The Relationship Between ICT Perceived Competence and Adolescents’ Digital Reading Performance: A Multilevel Mediation Study

Abstract

The age of information and communication technology (ICT) has witnessed the prevalence of computer-based reading. Previous studies yielded mixed results concerning the relationship between ICT perceived competence, ICT use and digital reading performance, and the underlying mechanism was rarely explored. This study aimed to explore the relationship between ICT perceived competence and adolescents’ digital reading performance and the potential mediating roles of three types of ICT use. Multilevel mediation analyses were conducted for 199,646 15-year-old students from 29 OECD countries/regions in the Programme for International Student Assessment (PISA) 2018, in which digital reading performance was assessed by multistage adaptive computer-based testing. The results revealed that a) ICT perceived competence and digital reading performance were positively correlated; b) ICT use significantly mediated the relationship; c) students with higher-level ICT perceived competence tended to use ICT for leisure at home more frequently, which led to better digital reading performance; d) the suppression effects were revealed, indicating the need to consider the interrelationship between ICT perceived competence and ICT use. These findings provided new evidence for the self-determination theory and the ICT engagement model, suggesting that appropriate ICT use driven by high-level ICT perceived competence might help improve adolescents’ digital reading performance.

Keywords

digital reading information and communication technology perceived competence information and communication technology use multilevel mediation PISA reading

Introduction

In the age of information and communication technology (ICT), ICT users and digital readers are on the rise (Reimer et al., 2021; Sage et al., 2019). Furthermore, online learning has become much more prevalent especially in the post coronavirus disease 2019 (post-COVID-19) pandemic era (Silva et al., 2021). Reading texts on screen has been found to be different from on paper (Ackerman & Goldsmith, 2011). Since online platforms involve novel types of texts and sociocultural practices (OECD, 2011; Yu et al., 2022), it is important that students master digital reading skills such as navigating through ambiguity and assessing text credibility (OECD, 2011). Having witnessed the expansion of the medium of reading from paper to digital devices and the new requirement for students to read smoothly on digital platforms, the Programme for International Student Assessment (PISA), a triennial international survey conducted by the Organisation for Economic Co-operation and Development (OECD) to assess 15-year-olds’ academic achievement, redefined reading performance in its most recent cycle, PISA 2018, and conducted a large-scale computer-based reading assessment. Accordingly, in this study, digital reading performance refers to “understanding, using, evaluating, reflecting on and engaging with digital texts in order to achieve one’s goals, to develop one’s knowledge and potential and to participate in society” (see OECD, 2019b, p. 26).

In studying the significant determinants of students’ digital reading performance, existing studies have highlighted ICT perceived competence and frequency of ICT use. However, prior research has yielded inconsistent findings on the relationship between ICT perceived competence, ICT use, and digital reading performance, mostly positive (e.g., Hu et al., 2018; Lim & Jung, 2019; Skryabin et al., 2015) but several negative (e.g., Chen et al., Xiao et al., 2019). Additionally, PISA 2018 has renewed the measurement of the ICT use variables due to the emergence of new digital reading media such as social media and chatting apps (Hu & Yu, 2021), and it also encompassed more participating countries/regions in the digital reading assessment. Therefore, the data used by previous studies need to be updated, and the sample size could be enlarged (e.g., Meng et al., 2019; Xiao et al., 2019). Most importantly, the mechanism underlying the relationship between ICT perceived competence and digital reading performance remains largely uninvestigated (Goldhammer et al., 2016; Lee & Wu, 2012).

Insights into the roles of ICT perceived competence and ICT use in the development of students’ digital reading performance could further enhance the understanding of the underlying mechanism, inform theory, assist educators in guiding adolescents to regard and use ICT properly, and provide pedagogical implications for technology-assisted language learning (Costa et al., 2019).

Therefore, this study aims to reveal the relationship between ICT perceived competence and 15-year-old students’ digital reading performance and the potential mediating effects of three types of ICT use, i.e., recreational use of ICT at home, academic use of ICT at home, and ICT use at school. Specifically, 1-1-1 multilevel mediation analyses with three parallel mediators are conducted for each of the 29 OECD countries/regions, controlling for demographic information, including gender and socioeconomic status.

Literature Review

Information and Communication Technology Perceived Competence and Digital Reading Performance

The concept of ICT perceived competence is grounded in perceived competence. The term “perceived competence” is often used interchangeably with “self-concept”, “self-efficacy”, and “self-confidence” in the extant literature (Christoph et al., 2015; Marsh et al., 2019; Monique et al., 2013; Zylka et al., 2015). As proposed in the self-determination theory (SDT) of Deci and Ryan (2000), perceived competence refers to “the belief that one is capable of interacting effectively with the environment in order to obtain valued outcomes” (Marshik et al., 2017, p. 43) during the process of goal pursuit (Burkley et al., 2015), and it has shown to be an essential predictor of adolescent students’ academic performance in previous educational psychology studies (e.g., Chen et al., 2021; Geng et al., 2022; Hsu et al., 2022; Huang & Mayer, 2019; Usher et al., 2019). ICT perceived competence, a particular type of perceived competence, is conceptualized as “the individual’s perception of his or her own knowledge about ICT and about how to use it” (p. 342) in the SDT-based ICT engagement model devised by Goldhammer et al. (2016).

The potential relationship between ICT perceived competence and digital reading performance are demonstrated in several theories. Generally, as proposed in SDT, satisfaction of the need for competence could enhance intrinsic motivation, which in turn generates better learning and performance (Deci & Ryan, 2000). Specifically, in the ICT engagement model (Goldhammer et al., 2016), ICT perceived competence at a high level might leads to greater performance in ICT tasks, which may further support the learning outcomes of more general learning tasks involving ICT use.

In the context of PISA reading research focused on adolescents, ICT perceived competence has been found to be related to the development of 15-year-old students’ digital reading performance. Lim and Jung (2019) found that students’ self-confidence in completing high-level ICT tasks was positively related to their digital reading performance. Similarly, studies utilizing the PISA 2015 (Hu et al., 2018) and PISA 2018 database (Xiao & Hew, 2022) revealed that students who perceived themselves as competent in using ICT performed better in digital reading tasks. Moreover, when students perceive ICT as useful but uncontrollable, their performance in digital reading assessment deteriorates (Naumann & Saelzer, 2017). While the aforementioned studies indicated that ICT perceived competence might play a positive role in fostering reading activities on digital devices, a relatively small number of studies yielded negative results for Singapore, New Zealand, Australia, Finland, and France in PISA 2015 (Xiao et al., 2019), and for China and Germany in PISA 2015 (Meng et al., 2019). While Xiao et al. (2019) didn’t provide any possible reason for such a phenomenon, Meng et al. (2019) speculated that the use of self-report and the cultural differences between Eastern and Western countries might partially explain it. Students in Eastern countries, though being modest about their competence, were found to perform better in academic assessment including digital reading assessment. On the contrary, Western countries often support students’ confidence in their uniqueness, so they are more likely to deem themselves as highly competent so that their self-perception even becomes inflated, i.e., higher than their real competence (Liu & Meng, 2010). These findings suggested that more large-scale studies involving more countries/regions are in need to explore the replicability of such cultural differences.

Considering the important role of ICT perceived competence, the currently inconsistent findings on the effects of this factor, and the scarcity of studies that probe into the mechanism underlying the influential pattern, further studies with updated data are needed to shed light on the relationship between ICT perceived competence and adolescents’ digital reading performance.

Information and Communication Technology Use and Digital Reading Performance

Information and communication technology use concerns students’ physical involvement in ICT (Goldhammer et al., 2016). A host of studies have revealed that the frequency of ICT use is a determining factor for the development of adolescents’ digital reading performance. The use of ICT has been found to affect students’ reading performance in different use contexts (i.e., at school and outside of school) (Hammond, 2020) and with different use purposes (i.e., recreational and academic purposes) (Eger et al., 2020). Most studies found that using ICT at school was negatively correlated with students’ reading performance (e.g., Chen et al., 2021; Hu et al., 2018; Petko et al., 2017; Skryabin et al., 2015); however, mixed correlations were found with respect to the use of ICT outside of school and digital reading performance— recreational use could be either positively (e.g., Hu et al., 2018; Skryabin et al., 2015) or negatively correlated with performance (e.g., Petko et al., 2017; Xiao et al., 2019). Similarly, the relationship between academic use of ICT and digital reading performance were found to be either positive (e.g., Petko et al., 2017; Skryabin et al., 2015; Srijamdee & Pholphirul, 2020) or negative (e.g., Hu & Yu, 2021; Hu et al., 2018).

Despite the substantial efforts made by previous researchers to reveal the relationship between ICT use and digital reading performance, the currently inconsistent findings need further corroboration using more updated data from a large sample, especially in the post-COVID-19 era when learning becomes increasingly ICT-dependent (Sage et al., 2021).

Information and Communication Technology perceived competence, Information and communication Technology Use and Digital Reading Performance

The aforementioned studies were focused on the relationship between digital reading performance and ICT perceived competence or ICT use separately. However, few have investigated the interrelationship between these three factors concurrently.

In fact, ICT perceived competence and ICT use have been found to be interrelated both theoretically and empirically. As proposed in the ICT engagement model (Goldhammer et al., 2016), ICT perceived competence was deemed as one of the maintaining and enhancing factors of self-determined use of ICT. This proposition is underpinned by the SDT (Deci & Ryan, 2000), which assumed that people would continue to show certain behaviours if their needs to feel competent were satisfied. Empirically, on the one hand, ICT perceived competence has been shown to be an important antecedent of students’ ICT use (Chen & Hu, 2020); on the other hand, computer anxiety was found to be one possible reason for the long-term avoidance of computers (Monique et al., 2013). Hence, the correlation between ICT perceived competence and ICT use lays the foundation for studying the competence-use-performance link.

Furthermore, a notable study by Lee and Wu (2012) provided the impetus for the present study. Their study revealed the relationship between ICT-related psychological and behavioral factors and reading performance, as well as the mediating role of ICT-related behavioral factors. Specifically, by analysing data from 42 countries/regions in the PISA 2009, they found that the ICT inner state, which included ICT confidence, was found to be related to reading performance through the mediating role of online reading activities. In other words, the effects of ICT confidence on the development of reading performance were transmitted through the frequency of online reading activities. Although their study provided inspiration for the current one, such as the exploration of the mediating path, some improvements could be made to provide a more comprehensive view of the potential factors that are closely linked to digital reading performance and the influential pattern. First, the mediating variables were different. While the study of Lee and Wu (2012) revealed how involvement in different types of digital reading activities might take effect, focusing on different types and purposes of ICT use might provide educators with new ideas for improving computer-assisted reading education, both in the formal and informal learning contexts. Second, the study of Lee and Wu (2012) explored the relationship with traditional print reading performance instead of the digital one, which might be due to the data availability in PISA 2009. Thus, this study aimed to make an extension of their study based on the latest PISA digital reading assessment data, which might help provide a better understanding of how ICT-related factors related to ICT-based reading performance.

The Present Study

The aforementioned studies investigated the relationship between ICT perceived competence and ICT use and adolescents’ digital reading performance; however, more research on this subject is needed for four reasons. First, prior research has yielded mixed results, thus lacking a definitive answer to whether ICT perceived competence and ICT use could foster the development of digital reading proficiency. Second, scales measuring ICT perceived competence and ICT use have been continuously updated in the PISA assessment with the rapid development of ICT, necessitating an updated investigation of the influential pattern utilizing more comprehensive data. Third, in the vast majority of the previous studies, ICT perceived competence and ICT use were both treated as predictors of digital reading performance, but their intrarelationships and concurrent relationships with digital reading performance were rarely investigated directly. Fourth, although previous researchers have revealed some cultural differences in the relationship between ICT perceived competence and digital reading performance, whether a similar discrepancy exists in the relationship between ICT perceived competence, ICT use and digital reading performance remained to be investigated.

Accordingly, three research questions were addressed in this study:

RQ 1: What is the relationship between ICT perceived competence and adolescents’ digital reading performance?

RQ 2: Whether and to what extent are the relationship between ICT perceived competence and adolescents’ digital reading performance mediated by the frequency of three types of ICT use, i.e., ICT use at home for leisure, ICT use at home for schoolwork, and ICT use at school?

RQ 3: Do the total effects and the mediating effects of ICT perceived competence on adolescents’ digital reading performance differ or are they similar across countries?

With respect to each research questions, three hypotheses were proposed.

H1: The relationship between ICT perceived competence and adolescents’ digital reading performance is significantly positive.

H2: A large proportion of the effects of ICT perceived competence on digital reading performance were mediated by ICT use at home for leisure, ICT use at home for schoolwork, and ICT use at school. In other words, a higher level of ICT perceived competence would lead to more frequent use of ICT, which would in turn result in greater digital reading performance.

H3: The total and mediating effects of ICT perceived competence on adolescents’ digital reading performance are different across countries.

Method

Participants

The data were extracted from the latest PISA database, i.e., the PISA 2018 database (URL: http://www.oecd.org/pisa/data/2018database/) released in December 2019. As a triennial international large-scale educational survey, PISA mainly assesses 15-year-old students’ reading, mathematics, and science competencies, with reading as the major domain in PISA 2018, which was assessed mainly through computers (OECD, 2019b). Within each participating country/region, two-stage sampling was conducted to obtain representative samples of students nearing the end of compulsory education (OECD, 2019c). With respect to the two-stage sampling, first, based on the participating country/region, schools were chosen in proportion to the number of 15-year-old students they had, with 150 such schools set as the minimum number. Second, within each selected school, at least 42 students with proper age were randomly selected as participants. To avoid sampling bias, survey weights for each student were involved in the dataset (OECD, 2019a).

To attain results based on comparable samples, this study focused on OECD countries/regions that participated in the computer-based reading assessment and completed the ICT Familiarity Questionnaire, which collected information on students’ ICT use and ICT perceived competence. Thus, after excluding countries/regions completely missing the required variables, 29 countries/regions with a total of 199,646 students (female = 99,554) from 7973 schools were selected (for demographic information, see Table 1).

Table 1.

Demographic Information of the Sample and the Descriptive Statistics of the Variables.

Country/Region	N (Schools)	N (Students)	Female %	Mean of 10 Plausible Values (PVs)		COMPICT		ENTUSE		HOMESCH		USESCH		Gender		ESCS
Country/Region	N (Schools)	N (Students)	Female %	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD
Australia	763	14273	50	502.8597	108.7629	0.1311	0.8178	0.0085	0.8326	0.1939	0.8467	0.3753	0.7252	0.5043	0.5000	0.2866	0.8727
Belgium	288	8475	50	495.4532	101.9122	0.0204	0.8226	0.0223	0.7828	−0.1564	0.8571	−0.1133	0.8503	0.4960	0.5000	0.0902	0.9175
Chile	228	5822	48	469.3730	93.4130	−0.0081	0.9585	0.0314	0.9001	−0.2024	0.9205	−0.1115	0.8511	0.5210	0.4996	−0.0015	0.9173
Czech Republic	254	7621	50	506.8557	100.3050	0.0557	0.8520	−0.0125	0.9306	0.0914	0.8689	0.0784	0.8131	0.4995	0.5000	−0.3090	1.0833
Denmark	333	7019	50	488.0781	94.4122	−0.1673	0.8425	−0.0083	0.9164	−0.0543	0.8726	0.0679	0.9318	0.4988	0.5000	−0.0854	0.8760
Estonia	348	7657	50	523.2746	93.0781	0.1574	0.7980	0.0562	0.7972	0.1678	0.7527	0.5135	0.6469	0.5016	0.5000	0.3727	0.8361
Finland	230	5316	50	520.3270	99.3134	−0.0285	0.8898	0.0701	0.8793	0.1913	0.8661	−0.1024	0.8962	0.5013	0.5000	0.0979	0.7924
France	214	5649	49	483.7900	105.4722	−0.0222	0.8830	−0.0291	0.7638	−0.2296	0.8462	0.1655	0.7040	0.5093	0.5000	0.2898	0.7888
Greece	252	6308	49	460.6195	96.3931	0.1369	0.9867	0.0518	0.9355	−0.0722	0.9184	−0.0188	0.8267	0.5120	0.4999	−0.1161	0.9662
Hungary	471	13818	51	482.9059	96.7600	0.1520	0.6905	0.0642	0.6169	0.0733	0.6191	0.0738	0.6016	0.4937	0.5000	0.2183	0.8660
Iceland	242	6403	50	473.3086	104.2633	0.0345	0.8792	0.2095	1.2215	0.1103	1.1270	−0.1039	1.1501	0.5037	0.5000	−0.0905	0.9084
Ireland	238	5132	51	517.7072	90.3151	0.0538	0.9337	0.1238	0.9396	0.0727	0.9390	−0.0932	0.9445	0.4924	0.5000	−0.0653	0.9120
Israel	157	5577	50	472.3269	123.5917	0.1429	0.8386	−0.0099	0.8073	−0.4014	0.8983	−0.3540	0.8941	0.5021	0.5000	0.1225	0.8614
Italy	142	3296	50	481.0910	94.1373	−0.0744	0.8692	−0.1715	0.8597	0.0251	0.8476	0.2393	0.7333	0.4976	0.5001	0.5208	0.8209
Japan	174	6623	54	503.4793	97.3200	−0.0386	0.8914	−0.1475	1.0695	−0.0123	0.9128	0.0334	0.8534	0.4649	0.4988	0.3238	0.9748
Korea	542	11785	48	515.7486	101.3614	−0.0729	0.8208	0.1097	0.9331	0.0567	0.8905	0.0551	0.8439	0.5180	0.4997	−0.2139	0.8777
Latvia	183	6109	51	476.1705	88.8775	−0.8102	0.9918	−0.2303	0.8104	−1.0774	0.8949	−1.0842	0.8084	0.4893	0.4999	−0.1087	0.7285
Lithuania	188	6650	48	471.0583	94.1684	−0.3170	0.9599	−0.0972	0.8279	−0.0282	0.9020	−0.7040	0.9937	0.5202	0.4996	0.0852	0.7716
Luxembourg	362	6885	49	470.5199	108.4390	0.0025	0.9930	0.1429	1.0714	0.2780	1.0061	0.1311	1.0032	0.5095	0.4999	0.0285	0.8478
Mexico	44	5230	50	470.5199	108.4390	−0.0146	0.9613	0.1315	0.9794	−0.1620	0.9807	−0.0104	0.9196	0.5040	0.5000	0.0074	1.1347
New Zealand	308	5303	51	508.2166	105.5848	−0.1125	0.8446	0.0741	0.9870	0.3117	0.8485	0.2675	0.8571	0.4937	0.5000	−0.0151	0.8159
Poland	286	7299	52	512.6324	96.4716	−0.0328	0.8518	−0.2120	0.9898	0.3005	0.8647	0.0669	0.8602	0.4758	0.4994	−0.9911	1.1800
Slovak Republic	192	6173	51	460.5737	100.8561	0.1408	0.8630	−0.0353	0.8039	0.1140	0.8477	0.3323	0.6930	0.4891	0.4999	0.1663	0.9580
Slovenia	240	5625	51	480.0999	93.7653	0.0107	0.8822	0.1719	1.0544	0.2474	0.9351	0.1366	1.0473	0.4921	0.5000	−0.1408	0.8425
Sweden	376	5965	50	505.5255	107.0829	−0.1739	0.8458	−0.0563	1.0692	0.1704	0.9322	0.1635	0.9660	0.4967	0.5000	−0.1771	0.9106
Switzerland	345	6401	47	484.4248	100.6693	−0.0223	0.9004	−0.1432	1.0561	0.0651	0.9316	0.0329	0.9470	0.5324	0.4990	−0.0140	0.7896
Turkey	223	5504	50	464.5909	87.4305	0.1965	0.9252	0.0444	0.8877	0.1487	0.8681	0.3620	0.8234	0.4980	0.5000	0.3442	0.8863
United Kingdom	186	6890	49	499.5880	97.9624	−0.1317	1.0188	−0.1125	1.3015	0.1115	0.9582	−0.1629	1.0368	0.5071	0.5000	−1.1665	1.1772
United States	164	4838	49	500.5677	108.0246	0.1002	0.8954	−0.0232	1.0016	0.2266	0.9569	0.3979	0.8299	0.5089	0.5000	0.0528	1.0098
Total	7973	199646	50	/	/	/	/	/	/	/	/	/	/	/	/	/	/

Note. Due to the limited space, only the mean of the ten PVs was presented in lieu of the value of each of the PVs. The values of all the PVs can be accessed through the publicly available database of PISA 2018 (http://www.oecd.org/pisa/data/2018database/).

Materials

Table 2 includes the descriptions of all the variables in this study. Additionally, the descriptive statistics of the dependent, independent, mediating, and control variables in each participating country/region was presented in Table 1.

Table 2.

Variable Description.

Variable Codes	Description	Sample Item	Scale
Dependent variable
READ	Digital reading performance
Independent variable
EFFICT	ICT perceived competence	I feel comfortable using my digital devices at home	1 = Strongly disagree
			2 = Disagree
			3 = Agree
			4 = Strongly agree
Mediating variables
ENTUSE	ICT use outside of school for entertainment	Reading news on the Internet	1 = Never or hardly ever
			2 = Once or twice a month
			3 = Once or twice a week
			4 = Almost every day
			5 = Every day
HOMESCH	ICT use outside of school for schoolwork	Browsing the Internet for schoolwork (e.g. for preparing an essay or presentation)	Same as above
USESCH	ICT use at school	Using email at school	Same as above
Control variables
Gender	Gender		1= Female
Gender	Gender		2 = Male
ESCS	Students’ socioeconomic and cultural status

Source. PISA 2018 database.

The dependent variable was students’ digital reading performance in PISA 2018, represented by 10 plausible values (PVs). Due to the subset design of the assessment, the PVs were computed to provide multiple estimations of students’ reading performance. Correspondingly, all 10 PVs were utilized in each run of the analysis in this study. Across OECD countries/regions, the PVs were calibrated with means of approximately 500 and standard deviations of approximately 100 (OECD, 2019b). PISA 2018 digital reading assessment evaluates adolescents’ digital reading proficiency in the following aspects: reading fluently, understanding, and evaluating and reflecting upon manifold text format via computer. This assessment simulated the online-reading tasks and scenarios that students in engaged in by clicking on the links, browsing the Internet to retrieve information and evaluating documents. Furthermore, it employed a multistage adaptive design with three consecutive stages, including Core (7–10 items), Stage 1 (12–15 items), and Stage 2 (12–15 items) with a total of 245 items. Consequently, most students were tested 35–39 items in an hour, and the students’ proficiency in the previous stage was correlated to the testlet they were going to finish in the next stage (OECD, 2019b).

The independent variable was students’ perceived competence in using ICT. Coded as COMPICT, this variable was derived from five items concerning students’ perception of their competence in conducting ICT-related activities and applying ICT-related knowledge. Students provided their responses to the statements (e.g., “When I come across problems with digital devices, I think I can solve them”) on a four-point Likert scale ranging from “strongly disagree” to “strongly agree”, after which a composite score was calculated using item response theory (IRT) scaling, with an OECD mean of 0 and an OECD standard deviation of 1. Furthermore, the cross-country comparability of the variable was ensured by employing the IRT scaling methodology (OECD, 2009). For the detailed item wordings, see Supplementary Files Table S1.

The three potential mediating variables concerned students’ ICT use in different contexts and for different purposes: ICT use outside of school for entertainment (ENTUSE), ICT use outside of school for schoolwork (HOMESCH), and ICT use at school in general (USESCH). 12, eleven, and ten questions were asked, respectively, regarding the frequency of students’ involvement in these activities conducted on different types of ICT platforms and using various tools such as social media, digital games, and mobile devices. Students were required to respond on a five-point Likert scale, with the options of “never or hardly ever” “once or twice a month” “once or twice a week” “almost every day” and “every day”. Similarly, IRT scaling was utilized to derive the variables. For the detailed item wordings, see Supplementary Files Table S1.

In addition, two control variables were used throughout the analysis, student gender (Gender) and the index of economic, social, and cultural status (ESCS), because both were found to be related to students’ digital reading performance (Wu, 2014; Xiao & Hu, 2019), academic perceived competence especially ICT perceived competence (Cai et al., 2017; Cusso-Calabuig et al., 2018; Franzen et al., 2022; Vekiri, 2010), and frequency of ICT use (OECD, 2010; Qazi et al., 2022). The gender variable was converted to a dummy variable in this study, with female students coded as 0 and male students coded as 1.

The scale reliabilities of all the ICT-related derived variables were computed by the PISA data validation staff and all values exceeded 0.7. For COMPICT, ENTUSE, HOMESCH, and USESCH, the average values of the 29 countries/regions ranged from 0.8 to 0.9 (see Supplementary Files Table S2). Thus, acceptable internal consistency was demonstrated.

Procedure

The data were pre-processed prior to the main analysis. First, missing values were imputed with the expectation maximization algorithm in SPSS 20.0, which produced a single data point for each missing value by rendering maximum-likelihood estimates of the model parameters (Dempster et al., 1977). Second, the correlation matrices were computed for each country/region (see Supplementary Files Tables S2-S60). Since none of the correlation coefficients exceeded 0.80, no multicollinearity problem was detected (Field, 2009), and all variables were retained. Third, to obtain unbiased population-level parameters, the final student weight was normalized and used in the modelling (see OECD, 2009).

To investigate the relationship between ICT perceived competence and students’ digital reading performance and the potential mediating roles of three aspects of ICT use, this study conducted 1-1-1 multilevel mediation analyses with multiple parallel mediators.

The mediation analysis refers to a regression-based investigation of the means by which an independent variable (X) affects a dependent variable (Y) through one or more mediating variables (M), which aims to reveal the mechanism underlying the effects of X on Y. As demonstrated in the mediation model (Figure 1), this study is primarily concerned with the total effect (path $c$ ) and the indirect effects, also called mediating effects (path $a_{1} b_{1}$ , $a_{2} b_{2}$ , and $a_{3} b_{3}$ ). Parallel multiple mediator models were adopted in lieu of separated runs of simple mediation analysis because the former reduced the possibility of parameter bias by including multiple mediating variables simultaneously and accounting for the intercorrelation between them (Hayes, 2009; Preacher & Hayes, 2008).

Figure 1.

The mediation model of the present study. For the simplicity of the diagram, the two control variables (Gender and ESCS) are not presented but were used throughout the analysis.

Furthermore, this study employed multilevel mediation to cope with the PISA data structure, i.e., the three-level data structure in which students are nested within schools and then within countries, resulting from the PISA two-stage sampling (OECD, 2009). While the nested structure of the data violates the independence assumption of ordinary least square regression, multilevel models could account for the substantial proportion of variance that exists between different schools, thus reducing the risk of statistical errors such as aggregation bias and unreliable standard errors (Hox, 2010). Furthermore, multilevel mediation has been employed in educational studies (e.g., Chiao & Chiu, 2018; Ieong et al., 2016; Liu et al., 2015) and found to be feasible. Specifically, 1-1-1 multilevel mediation analyses were conducted, in which all the variables including X, M, and Y, were measured at level 1 (i.e., the student level). In addition, to prevent potential confounding, the data were CWC (M) in the multilevel analyses; i.e., X and M were “centred within context”. In other words, the group mean was centered at the student level with the aggregated mean reintroduced to the school-level equations (Zhang et al., 2009).

Thus, prior to model building, the appropriateness of data aggregation is examined by calculating, $r_{w g (j)}$ , an index of agreement within groups, and the intraclass correlation coefficient (ICC). Additionally, the $x^{2}$ of the between-school variance of the mediating variables are calculated. To support data aggregation, it is recommended that $r_{w g (j)}$ > 0.70, ICC >0.05 (LeBreton & Senter, 2008), and $x^{2}$ of between-school variance be statistically significant (p < 0.05) (Mak et al., 2017). In the present study, the $r_{w g (j)}$ for each of the mediators in each of the schools has a value of 1. With the exception of the ICC of ENTUSE and the $x^{2}$ of ENTUSE in three countries/regions, the results support data aggregation of the mediators for further analyses. The multilevel mediation effects of ENTUSE, if significant, should be interpreted with caution (see Supplementary Files Table S61).

The models were built using the software Hierarchical Linear & Nonlinear Modeling 8 (Raudenbush et al., 2019) (License ID: 1686), which can conveniently handle PVs and student weights. After all model assumptions were checked (Hox, 2010), stepwise model building, adapted from Zhang et al. (2009), was applied to each of the 29 countries/regions separately, as presented in Table 3. First, the null model was built, with digital reading performance as the dependent variable and no independent variable, so that the necessity of multilevel modelling was examined by computing the ICC. Second, the total effect model was built, with digital reading performance as the dependent variable, ICT perceived competence as the independent variable, and gender and ESCS as the control variables. In this step, the total effect (path c) was computed. In the third step, three intermediate models were constructed, with the three ICT use variables as the dependent variables, ICT perceived competence as the independent variable, and gender and ESCS as the control variables, so that the coefficients of path a were computed. Then, the last model was built, with digital reading performance as the dependent variable, ICT perceived competence as the independent variable, the three types of ICT use as the mediating variables, and gender and ESCS as the control variables, so that the direct (c’) and indirect effects (ab) were computed. All the aforementioned steps were applied to each of the 29 countries/regions separately. Each PV was employed as the dependent variable in every single run of the model, after which the mean of the 10 PVs was computed (OECD, 2009).

Table 3.

1-1-1 Multilevel Mediation Models with Three Parallel Mediators.

Null Model	Step 1: Total Effect Model	Step 2: Models of the Independent variable’s Effects on the Mediators	Step 3: Direct and Indirect Effect Model
$Y_{i j} = β_{0 j} + r_{i j}$	$Y_{i j} = β_{0 j}^{(1)} + β_{1 j}^{(1)} (X_{i j} - X_{\cdot j}) + β_{2 j}^{(1)} (C_{i j} - C_{\cdot j}) + r_{i j}^{(1)}$	${M 1}_{i j} = β_{0 j}^{(2.1)} + β_{1 j}^{(2.1)} (X_{i j} - X_{\cdot j}) + β_{2 j}^{(2.1)} (C_{i j} - C_{\cdot j}) + r_{i j}^{(2.1)}$	$Y_{i j} = β_{0 j}^{(3)} + β_{1 j}^{(3)} (X_{i j} - X_{\cdot j}) + β_{2 j}^{(3)} ({M 1}_{i j} - {M 1}_{\cdot j}) + {β_{3 j}^{(3)} ({M 2}_{i j} - {M 2}_{\cdot j}) + β_{4 j}^{(3)} ({M 3}_{i j} - {M 3}_{\cdot j}) + β_{5 j}^{(3)} (C_{i j} - C_{\cdot j}) + r}_{i j}^{(3)}$
$β_{0 j} = γ_{00} + u_{0 j}$	$β_{0 j}^{(1)} = γ_{00}^{(1)} + γ_{01}^{(1)} X_{\cdot j} + γ_{02}^{(1)} C_{\cdot j} + u_{0 j}^{(1)}$	$β_{0 j}^{(2.1)} = γ_{00}^{(2.1)} + γ_{01}^{(2.1)} X_{\cdot j} + γ_{02}^{(2.1)} C_{\cdot j} + u_{0 j}^{(2.1)}$	$β_{0 j}^{(3)} = γ_{00}^{(3)} + γ_{01}^{(3)} X_{\cdot j} + γ_{02}^{(3)} {M 1}_{\cdot j} + γ_{03}^{(3)} {M 2}_{\cdot j} + γ_{04}^{(3)} {M 3}_{\cdot j} + γ_{05}^{(3)} C_{\cdot j} + u_{0 j}^{(3)}$
	$β_{1 j}^{(1)} = γ_{10}^{(1)}$	$β_{1 j}^{(2.1)} = γ_{10}^{(2.1)}$	$β_{1 j}^{(3)} = γ_{10}^{(3)}$
	$β_{2 j}^{(1)} = γ_{20}^{(1)}$	$β_{2 j}^{(2.1)} = γ_{20}^{(2.1)}$	$β_{2 j}^{(3)} = γ_{20}^{(3)}$
		${M 2}_{i j} = β_{0 j}^{(2.2)} + β_{1 j}^{(2.2)} (X_{i j} - X_{\cdot j}) + β_{2 j}^{(2.2)} (C_{i j} - C_{\cdot j}) + r_{i j}^{(2.2)}$	$β_{3 j}^{(3)} = γ_{30}^{(3)}$
		$β_{0 j}^{(2.2)} = γ_{00}^{(2.2)} + γ_{01}^{(2.2)} X_{\cdot j} + γ_{02}^{(2.2)} C_{\cdot j} + u_{0 j}^{(2.2)}$	$β_{4 j}^{(3)} = γ_{40}^{(3)}$
		$β_{1 j}^{(2.2)} = γ_{10}^{(2.2)}$	$β_{5 j}^{(3)} = γ_{50}^{(3)}$
		$β_{2 j}^{(2.2)} = γ_{20}^{(2.2)}$
		${M 3}_{i j} = β_{0 j}^{(2.3)} + β_{1 j}^{(2.3)} (X_{i j} - X_{\cdot j}) + β_{2 j}^{(2.3)} (C_{i j} - C_{\cdot j}) + r_{i j}^{(2.3)}$
		$β_{0 j}^{(2.3)} = γ_{00}^{(2.3)} + γ_{01}^{(2.3)} X_{\cdot j} + γ_{02}^{(2.3)} C_{\cdot j} + u_{0 j}^{(2.3)}$
		$β_{1 j}^{(2.3)} = γ_{10}^{(2.3)}$
		$β_{2 j}^{(2.3)} = γ_{20}^{(2.3)}$

Note. Y refers to the dependent variable, students’ digital reading performance. X refers to the independent variable, students’ ICT perceived competence. M1, M2 and M3 refer to the three mediators ENTUSE, HOMSCH and USESCH. For the simplicity of the model, C stands for the two control variables, ESCS and Gender.

The mediating effects were tested by joint significance (Hayes, 2009) accompanied by the computation of Monte Carlo confidence intervals (Preacher & Selig, 2012; Veas et al., 2019) using the RMediation (Tofighi & MacKinnon, 2011) package in R (R Core Team, 2019). Regarding the effect size of the mediators, the proportion mediated (PM), i.e., the ratio of the indirect effect to the total effect, was computed as the main indicator (Wen & Fan, 2015).

Results

Necessity of Multilevel Modelling

Prior to the main analyses, an intercept-only null model, with digital reading performance as the dependent variable and no independent variable, was built for each country/region to examine the necessity of multilevel modelling. Specifically, the ICC was obtained by computing the proportion of the variance explained by the student-level and school-level variables (Hox, 2010). According to Cohen (1988), an ICC higher than 0.059 and lower than 0.138 indicates medium intraclass correlation, and an ICC higher than 0.138 indicates high intraclass correlation. Thus, the ICCs in this study, ranging from 0.08 to 0.58 (see Table 4), confirmed the significant proportion of variances in digital reading performance across schools and the subsequent need for multilevel modelling.

Table 4.

ICC for Each of the 29 Countries/Regions.

Country/Region	Within-School Variance	Between-School Variance	ICC
Australia	9642.1000	2182.0635	0.1845
Belgium	5781.7304	5109.3801	0.4691
Chile	5764.4587	3432.5452	0.3732
Czech Republic	5059.0920	4325.6121	0.4609
Denmark	7276.9548	1281.8081	0.1498
Estonia	6965.9920	1706.0418	0.1967
Finland	9178.7872	772.58880	0.0776
France	5622.1992	4929.2451	0.4672
Greece	6165.3383	4173.4034	0.4037
Hungary	4233.4978	5849.7368	0.5801
Iceland	10127.6705	832.86900	0.0760
Ireland	7106.1265	1236.9685	0.1483
Israel	8153.2903	8193.0806	0.5012
Italy	5227.0047	4625.8880	0.4695
Japan	5859.4152	3599.1465	0.3805
Korea	7768.5485	2599.2265	0.2507
Latvia	6068.4906	2028.2857	0.2505
Lithuania	5608.2340	3717.7297	0.3986
Luxembourg	8190.9079	3489.2668	0.2987
Mexico	4284.8727	2834.8943	0.3982
New Zealand	9585.5015	1977.8589	0.1710
Poland	7665.7140	1810.4368	0.1911
Slovak Republic	5344.1128	4215.7161	0.4410
Slovenia	4671.5317	4197.0756	0.4733
Sweden	9675.9067	2013.7697	0.1723
Switzerland	6562.1581	3908.4890	0.3733
Turkey	3390.1267	4652.5096	0.5785
United Kingdom	8261.4444	1760.0676	0.1756
United States	9668.3688	2055.8864	0.1754

Total Effect

First, a random intercept and fixed slope model was built to investigate the total effect. The model, with digital reading performance as the independent variable, COMPICT as the dependent variable, and Gender and ESCS as the control variables, was constructed for each country/region.

As shown in Table 5, in terms of the within-school total effects, the model coefficients are significantly positive for all of the countries/regions except for Japan and Korea, ranging from 2.3485 to 14.2787, indicating that with a 1-unit increase in students’ ICT perceived competence, their digital reading performance increases by 2.3485–14.2787 points. Regarding the between-school effects, the model coefficients are significantly positive for 20 countries/regions and range from 24.5770 to 69.0838 (see Supplementary Files Table S62).

Table 5.

Within-School Total and Direct Effects of ICT Perceived Competence on Digital Reading Performance.

Country/Region	Total Effects		Direct Effects
Country/Region	B	SE	B	SE
Australia	12.8507	1.2883^***	19.8346	1.4343^***
Belgium	5.5353	1.2395^***	9.4429	1.2534^***
Chile	9.4814	1.2575^***	13.4120	1.3752^***
Czech Republic	5.2955	1.3307^***	7.4580	1.4093^***
Denmark	11.3795	1.6941^***	16.4986	1.8416^***
Estonia	11.9345	1.7036^***	14.0141	1.7741^***
Finland	9.4875	1.5577^***	13.5381	1.6076^***
France	2.3485	1.1206^*	5.0252	1.1753^***
Greece	11.3040	1.3352^***	13.5250	1.4023^***
Hungary	7.7343	1.2550^***	10.4751	1.3166^***
Iceland	14.1479	1.9329^***	18.9982	2.1455^***
Ireland	11.2840	1.5565^***	13.3718	1.5962^***
Israel	10.9428	1.5275^***	14.2517	1.6349^***
Italy	9.3115	1.2868^***	10.0787	1.3424^***
Japan	2.0813	1.2603	5.1313	1.3008^***
Korea	1.3110	1.4593	3.2808	1.5920^*
Latvia	14.0569	1.3595^***	15.5287	1.4413^***
Lithuania	14.2787	1.0579^***	16.7054	1.0550^***
Luxembourg	7.4888	1.5545^***	12.401	1.5638^***
Mexico	9.3619	1.2319^***	10.0828	1.2978^***
New Zealand	11.4330	1.6583^***	17.5742	1.7732^***
Poland	13.0843	1.6312^***	17.5228	1.6247^***
Slovak Republic	9.6539	1.4452^***	11.5003	1.5547^***
Slovenia	6.8801	1.3297^***	9.3629	1.4297^***
Sweden	7.5797	1.5226^***	12.9398	1.7087^***
Switzerland	5.1357	1.4080^**	8.5365	1.4966^***
Turkey	7.7539	0.7922^***	8.5083	0.7961^***
United Kingdom	9.5745	1.5019^***	15.7173	1.5436^***
United States	11.4034	2.0223^***	16.2414	1.9493^***

Note. *: p < .05, **: p < .01, ***: p < .001.

In addition, the effects of the control variables are presented in Supplementary Files Table S63. For student gender, the coefficients range from −52.2360 to −8.2990, all of which are significantly negative, indicating that male adolescents in general have a lower level of digital reading performance than their female counterparts. For ESCS, the coefficients range from 3.2952 to 34.6525, all of which are significantly positive, suggesting that students with higher levels of ESCS perform better in the digital reading assessment.

Indirect and Direct Effects

Following the null and total effect models, two sets of models were subsequently constructed to examine the direct and indirect effects of adolescents’ ICT perceived competence and their digital reading performance.

As for the direct effects, the coefficients for the within-school are significantly positive for all countries/regions, ranging from 3.2808 to 19.8346 (see Table 5). The between-school coefficients were significantly positive for 22 out of 29 countries/regions, ranging from 35.1991 to 110.1892 (see Supplementary Files Table S62).

As suggested by the model coefficients and the effect size of the within-school indirect effects presented in Table 6, ENTUSE is a significant mediator for 16 countries/regions, HOMESCH for 27 countries/regions, and USESCH for all 29 countries/regions. Primarily, to compute the effect size, PM was used for partial and full mediation. Nevertheless, since the suppression effects, i.e., a particular situation of mediation when the direct and indirect effects have opposite signs (MacKinnon et al., 2000), were detected in this study, the absolute value of the ratio of the indirect effect to the direct effect was used as the indicator of the effect size (Wen & Ye, 2014). According to Cohen (1988), the thresholds of the effect sizes for small, medium, and large effects are 0.10, 0.30, and 0.50, respectively. As shown in Table 6, the effect sizes range from 5.51% to 59.99% with a mean of 28.59%; these relatively medium effect sizes indicate that a large proportion of the effect of ICT perceived competence on digital reading performance is mediated through the three types of ICT use. Regarding the between-school indirect effects (see Supplementary Files Table S64), ENTUSE, HOMESCH, and USESCH are significant mediators for 9, 3, and three countries/regions, respectively, indicating that the three types of ICT use are not significant mediators at the school level.

Table 6.

Within-School Indirect Effects of ICT Perceived Competence on Digital Reading Performance and the Effect Size.

Country/Region	ENTUSE		HOMESCH		USESCH		Effect Size
Country/Region	$a_{1} b_{1}$ (95%CI)	SE	$a_{2} b_{2}$ (95%CI)	SE	$a_{3} b_{3}$ (95%CI)	SE	Effect Size
Australia	−1.8368 (-2.9541, -0.7389)	0.5650	−1.6716 (-2.4440, -0.9475)	0.3796	−3.4794 (-4.3737, -2.6410)	0.4414	0.3523
Belgium	−0.3475 (−1.2882, 0.5977)	0.4801	−1.9253 (-2.6072, -1.3163)	0.3298	−1.6335 (-2.2205, -1.1062)	0.2847	0.4137
Chile	2.3446 (1.3251, 3.4121)	0.5307	−2.9771 (-3.9950, -2.0411)	0.4964	−3.2974 (-4.2213, -2.4626)	0.4509	0.2930
Czech Republic	1.9848 (0.9225, 3.0931)	0.5545	−1.6833 (-2.3698, -1.0534)	0.3356	−2.4632 (-3.2933, -1.7220)	0.4014	0.2899
Denmark	0.2313 (−1.1699, 1.6360)	0.7116	−4.0953 -(5.5969, -2.7729)	0.7240	−1.2606 (-2.4728, -0.1282)	0.5947	0.3106
Estonia	1.1557 (−0.0270, 2.3658)	0.6091	−1.6642 (-2.4612, -0.9549)	0.3844	−1.5779 (-2.2618, -0.9831)	0.3276	0.1489
Finland	1.1735 (−0.0331, 2.4142)	0.6235	−1.6842 (-2.4243, -1.0196)	0.3586	−3.5425 (-4.6498, -2.5298)	0.5423	0.2994
France	−0.3197 (−1.1350, 0.4877)	0.4151	−1.7102 (-2.3064, -1.1784)	0.2878	−0.6484 (-1.0079, -0.3271)	0.1735	0.5330
Greece	1.6307 (0.7957, 2.5031)	0.4363	−2.6837 (-3.5159, -1.9384)	0.4032	−1.1688 (-1.8042, -0.5709)	0.3136	0.1643
Hungary	1.1555 (0.2791, 2.0534)	0.4524	−1.9161 (-2.6594, -1.2582)	0.3577	−1.9805 (-2.6322, -1.3962)	0.3145	0.2617
Iceland	1.0467 (−0.4968, 2.6346)	0.7938	−2.3823 (-3.7835, -1.1981)	0.662	−3.5167 (-5.1584, -2.1162)	0.7791	0.2554
Ireland	0.1910 (−0.8852, 1.2670)	0.5493	−1.4755 (-2.1152, -0.9049)	0.3107	−0.7989 (-1.3153, -0.3546)	0.2456	0.1558
Israel	1.2932 (0.1374, 2.4650)	0.5941	−1.7811 (-2.6933, -0.9421)	0.4472	−2.8167 (-3.7187, -1.9923)	0.4425	0.2319
Italy	2.5868 (1.3541, 3.8751)	0.6441	−1.3190 (-2.0155, -0.6974)	0.3370	−2.0390 (-2.8732, -1.3183)	0.3979	0.0765
Japan	−1.1870 (-2.2275, -0.1587)	0.5277	−1.1889 (-1.8985, -0.5262)	0.3497	−0.6746 (-1.1740, -0.2189)	0.2434	0.5945
Korea	0.4208 (−0.7211, 1.5742)	0.5869	−0.1943 (−1.0786, 0.6856)	0.4498	−2.1947 (-2.8480, -1.6068)	0.3172	0.5999
Latvia	2.6642 (1.6269, 3.8024)	0.5548	−1.4218 (-2.2177, -0.7187)	0.3830	−2.7216 (-3.6205, -1.8949)	0.4410	0.0953
Lithuania	1.3501 (0.7171, 2.0205)	0.3315	−1.6517 (-2.3084, -1.0635)	0.3173	−2.1295 (-2.8690, -1.4489)	0.3639	0.1455
Luxembourg	−0.2496 (−1.3302, 0.8271)	0.5517	−1.9029 (-2.8441, -1.1075)	0.4448	−2.7546 (-3.7803, -1.8471)	0.4924	0.3957
Mexico	2.8917 (1.8457, 3.9861)	0.5451	−1.4205 (-2.1770, -0.6995)	0.3776	−0.9558 (-1.5056, -0.4597)	0.2666	0.0551
New Zealand	−2.2653 (-3.5207, -1.0649)	0.6266	−1.4873 (-2.1961, -0.8461)	0.3431	−2.3951 (-3.3880, -1.4797)	0.4860	0.3498
Poland	1.5216 (0.4670, 2.6112)	0.5440	−2.3377 (-3.2373, -1.5209)	0.4366	−3.6268 (-4.5668, -2.7553)	0.4632	0.2536
Slovak Republic	2.7224 (1.5924, 3.8921)	0.5865	−2.0371 (-2.8742, -1.2801)	0.4054	−2.5294 (-3.3156, -1.8187)	0.3827	0.1604
Slovenia	0.8244 (−0.1021, 1.7790)	0.4792	−1.5403 (-2.2508, -0.9057)	0.3443	−1.7648 (-2.4694, -1.1485)	0.3382	0.2650
Sweden	0.0562 (−1.1445, 1.2582)	0.6112	−3.0694 (-3.7563, -2.3832)	0.3514	−2.3559 (-3.3359, -1.4745)	0.4750	0.4149
Switzerland	−0.4589 (−1.4790, 0.5428)	0.5146	−2.0877 (-2.8930, -1.3799)	0.3875	−0.8561 (-1.3465, -0.4479)	0.2305	0.3986
Turkey	1.5724 (1.0064, 2.1595)	0.2941	−0.9661 (-1.4813, -0.4846)	0.2532	−1.3612 (-1.7837, -0.9816)	0.2055	0.0887
United Kingdom	−3.0299 (-4.2336, -1.8894)	0.5963	−0.4055 (−1.0541, 0.2168)	0.3210	−2.7085 (-3.6122, -1.9035)	0.4353	0.3909
United States	−0.6124 (−2.0951, 0.8619)	0.7535	−2.1419 (-3.3309, -1.0471)	0.5821	−2.0892 (-3.2027, -1.0463)	0.5490	0.2982

Note. Confidence intervals that contain zero are deemed nonsignificant. The significant indirect effects are highlighted in bold.

Discussion

This study aimed to investigate the relationship between adolescents’ ICT perceived competence and their digital reading performance and to examine the extent to which this relationship was mediated by three types of ICT use, including ICT use at home for leisure, ICT use at home for schoolwork, and ICT use at school. Accordingly, three research questions were addressed.

The Total Effects of ICT Perceived Competence

Research question 1 explored the relationship between ICT perceived competence and adolescents’ digital reading performance. This study revealed a significantly positive correlation between ICT perceived competence and digital reading performance in the vast majority of the sample, i.e., 27 out of 29 countries/regions. Theoretically, such a finding is congruent with the ICT engagement model (Goldhammer et al., 2016), which proposed that high ICT perceived competence might improve performance in learning tasks involving ICT use, in this case, reading activities via digital platforms. Empirically, the positive results were in alignment with the studies of Hu et al. (2018), Lim and Jung (2019), and Naumann and Saelzer (2017), all of which revealed better digital reading performance among the students who perceived themselves as competent in using ICT than those who were less confident in using digital devices or solving problems with computers, etc.

The finding that a higher level of ICT perceived competence might generate better digital reading performance could be explained in two ways. First, a substantial number of studies have revealed that perceiving oneself as competent in performing ICT-related tasks is predictive of performance in ICT tasks (e.g., Christoph et al., 2015; Hatlevik et al., 2015; Naumann & Saelzer, 2017; Senkbeil, 2022). Moreover, the mastery of basic ICT skills such as using hyperlinks or other navigation tools is indispensable for successful digital reading (OECD, 2019c) because students will frequently encounter these features while reading on digital devices. Thus, ICT perceived competence might foster digital reading performance by improving ICT-related task performance. Second, students perceiving themselves to be ICT competent might be less anxious in the PISA digital reading assessment environment, enabling them to be more focused on the reading task itself than on the test environment. A study of computer-based science assessment in the PISA 2000 in the U.S.A. (Papanastasiou et al., 2003) indicated that students with computer anxiety might be overwhelmed by their anxiety and the unfamiliar technology and thus be less focused on the computer-based task itself. A similar phenomenon was found in higher education, as the study of Deutsch et al. (2012) suggested that students’ academic performance in computer-based assessment might vary for students with different perceptions of their computer skills. Such impacts, however, might be mitigated by acquainting the students with the computer-based test-taking environment.

The Mediating Effects of ICT Use

Research question 2 explored whether and to which extent the effects of ICT perceived competence on adolescents’ digital reading performance were mediated by the frequency of three types of ICT use. The effect sizes of the within-school indirect effects indicated that mediation did exist and that a large proportion of the effect of ICT perceived competence on digital reading performance was mediated through ICT use. A closer investigation shows that using ICT at home for leisure, at home for schoolwork, and at school all significantly mediates the effects of ICT perceived competence on digital reading performance for most countries/regions.

Furthermore, a special type of mediating effect, the suppression effect, was detected in the present study. While adolescents considering themselves as competent in performing ICT-related tasks were more likely to perform better in the digital reading assessment, when the effects of ICT use on digital reading performance were considered simultaneously, the positive effects of ICT perceived competence tended to decrease. In other words, the positive effects of ICT perceived competence were offset by particular types of ICT use. While a higher level of ICT perceived competence led to frequent use of ICT, particular types of ICT use, especially ICT use outside of school for schoolwork and ICT use at school in general, in contrast, were found to result in poorer digital reading performance. This paradox could be explained by two reasons. First, the negative relationship between these types of ICT uses and digital reading performance are consistent with previous findings (e.g., Hu et al., 2018; Petko et al., 2017; Skryabin et al., 2015; Xiao et al., 2019; Zhu & Mok, 2020). Second, although students who are more confident in their ICT skills might be more active in engaging in ICT-related activities, students might unconsciously inflate their ICT abilities (Aesaert et al., 2017). Such overestimation indicates that actually, students might not fully master the ICT skills for digital reading (OECD, 2019c), which leads to poor digital reading performance.

Cross-Cultural Comparison

The present study answered research question three by revealing that the total effects of ICT perceived competence on adolescents’ digital reading performance were consistently positive for the vast majority of the sample, i.e., 27 out of 29 countries/regions. With respect to Japan and Korea, which showed nonsignificant results, one possible reason is that cultural differences do exist in the relationship between ICT perceived competence and digital reading performance. Such a finding partly resonates with the findings of Meng and partners (2019), which suggested that ICT perceived competence, as a characteristic construct of students’ self-concepts, was culturally specific. Specifically, since students in Eastern cultures tend to nurture dispositions that pursue modesty and in-depth knowledge exploration (Schutte, 2015), they were found to have better academic performance than students in Western countries, even though they did not perceive themselves to be competent (Liu & Meng, 2010). It is recommended that more studies be conducted in the future to explore this cultural difference in digital reading performance.

In addition, the results showed that the suppression effects of using ICT at home for academic purposes and using ICT at school were highly consistent for all of the countries/regions, with significant indirect effects, suggesting that the present study revealed an influential pattern with considerable generalizability. The mediating effects of using ICT at home for recreational purposes in the relationship between ICT perceived competence and digital reading performance were positive for 10 countries/regions and negative for 4 countries/regions. Such mixed results have already been shown in previous studies (e.g., Hu et al., 2018; Petko et al., 2017; Skryabin et al., 2015; Xiao et al., 2019) and the varying relationship might be due to different study samples and the relevant cultural differences.

Educational Implications

These findings have several pedagogical implications for improving digital reading education for adolescents. First, this study confirmed that when adolescents differ greatly in their self-perceived competence in using ICT, such differences may further deepen the achievement gaps in digital reading. Thus, appropriate steps need to be taken at school and at home to enhance students’ confidence in their ICT competence. On the one hand, schools are recommended to build a competence-supportive environment, for instance, to give positive feedback (Deci & Ryan, 2000), which could help satisfy students’ basic psychological needs for competence (Areepattamannil & Santos, 2019). Moreover, providing more quality-oriented support for technology-integrated instruction at school are likely to strengthen the positive effects of ICT perceived competence on students’ digital reading performance (Xiao & Hew, 2022). On the other hand, positive attitudes can derive from school use as well as from home use (Petko et al., 2017). Since students’ ICT use at home, either for recreational or academic purposes, might impact how well they develop their digital reading skills, parents’ support is also deemed helpful for the cultivation of ICT perceived competence at home. To improve adolescents’ digital reading performance through facilitating ICT perceived competence, students’ significant others, such as teachers and parents, are encouraged to be aware of students’ psychological needs of recognition from seniors and provide affirmative remarks constantly (Park, 2011).

Second, the current study, demonstrating the suppression effects, newly found that to improve digital reading performance simply boosting ICT perceived competence is not enough. Instead, the underlying effects of ICT use need cautious treatment. It is suggested that students with a high level of ICT perceived competence be encouraged to use ICT at home for leisure more frequently, as more frequent use was found to be positively correlated with digital reading performance, in line with previous studies (Hu et al., 2018; Skryabin et al., 2015; Srijamdee & Pholphirul, 2020). In contrast, further studies could explore the reasons for the negative effects of ICT use at home for schoolwork and ICT use at school to develop relevant measures to counteract such effects. For instance, efforts could be made to monitor students’ use of ICT (e.g., Holte, Giesen, & Ferraro, 2021; Tomte & Hatlevik, 2011), special attention is suggested to be paid to various ICT-based communication modes (Yu & Hu, 2022), the use of digital technology in classrooms need to be carefully “orchestrated” (Petko et al., 2017), and parents are suggested to guide students’ ICT use at home (Steinfeld, 2021). In this way, the positive relationship between ICT perceived competence and digital reading performance could be maximized. Overall, by exploring the relationship between ICT perceived competence and digital reading performance and the mediating role of ICT use, this study suggested that when striving to help adolescents perform better in digital reading, it is appropriate ICT use driven by a high-level perception of competence that matters.

Conclusions

Amid the post-COVID-19 era which witnesses more frequent ICT use and digital reading activities, this study investigated the relationship between ICT perceived competence and adolescents’ digital reading performance and the mediating effects of the frequency of three kinds of ICT use by conducting multilevel mediation analyses for 29 OECD countries/regions in the PISA 2018 computer-based reading assessment, encompassing 199,646 15-year-old students. The study revealed a significantly positive correlation between ICT perceived competence and adolescents’ digital reading performance for the vast majority of the countries/regions. Moreover, mediation did exist and a large proportion of the effect of ICT perceived competence on digital reading performance was mediated through the three types of ICT use, i.e., ICT use at home for leisure, ICT use at home for schoolwork, and ICT use at school. Specifically, one new finding is the “suppression effects” for a large proportion of the sample, suggesting that merely encouraging students to perceive themselves as competent in ICT tasks is not enough to improve digital reading performance; instead, it is appropriate ICT use driven by high ICT perceived competence that counts.

This study makes some contributions to the educational community especially the research field of digital reading. First, it adds to previous findings on the relationship between students’ ICT perceived competence and digital reading performance by providing comprehensive evidence utilizing updated data from a relatively large sample and revealing cross-cultural commonalities and disparities. Second, this study offers an in-depth exploration of the underlying mechanism of how the self-perception of ICT competence related to digital reading performance by examining the interrelation between the ICT perceived competence and ICT use. Third, the empirical findings of this study provide additional corroboration for existing theoretical models, including the SDT (Deci & Ryan, 2000) and the ICT engagement model (Goldhammer et al., 2016), which proposed a positive link between perceived competence and learning outcomes. Moreover, the findings build on SDT and delve deeper to link psychological and behavioral factors by showing that the effects of ICT perceived competence on digital reading performance are transmitted through ICT use in several settings and for various purposes. Last, based on large-scale quality data and multidimensional variables, the results might provide thorough insights to provoke efficient educational practices involving the use of ICT (e.g., Teo et al., 2019; Yang et al., 2022), especially in the post-COVID-19 era when ICT users and digital readers will be on the rise.

The limitations of this study are expected to inspire future studies. First, since this study focused on a particular psychological factor, ICT perceived competence, other factors at the cognitive-motivational level of the ICT engagement model, such as ICT interest and ICT perceived autonomy can be studied in the future. Second, although this study helps further the understanding of the underlying mechanism of the relationship between ICT perceived competence and digital reading performance, future researchers could search for other potential mediators, such as intrinsic motivation (Deci & Ryan, 2000), which could help build a more comprehensive model of the influential pattern supported by empirical findings.

Supplemental Material

Supplemental Material - The Relationship Between ICT Perceived Competence and Adolescents’ Digital Reading Performance: A Multilevel Mediation Study

Supplemental Material for The Relationship Between ICT Perceived Competence and Adolescents’ Digital Reading Performance: A Multilevel Mediation Study by Rushi Yu, Meishu Wang, and Jie Hu in Journal of Educational Computing Research

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Social Science Fund of China, “Construction and Research on the Multidimensional Evaluation of the Database of Chinese Second Language Students’ Reading Literacy” [grant number: 21BYY024].

Data Accessibility

The datasets generated and analysed during the current study are available in the Programme for International Student Assessment (PISA) repository which is publicly available,

ORCID iD

Jie Hu

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Rushi Yu is a Master’s student at the Department of Linguistics, the School of International Studies, Zhejiang University, Hangzhou 310058, China. Her research interests lie in e-learning, technology-assisted language learning and educational assessment.

Meishu Wang is a PhD student at the Department of Linguistics, the School of International Studies, Zhejiang University, Hangzhou 310058, China. Her research interests lie in educational data mining, technology-assisted language learning and educational technology.

Jie Hu is a professor working at the Department of Linguistics, the School of International Studies, Zhejiang University, Hangzhou 310058, China. She has been focusing on English language education for more than 10 years after her PhD graduation from the University of Warwick, U.K. Her research interests include ICT-based English language education, second language acquisition, educational data mining and learning analysis.

References

Ackerman

Goldsmith

(2011). Metacognitive regulation of text learning: On screen versus on paper. Journal of Experimental Psychology-Applied, 17(1), 18–32. https://doi.org/10.1037/a0022086

Aesaert

Voogt

Kuiper

van Braak

(2017). Accuracy and bias of ICT self-efficacy: An empirical study into students' over- and underestimation of their ICT competences. Computers in Human Behavior, 75, 92–102. https://doi.org/10.1016/j.chb.2017.05.010

Areepattamannil

Santos

I. M.

(2019). Adolescent students' perceived information and communication technology (ICT) competence and autonomy: Examining links to dispositions toward science in 42 countries. Computers in Human Behavior, 98, 50–58. https://doi.org/10.1016/j.chb.2019.04.005

Burkley

Curtis

Burkley

Hatvany

(2015). Goal fusion: The integration of goals within the self-concept. Self and Identity, 14(3), 348–368. https://doi.org/10.1080/15298868.2014.1000959

Cai

Z. H.

Fan

X. T.

J. X.

(2017). Gender and attitudes toward technology use: A meta-analysis. Computers & Education, 105, 1–13. https://doi.org/10.1016/j.compedu.2016.11.003

Chen

Zhang

(2021). Synergistic effects of instruction and affect factors on high- and low-ability disparities in elementary students’ reading literacy. Reading and Writing: An Interdisciplinary Journal, 34(1), 199–230. https://doi.org/10.1007/s11145-020-10070-0

Chen

(2020). ICT-related behavioral factors mediate the relationship between adolescents' ICT interest and their ICT self-efficacy: Evidence from 30 countries. Computers & Education, 159, Article 104004. https://doi.org/10.1016/j.compedu.2020.104004

Chiao

Chiu

C. H.

(2018). The mediating effect of ICT usage on the relationship between students' socioeconomic status and achievement. Asia-Pacific Education Researcher, 27(2), 109–121. https://doi.org/10.1007/s40299-018-0370-9

Christoph

Goldhammer

Zylka

Hartig

(2015). Adolescents' computer performance: The role of self-concept and motivational aspects. Computers & Education, 81, 1–12. https://doi.org/10.1016/j.compedu.2014.09.004

10.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Eribaum.

11.

Costa

Hammond

Younie

(2019). Theorising technology in education: An introduction. Technology, Pedagogy and Education, 28(4), 395–399. https://doi.org/10.1080/1475939X.2019.1660089

12.

Cusso-Calabuig

Farran

X. C.

Bosch-Capblanch

(2018). Effects of intensive use of computers in secondary school on gender differences in attitudes towards ICT: A systematic review. Education and Information Technologies, 23(5), 2111–2139. https://doi.org/10.1007/s10639-018-9706-6

13.

Deci

E. L.

Ryan

R. M.

(2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268. https://doi.org/10.1207/S15327965PLI1104_01

14.

Dempster

A. P.

Laird

N. M.

Rubin

D. B.

(1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society Series B-Methodological, 39(1), 1–38. https://doi.org/10.1111/j.2517-6161.1977.tb01600.x

15.

Deutsch

Herrmann

Frese

Sandholzer

(2012). Implementing computer-based assessment - a web-based mock examination changes attitudes. Computers & Education, 58(4), 1068–1075. https://doi.org/10.1016/j.compedu.2011.11.013

16.

Eger

Tomczyk

Klement

Pisonova

Petrova

(2020). How do first year University students Use ICT in their leisure time and for learning purposes? International Journal of Cognitive Research in Science Engineering and Education, 8(2), 35–52. https://doi.org/10.5937/ijcrsee2002035e

17.

Field

(2009). Discovering statistics using SPSS (3rd ed.). Sage publications.

18.

Franzen

Arens

A. K.

Greiff

Niepel

(2022). Student profiles of self-concept and interest in four domains: A latent transition analysis. Learning and Individual Differences, 95, Article, 102139. https://doi.org/10.1016/j.lindif.2022.102139

19.

Geng

S. N.

Y. Q.

Shu

(2022). Cross-cultural generalizability of expectancy-value theory in reading: A multilevel analysis across 80 societies. Current Psychology. Advance online publication. https://doi.org/10.1007/s12144-022-03014-0

20.

Goldhammer

Gniewosz

Zylka

(2016). ICT engagement in learning environments. In Kuger

Klieme

Jude

Kaplan

(Ed.), Assessing contexts of learning (pp. 331–351). Springer.

21.

Hammond

(2020). What is an ecological approach and how can it assist in understanding ICT take-up? British Journal of Educational Technology, 51(3), 853–866. https://doi.org/10.1111/bjet.12889

22.

Hatlevik

O. E.

Ottestad

Throndsen

(2015). Predictors of digital competence in 7th grade: A multilevel analysis. Journal of Computer Assisted Learning, 31(3), 220–231. https://doi.org/10.1111/jcal.12065

23.

Hayes

A. F.

(2009). Beyond baron and kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408–420. https://doi.org/10.1080/03637750903310360

24.

Holte

A. J.

Giesen

D. T.

Ferraro

F. R.

(2021) Color me calm: Grayscale phone setting reduces anxiety and problematic smartphone use. Current Psychology. Advance online publication. https://doi.org/10.1007/s12144-021-02020-y

25.

Hox

J. J.

(2010). Multilevel analysis: Techniques and applications (2nd ed.). Routledge.

26.

Hsu

C. C.

Chen

Y. L.

Lin

C. Y.

Lien

W. C.

(2022). Cognitive development, self-efficacy, and wearable technology use in a virtual reality language learning environment: A structural equation modeling analysis. Current Psychology, 41(3), 1618–1632. https://doi.org/10.1007/s12144-021-02252-y

27.

(2021). The effects of ICT-based social media on adolescents' digital reading performance: A longitudinal study of PISA 2009, PISA 2012, PISA 2015 and PISA 2018. Computers & Education, 175, Article 104342. https://doi.org/10.1016/j.compedu.2021.104342

28.

Gong

Lai

Leung

F. K. S.

(2018). The relationship between ICT and student literacy in mathematics, reading, and science across 44 countries: A multilevel analysis. Computers & Education, 125, 1–13. https://doi.org/10.1016/j.compedu.2018.05.021

29.

Huang

X. X.

Mayer

R. E.

(2019). Adding self-efficacy features to an online statistics lesson. Journal of Educational Computing Research, 57(4), 1003–1037. https://doi.org/10.1177/0735633118771085

30.

Ieong

M. K.

Cheung

K. C.

Sit

P. S.

Soh

K. C.

Mak

S. K.

(2016). Effects of economic, social and cultural status on mathematics performance: A multilevel mediation analysis of self-regulated learning processes. Contemporary Educational Research Quarterly, 24(4), 109–143. https://doi.org/10.6151/CERQ.2016.2404.05

31.

LeBreton

J. M.

Senter

J. L.

(2008). Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods, 11(4), 815–852. https://doi.org/10.1177/1094428106296642

32.

Lee

Y.-H.

J.-Y.

(2012). The effect of individual differences in the inner and outer states of ICT on engagement in online reading activities and PISA 2009 reading literacy: Exploring the relationship between the old and new reading literacy. Learning and Individual Differences, 22(3), 336–342. https://doi.org/10.1016/j.lindif.2012.01.007

33.

Lim

H. J.

Jung

(2019). Factors related to digital reading achievement: A multi-level analysis using international large scale data. Computers & Education, 133, 82–93. https://doi.org/10.1016/j.compedu.2019.01.007

34.

Liu

H. Q.

Van Damme

Gielen

Van Den Noortgate

(2015). School processes mediate school compositional effects: Model specification and estimation. British Educational Research Journal, 41(3), 423–447. https://doi.org/10.1002/berj.3147

35.

Liu

S. J.

Meng

L. Q.

(2010). Re-examining factor structure of the attitudinal items from TIMSS 2003 in cross-cultural study of mathematics self-concept. Educational Psychology, 30(6), 699–712. https://doi.org/10.1080/01443410.2010.501102

36.

MacKinnon

D. P.

Krull

J. L.

Lockwood

C. M.

(2000). Equivalence of the mediation, confounding and suppression effect. Prevention Science: The Official Journal of the Society for Prevention Research, 1(4), 173–181. https://doi.org/10.1023/a:1026595011371

37.

Mak

S.-k.

Cheung

K.-c.

Soh

Sit

P.-s.

Ieong

M.-k.

(2017). An examination of student- and across-level mediation mechanisms accounting for gender differences in reading performance: A multilevel analysis of reading engagement. Educational Psychology, 37(10), 1206–1221. https://doi.org/10.1080/01443410.2016.1242712

38.

Marsh

H. W.

Pekrun

Parker

P. D.

Murayama

Guo

Dicke

(2019). The murky distinction between self-concept and self-efficacy: Beware of lurking jingle-jangle fallacies. Journal of Educational Psychology, 111(2), 331–353. https://doi.org/10.1037/edu0000281

39.

Marshik

Ashton

P. T.

Algina

(2017). Teachers' and students' needs for autonomy, competence, and relatedness as predictors of students' achievement. Social Psychology of Education, 20(1), 39–67. https://doi.org/10.1007/s11218-016-9360-z

40.

Meng

Qiu

Boyd-Wilson

(2019). Measurement invariance of the ICT engagement construct and its association with students’ performance in China and Germany: Evidence from PISA 2015 data. British Journal of Educational Technology, 50(6), 3233–3251. https://doi.org/10.1111/bjet.12729

41.

Monique

Sylvie

V.-H.

Jasmin

(2013). The computer-related self concept: A gender-sensitive study. International Journal of Social and Organizational Dynamics in IT, 3(3), 1–16. https://doi.org/10.4018/ijsodit.2013070101

42.

Naumann

Saelzer

(2017). Digital reading proficiency in German 15-year olds: Evidence from PISA 2012. Zeitschrift Fur Erziehungswissenschaft, 20(4), 585–603. https://doi.org/10.1007/s11618-017-0758-y

43.

OECD . (2009). PISA data analysis manual: SPSS (2nd ed.). OECD Publishing. https://doi.org/10.1787/9789264056275-en

44.

OECD (2010), Are the new millennium learners making the grade?: Technology use and educational performance in PISA 2006. OECD Publishing. https://doi.org/10.1787/9789264076044-en.

45.

OECD . (2011). PISA 2009 results: Students on line: Digital technologies and performance (Vol. VI). OECD Publishing. https://doi.org/10.1787/9789264112995-en

46.

OECD . (2019a). PISA 2015 technical report. PISA: OECD Publishing. https://www.oecd.org/pisa/data/2015-technical-report/

47.

OECD . (2019b). PISA 2018 assessment and analytical framework. OECD Publishing. https://doi.org/10.1787/b25efab8-en

48.

OECD . (2019c). PISA 2018 results: What students know and can do (Vol. I). OECD Publishing. https://doi.org/10.1787/5f07c754-en

49.

Papanastasiou

E. C.

Zembylas

Vrasidas

(2003). Can computer use hurt science achievement? The USA results from PISA. Journal of Science Education and Technology Pedagogy and Education, 12(5), 325–332. https://doi.org/10.1023/A:1025093225753

50.

Petko

Cantieni

Prasse

(2017). Perceived quality of educational technology matters: A secondary analysis of students' ICT use, ICT-related attitudes, and PISA 2012 scores. Journal of Educational Computing Research, 54(8), 1070–1091. https://doi.org/10.1177/0735633116649373

51.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. https://doi.org/10.3758/brm.40.3.879

52.

Park

(2011). How motivational constructs interact to predict elementary students' reading performance: Examples from attitudes and self-concept in reading. Learning and Individual Differences, 21(4), 347–358. https://doi.org/10.1016/j.lindif.2011.02.009

53.

Preacher

K. J.

Selig

J. P.

(2012). Advantages of Monte Carlo confidence intervals for indirect effects. Communication Methods and Measures, 6(2), 77–98. https://doi.org/10.1080/19312458.2012.679848

54.

Qazi

Hasan

Abayomi-Alli

Hardaker

Scherer

Sarker

(2022). Gender differences in information and communication technology use & skills: A systematic review and meta-analysis. Education and Information Technologies, 27(3), 4225–4258. https://doi.org/10.1007/s10639-021-10775-x

55.

Raudenbush

S. W.

Bryk

A. S

Cheong

Y. F.

Congdon

(2019). HLM 8 for windows (Version 8.0). Scientific Software International, Inc. https://ssicentral.com/index.php/products/hlm-general/

56.

R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/

57.

Reimer

Smith

Andersen

I. G.

Sortkaer

(2021). What happens when schools shut down? Investigating inequality in students' reading behavior during covid-19 in Denmark. Research in Social Stratification and Mobility, 71, Article 100568. https://doi.org/10.1016/j.rssm.2020.100568

58.

Sage

Augustine

Shand

Bakner

Rayne

(2019). Reading from print, computer, and tablet: Equivalent learning in the digital age. Education and Information Technologies, 24(4), 2477–2502. https://doi.org/10.1007/s10639-019-09887-2

59.

Sage

Jackson

Fox

Mauer

(2021). The virtual COVID-19 classroom: Surveying outcomes, individual differences, and technology use in college students. Smart Learning Environments, 8(1), Article 27. https://doi.org/10.1186/s40561-021-00174-7

60.

Schutte

(2015). Science self-concept and valuing science: A cross-cultural analysis of their relation among students from western and East Asian countries. Social Psychology of Education, 18(4), 635–652. https://doi.org/10.1007/s11218-015-9311-0

61.

Senkbeil

(2022). ICT-related variables as predictors of ICT literacy beyond intelligence and prior achievement. Education and Information Technologies, 27(3), 3595–3622. https://doi.org/10.1007/s10639-021-10759-x

62.

Silva

P. G. D.

de Oliveira

C. A. L.

Borges

M. M. F.

Moreira

D. M.

Alencar

P. N. B.

Avelar

R. L.

(2021). Distance learning during social seclusion by COVID-19: Improving the quality of life of undergraduate dentistry students. European Journal of Dental Education, 25(1), 124–134. https://doi.org/10.1111/eje.12583

63.

Skryabin

Zhang

Liu

Zhang

(2015). How the ICT development level and usage influence student achievement in reading, mathematics, and science. Computers & Education, 85, 49–58. https://doi.org/10.1016/j.compedu.2015.02.004

64.

Srijamdee

Pholphirul

(2020). Does ICT familiarity always help promote educational outcomes? Empirical evidence from PISA-Thailand. Education and Information Technologies, 25(4), 2933–2970. https://doi.org/10.1007/s10639-019-10089-z

65.

Steinfeld

(2021). Parental mediation of adolescent Internet use: Combining strategies to promote awareness, autonomy and self-regulation in preparing youth for life on the web. Education and Information Technologies, 26(2), 1897–1920. https://doi.org/10.1007/s10639-020-10342-w

66.

Teo

Sang

G. Y.

Mei

Hoi

C. K. W.

(2019). Investigating pre-service teachers' acceptance of web 2.0 technologies in their future teaching: A Chinese perspective. Interactive Learning Environments, 27(4), 530–546. https://doi.org/10.1080/10494820.2018.1489290

67.

Tofighi

MacKinnon

D. P.

(2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43(3), 692-700. https://doi.org/10.3758/s13428-011-0076-x

68.

Tomte

Hatlevik

O. E.

(2011). Gender-differences in self-efficacy ICT related to various ICT-user profiles in Finland and Norway. How do self-efficacy, gender and ICT-user profiles relate to findings from PISA 2006. Computers & Education, 57(1), 1416–1424. https://doi.org/10.1016/j.compedu.2010.12.011

69.

Usher

E. L.

C. H. R.

Butz

A. R.

Rojas

J. P.

(2019). Perseverant grit and self-efficacy: Are both essential for children' academic success? Journal of Educational Psychology, 111(5), 877–902. https://doi.org/10.1037/edu0000324

70.

Veas

Castejon

J. L.

Minano

Gilar-Corbi

(2019). Relationship between parent involvement and academic achievement through metacognitive strategies: A multiple multilevel mediation analysis. British Journal of Educational Psychology, 89(2), 393–411. https://doi.org/10.1111/bjep.12245

71.

Vekiri

(2010). Socioeconomic differences in elementary students' ICT beliefs and out-of-school experiences. Computers & Education, 54(4), 941–950. https://doi.org/10.1016/j.compedu.2009.09.029

72.

Wen

Fan

(2015). Monotonicity of effect sizes: Questioning kappa-squared as mediation effect size measure. Psychological Methods, 20(2), 193–203. https://doi.org/10.1037/met0000029

73.

Wen

(2014). Analyses of mediating effects: The development of methods and models. Advances in Psychological Science, 22(5), 731–745. https://doi.org/10.3724/SP.J.1042.2014.00731

74.

J-Y.

(2014). Gender differences in online reading engagement, metacognitive strategies, navigation skills and reading literacy. Journal of Computer Assisted Learning, 30(3), 252–271. https://doi.org/10.1111/jcal.12054

75.

Xiao

Hew

K. F.

(2022). The relationships among ICT-related psychological factors, school contextual factors and secondary students' reading performance: A multilevel analysis across 47 Economies. Journal of Educational Computing Research, 60(5), 1166–1196. https://doi.org/10.1177/07356331211070975

76.

Xiao

(2019). The moderation examination of ICT use on the association between Chinese mainland students’ socioeconomic status and reading achievement. International Journal of Emerging Technologies in Learning, 14(15), 107–120. https://doi.org/10.3991/ijet.v14i15.10494

77.

Xiao

Liu

(2019). Regression analysis of ICT impact factors on early adolescents' reading proficiency in five high-performing countries. Frontiers in Psychology, 10, Article 1646. https://doi.org/10.3389/fpsyg.2019.01646

78.

Yang

Zhou

(2022). Students’ preferences for seating arrangements and their engagement in cooperative learning activities in college English blended learning classrooms in higher education. Higher Education Research & Development, 41(4), 1356–1371. https://doi.org/10.1080/07294360.2021.1901667

79.

(2022). A multilevel regression analysis of computer-mediated communication in synchronous and asynchronous contexts and digital reading achievement in Japanese students. Interactive Learning Environments. Advance online publication. https://doi.org/10.1080/10494820.2022.2066136

80.

Zhou

Yang

(2022). Mobile-assisted or paper-based? The influence of the reading medium on the reading comprehension of English as a foreign language. Computer Assisted Language Learning, 35(1–2), 217–245. https://doi.org/10.1080/09588221.2021.2012200

81.

Zhang

Zyphur

M. J.

Preacher

K. J.

(2009). Testing multilevel mediation using hierarchical linear models problems and solutions. Organizational Research Methods, 12(4), 695–719. https://doi.org/10.1177/1094428108327450

82.

Zhu

J-X.

Mok

M. M. C.

(2020). Predictors of students' participation in internet or computer tutoring for additional instruction and its effect on academic achievement. Journal of Computer Assisted Learning, 36(5), 729–740. https://doi.org/10.1111/jcal.12440

83.

Zylka

Christoph

Kroehne

Hartig

Goldhammer

(2015). Moving beyond cognitive elements of ICT literacy: First evidence on the structure of ICT engagement. Computers in Human Behavior, 53, 149–160. https://doi.org/10.1016/j.chb.2015.07.008

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.93 MB