The first UTAUT theory presented by Venkatesh, et al. (2003) proposed the study of the influence of the moderating variables: Gender, Age, Experience and Voluntariness of use on the independent variables of the model: Performance Expectancy, Effort Expectancy, Social Influence and Facilitating Conditions. These relationships between independent variables and moderating variables, in turn, influence the dependent variable Behavioral Intention. Behavioral Intention and Facilitating Conditions are directly related to the final variable: Use Behavior of the studied technology. These relationships can be viewed in Figure 1.

Figure 1 Source: Venkatesh et al. (2003)

The Variable Performance Expectancy refers to the degree to which individuals believe that technology will assist them in their tasks and is an influencer of Behavioral Intention, and the relationship is moderated by the variables of Gender and Age. The Variable Effort Expectancy refers to the difficulty associated with the use of the system and is an influencer of Behavioral Intention, this relationship being moderated by the variables of Gender, Age and Experience. The variable Social Influence refers to the level at which the user thinks that the technology is important for other people who are in their familiar circle and is influential on Behavioral Intention and use of technology, this relationship being moderated by the variables Gender, Age, Experience and Voluntariness of use. The Facilitating Conditions variable refers to the degree to which the individual believes that the technology development organization supports the problems related to its use. This variable influences Behavioral Intention and Use Behavior of technology, these relations being moderated by the variables of Gender, Age, and Experience.

The variable Behavioral Intention can be defined as being the will that the individual has to try and use a certain technology. In their study, Venkatesh, et al. (2003) state that the set of reactions of each individual in relation to these technologies, influences both their intention to use certain technology and their Use Behavior (Actual Use), both before and after the first contact with technology. This relationship can be observed in Figure 2, below:

Figure 2 Source: Adapted from Venkatesh et al. (2003)

According to Williams et al. (2011), UTAUT is cited worldwide and used as a reference in several ICT studies. These authors developed a systematic review of the UTAUT citations made by other authors and their constructs that are a reference for different researches, concluding that the main motivation for the use of UTAUT is the reasoning of arguments and discussions on several subjects related to the area of technology, in addition to having its constructs partially or totally used in the development of new applications of the model.

After applying the UTAUT model (VENKATESH et al., 2003) in an organizational context in which the adoption of a system by the employees was mandatory, UTAUT evolved in 2012 to UTAUT2. It should be emphasized that Venkatesh et al. (2003) used the longitudinal perspective to study a technology of mandatory adoption, in which employees of organizations periodically evaluated - from 3 temporal stages of adoption - their degree of training and use of the technology adopted.

In the 2012 proposal, Venkatesh, Thong and Xu studied mobile Internet technology. Its adoption was voluntary, that is, the users looked for and used the technology because they were motivated to do so. As in the mandatory adoption study, the authors chose to conduct a longitudinal study, which presented the interviewees with two steps. After answering the first step - a questionnaire regarding the intention to use mobile Internet, the users were contacted again four months later in the second step, for a new information collection.

As with Venkatesh, Thong and Xu (2012), this study deals with a voluntary adoption of technology (e-book), in which people are not trained for the use of technologies, but rather they seek it of their own volition. Although the 2012 study uses data collection at different times, in this research the collection was performed at a single moment due to the amount of data collected and treated, which made it impossible to follow the evolution of the relationship between the interviewees and the technology studied.

In 2012, Venkatesh, Thong and Xu proposed the inclusion of new variables in the model, indicating new relations and also the exclusion of the moderator variable Voluntariness of Use, since the technological context that makes possible the employment of UTAUT2 will always be an optional adoption of a technology by the user. The 2012 study was conducted in Hong Kong, and applied to the study of the acceptance of mobile Internet technology. The extension of UTAUT to UTAUT2 suggested the inclusion of independent variables that are direct influencers of the Behavioral Intention and Use Behavior of the technology. The variables included were: 1) Habit, with which the individual tends to perform automatic behaviors when dealing with technology. Kim and Malhotra (2005) defined it as automation. Thus, Habit would influence Behavioral Intention and the Use Behavior of technology and the relation would be moderated by the variables Gender, Age and Experience; 2) the Price Value variable was also added, whose perceived value from technology adoption compensates the expense due to the positive factors provided. This variable would influence Behavioral Intention and Use Behavior of technology and the relationship would be moderated by the variables Gender and Age; 3) the variable Hedonic Motivation, also added to the model, corresponds to the pleasure or delight that the technology could provide to the individual, influencing Behavioral Intention to use technology and being moderated by the variables Gender, Age and Experience. Thus, UTAUT2 (2012) extended the initial UTAUT (2003) and is shown in Figure 3, below.

Figure 3 Source: Venkatesh, Thong e Xu (2012)

The UTAUT (2003) and UTAUT2 (2012) models, since they are a proposal of unification, cover the collection of theories and studies on issues of acceptance, diffusion and use of technologies, which makes these models important tools for studies in this area.

The use of the UTAUT2 model inserted in the scope of the adoption of e-books, brings to the research the possibility of studying the relationships that motivate users to adopt the technology, based on a unified and robust model.

In order to help the development of the research results in a clear and structured way, the following hypotheses were generated, and each hypothesis below represents a relation between variables presented in the model:

H1: Performance Expectancy (PE) influences Behavioral Intention (BI) of e-book use, and this relationship can be moderated by the age and gender of the subjects.

We received a 1,013 questionnaires, of which 346 (34.16 % of the total) were answered by men and 667 (65.84 %) were answered by women.

The age of the interviewees ranged from 18 to 68 years old and, in relation to the level of schooling, 3 (0.3 %) had completed elementary education, 29 (2.86 %) had completed secondary education, 287 (28.33 %) had not completed higher education, 288 (28.43 %) had completed higher education, 327 (32.28 %) had postgraduate degree courses, 64 (6.32 %) had a master's degree, and 15 (1.48 %) had a doctorate.

Regarding reading experience, 92 (9.08 %) read at least one book per year, 390 (38.50 %) read from 2 to 4 read books per year, 226 (22.31 %) read from 5 to 7 books per year, 132 (13.03 %) read from 8 to 10 books per year, and 173 (17.08 %) read more than 10 books per year.

On reading e-books, 536 (52.91 %) respondents reported that only one of the books read was an e-book, 377 (37.22 %) reported that 2 to 4 books were e-books, 52 (5.13 %) reported that from 5 to 7 books were e-books, 23 (2.27 %) reported that 8 to 10 books were e-books, and 25 (2.47 %) reported that more than 10 of these books were e-books.

The data collected on the reading of e-books were used to measure the moderating variable Experience (EXP), according to the number of e-books read by the subjects in the period of one year. This variable (EXP) ranked the respondents as follows: reader of only one e-book - poorly experienced; reader of 2 to 4 e-books - familiar; reader of 5 to 7 e-books - experienced; reader of 8 to 10 e-books - very experienced; reader of more than 10 e-books - dominates the technology.

In analyzing the data collected in relation to the profile of the respondents, we noted that there were a significant number of questionnaires answered by women, consisting of 667 of the 1013 questionnaires (approximately 65.84 %). This fact was totally random, given that the database used had 118,456 e-mail addresses and the respondents were selected in order of response (1,013 questionnaires were obtained and the collection was concluded). Although gender plays a key role in accepting technologies, the study is not focused on differences by gender categories, but rather on the moderation relations between gender and other variables in the model.

Having presented the profile of the research participants, the results of the data analysis will be presented as follows.

In the questionnaire presented by Venkatesh, Thong and Xu (2012), the author measured the variable Use Behavior, through a composite formed by both the frequency at which the interviewees use the mobile Internet service and the variety of services and ICTs that have their use related to this technology (mobile Internet). As adapted from the 2012 study, we analyzed ICTs with which the users of digital books could also deal with. As a way to better understand the profile of these users and to outline characteristics of e-book use, we collected data on how many digital books the subjects of the research had read in a 12-month period.

Additionally, in the questionnaire of this research, we added an input question where the respondents, in order to continue responding to the questionnaire, should first inform if they have already had any contact with digital books, thus distinguishing users from nonusers.

To measure the number of digital books read in a 12-month period, the user should first state the number of books they had read during this period, subsequently showing how many of these books had been in digital format.

In order to measure the mean frequency of use of ICTs related to e-books, the respondents selected on a scale from 0 (never) to 10 (many times a day), the degree at which they use the following tools: tablets, SMS (text messaging), Internet browser, e-books, and electronic games, as shown in Table 1.

As can be seen, the use of the Internet, e-books and SMS represents a significant frequency among the subjects of the research. These data, besides better consolidating the profile of the respondents, also help to understand their behavior of use of both digital books and ICTs, which have a direct relationship with e-books.

Initially, we verified the means of the values obtained on the 0 to 100 scale related to assertions PE1, PE3 and PE4 presented in Table 2, which deal with Performance Expectancy, presenting the following means:

The effect of the direct relationship between Performance Expectancy (PE) and Behavioral Intention (BI) of e-books use (0.000782) was not significant at p <0.01. The relationship, when moderated by Age, had a significant effect (p <0.01) corresponding to 3.396414. When the relationship was moderated by Gender, the effect was 8.101312 (p <0.01).

For the variable Effort Expectancy, we formulated the assertions EE1, EE2, EE3 and EE4, whose response means are presented in Table 3, with a scale varying from 0 to 100 points.

Data analysis confirmed the UTAUT2 model, showing significant effect (-0.007249) between Effort Expectancy (EE) and Behavioral Intention (BI). The relationship, when moderated by Age, produced an effect of 4.711810. When the relationship was moderated by Gender the effect was 10.472413. The effect exerted by the moderating variable Experience is 82.706993. All tests presented significance (p <0.01).

For the variable Social Influence (SI), we formulated the assertions SI1, SI2 and SI3, whose response means are presented in Table 4, with a scale varying from 0 to 100 points.

Data analysis showed an effect (0.001471) between Social Influence (SI) and Behavioral Intention (BI), albeit not significant at p <0.01. The relation will have a significant effect when moderated by Age (effect 3.5558921; p <0.01); Gender (effect 7.3730093, p <0.01) and Experience (effect 62.776214; p <0.01).

We formulated the variables FC1, FC2, FC3 and FC4 for Facilitating Conditions (FC), whose response means are presented in Table 5, with a scale ranging from 0 to 100 points.

The analysis did not confirm significance (p <0.01) in the negative effect (-0.000868) found in the relation of influence of Facilitating Conditions (FC) on Behavioral Intention (BI). The relationship will have a significant effect, when moderated by Age (effect 4.7323898; p <0.01) and Experience (effect 78.463796; p <0.01).

The variable Facilitating Conditions (FC) has a significant effect (0.125277; p <0.01) on the relation of direct influence on Use Behavior (USE) of e-books.

For the Hedonic Motivation variable (HM), we formulated the assertions HM1, HM2 and HM3, whose response means are presented in Table 6, with a scale ranging from 0 to 100 points.

The analysis did not confirm significance (p <0.01) in the negative effect (-0.003262) between Hedonic Motivation (HM) and Behavioral Intention (BI). The relationship will have a significant effect when moderated by Age (effect 3.672709; p <0.01); Gender (effect 8.0018696; p <0.01) and by Experience (effect 63.258582; p <0.01).

For the variable Price Value (PV), we formulated the assertions PV1, PV2 and PV3, whose response means are presented in Table 7, with a scale ranging from 0 to 100 points.

The analysis did not confirm significance (p <0.01) in the negative effect (-0.000167) between Price Value (PV) and Behavioral Intention (BI). The relationship will have a significant effect when moderated by Age (effect 3.584430; p <0.01) and by Gender (effect 7.795411; p <0.01).

For the variable Habit (HT), we formulated the assertions HT1, HT2 and HT3, whose response means are presented in Table 8, with a scale ranging from 0 to 100 points.

The results were significant, as can be observed in the positive effect (0.007280018) between Habit (HT) and Behavioral Intention (BI). The relationship, when moderated by Age, had an effect of 3.5126229. When the relationship was moderated by Gender, the effect was 7.4181099. The effect exerted by the moderator Experience is 61.265890. In all tests, significance was set at p <0.01

The Habit (HT) variable also has a significant direct relation with the dependent variable Use Behavior (USE) of e-books. The effect of this ratio was 0.01640. The effect of the moderating variable Age with the relation between Habit (HT) and Use Behavior (USE) was 5.934185. The effect of Gender on the relation between Habit and Use Behavior was 13.015797. The effect of Experience was 104.325102. All tests presented significance p<0.01.

For the Behavioral Intention (BI) variable, we formulated assertions BI1, BI2 and BI3, whose response means are presented in Table 9, with a scale ranging from 0 to 100 points.

The variable Behavioral Intention (BI) is influenced by the seven other input variables presented in the model and, in turn, influences Use Behavior, and is also moderated by Experience. The effect of this relationship has a value of 0.012706 (p <0.01) and the effect of the Experience moderator with this relation is 102.749286 (p <0.01).

The effect of Habit on Behavioral Intention is 0.00728002 (p <0.01) and that of Social Influence is 0.001471, but in this case, it is not significant at p <0.01. Effort Expectancy has a significant effect of 0.007249 (p <0.01) on Behavioral Intention.

When the variables that have a direct influence relation to the use of e-books are analyzed, Facilitating Conditions has a significant effect of 0.1252769 (p <0.01), Habit also has a significant effect (0.016401; p < 0.01), as well as Behavioral Intention (0.012706; p <0.01). This result is corroborated by the study by Hsu et al. (2017), developed in Taiwan with a sample of 343 questionnaires, whose objective was to examine key factors that affect the adoption of e-books based on the UTAUT theory. The results of the authors' research showed that the adoption of e-books is determined by Performance expectancy (PE), Effort Expectancy (EE), Social Influence (SI) and Facilitating Conditions (FC), covered in this research, as well as other influencing variables, according to the authors, but not employed in this study, namely: perception of benefit, trust of benevolence and environmental concern.

Table 10 summarizes the effects of the independent direct variables on the variables Behavioral Intention and Use Behavior of e-books, as verified in this research.

From the results obtained with the research, we summarize the rejection or non-rejection of the hypotheses of the study, as shown in Table 11.

Table 11 (Cont.)

Table 11 (Cont.)

It should be emphasized that when moderated by Gender, Age, and Experience, the relations presented in hypotheses that test the influence of independent variables on the dependent variable Use Behavior (H1 to H7) cannot be rejected. This finds support in the seminal study by Venkatesh, Thong and Xu (2012). However, this is not the case in Maduku (2015), a study conducted in South Africa with 544 students about gender differences in the behavioral antecedents of the intention to use e-books. When Gender was included as moderator of these relations, Gender did not significantly moderate these relationships, except in the situation partially represented in H4, in which Facilitating Conditions (FC) influence Behavioral Intention (BI), and this relationship can be moderated by Gender.

An important focus given to data discussion and hypotheses testing was the analysis of the effect that the moderating variables Age, Gender and Experience exert on the direct influence relation of the independent variables to Behavioral Intention and Use Behavior of the studied technology. As shown in Table 11, if the test is only to verify the significance of the effects related to direct relationships between independent variables and Behavioral Intention and Use Behavior of e-books, five of the hypotheses are rejected and five cannot be rejected. However, as can be seen in the second column of Table 11, when the age, gender and experience of the subjects are included, then all ten hypotheses formulated in this study cannot be rejected at p <0.01.

Three tables (Tables 12, 13 and 14) present the effect of the moderating variables (Age, Gender and Experience) on the influence relations between the direct influencing variables of Behavioral Intention and Use Behavior of e-books.

The effect of the variable "Age", when moderating the relations between independent variables (direct influencers) proposed in the UTAUT2 model (VENKATESH; THONG; XU, 2012) and Behavioral Intention and Use Behavior of e-books, has always been significant p <0.01). The relationships and the respective effects of the moderating variable Age are shown in Table 12.

The effect of the variable "Gender", when moderating the relations between independent variables (direct influencers) proposed in the UTAUT2 model (VENKATESH; THONG; XU, 2012) and Behavioral Intention and Use Behavior of e-books, has always been significant p <0.01). The relations and respective effects of the moderating variable Gender are shown in Table 13 below:

Maduku (2015), as previously mentioned, concluded that although Gender appears to moderate the relationship between Facilitating Conditions (FC) and Behavioral Intention (BI), surprisingly, we found no significant gender moderating effects in Performance Expectancy (PE), Effort Expectancy (EE) and Social Influence (SI), which is different to what occurred in this study.

Nurkhin, Rohman and Mukhibad (2017) developed a study in Indonesia with 259 university students, seeking to analyze the influence of these variables of the UTAUT model as proposed in 2003 by Venkatesh et al. (2003) on the intention to use and in the behavior of use of electronic journals (e-journals). The results of the study showed that Performance Expectancy (PE) and Effort Expectancy (EE) for the use of e-journals significantly impacted Behavioral Intention (BI) (p <0.05), even when relations are moderated by Gender, while Social Influence (SI) did not show a significant influence on BI (p> 0.05), even when the relation was moderated by Gender. Other profile variables of the subjects were not studied by the researchers. The research of the mentioned authors demonstrates total convergence with the result found in this study, according to Hypothesis 3 (H3 - see Table 11).

The effect of the variable "Experience", when moderating relations between independent variables (direct influencers) proposed in the UTAUT2 model (VENKATESH; THONG; XU, 2012) and Behavioral Intention and Use Behavior of e-books, has always been significant p <0.01). The relations and respective effects of the moderating variable Gender are shown in Table 14 below:

Finally, from the presentation and discussion of the results found in the application of UTAUT2 to the context of e-book users, it is clear that at p <0.01, all influence relations of the model on Behavioral Intention and Use Behavior of e-books have a significant effect, provided that the profile variables of the subjects, such as age, gender, and experience, are included as moderators. When the subjects' profiles are disregarded, only 50 % of the hypotheses are supported by the tests (p <0.01). This may explain the importance placed by Venkatesh and colleagues in all their studies (VENKATESH et al., 2003; VENKATESH; THONG; XU, 2012) on the profile variables of technology users, as moderators of the relations studied. However, most recent studies also demonstrate the strength of Gender as a moderating variable in these studies, irrespective of geographic distances between the loci of the research, and regardless of the significance of age among others (MADUKU, 2015; NURKHIN; ROHMAN; MUKHIBAD, 2017).

Additionally, as pointed out in this research, Habit, Effort Expectancy, Behavioral Intention, and Facilitating Conditions are significant variables in the sense of stimulating individuals to use digital books, directly contributing to users' behavior; thus, an increase in the use of this book technology in the coming years is expected.

According to the study by Van der Velde and Ernst (2009), students and researchers were convinced of the success of digital book technology for its ease of research and access. Both in universities and libraries, the collection of many of these institutions already presents this tendency, where some of the books are in digital format, as are academic articles and other types of publications. According to Ashcroft (2011), having a collection of digital books has become an important factor for several libraries and universities, with some of these institutions having more registrations after making digital books available to their users. Remote and continuous access to archives are some of the advantages of new digital libraries.

Despite the growing use of e-books, a substitution of book technology by digital book technology is not expected. Rather, the complementary use of both technologies is expected, because access to large digital book databases makes it easier to access specific data and materials, and the user has the option of finding the necessary book in digital format and then deciding whether to purchase the printed version, download it in the digital format, or even read it online.

Figure 4 below provides the general picture of the effects obtained in the relations studied.

Figure 4

Significance p <0.01 for effects, where significant.