Descriptive analysis was carried out by collecting and presenting data to facilitate understanding. This evaluation described the conditions, symptoms, and problems. The descriptive analysis used data on a group to explain or draw conclusions about the group. This technique was used to analyze data by describing the collected data without intending to make generalized conclusions. In addition, it was used to assess the Reading Interest and Digital Literacy of UIN Suska Riau students. The scoring criteria for students' Reading Interest and Digital Literacy used the Likert scale, including Strongly Agree (SS), Agree (S), Disagree Less (KS), and Disagree (TS).

To test the hypothesis and produce a viable model (fit), this study used Structural Equation Modeling (SEM) with a variancebased or component-based approach using Partial Least Square (PLS). In PLS structural models, the relationship between latent variables was called the inner model, while the measurement model (reflexive or formative) was referred to as the outer model. SEM was not clearly distinguished between the two methods, but there was an overall model, which was a combination of structural models of relationships between latent variables and measurement models with confirmatory factor analysis (CFA).

The Partial Least Squares (PLS) method served as a testing tool designed to evaluate models or prediction examples, distinctively free from reliance on diverse opinions. In addition, it proved particularly valuable in determining the suitability of models using weaker theories. PLS was also well-suited for analyzing data afflicted by issues, such as non-normal contributions, multicollinearity, and autocorrelation. The method could also accommodate small sample sizes and was adaptable for formative and reflective constructs (Abdillah, 2009). Given these characteristics, PLS emerged as the most appropriate tool for the present study. The predictive capabilities made it well-suited to assess and analyze the presence or absence of independent latent variables and dependent latent variables within the context of this study. The sample size, consisting of only 82 students, was in line with PLS's versatility in handling smaller datasets effectively.

The steps in PLS analysis used in this study included: (a) Development of a theory-based model or inner model to analyze the relationship between exogenous and endogenous variables elaborated in a conceptual framework. (b) Development of a flowchart (Path Diagram). The theoretical model that had been built in a conceptual framework was then drawn in a flowchart showing the relationship between exogenous and endogenous variables.

Interest in reading was individuals' desire or attention that was accompanied by effort and was made a need and interest in reading sources with a feeling of happiness. Students as one of the subjects in learning need an interest that spurs them to learn both from within and from outside. Investigators in this case obtained data on students' reading interests which had been shared with 82 students.

The student reading interest questionnaire consists of 19 questions packaged from 5 indicators of reading interest, namely: 1). Feelings of Pleasure, 2). Awareness of Reading, 3). Frequency of Reading, 4). Duration and 5). Quantity of Reading sources. Each question indicator was given an alternative 4 answers with a score of 1,2,3,4 so that the maximum score of all questions was 76 and a minimum score of 19.

To determine the value of students' reading interest intervals, investigators divided them into five categories, namely: 1) Not Good, 2) Less Good, 3) Good Enough, 4) Good, and 5) Very Good. The results of the analysis of student reading interest would be presented in Table 2 as follows:

Based on Table 2 above, it was found that 41 students were in the interval 55-66, or 50% with a good reading interest category, 21 people were in the 67-76 interval, or 25.6% with a very good reading interest category. 16 people were at intervals of 43-54 or 19.5% with moderately good categories and 4 people at intervals of 31-42 or 4.9%. Therefore, data was obtained that the interest of UIN Suska Riau students was in the "Good" category.

Digital literacy was the ability to access, select, and expand information related to digital skills that students applied in their daily lives using technological devices. Investigators in this case obtained data on digital literacy through questionnaires that had been distributed to 82 students. The questionnaire on student digital literacy consisted of 18 question items packaged from 1). Ability to use technology, 2) Creativity, 3). Communication Collaboration, 4). Ability to search for information, 5). Critical thinking and 6). E-Safety was each given 4 answers with a score of 1,2,3,4 so the maximum score for all questions was 72 and a minimum score of 18.

To determine the value of students' digital literacy intervals, investigators divided them into five categories, namely: 1) Not Good, 2) Less Good, 3) Good Enough, 4) Good, and 5) Very Good. The results of the student digital literacy analysis are be presented in Table 3 as follows:

Based on Table 3 above, it was found that 41 students were in the 66-72 interval as many as 41 people, or 50% in the Very Good category, 35 people were in the 54-65 interval, or 42.7% in the good category, and the 42-53 interval as many as 6 people or 7.3% with the good enough category. Therefore, data was obtained that the Digital Literacy of UIN Suska Riau students was in the "Very Good" category.

There were three criteria in using the technique: data analysis with SmartPLS to assess the outer model, namely Convergent Validity, Discriminant Validity, and Composite Reliability. The convergent validity of measurement models with reflexive indicators was assessed based on the correlation between item scores and component scores. An individual's reflexive measure was said to be high when it correlated more than 0.70 with the variable i.e. measured. (Imam, 2008) said that for early stage study, a measurement scale for loading values of 0.5 to 0.6 was considered adequate. In this study would be used a loading factor limit of 0.5. The results of processing using PLS could be seen in Table 4. The outer value of the model or the correlation between variables and variables had met convergent validity because all of them have a loading factor value above 0.50.

Table 4 showed the Loading Factor value for each construct of the variables, and all constructs were declared valid. The next step was the testing of Average Variance Extraction (AVE) to further strengthen the result of Convergent validity. When the AVE value was > 0.5, the construct used in the study was said to be valid. Fornell and Lacker (Cheng et al., 2013) stated that when AVE < 0.5 and Composite Reliability > 0.6, the construct was still acceptable.

Based on Table 5, the results of convergent validity could be found using the value of the average variance extracted. These results showed that all latent variables had AVE values smaller than 0.5, namely digital literacy of 0.428 and reading interest of 0.416. However, the composite reliability value was > 0.6, namely Digital Literacy of 0.926 and reading Interest of 0.919, indicating that all constructs were declared valid. These findings indicated that the indicators that inherent in the latent construct had good convergent validity when viewed from the value of average variance extracted.

Structural model testing was carried out to see the relationship between variables, significance values, and R Square from the study model. The structural model was evaluated using R Square for the dependent variable T-test and the significance of the structural path coefficient. The following was an overview of the structural model in this study.

Based on Figure 1, the covariance of indicator measurements was influenced by latent constructs and reflected variations from construct to indicator. The model hypothesized that changes in latent construction affected changes in indicators. In the model, there was 1 exogenous variable, namely learning satisfaction, and 2 endogenous variables, including learning motivation and learning achievement. Model assessment with PLS began by looking at the R Square for each dependent latent variable. In addition, changes in the value of R Square could be used to assess the effect of certain exogenous latent variables on endogenous latent variables that had a significant influence. The figure above showed that changes in digital literacy scores contributed 22.2% to changes in reading interest scores, and the remaining 77.8% was influenced by other factors not analyzed in this model.

The significance of the estimated parameters provided very useful information about the relationship between the reported variables. The basis used in hypothesis testing was the value contained in the output results for inner weight. The following table presented the output results for the structural model testing:

Hypothesis Testing:

H0: Digital Literacy has no significant effect on the Reading Interest of UIN Suska Riau Students

H1: Digital Literacy has significant rffect on the Reading Interest of UIN Suska Riau Students

Based on hypotheses above, a hypothesis test was carried out with the bootstrapping method using SmartPLS software, and obtained the following values:

The path coefficient value of the original sample estimate (beta) was positive, which was 0.471. This indicated that the direction of the relationship between Digital Literacy and Reading Interest was positive or unidirectional. When Digital Literacy increased, the Reading Interest was likely to increase, and vice versa. The effect of Digital Literacy on Reading Interest was significant in the 2-tailed test (t table = 1.97) with a T-statistic value of 8.105 greater than the t table, and a p-value smaller than alpha 5% (0.000 < 0.05). Therefore, H1 was accepted, indicating that Digital Literacy had a significant effect on the Reading Interest of UIN Suska Riau Students.

The results of this study were in line with a previous report that digital literacy had a strong influence on students' reading interest. This was shown by grade correlation coefficient (R) of 0.795, which showed that the degree of relationship between both variables was 79.5%. Based on these findings, the better the digital literacy of students, the better the interest in reading (Siroj et al., 2022).

Several studies had also reported similar findings including 1) the implementation of digital literacy program at MI Islamiyah Banjarpoh Ngoro considered 3 stages of literacy, namely habituation, development, and learning, 2) student reading interest before the implementation of digital literacy was still relatively moderate, but after the program, the variable became high. 3) the application of a digital-based 30-day literacy program to the reading interest of MI Islamiyah Banjarpoh Pulorejo Ngoro Jombang students was effective because the media used was well processed and more interesting to read (Khusaini & Baca, 2022).

Various reports had also shown the influence of digital libraries on students' reading interest. The data processing and statistical data analysis showed that students' reading interest was higher when using digital libraries, showing the presence of a significant influence. Good facilities in the digital library could attract learners and increase the ease of looking for reference books, doing assignments, or studying. Therefore, it could be concluded that the use of digital libraries affected reading interest ((Qoolili & Husni, 2022);(Awaludin et al., 2022)). This was consistent with a previous study that Gadgets had a major impact on reading interest due to the lack of supervision from parents or teachers during usage. The study concluded that positive impacts or negative impacts can be addressed depending on the use of the gadgets (Zuhria et al., 2022).

A previous study also explored the application of media Literacy Cloud Useful for fostering students' interest in reading both at school and at home. The positive impacts that occurred included reading endurance becoming longer than usual, increasing focus and concentration, enhancing interest in obtaining new information when studying, and fostering the desire to read (Fina & Susanto, 2023). Based on the finding, increasing interest in reading through the application of digital literacy was very necessary for school students even though learning activities were carried out remotely (Aisha, 2022).

Another study suggested a significant impact of the School Literacy Movement on students' reading interests. This influence was substantiated by the results of a correlation test, where the significance value was less than 0.05, and the F-calculate value exceeded the table t value. According to the findings at UPT SPF SD Inpres Lanraki 2 Makassar City by (Inzani et al., 2022), there was a statistically significant correlation between the variables of the school literacy movement (X) and students' reading interest (Y). In a separate study, the UPN Veteran Yogyakarta library conducted a study to assess the impact of information source exploration on master and doctorate students. The search for information comprised utilizing EBSCO journals and ebooks, engaging 50 participants. This initiative aimed to address the fact that a considerable portion of students, namely 59.5%, were unfamiliar with and had never accessed digital collections in the library. The training provided by the library proved highly effective, with 73.8% of participants acknowledging its value in introducing them to the library's collections and guiding them on accessing digital resources. This heightened awareness contributed to increased interest among students, with 69% expressing that the training significantly aided them in their final projects and scientific publications (Alifah et al., 2019). The findings also showed that 64.3% of participants expressed a heightened interest in accessing digital collections more frequently following their participation in the training. These results underscored the positive impact of library-led initiatives on students' engagement with digital resources and their subsequent academic endeavors.

The findings of a study showed that Electronic Books (e-books) served as a Please avoid use this viable alternative for students to cultivate a heightened interest in reading. This recommendation stemmed from the undeniable reality that students now possessed more mobile phones than physical books. The practical and portable nature of digital books allowed individuals the convenience to read at their discretion, fostering an environment where reading could take place anytime and anywhere. Consequently, the utilization of digital books was anticipated to elevate interest in reading, providing a more effective and efficient means of engaging with literary material (Ruddamayanti, 2019). This conclusion was supported by related report, indicating a significant positive correlation between digital literacy activities among adolescents aged 16-17 and reading comprehension. The data suggested that an increase in digital literacy activities corresponded to an improvement in reading comprehension, and vice versa. This interplay underscored the potential impact of enhancing digital literacy activities, achieved through purposeful engagement with digital technologies Setyawati et al., 2021). Meanwhile, an alternative perspective emerged from a study focusing on students' reading habits, responses, and questioning behaviors. Despite evaluating 50% of the class, the results consistently showed a lack of interest in reading, leading to a parallel absence of enthusiasm in participating in discussions (Ningsih, 2020).