4. What is the effect of social factors on
5. What is the effect of behavioral
intention on usage?
4.1 Is this influence moderated by nationality so that the effect is stronger for
Taiwanese citizens?
of different technologies including Web‐enhanced instructional elements43, digital libraries44
43 Landry, B.J. (2003). Student reactions to Web‐enhanced industrial elements. Retrieved from Dissertations &
Theses: Full Text database. (AAT3134932)
29
and so forth. These studies used the basic concepts of perceived usefulness and perceived ease of use to analyze the usage of the technology being studied.
The use of the TAM in these studies has demonstrated the validity and reliability of this model to conduct the research proposed by this study. As with other studies using the TAM, the objective of this study was to evaluate the perceived usefulness and ease of use relative to usage and behavioral intention. The reliability of the modified TAM instrument used in this study was evaluated using Cronbach’s alpha. The result of the reliability statistics are shown in Table 4.
Table 4. TAM reliability statistics
TAM Research Question Cronbach’s alpha Perceived Usefulness 0.722
Relative Advantage 0.803 Perceived Ease of Use 0.818 Social Factors 0.761
According to most researchers, it is recommended that instruments used in basic research have a reliability of 0.70 or better. The explanation for the reliability statstics indicated in Table 4 being above 0,70 is likely the result of the research questions analyzing what they were intended to analyze. There is strong internal consistency between the four different identifed constructs.
Results
44 Hong, W., Thong, J.Y., Wong, W., & Tam, K.Y. (2001 – 2002, Winter). Determinants of user acceptance of digital libraries: An empirical examination of individual differences and system characteristics. Journal of Management Information Systems, 18(3), 97 – 124.
30 Descriptive Analysis
In the end, 106 questionnaires were collected. The following graphs sum up the situation:
Table 5. Table concluding gender distribution
Gender
Frequency Percent Valid Percent
Cumulative Percent
Valid Female 75 70.8 70.8 70.8
Male 31 29.2 29.2 100.0
Total 106 100.0 100.0
Figure 2. Graph illustrating gender distribution
Of all the collected questionnaires, 29.2% (31) were completed by males and 70.8% (75) were completed by females
31
Table 6. Table concluding distribution according to nationality
Nationality
Frequency Percent Valid Percent
Cumulative Percent
Valid Foreigner 36 34.0 34.0 34.0
Taiwanese 70 66.0 66.0 100.0
Total 106 100.0 100.0
Table 7. Table concluding distribution according to current and future e‐book users.
Users
Frequency Percent Valid Percent
Cumulative Percent
Valid Current users 34 32.1 32.1 32.1
Future users 72 67.9 67.9 100.0
Total 106 100.0 100.0
Figure 3. Graph illustrating distribution according to nationality as well as users
Of all the collected questionnaires, 66% (70) were completed by Taiwanese citizens and 34% (36) were completed by foreigners who currently work and live in Taiwan.
32
Of all the collected questionnaires 67.9% (72) were completed by future e‐book users and 32.1% (34) were completed by current e‐book users. These percentages could be split up further with regards to nationality: 44 out of the 72 future users (62.1%) were Taiwanese and 28 out of the 72 (37.9%) were foreigners. Furthermore 26 out of the 34 current users were Taiwanese (76.5%) and 8 out of the 34 were foreigners (23.5%).
Table 8. Table distinguishing between the different age groups
AgeRange
Frequency Percent Valid Percent
Cumulative Percent
Valid 15-24 12 11.3 11.3 11.3
25-34 58 54.7 54.7 66.0
35+ 36 34.0 34.0 100.0
Total 106 100.0 100.0
Figure 4. Graph distinguishing between the different age groups as well as users.
Of all the collected questionnaires, 13.2% (14) were completed by participants between ages 15
33
Variable Perceived Relative Perceived Ease Social
Usefulness Advantage of use Factors
Have you read an e‐book
within the last 6 months? 0.742
It is easy for me to find my favorite books in e‐book
format. 0.450
I obtained an e‐book because it is so portable
and convenient. 0.516
I finish reading an e‐book faster than I finish reading
a paperback book. 0.449
34 People around me talk about
new e‐books or e‐book
reader devices. 0.594
Eigenvalue 2.673 1.300 1.607 1.13
Percentage of total
variance explained 22.274 10.830 13.392 9.450
The factor analysis is statistical proof that the questions designed in the questionnaire are testing the constructs that they are supposed to test given the research context. Each of these numbers does not necessarily have to be greater than 0,5, because the factor analysis gives a relative result as to which category the question belongs to. Thus the logical conclusion is that even if a number is below 0.5, this is of no importance, because the column in which the question has the highest ‘loading’ indicates the construct it is bound to test.
The factor analysis indicated that both Question 5 (which was designed to test relative advantage), as well as question 10 (which was designed to test social factors) should not be a part of the analysis because of a too low ‘loading’. Therefore they have been deleted from the inferential analysis and only the questions which could be statistically verified were used to execute the inferential analysis.
The table above indicates the Eigenvalues and these values are significant if they are greater than one. Therefore it is evident that the statistical is reliable as all of these values are greater than 1 and it could be concluded that the data explains a significant portion of the data variability. The percentage of the total variance explained serves the same function as the traditional R² and in this case the sum of the percentage of total variance is greater than 55%
(0.55946) and indicates not only the quality of the regression, but also the distribution of the independent variables.
35
Table 10. Regression Analysis
The best way in which to validate the interaction between the independent variables, dependent variables and moderators was to run a binary logistic regression in SPSS. The following results explain the relevance of the hypotheses which are linked to the research questions.
Table 10.1
Influence of perceived usefulness on behavioral intention (moderated by age).
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a PerceivedUsefulness -1.008 .695 2.104 1 .147 .365
Constant 1.907 2.054 .862 1 .353 6.732
a. Variable(s) entered on step 1: PerceivedUsefulness.
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a PerceivedUsefulness .954 .365 6.817 1 .009 2.595
Constant -3.502 1.349 6.744 1 .009 .030
a. Variable(s) entered on step 1: PerceivedUsefulness.
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a PerceivedUsefulness .499 .311 2.582 1 .108 1.647
Constant -1.117 1.016 1.208 1 .272 .327
a. Variable(s) entered on step 1: PerceivedUsefulness.
36
Research question 1: What is the influence of perceived usefulness on behavioral intention?
1.1 Is this influence moderated by age so that the effect is stronger for the younger generation?
Hypothesis: the influence of perceived usefulness on behavioral intention is moderated by age such that the effect is stronger for the younger generation.
The regression correlation coefficients of ‐1,088, 0,954 and 0,499 are indicators of how ‘big’
the influence of the ‘risk factor’ or moderator on the dependent variable is. These correlations reflect the three different age groups in question, 15‐24, 25‐34 and 35+
respectively. In terms of the youngest group the correlation between perceived usefulness and behavioral intention while moderated by age shows a very strong negative relationship.
This implies that within this group, perceived usefulness and behavioral intention are negatively related. This is most likely the result of the small number of participants within this group. With regards to the second and third group, there are a very strong positive relationship and a weak positive relationship. These regression coefficients indicate that for the middle and older generations, perceived usefulness increases as behavioral intention increases. The statistical result proves that the first hypothesis is indeed statistically
validated and that the effect is stronger for the younger generation (the younger generation being defined as those participants younger than 34).
Table 10.2
Influence of relative advantage on behavioral intention (moderated by age).
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a RelativeAdvantage .843 .843 1.000 1 .317 2.323
Constant -3.908 3.061 1.630 1 .202 .020
a. Variable(s) entered on step 1: RelativeAdvantage.
37
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a RelativeAdvantage .654 .314 4.345 1 .037 1.923
Constant -1.839 .864 4.530 1 .033 .159
a. Variable(s) entered on step 1: RelativeAdvantage.
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a RelativeAdvantage .740 .431 2.943 1 .086 2.095
Constant -1.367 1.098 1.550 1 .213 .255
a. Variable(s) entered on step 1: RelativeAdvantage.
Research question 2: What is the influence of relative advantage on behavioral intention?
2.1 Is this influence moderated by age so that the effect is stronger for the younger generation?
Hypothesis: The influence of relative advantage on behavioral intention is moderated by age such that the effect is stronger for the younger generation.
The research shows that there is an undeniable positive correlation between relative advantage and behavioral intention (see the full statistical analysis in the appendix. The regression
coefficients for the three different age groups are 0,843, 0,654 and 0,740 respectively. This indicates that for all the age groups, there is a strong positive relationship between relative advantage and behavioral intention. This comes as no surprise, but proves the hypothesis to be statistically validated, given that the relationship is just as strong for the younger generation than for the older generation. Even though the effect is stronger for males than females (other than assumed), it proves to be extremely useful due to the statistical validation.
38
Table 10.3
Influence of perceived ease of use on behavioral intention (moderated by gender).
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a PerceivedEaseOfUse .061 .262 .055 1 .815 1.063
Constant -.068 1.116 .004 1 .951 .934
a. Variable(s) entered on step 1: PerceivedEaseOfUse.
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a PerceivedEaseOfUse .292 .463 .398 1 .528 1.340
Constant -1.872 2.070 .818 1 .366 .154
a. Variable(s) entered on step 1: PerceivedEaseOfUse.
Research question 3: What is the effect of perceived ease of use on behavioral intention?
3.1 Is this effect moderated by gender so that the effect is stronger for females?
Hypothesis: the influence of perceived ease of use on behavioral intention is moderated by gender to such an extent that the effect is stronger for females.
It is undeniable that no‐one who perceives a device as difficult to use would be interested in buying it – with other words the relationship between perceived ease of use and behavioral intention is inevitable. However, the correlation coefficients of 0,061 and 0,292 for females and males respectively indicate an extremely weak positive correlation. Even though the
relationship is confirmed to be very weak according to the regression coefficients, it proves the hypothesis to be statistically valid. Ever so slightly, the effect of perceived ease of use on behavioral intention is stronger for males than for females (which means that the assumption of the moderator was not correct). The statistical validity proves to be extremely useful.
39
Table 10.4
Influence of social factors on usage (moderated by nationality).
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a SocialFactors .533 .541 .970 1 .325 1.704
Constant .366 .936 .152 1 .696 1.441
a. Variable(s) entered on step 1: SocialFactors.
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a SocialFactors -.494 .326 2.303 1 .129 .610
Constant 1.552 .730 4.516 1 .034 4.721
a. Variable(s) entered on step 1: SocialFactors.
Research question 4: What is the effect of social factors on usage?
4.1 Is this effect moderated by nationality so that the effect is stronger for Taiwanese citizens?
Hypothesis: the effect of social factors on usage is moderated by age to such an extent that the effect is stronger for Taiwanese citizens.
The regression coefficients of 0,533 and ‐0,494 for foreigners and Taiwanese citizens
respectively, indicate that the effect is about just as strong for both nationalities. This proves the hypothesis statistically valid and results in valuable information, because it is obvious that there is a moderately strong positive correlation between social factors and usage for
foreigners, whereas the opposite is true for Taiwanese citizens. Even though the effect is about just as strong for both nationalities, it is in opposite directions.
40
Table 10.5
Influence of behavioral intention on usage
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a BehavioralIntention 1.017 .437 5.416 1 .020 2.765
Constant .298 .275 1.176 1 .278 1.348
a. Variable(s) entered on step 1: BehavioralIntention.
Research question 4: What is the effect of behavioral intention on usage?
Hypothesis: there is a strong positive correlation between behavioral intention and usage.
It is obvious that behavioral intention and usage go hand in hand. There is a direct correlation between a person’s intent to use an e‐book reader device for recreational purposes and being a current or a future user. The regression coefficient of 1,017 indicate an extremely strong
positive relationship between behavioral intention and usage and proves the hypothesis to be statistically valid and true.
41
Correlations
PerceivedUsefulness
RelativeAd-vantage
PerceivedEase
OfUse SocialFactors PerceivedUsefulness Pearson
Correlation Sig. (2-tailed)
N 106
RelativeAdvantage Pearson Correlation
.410**
Sig. (2-tailed) .000
N 106 106
PerceivedEaseOfUse Pearson Correlation
.191* .038
Sig. (2-tailed) .050 .702
N 106 106 106
SocialFactors Pearson
Correlation
.118 .041 .066
Sig. (2-tailed) .228 .676 .503
N 106 106 106 106
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
The correlation analysis indicates how the different constructs correlate with one another. If the number is negative, it indicates a negative correlation, if it is between 0 and 0.3 it refers to a weak positive correlation, between 0.3 and 0.7 a moderate strong positive correlation and greater than 0.7 a strong positive correlation. It is thus obvious that all the constructs are positively related to one another and the correlations worth mentioning are that between
42
relative advantage and perceived usefulness (0.411 implies a moderately strong positive correlation). Another value worth mentioning is that between perceived usefulness and
perceived ease of use (0.191). Even though this implies a weak positive correlation, it states the fact that people ascribe value to those technological devices which they can easily operate.
Assumptions and limitations
The basis of this study is the application of the TAM research methodology developed by Davis in 1989. The TAM has been used, modified and extended in numerous studies examining and exploring the acceptance, usefulness, ease of use, and usage of different forms of technology.
The assumptions of this study include (a) the responses to the survey will be honest and, (b) the results of this study will provide applicable results to e‐book publishers and other key important players in the e‐book market place considering non‐academic e‐books for leisure purposes.
The limitations of this study are based on using a population limited to one city in Taiwan and (a) this study does not take the way of obtaining an e‐book into consideration (e.g. the black
market), (b) this study will not evaluate e‐book reader devices and (c) it will only be applicable to Taiwan considering the variability of the consistency of other populations and demographics.
In my opinion the biggest limitation of the study is crossing the language barrier: to get valuable feedback from Taiwanese citizens with good English speaking and understanding ability limited the research to mostly English teachers and therefore also mostly females.
Another limitation of this study is that it will not gain input from all stakeholders beyond the general public, such as e‐book publishers, e‐paper manufacturers, and so forth to more fully explore the topic.
Conclusion and recommendations
This study addressed the lack of empirical knowledge related to specific dimensions of the unexplored e‐book market segment for recreational e‐books. The lack of knowledge has been