored a key question that affects the success of Internet television. He asks, “Do audiences want to watch video on a personal computer and is there a demand for all of the associated features that could be provided, such as interactive television, customization of video, and two–way video telephone calls?” Carey (2004) found that older consumers with high–end home theatre TVs were less enthusiastic about video over the Web. Similarly, Lin (2004) explored the adoption of webcasting from communication research paradigms. She examined whether users’ gratifications, motivations, media substitution, and fluidity of the Internet can predict audience’s interests in webcasting. Her research found that diversion/escape and perceived fluidity significantly predicted interest.

The shift of research on consumer needs addresses a critical dimension about the diffusion of Internet television. The relationships between Internet television and consumer adoption are important and need further exploration. Without extensive diffusion and adoption, Internet television will likely follow the same fate of many failed technologies in the past 20 years. This study examines factors influencing Internet users’ adoption of Internet television in Taiwan. Specifically, this study asks the following questions:

Research Question 1: What are Internet users’ beliefs about Internet television?
Research Question 2: How will Internet users’ beliefs influence their adoption intention of Internet television?
Research Question 3: How will demographics, innovativeness, and use motivations influence Internet users’ adoption intention of Internet television?

 

++++++++++

Literature review

Internet use motivations

Motivations relative to the use of the Internet have been well–researched (Joines, et al., 2003; Parker and Plank, 2000; Stafford, et al., 2004). Earlier uses and gratifications studies mainly provided a typology of use motivations by extending the uses and gratifications framework to study emerging new media (Kaye, 1998; Lin, 2004; Morris and Ogan, 1996; Parker and Plank, 2000; Stafford, et al., 2004). For example, Joines, et al. (2003) explored motivations for consumer Web use and identified two types of Web uses. Stafford, et al. (2004) also factor analyzed Internet users’ gratifications and extracted three factors motivating Internet use behavior: process, content, and social gratifications. Nevertheless, Stafford, et al. (2004) criticized that present uses and gratifications have failed to develop new Internet–specific gratifications.

The majority of uses and gratifications research has focused on the effects of consumer motivational factors on media use pattern and behavior, media effects, and technology choice and adoption. Digressing from traditional uses and gratifications research, Stafford, et al. (2004) integrated consumer media use with their adoption of technology by combining uses and gratification with diffusion theory. On the basis of their summary of previous technology adoption and acceptance research [5], consumer perception of new technologies will affect their willingness to adopt a given technology. For example, the technology acceptance model [6] postulated that consumer perceived usefulness (PU) and perceived ease of use (PEOU) of a specific technology influence their attitudes toward and adoption intention of a technology. Similarly, Lin (2004) found that audience’s perceived fluidity of the Internet medium significantly predict people’s viewing interest of webcasting programs. Moreover, innovation diffusion theory [7] also argued that consumer perceived relative advantage, compatibility, complexity, and observability will predict the diffusion of a new innovation. Stafford and Stafford [8] proposed that consumers are motivated by both external and internal factors to adopt new technologies. External factors included normative influences, while internal factors included personal goals and desires (Stafford, et al., 2004). We argued that consumer motivational factors would influence their perception and adoption of Internet television. Therefore, we proposed the following two research questions:

Research Question 4: Do Internet use motivations predict Internet users’ intention to adopt Internet television?
Research Question 5: Which Internet use motivation predicts Internet intention to adopt Internet television?

Demographics

Demographic variables have been found to affect Internet users’ subscription intention of interactive television (Leung and Wei, 1998). Leung and Wei (1998) found that male respondents tend to have more positive attitudes toward interactive television and have a higher subscription intention, compared with female respondents. Furthermore, respondents’ education level also positively affects the above variables. However, given the inconsistent results reported about demographic variables in influencing consumers’ adoption behavior (Leung and Wei, 1998; Lin, 2004), we proposed the following research question:

Research Question 6: Do Internet users’ demographic variables predict Internet users’ intention to adopt Internet television?

Innovativeness

We include another individual characteristic, innovativeness, in this study because the variable was expected to influence consumers’ beliefs about and adoption intention of Internet television, which is an innovative media application on the Internet. Leung and Wei (1998) also reported that consumer innovativeness is positively related to their adoption decision of various media technologies. Innovative individuals have been also found to be dynamic, communicative, curious, venturesome, and stimulation–seeking. Moreover, Internet users have been often considered to be innovators [9]. Other diffusion studies [10] also confirmed that innovativeness is related to consumer adoption behavior. Based on these studies about this important variable, we proposed the following research question:

Research Question 7: Internet users’ innovativeness affects their intention to adopt Internet television.

Internet Usage Behavior

Rogers (1995) argued “the adoption of one new idea may trigger the adoption of several others in a cluster which consists of one or more distinguishable elements of technology that are perceived as being interrelated.” Leung and Wei (1998) observed that adoption of new technologies can be best predicted by consumer past adoption of functionally similar technologies. The concept has been called technology cluster, which was used to study consumer adoption of new technologies such as ICQ (Leung, 2001), interactive television and video–on–demand (Leung and Wei, 1998), and electronic commerce activities (Eastin, 2002). Therefore, given the similarities between Internet television and other Internet uses, we thus hypothesized that prior adoption behavior, knowledge about innovations, the adoption of technology cluster product influence consumer adoption behavior of Internet television:

Research Question 8: Internet experience affects Internet users’ intention to adopt Internet television.
Research Question 9: Knowledge about the Internet affects Internet users’ intention to adopt Internet television.

We proposed the following theoretical model to organize the hypothesized relationships between variables as discussed above (see Figure 1).

 

Factors influencing the adoption of Internet television

Figure 1: Factors influencing the adoption of Internet television.

 

++++++++++

Methodology

Sampling

The sample was recruited, through a convenience sampling method, from two metropolitan areas (i.e., Taipei and Kaohsiung) in Taiwan where there were the highest Internet penetration rates. Several research assistants were trained and dispatched to locations (such as train stations, bus terminals, and department stores) that often attracted large crowds to facilitate this research.

Taiwan is one of the most wired nations in Asia. As of September, 2004, the Internet penetration was about 40 percent of the island population of 23 million (Find, 2004a). The Internet users reached 9.05 million as of 2004 (Find, 2004a). The number of host computer ranked number 10 worldwide and number 2 among Asian countries Find, 2004b).

Sample characteristics

Total valid respondents for the study were 1,492. Over 47.7 percent (n=711) of our sample were male, while 52.3 percent (n=781) were female. Those whose ages were between 15 and 19 years old account for 15.2 percent (n=227), while those who fall between 20 and 24 years old account for 77.2 percent (n=1,163) of the total sample. Overall, younger respondents (aged between 15 and 29 years old) account for 98.9 percent of the sample. The demographic composition was reflective of current Internet population profile in Taiwan. According to Yam’s survey (2000), over 75 percent of the Taiwanese Internet users fell between 15–29 years old.

Most of the respondents in the sample used the Internet for less than three years (69.8 percent, n=1,036). Moreover, over 30.2 percent (n=448) in this survey used the Internet for more than three years. As to their knowledge about the Internet, the average rating was 4.52 (SD=1.858) on a scale of 10, with 1 representing novice and 10 representing expert.

 

Table 1: Demographic profiles of sample
   
Cases
Percentage
Gender
Male
711
47.7
Female
781
52.3
 
Age
15–19 years old
227
15.2
20–24 years old
1,163
77.2
25–29 years old
97
6.5
30 years old and above
10
0.9
 
Educational levels
Post–graduate level
25
1.7
University level
1,323
89.0
Junior college level
129
8.7
Senior high and vocational school level
10
0.7
 
Income
Below NTD$5,000
585
39.3
NTD$5,001–10,000
630
42.3
NTD$10,001–15,000
159
10.7
NTD$15,001–20,000
61
4.1
NTD$20,001–25,000
42
2.8
NTD$25,001–30,000
3
0.2
NTD$30,001–35,000
2
0.1
Above NTD$35,001
7
0.5
 
Internet experience (years)
Below 6 months
105
7.1
6 months to 1 year
241
16.2
1–3 years
690
46.5
3–5 years
316
21.3
Above 5 years
132
8.9
 
Knowledge about the Internet
1 (Novice)
88
5.9
2
127
8.5
3
239
16.0
4
264
17.7
5
250
23.5
6
210
14.1
7
128
8.6
8
64
4.3
9
14
0.9
10 (Expert)
8
0.5
Mean=4.52, SD=1.858, Median=5

 

Instrumentation

The data were taken from a large scale study about Internet users’ attitudes toward a variety of emerging Internet applications. We reported the data set about consumers’ adoption behavior of Internet television in this paper. The survey instrument for this research consisted of multi–item scales to measure: (1) Internet users’ motivations to use the Internet (17 five–point Likert statements) (Joines, et al., 2003; Kaye, 1998; Parker and Plank, 2000; Stafford and Stafford, 2004); (2) beliefs about Internet television (11 five–point Likert statements); (3) adoption intention of Internet television (2 five–point Likert statements); (4) control variables including demographics, Internet use behavior, and innovativeness (8 five–point Likert statements) characteristics of the Internet users. The innovative scale was measured by 8 five–point Likert items and had a Cronbach’s alpha value of 0.720. Individual items were aggregated for later analyses.

Cronbach’s coefficient alpha was employed to estimate the internal consistency of the multi–item scales used in the present study (Alsawalmeh and Feldt, 1999). The reliability coefficient helped assess the internal consistency for each of the scales identified by the factor analysis procedure. Nunnally [11] suggested a reliable instrument for a preliminary research needs to attain a Cronbach’s alpha value between 0.50 and 0.60. In this study, scales employed had attained the minimum reliability requirement for an exploratory study (Nunnally, 1967).

 

++++++++++

Discussion

Internet use motivations

We conducted an exploratory factor analysis to uncover the dimensions of consumer motivations to use the Internet. Five orthogonal factors with eigenvalues greater than 1.0 were extracted from the analysis. Despite a potential problem of overestimating or underestimating the number of factors retained for later statistical analyses, the eigenvalue criterion has been one of the commonly used standards in the factor analysis procedure (Tucker, et al., 1969; also see Park, et al., 2002). We also used a scree test to supplement the process of identification and selection of factors (Park, et al., 2002). Both procedures generated the same number of factors to be retained for later statistical analyses. In the end, we extracted five motivational factors, including “Entertainment factor”, “Habit factor”, “Social factor”, “Information factor”, and “Escapism factor”. These five factors accounted for 60.426 percent of the variance within the data. These results were reported with their Cronbach’s alpha values for each factor.

 

Table 2: Results of factor analysis of Internet use motivations
Factor 1: Entertainment factor
(Eigenvalue=2.666, Variance explained=14.813%, alpha=0.8009)
Because it relaxes me.
0.805
Because it’s enjoyable.
0.750
Because it allows me to unwind.
0.749
Because it entertains me.
0.727
 
Factor 2: Habit factor
(Eigenvalue=1.872, Variance explained=10.402%, alpha=0.5800)
I feel restless if I do not use the Internet.
0.784
I make a habit of using the Internet.
0.775
Because it gives me something to occupy my time.
0.568
 
Factor 3: Social factor
(Eigenvalue=1.820, Variance explained=10.111%, alpha=0.5897)
So I can chat with my friends.
0.730
So I can feel less lonely.
0.711
So I can make new friends.
0.560
 
Factor 4: Information factor
(Eigenvalue=1.552, Variance explained=8.620%, alpha=0.6752)
Because it helps me learn about myself.
0.809
Because it helps me learn about others.
0.789
 
Factor 5: Escapism factor
(Eigenvalue=2.966, Variance explained=16.480%, alpha=0.8195)
So I can forget about what happens at work.
0.806
So I can forget about what happens at home.
0.805
So I can forget about what happens at school.
0.774
So I can get away from the rest of the family or others.
0.707
So I can get away from what I’m doing.
0.577

 

The first research question explored beliefs and belief factors underlying Internet users’ perceptions of Internet television. Following the same exploratory factor analysis procedure, the belief factors about Internet television were presented in Table 3. Four factors were extracted from the principal component analysis using varimax rotation. These four factors accounted for 69.961 percent of variance and were as follows: “Program quality factor”, “Pro–regulation factor”, “Negative programming factor”, and “Media impact factor”. These results were reported with their Cronbach’s alpha values for each factor.

 

Table 3: Results of factor analysis — Beliefs about Internet television
Factor 1: Program quality factor
(Eigenvalue=2.331, Variance explained=21.194%, alpha=0.750)
Network transmission speed affects the quality of Internet television.
0.843
Network bandwidth affects the quality of Internet television.
0.801
Programming quality affects the success of Internet television.
0.749
Internet television should contain programming not seen in traditional television.
0.608
 
Factor 2: Pro–regulation factor
(Eigenvalue=2.059, Variance explained=18.718%, alpha=0.777)
Internet television should apply for broadcasting licenses.
0.802
Government should control Internet television programming.
0.801
Internet television should be regulated by existing broadcasting laws and regulations.
0.775
 
Factor 3: Negative programming factor
(Eigenvalue=1.660, Variance explained=15.092%, alpha=0.790)
Internet television will have a lot of pornographic contents.
0.903
Internet television will have a lot of violent contents.
0.882
 
Factor 4: Media impact factor
(Eigenvalue=1.535, Variance explained=13.957%, alpha=0.674)
Internet television will affect my media use behavior.
0.855
Internet television will affect my life.
0.848

 

Two 5–point Likert statements were used to measure consumer intention to adopt Internet television. These statements include “I think Internet television will become a commercial success”, and “I will adopt Internet television”. Table 4 reported factor loadings, eigenvalues, variances accounted for, and Cronbach’s alpha coefficient for each extracted factor.

 

Table 4: Results of factor analysis — Adoption intention of Internet Television
Factor: Adoption intention
(Eigenvalue=1.283, Variance explained=64.148%, alpha=0.836)
I think Internet television will become a commercial success.
0.801
I will adopt Internet television.
0.801

 

Predictors of Internet users’ intention to adopt Internet television

The second research question examined what belief factors can be used to predict Internet users’ intention to adopt Internet television. We conducted regression analyses to determine the relationships between these belief factors and respondents’ adoption intention. Mansfield and Helms (1982) argued that multicollinearity test should be performed before any multiple regression analysis. Therefore, we conducted the variance inflation factor (VIF) procedure to examine this problem in the regression models. None of the VIF values surpassed the threshold of 5, as proposed by Bernstein (2001) which he suggested as a general rule of thumb is that severe multicollinearity exists if a VIF is larger than 5. All VIFs ranged from 1.108 to 1.405. As a result, the initial analysis indicated that multicollinearity was not a problem for the regression model.

Three out of four belief factors about Internet television contributed significantly to the prediction of users’ intention to adopt this technology (see Table 5). These predictor factors altogether explained 26.2 percent of the variance in users’ adoption intention of Internet television. These factors were: Programming quality factor (beta.gif=0.307, t=11.635, p<0.001), Pro–regulation factor (beta.gif=0.086, t=3.293, p<0.001), Media impact factor (beta.gif=0.264, t=11.249, p<0.001), and Negative programming factor (beta.gif=0.040, t=1.692, p>0.05).

 

Table 5: Regression of beliefs about Internet television
Multiple R: 0.512
R Square: 0.262
Adjusted R square: 0.260
Standard Error: 0.569
F Value: 132.431***
Durbin–Watson=1.249
 
df
Sum of squares
Mean square
Regression
4
171.668
42.917
Residual
148
9
482.540
0.324
Total
149
3
654.208
 
 
Factors
Unstandardized coefficients
Standardized coefficients
t
Programming quality factor
0.301
0.307
11.635***
Pro–regulation factor
0.072
0.086
3.293***
Negative programming factor
0.037
0.040
1.692
media impact factor
0.242
0.264
11.249***
 
Notations: * p<0.05 ** p<0.01 *** p<0.001

 

Three significant belief factors have shown positive signs in beta.gif coefficients, meaning that higher scores predicted higher adoption intention of Internet television. In other words, those who believed in the programming quality of Internet television were more likely to adopt Internet television. Similarly, respondents who were in favor of government regulation were also more likely to adopt Internet television. Furthermore, Internet users who expected more media impacts caused by Internet television also had a higher adoption intention. As expected, Internet users who did not like potentially harmful media contents were less likely to adopt this technology. Overall, the findings supported the linkage between belief factors and users’ adoption of a new technology.

Previous studies of users’ adoption of new technologies have relied on individual perception of a technology. For example, innovation diffusion theory (Rogers, 1995) emphasized the social factors and their impacts on people’s perception and evaluation of innovation characteristics. Intended as a universal model to understand innovation choice and adoption, Rogers’ (1995) model failed to examine innovation–specific characteristics that can influence consumer decision to adopt a new technology. Similarly, Lin (2004) reported that audience’s perceived fluidity of the Internet significantly predict their viewing interest. In this study, we have identified innovation–specific factors related to the adoption of Internet television. Factors such as programming quality, pro–regulation, and media impacts have been found to significantly predict consumer intention to adopt Internet television. We thus argue that these second order and innovation–specific variables would supplement first–order adoption variables (e.g., compatibility, trialability, complexity, relative advantage, and observability).

To determine if belief factors about Internet television could predict respondents’ adoption intention after taking into consideration other confounding variables, we conducted several hierarchical regression analyses (RQ3). The results of hierarchical regression analyses were used to estimate the incremental and total variance associated with the variable blocks and were reported in Table 6. We also reported the final betas for individual variables from final regression equation in Table 6.

 

Table 6: Hierarchical regression predicting Internet users’ adoption of Internet television
 
R square
Standardized beta.gifa
t
Model 1: F=26.613, df=5/1457, p<0.001
Demographics
Gendera (male)  
-0.068
-2.700**
Education levelb  
0.049
1.894
Agec  
0.005
0.189
Incomed  
-0.116
-4.555***
Innovativeness  
0.244
9.617***
R after step 1
0.289
 
R square after step 1
0.084
 
 
Model 2: F=19.629, df=7/1455, p<0.001
Gendera (male)  
-0.065
-2.521**
Education levelb  
0.043
1.661
Agec  
-0.001
-0.038
Incomed  
-0.113
-4.406***
Innovativeness  
0.238
0.926***
Internet use behavior    
Internet usage historye  
0.047
1.749
Knowledge about Internet  
0.013
0.458
Incremental R square for Group 2
0.003
 
R after step 2
0.294
 
R square after step 2
0.086
 
 
Model 3: F=24.396, df=12/1450, p<0.001
Gendera (male)  
-0.083
-3.323
Education levelb  
0.042
1.674
Agec  
0.011
0.455
Incomed  
-0.096
-3.886***
Innovativeness  
0.144
5.584***
Internet usage historye  
0.035
1.361
Knowledge about Internet  
-0.014
-0.508
Internet use motivation    
Escapism factor  
-0.032
-1.211
Entertainment factor  
0.232
8.376***
Habit factor  
-0.020
-0.701
Social factor  
0.139
4.993***
Information factor  
0.046
1.703
Incremental R square for Group 3
0.082
 
R after step 3
0.410
 
R square after step 3
0.168
 
 
Model 4: F=39.020, df=16/1446, p<0.001
Gendera (male)  
-0.055
-2.346**
Education levelb  
0.022
0.933
Agec  
0.007
0.306
Incomed  
-0.020
-0.830
Innovativeness  
0.047
1.922
Internet usage historye  
0.027
1.129
Knowledge about Internet  
-0.044
-1.787
Escapism factor  
-0.001
-0.025
Entertainment factor  
0.139
5.326***
Habit factor  
-0.030
-1.147
Social factor  
0.078
3.025**
Information factor  
0.044
1.776
Beliefs about Internet television
Programming quality factor  
0.258
8.949***
Pro–regulation factor  
0.067
2.534**
Negative programming factor  
0.023
0.980
Media impact factor  
0.221
0.917**
Incremental R square for Group 4
0.134
 
R after step 4
0.549
 
R square after step 4
0.302
 
a. Coded as 0=female, 1=male.
b. Coded as 0=less educated (from senior high school to junior college level), 1=more educated (from university to post graduate level)
c. Coded as 0=young (15–19 years old), 1=old (above 20 years old)
d. Coded as 0=less affluent (below NTD10,000), 1=affluent (above NTD$10,000)
e. Coded as 0=less experienced (below 1 year), 1=more experienced (more than 1 year)
f. * p<0.05 ** p <0.01 *** p <0.001

 

A total of sixteen variables were selected and grouped into four blocks separately in the hierarchical regression analyses. Demographic variables (e.g., gender, education, income level, age, and innovativeness), Internet use behavior (e.g., Internet usage history and knowledge about Internet), and Internet use motivations were entered in the first three blocks. Belief factors were entered in the last block. This approach offered the most conservative test possible and ensured that any effects attributed to users’ adoption intention of Internet television would not be due to their relationship with other factors included in the hierarchical regression model.

Demographic variables as a group only accounted for 8.4 percent of the variance in users’ adoption intention of Internet television. As indicated in Table 6, variables such as educational level (beta.gif=0.049, t=1.894, p>0.05) and age (beta.gif=0.005, t=0.189, p>0.05) were not significant predictors in the final model. However, gender (beta.gif=-0.068, t=-2.700, p<0.001), income level (beta.gif=-0.116, t=-4.555, p<0.001), and innovativeness (beta.gif=0.244, t=9.617, p<0.001) significantly predicted consumer adoption intention as indicated in the first model. Findings suggested that not all demographic variables predicted Internet users’ adoption intention (RQ 6). Two of these three significant predictor factors had shown negative signs in beta.gif coefficients, meaning that female respondents were less likely to adopt Internet television. The results concurred with previous interactive television study conducted in Hong Kong (Leung and Wei, 1998) that showed that female respondents were less likely to adopt similar television innovations. The gender–related adoption behavior is common in Hong Kong and Taiwan because males are taught and expected to be more tech–savvy, culturally speaking. Further studies on the effects of social and cultural factors may shed light on the complexity of consumer decision–making processes.

We also found that those who had a lower income level were more likely to adopt Internet television than those with a higher income level. While this finding may be unexpected at the beginning, it is possible that lower income respondents tend to give socially approved answers. Because the adoption of Internet television carries certain social, educational, and financial status, it is likely that individuals in the lower income category tend to signal positive intentions. Because the present study only measured adoption intention, instead of actual adoption, we expected some degree of overestimation. Another important demographic variable, innovativeness, was found to be significant statistically. Furthermore, the positive sign in beta.gif coefficient in the innovativeness variable also suggested that the more innovative respondents were, the more likely they would adopt Internet television (RQ7). Similar to previous diffusion studies, innovativeness plays a critical role in predicting consumer adoption of new media technologies (Leung and Wei, 1998).

Variables related to Internet use behavior were considered next in the hierarchical regression model. The incremental R square is 0.003. Internet usage history (beta.gif=0.047, t=1.749, p>0.05) and knowledge about Internet (beta.gif=0.013, t=0.458, p>0.05) were significant (RQ8, RQ9).

Incremental R square for Internet use motivation factors was 0.082. Among five Internet use motivation factors, two motivation factors, Entertainment Factor (beta.gif=0.232, t=8.376, p<0.001) and Social Factor (beta.gif=0.139, t=4.993, p<0.001) were significant predictors of users’ adoption intentions. However, three other motivation factors were not significant: Habit factor (beta.gif=-0.020, t=-0.701, p>0.05), Information factor (beta.gif=0.046, t=1.703, p>0.05), and Escapism factor (beta.gif=-0.032, t=-1.211, p>0.05). The positive signs in beta.gif coefficient in Entertainment and Social factors also suggested that if more respondents used the Internet for entertainment and social reasons, they would be more likely ready to adopt Internet television (RQ4, RQ5).

Incremental R square for belief factors was 0.134. Belief factors explained the greatest amount of the variance in all four blocks of variables included in the hierarchical regression model. Three out of four belief factors were significant in the final model. While Programming quality, Pro–regulation, and Media impact factors were significant predictors, The Negative programming factor was not significant. Compared with other groups of variables, we observed that consumer beliefs consistently predicted intentions to adopt Internet television. Given the close relationship between beliefs about a new technology and intention to adopt, the importance of beliefs is heightened in the adoption of new media technologies. This finding concurred with previous adoption behavior research that indicated users’ attitudes toward technology affect their adoption behavior [12].

The full model explained 30.2 percent of total variance in consumer adoption intention of Internet television. After controlling all possible confounding variables, the predictive impacts of belief factors on users’ adoption intention remained stable.

 

++++++++++

Conclusion

The extension of using Internet and telecommunications technologies to broadcasting industry leads to the development of Internet television, which Noam and Gerberg (2004) claimed “the quintessential digital convergence medium” that will influence television industry. Although Noll (2004) argued that the definition of Internet television is still evolving and unclear, it can be broadly defined as the delivery of television programming through either broadband or narrowband Internet channels (Noam and Gerbarg, 2004). The emergence of this new medium was also viewed to influence telecommunications infrastructures, network business models and strategies, government regulations, and culture (Noam and Gerbarg, 2004).

Our study extends the theoretical frameworks and measurements previously developed for the study of traditional and Internet media. The study suggests a model that can be used to predict adoption intentions for Internet television. The importance of consumer perception to predict their adoption intention is also confirmed in Lin’s (2004) study that she found perceived fluidity to be a significant predictor of viewing interest in webcasting programs. While belief factors are useful predictors, the study also found that the effects of belief factors were affected by use motivations and demographic variables.

Limitations

Several limitations of this study should be taken into consideration. First, the unique characteristics of a convenience sample should be considered. Although the sample in this study has been comparable to the characteristics of Internet users as presented in other national surveys (Yam, 2000), a convenience sample of Internet users definitely limits the generalizability of these findings. Furthermore, our sample showed a tilt toward college students, which has made the findings of the present study more appropriate to future research focusing on student populations.

Secondly, the study examined Internet users’ future adoption of Internet television. College students are likely to provide socially approved answers, which may tilt toward pro–technology responses. As previous studies of new technology adoption have cautioned, pro–technology responses are likely to overestimate the adoption likelihood of new technologies.

In spite of these above limitations, this research has contributed to our understanding of adoption behavior for Internet television. These findings may have implications for models that attempt to explain factors influencing adoption behaviors of future emerging applications on the Internet. End of article

 

About the authors

Kenneth C.C. Yang (Ph.D., The Ohio State University) is an Associate Professor, Department of Communication, University of Texas at El Paso. His research interests focus on adoption behavior of new information communication technologies (ICTs), new media advertising, and telecommunications policy.

Yowei Kang (MFA) is a doctoral student at the English Department, University of Texas at El Paso. His research interests are visual rhetoric, new technology and rhetoric.

 

Notes

1. Gerberg and Noam, 2004, p. xxi.

2. Owen, 1999, cited in Waterman, 2004, p. 62.

3. See Waterman, 2004, pp. 62–63 for a review.

4. Except in Leung and Wei, 1998; Lin, 2004.

5. See Stafford, et al., 2004, p. 264 for review.

6. Vankatesh and Davis, 2000, cited in Stafford, et al., 2004, p. 264.

7. Rogers, 1995, cited in Stafford, et al., 2004, p. 264.

8. Stafford and Stafford, 2001, cited in Stafford, et al., 2004, p. 264.

9. Goldsmith, 1983; Raju, 1980, cited in Wood and Swait, 2002, p. 2.

10. LaRose and Atkin, 1988; Leung, 1988; Rogers, 1995, cited in Leung and Wei, 1998, p. 129.

11. Nunnally, 1967, p. 226.

12. See Leung and Wei, 1998, p. 141 for review.

 

References

Y.M. Alsawalmeh and L.S. Feldt, 1999. “Testing the equality of independent Alpha coefficient adjusted for test length,” Educational and Psychological Measurement, volume 59, number 3, pp. 373–383. http://dx.doi.org/10.1177/00131649921969929

J. Bernstein, 2001. “PA819 Quantitative methods for public policy lecture 8: multicollinearity,” at http://polisci.wisc.edu/~PMANNA/pa819/Lecture_and_Section/Lec8_Multicollinearity.pdf.

J. Carey, 2004. “Content models: Will IPTV be more of the same, or different?” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.

M.S. Eastin, 2002. “Diffusion of e–commerce: An analysis of the adoption of four e–commerce activities,” Telematics and Informatics, volume 19, pp. 251–267. http://dx.doi.org/10.1016/S0736-5853(01)00005-3

M.A. Einhorn, 2004. “Network business models and strategies: The role of public service broadcasting,,” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.

Find, 2004a. “xDSL users reach 3 million in Taiwan,” (29 November), at http://www.find.org.tw/0105/howmany/howmany_disp.asp?id=89.

Find, 2004b. “As of January 1, the number of host computers in Taiwan reaches 6 million,” (30 April), at http://www.find.org.tw/0105/howmany/howmany_disp.asp?id=72.

D. Gerbarg and E. Noam, 2004. “Internet television: Definition and prospects,” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.

J.L. Joines, C.W. Scherer, and D.A. Scheufele, 2003. “Exploring motivations for consumer Web use and their implications for e–commerce,” Journal of Consumer Marketing, volume 20, number 2, pp. 90–108. http://dx.doi.org/10.1108/07363760310464578

B.K. Kaye, 1998. “Uses and gratifications of the World Wide Web: From couch potato to web potato,” New Jersey Journal of Communication, volume 6, number 1, pp. 21–40. http://dx.doi.org/10.1080/15456879809367333

M.L. Katz, 2004. “Business models and program content,” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.

L. Leung, 2001. “College student motives for chatting on ICQ,” New Media & Society, volume 3, number 4, pp. 483–500. http://dx.doi.org/10.1177/14614440122226209

L. Leung and R. Wei, 1998. “Factors influencing the adoption of interactive TV in Hong Kong: Implications for advertising,” Asian Journal of Communication, volume 8, number 2, pp. 124–147. http://dx.doi.org/10.1080/01292989809364766

C.A. Lin, 2004. “Assessing audience viewing interest in webcasting,” In: P.S.N. Lee, L. Leung, and C.Y.K. So (editors). Impact and issues in new media: Toward intelligent societies. Cresskill, N.J.: Hampton Press, pp. 131–148.

E.R. Mansfield and B.P. Helms, 1982. “Detecting multicollinearity,” American Statistician, volume 36, number 3, pp. 158–160. http://dx.doi.org/10.2307/2683167

M. Morris and C. Ogan, 1996. “The Internet as mass medium,” Journal of Communication, volume 46, number 1, pp. 39–50. http://dx.doi.org/10.1111/j.1460-2466.1996.tb01460.x

J.C. Nunnally, 1967. Psychometric theory. New York: McGraw–Hill.

A.M. Noll, 2004. “Industry structure and competition absent distribution bottlenecks,” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.

H.S. Park, R. Dailey, and D. Lemus, 2002. “The use of exploratory factor analysis and principal components analysis in communication research,” Human Communication Research, volume 28, number 4, pp. 562–577. http://dx.doi.org/10.1111/j.1468-2958.2002.tb00824.x

B.J. Parker and R.E. Plank, 2000. “A uses and gratifications perspective on the Internet as a new information source,” American Business Review, volume 18, number 2, pp. 43–49.

E.M. Rogers, 1995. Diffusion of innovations. Fourth edition. New York: Free Press.

T.F. Stafford, M.R. Stafford, and L.L. Schkade, 2004. “Determining uses and gratifications for the Internet,” Decision Sciences, volume 35, number 2, pp. 259–288. http://dx.doi.org/10.1111/j.00117315.2004.02524.x

R.F. Tucker, R.E. Koopman, and R.L. Linn, 1969. “Evaluation of factor analytic research procedures by means of stimulated correlation matrices,” Psychometrika, volume 34, pp. 421–459. http://dx.doi.org/10.1007/BF02290601

S.L. Wood and J. Swait, 2002. “Psychological indicators of innovation adoption: Cross–classification based on need for cognition and need for change,” Journal of Consumer Psychology, volume 12, number 1, pp. 1–13. http://dx.doi.org/10.1207/S15327663JCP1201_01

D. Waterman, D. 2004. “Broadcasters’ Internet engagement: From being present to becoming successful,” In: E. Noam, J. Groebel, and D. Gerberg (editors) Internet television. Mahwah, N.J.: Lawrence Erlbaum.


Editorial history

Paper received 20 July 2005; accepted 15 February 2006.


Copyright ©2006, First Monday

Copyright ©2006, Kenneth C.C. Yang and Yowei Kang

Exploring factors influencing Internet users’ adoption of Internet television in Taiwan by Kenneth C.C. Yang and Yowei Kang
First Monday, Volume 11, Number 3 - 6 March 2006
http://firstmonday.org/ojs/index.php/fm/article/view/1319/1239





A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2016. ISSN 1396-0466.