Information and Communication Technologies in Peruvian University Students: A Confirmatory Analysis of Their Frequency and Extent of Use

Abstract

Pedagogical strategies with technology have an impact on university academic teaching. Hence, there is a need to develop competencies that allow the efficient use of various digital resources. The present study validated through a confirmatory factor analysis the constructs proposed in the CUTIC-28 in a sample of 318 Peruvian university students. It was a quantitative approach research at a descriptive level and based on a non-experimental design. The results demonstrated, as reflected in each metric of the confirmatory factor analysis (CFA), the theoretical and empirical sustainability of the original questionnaire to assess the frequency and extent of ICT use in Peruvian university students; the set of data reported offers the certainty that it is a defensible and sustainable factorial model. The covariances and correlations between the dimensions and subdimensions are highly significant and positive, and, therefore, the factorial structure is confirmed by the sample data. The confirmed scale has adequate properties that allow it to be considered a valid and reliable measure in future research, even after adding other variables, such as gender, age, and type of university, among the variables of interest that show significant differences. The results also show that there is still a knowledge gap to be covered.

1. Introduction

Around the globe, where the fourth industrial revolution is unfolding and affecting key aspects of people’s daily lives, information and communication technologies (ICT) have become the cornerstone of human development. At the international level, it is known that 58% of European Union citizens have basic digital skills, ranging from almost 80% in Finland to just over 30% in Italy. Similarly, the predominance of smartphones as a digital medium (99.5%) is noted in countries such as Spain, where there are more Internet users (93.2%) than computer users (81.4%) [1].

On the other hand, in Latin American households, there is little access to computers (35.5%), especially considering that 82% of European households have computers, potentially representing an obstacle in the acquisition of digital skills [2]. Similarly, an exponential growth has developed in Latin American countries in relation to access to cell phones, although with significant gaps [3].

In the Peruvian context, 70% of the population does not have internet access at home, highlighting that there are very marked gaps with respect to urban and rural areas [4]. According to the National Institute of Statistics and Informatics, [5], 95.4% of Peruvian households use at least one ICT. Specifically, 97.4% of Peruvian households have at least one member with a cell phone, 40.8% with at least one computer, and 79% with a radio and 84.6% with a television. Young university students are the ones who access the internet the most (95.9%). Smart cell phones without mobile data are the most used means (49.8%) by Peruvians, while communication (email, chat, etc.) and searching for information or entertainment (95%, 78.4%, and 82.3%, respectively) are the main activities performed by the population with internet access.

Ref. [6] states that pedagogical strategies with technology have an impact on university academic teaching; for this, it is proposed not only to generate students with mastery of internet access, but also to have an important domain in digital platforms and applications. It is crucial to note that, through this virtual environment, students must have a good cognitive and emotional state, the latter highlighting the motivation and desire for improvement in process. It is an important task not only to highlight the student role, but also the teaching role to educate students how to access tools depending on the characteristics of the students.

It is pivotal to mention that the ICT gaps in Peru are still notorious, especially since the beginning of the pandemic. Following this logic, [7] suggests that the increase in digitalization at a global level in the academic, work, and communication environments has generated changes in society. There are differences in the ease of internet use in rural areas, along with the need for a change in the proportion of telecommunication infrastructure to allow greater coverage.

Thus, the present study aimed to validate through a confirmatory factor analysis the constructs proposed in the CUTIC-28 in Peruvian university students. The constructs are composed of dimensions of frequency of ICT use, behavior/emotion generated by ICT, and perceived usefulness of ICT, by medium or support (computer/laptop or tablet versus cell phone).

There have been previous reviews, such as the review conducted by [8], which objective was to determine how future Spanish and Norwegian teachers perceive concepts such as privacy, cyberbullying, and digital content evaluation, and how the concepts under study relate to each other. The following conclusion can be made: in both countries, the three terms are distinct and there is a positive relationship between the perceived understandings of the concepts by prospective teachers, implying that these should be taught as separate components of professional digital competence.

On the contrary, we have [9] who conducted a study translated into English as “Types of ICT applications used and competency level of nursing students in higher education: a cross-sectional study”. The objective of this study was to explore the various types of ICT applications used and the level of knowledge of nursing students at a South African university. It concluded that the use of technology in nursing education is essential when training future nurses, especially in an IT-rich workplace. The author also highlighted the use of programs, such as MS Word, Ms PowerPoint, or Moodle.

In addition, [10] executed a scientific study which objective was to examine the relationship between ICT and students’ mathematical, reading, and scientific literacy using the latest PISA data. The authors concluded that most of the ICT-related factors significantly influence students’ learning outcomes, while some of them have negative influences. Therefore, on the one hand, it is necessary to integrate ICT in education as it can be a possible solution to some puzzling educational problems; on the other hand, ICT must be used with care in educational settings, and further measures to improve the quality of ICT integration in education are essential.

It is important to mention the work of [11] who published a scientific article, which objective was to analyze the impact of ICT use in school on the results of students in compulsory secondary education in mathematics, reading, and science, using data from three rounds of PISA (2009, 2012, and 2015) for Spanish regions (autonomous communities). The author concluded that an increase in the use of ICT at a school in an Autonomous Community does not have positive effects on PISA scores in mathematics and reading, while there is a positive effect on PISA scores in science.

Finally, it is pertinent to mention [12] who published a scientific article. Their objective was to analyze the factors affecting the perception of ICT policy implementation in education in the empirical context of higher education in Namibia from the students’ perspective. This study concluded that the implementation of ICT policy in higher education in Namibia is mainly affected by the lack of ICT literacy and limited access to learning and training content.

It can be mentioned that information and communication technologies (ICTs), which can be defined as technological devices (hardware or software) that allow the editing, production, storage, and exchange and transmission of data through different information systems, have common computer protocols [13]. Nowadays, ICTs are increasingly integrated in society; computers, laptops, and smartphones with internet access are necessary resources for the successful development of young population in various countries [14].

Therefore, there is a need to develop competencies that enable the efficient use of digital resources. Digital literacy, or ICT literacy, is a concept that encompasses an individual’s interest, attitude, and ability regarding the appropriate use of digital technology and communication tools that enable effective participation in society [15]. Although digital literacy is developed under a general perspective, it is often discussed that there are demographic variables, such as gender [16] and age [17], that condition the acquisition of digital skills.

Specifically, new technologies have notably changed several aspects of everyday life, modifying how people communicate, work, study, or spend their free time [18,19,20] and reaching a detrimental level on some occasions [21,22,23]. In the perspective of [24] the growth of ICT has changed how people interact with their social environment; people use the internet for a variety of reasons, including seeking information, mere entertainment, socialization, and expressing both opinions and ideas.

While there is a tendency to attribute negative characteristics to internet use (e.g., dependence, isolation, and deterioration of family relationships), young people view the development of computers and cyberspace with optimism [25]. Certainly, ICTs have the potential to positively influence students’ lives; for example, it is said that social networks can serve to foster interpersonal relationships with peers [26], facilitating communication and interaction.

Generally, ICTs are developed in two ways in the life of university students. On the one hand, they serve as tools for study and even work, while, on the other hand, they are very useful for leisure, especially as a medium that allows precise and fast communication [27] Young people are said to allocate large amounts of time to social networks [28], as these websites allow users to create profiles and view or create content that can be shared with a large number of internet users [29]. Likewise, ICTs have caused and promoted, especially in young population, a growing interest in video games, which, although they are usually considered entertainment, have been acquiring competitive nuances [30].

Ref. [13] mentions that the integration of ICT has generated a substantial change in education. Therefore, such technologies have become a topic of interest for research educators, as they can improve the effectiveness and productivity of teaching [31]. At the same time, they could produce adverse effects if they are not integrated in the right way [10].

The involvement of ICT in education has resulted in the modification of learning processes, giving rise to concepts such as e-education, mobile learning, game-based learning, and ubiquitous learning [32]. Similarly, ICTs have become commonly used tools in the learning of students and teachers at the higher education level [33], highlighting notably how important teachers’ knowledge is when implementing such technologies [34].

Currently, ICTs are developed in a multidimensional way in education, involving multiple uses in a classroom [35]. That is, they can serve as a tool for acquiring information or a platform for communication and teaching. Ref. [36] added that the involvement of ICTs has brought about an advance in teaching. An example of this is the “flipped classroom” educational methodology, which is based on learning developed by the student themselves, who through their own means (e.g., videos, audios, and books) self-instructs at home, while the educator has a secondary role that involves the reinforcement and discussion of the knowledge acquired at home [37].

It is appropriate to emphasize that, in order to effectively take advantage of the virtues of ICTs in education, it is necessary to foster information literacy among students. This concept includes the ability to recognize what type of information is sought, how to access it, as well as the ability to evaluate the data and its sources and, subsequently, to classify, manipulate, and store the selected information [38].

Currently, the use of social networks has become widespread, growing rapidly around the world [39]. The most popular networks are Facebook, YouTube, Instagram, Twitter, and LinkedIn [40]. Added to these is the recent TikTok, which has about 1 billion users and is available in 150 countries [41]. University students are said to use social networks, such as Facebook, to maintain constant communication with close ones, while building and sustaining new social relationships, i.e., their use is motivated by rewards of a social nature [42].

Ref. [29], integrating the ideas of other authors, mentioned that the popularity of social networks in the young population has caused some concern in various authors, especially for the so-called “Facebook depression”, which is conceptualized as a phenomenon where a large exposure to social networks generates depressive symptoms in minors. On the other hand, ref. [28] provided a positive perspective by mentioning that children and young people who spend much of their time on social networks express their personality, in addition to the development of a sense of identity.

In general, it can be said that the effects of the Internet on people are strongly linked to the specific environment where it occurs and to both the characteristics and specific personal conditions of each individual, which are equally comprised of their own psychological state and their own perception of well-being [43]. Likewise, personality has a significant influence on the sensations that a stimulus can arouse in an individual; in the same way, it is said that personality serves as a predictor, directly or indirectly, of how people use ICTs [44].

The impact that ICTs have had on various areas of daily life is evident, even being linked to the mental health of their users. In recent decades, the proliferation of technologies and the Internet has caused the emergence of both benefits and psychological problems linked to online activities [45]. For example, social networks can have a significant impact on the behavior of their users; they can serve to strengthen social ties that improve self-esteem, but they have also been linked to poor academic performance [46] and even to a depressive disorder.

The prolonged confinement imposed by the COVID-19 pandemic closures resulted in the imposition of online education for two years. It required a high level of adaptation by most students. Although there was a significant preference for face-to-face approach, yet there was a high level of adaptation by students in reported performance; this highlights the difference between students’ learning styles and the influence of their personality profiles [47].

The COVID-19 pandemic has caused a notable increase in the use of ICTs by university teachers. In general terms, it can be seen that the digital tools most used during the pandemic period have become the most used after the pandemic, with notable increases in the frequency of use of digital resources for tutorials and discussions with students. On the other hand, the area of knowledge is an influential variable in terms of the impact of the pandemic on ICT usage habits in the Latin American and Caribbean regions, as there is a greater use of ICTs by social sciences teachers compared to other areas of knowledge [48].

In the Latin American and Caribbean areas, university teachers feel that their digital competence and their ability to adapt to the digital educational environments are intermediate-high. However, they feel that the available technical resources do not allow them to develop easily. It is noted that the level of stress caused by the digital educational environments is intermediate, but it is not conditioned by levels of digital skills or by the global innovation index of the countries of origin; thus, it is thought that there must be socio-demographic, cultural, or political factors that affect this type of stress. Finally, the influence of the demographic characteristics of teachers is highlighted, with females showing greater digital competence and lower digital pandemic stress, while older teachers showing lower levels of digital competence and lower levels of stress; with regard to the area of knowledge, it can be seen that those most affected by digital pandemic stress are teachers in the areas of health and humanities [49].

Naturally, the transition from traditional to digital teaching has increased the use of ICTs, providing new educational opportunities even with the limitations caused by an unforeseen situation. Among these limitations are how the instructor serves as a guide and a co-producer of the teaching process, the need to analyze educational processes leading to a planned integration of ICTs, the adequacy of the evaluation process to assess the digital environment, and good communication that addresses the gap caused by the absence of a physical classroom [50].

2. Materials and Methods

This is a quantitative, descriptive, and non-experimental design research. The sample was constituted of university students of the Metropolitan Lima in Peru, which total (as of 2021) +380,000 undergraduate students from the 31 licensed universities of the capital of Peru. The sample was recruited using non-probabilistic convenience sampling, with 319 students.

The technique used was a survey, and the instrument was a questionnaire validated by [51] to measure the frequency and scope of ICT use, which he called the CUTIC-28, a valid and reliable instrument to measure the frequency and extent of ICT use in different contexts in young university students in Madrid, Spain.

This questionnaire had been studied for the adequacy of its items based on a unidimensional exploratory factorial structure, discarding those items with weak factorial weights (less than 0.30). As a result, the questionnaire was reduced to 28 items on a 5-choice ordinal scale (from “never” to “always”), with a good internal consistency reliability as evident by an alpha coefficient of 0.86 [51]. At the end, the items were grouped into three dimensions, and each dimension, in turn, included two specific dimensions that matched the support used (computer/laptop or tablet versus cell phone). These dimensions included frequency of ICT use, behavior/emotion generated by ICT, and perceived usefulness of ICT.

The data analytical method of the present study involved a confirmatory factor analysis to test the theoretical validity of the constructs of the aforementioned exploratory study.

3. Results

The first step was the detection and elimination of outlier responses. In order to avoid bias in the parameter estimates of the structural model (AFC), the first step was to detect surveys with atypical responses and then remove them from the analysis.

In this case, only one survey was found through an anomaly indicator calculated on the basis of the 28 items of the questionnaire. Figure 1 shows the distribution of the anomaly index in contrast to the normal distribution, and the anomaly indicator of the outlier survey is far away from that of the right tail.

Table 1. CUTIC-28 Model.

Dimension 1 (F1): Frequency of ICT Use for Gaming, Messaging, and Social Networking (RRSS).
FCLT1	Dimension 1.A. Frequency of ICT use for gaming, messaging, and social networking on computer, laptop, or tablet.
FCLT2
FCLT3
FCLT4
FCLT5
FCLT6
FCLT7
FC1	Dimension 1.B. Frequency of ICT use for gaming, messaging, and social networking on cell phones.
FC2
FC3
FC4
FC5
FC6
FC7
Dimension 2 (F2): ICT-Generated Behavior/Emotion
CECLT1	Dimension 2.A Behavior/emotion generated by ICT on computer, laptop, or tablet.
CECLT2
CECLT3
CEC1	Dimension 2.B Behavior/emotion generated by ICT on cell phones
CEC2
CEC3
Dimension 3 (F3): Utility of ICTs in the Educational Environment
UCLT1	Dimension 3.A. Utility of ICTs in the educational environment on computer, laptop, or tablet.
UCLT2
UCLT3
UCLT4
UC1	Dimension 3.B. Utility of ICTs in the educational environment on cell phones
UC2
UC3
UC4

Source: questionnaire validated by [51].

shows the CUTIC-28 model proposed by [51], including three major dimensions or factors and two sub-dimensions for each factor; this is due to the component of the medium used: computer, laptop, or tablet versus cell phone.

The first column reflects the coding assigned to the items that make up the instrument, expressed as follows:

• FCLT = Frequency of use on computer, laptop, or tablet.
• FC = Frequency of cell phone use.
• BECLT = Behavior/emotion on computer, laptop, or Tablet.
• BC = Behavior/emotion on cell phone.
• UCLT = Utilization in the educational environment on computer, laptop, or tablet.
• UC = Utilization in the educational environment on cell phone.

The internal consistency of the 28 items, calculated as Cronbach’s alpha, indicates that all items are positively correlated, which suggests a unidimensionality of the questionnaire. At this moment, the question is to find out whether this unidimensionality underlies the dimensions representing the study variables (three study variables).

Table 2. Overall reliability—28 items.

Alfa de Cronbach	Number of Elements
0.863	28

Source: data collected and IBM SPSS.

shows the cronbach’s alpha for the total number of questions in the questionnaire, reporting a score of 0.863.

Table 3. FCLT scale—7 items.

Alfa de Cronbach	Number of Elements
0.739	7

Source: data collected and IBM SPSS.

shows the cronbach’s alpha (0.739) for dimension 1.A. (see Table 1) with 7 questions.

Table 4. FC scale—7 items.

Alfa de Cronbach	Number of Elements
0.730	7

Source: data collected and IBM SPSS.

shows the cronbach’s alpha (0.730) for the 7 questions of dimension 1.B.

Table 5. CECLT scale—3 items.

Alfa de Cronbach	Number of Elements
0.653	3

Source: data collected and IBM SPSS.

shows the value of cronbach’s alpha (0.653) for the 3 questions of dimension 2. A.

Table 6. CEC scale—3 items.

Alfa de Cronbach	Number of Elements
0.737	3

Source: data collected and IBM SPSS.

shows the value of cronbach’s alpha (0.737) for the 3 questions of dimension 2. B.

Table 7. UCLT scale—4 items.

Alfa de Cronbach	Number of Elements
0.895	4

Source: data collected and IBM SPSS.

and

Table 8. UC scale—4 items.

Alfa de Cronbach	Number of Elements
0.931	4

Source: data collected and IBM SPSS.

present the cronbach’s alpha values (0.895 and 0.931, respectively) for each of the 4 questions of dimensions 3.A. and 3.B. (see Table 1).

The fact that the Cronbach’s alphas of the subdimensions of the study variables are within the expected range (α > 0.65) and verify the unidimensionality means that the sample data provide evidence that the variables are positively correlated.

Figure 2 and Figure 3 represent the initial and final confirmatory models, which provide the metrics and indices that are considered appropriate for model fit, testing and empirical sustainability.

To monitor the fit of the data to the AFC model, the fit indices highlighted in

Table 9. Indicators of fit of the structural model to determine the frequency and amplitude of use in university students.

Indicator	Favorable Reference Values	Initial Values	Final Values
Ratio Chi-square/gl or CMIN/DF	<2	4.45	1.88
Goodness of fit index (GFI)	≥0.90	0.75	0.88
Residual square root (RMR)	≤0.05	0.08	0.05
Adjusted Goodness-of-Fit Index (AGFI)	>0.90	0.70	0.85
Parsimony Goodness-of-Fit Index (PGFI)	0.50–0.70	0.63	0.70
Normalized fit index (NFI)	≥0.90	0.70	0.88
Comparative Fit Index (CFI)	≥0.90	0.75	0.94
Tucker–Lewis index (TLI)	≥0.90	0.72	0.93
Root mean square error of approximation (RMSEA)	0.05–0.08	0.10	0.05

Source: Benchmark favorable values taken from [52,53].

are considered. The most important indices are CMIN/DF < 2, TLI and CFI > 0.9, and RMSEA < 0.08.

A model has a tolerable fit if the CMIN/DF values are less than two, even being accepted in some cases with limits up to five. The Tucker–Lewis Index (TLI) ranges from 0 to 1, with values greater than or equal to 0.90 being recommended [54]. The CFI (Comparative Fit Index) was developed by [55] from a previous index (BFI) that corrects for values beyond the 0–1 range. The CFI compares the χ² of two models: an independent model that claims there is no relationship between the variables in the model, and the model proposed by the researcher. Usually, it is accepted that the CFI should be around 0.90 to consider the model as fitting the data appropriately. The Root Mean Square Error of Approximation (RMSEA) represents the anticipated fit to the total population value and not to the sample. If RMSEA is less than or equal to 0.05, it indicates an error of approximation of the model with reality.

After verifying the good fit of the model, we proceeded to verify whether the sample is able to confirm the latent factors (dimensions and sub-dimensions) of the model.

Table 10. Estimated parameters of the structural model to determine the frequency and extent of use in university students.

Subdimensions and Items	Dimensions		Estimate	S.E.	C.R.	P	Standardized Regression Weights
FCLT		F1	1				0.731
FC		F1	0.629	0.156	4.026	***	0.692
UC		F3	1				0.741
UCLT		F3	0.683	0.122	5.611	***	0.801
CECLT		F2	1				0.927
CEC		F2	1.066	0.093	11.446	***	0.935
FCLT7		FCLT	1				0.482
FCLT6		FCLT	1.997	0.248	8.055	***	0.655
FCLT5		FCLT	1.797	0.222	8.084	***	0.695
FCLT4		FCLT	1.567	0.197	7.939	***	0.714
FCLT3		FCLT	0.915	0.158	5.774	***	0.372
FCLT2		FCLT	0.974	0.138	7.037	***	0.565
FCLT1		FCLT	0.627	0.129	4.845	***	0.299
UCLT1		UCLT	1				0.7
UCLT2		UCLT	1.096	0.07	15.66	***	0.742
UCLT3		UCLT	1.21	0.082	14.735	***	0.899
UCLT4		UCLT	1.174	0.079	14.87	***	0.915
CECLT3		CECLT	1				0.884
CECLT2		CECLT	0.556	0.067	8.267	***	0.465
CECLT1		CECLT	0.603	0.071	8.549	***	0.49
FC1		FC	1				0.31
FC2		FC	1.753	0.305	5.757	***	0.711
FC3		FC	1.324	0.297	4.463	***	0.326
FC4		FC	2.446	0.412	5.938	***	0.847
FC5		FC	2.413	0.426	5.669	***	0.626
FC6		FC	2.977	0.519	5.739	***	0.628
FC7		FC	1.485	0.3	4.949	***	0.431
UC1		UC	1				0.886
UC2		UC	1.037	0.041	25.521	***	0.946
UC3		UC	0.923	0.043	21.711	***	0.865
UC4		UC	0.935	0.051	18.216	***	0.791
CEC3		CEC	1				0.847
CEC2		CEC	0.635	0.064	9.943	***	0.559
CEC1		CEC	0.744	0.07	10.579	***	0.602

*** Values are highly significant at 0.01. Source: Data collected and IBM SPSS.

shows that all latent factors of the model are confirmed significantly at the 5% level.

The standardized regression weights represent the factor loadings. The first six rows measure the contribution of the dimensions to their sub-dimensions, and the subsequent rows represent the contribution of the dimensions to the items. The regression weights are greater than 0.50 (except for eight items, which do not condition the result), reaching more than 0.80; this shows the robustness of the model.

As shown in

Table 11. Estimated parameters of covariance and correlations of the factorial model.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.112	0.026	4.344	***	0.571
F3	<-->	F2	0.074	0.036	2.043	0.041	0.156
F1	<-->	F2	0.099	0.025	3.88	***	0.388

*** Values are highly significant at 0.01. Source: Data collected and IBM SPSS.

, the study variables are positively and significantly correlated at the 5% level, indicating that there is sufficient evidence in the data to support the factor structure of the measurement instrument. The covariances between the constructs are highly significant and positive, as expected. Therefore, the structure is confirmed by the data.

The main objective of this study was to validate and confirm the factorial structure of the constructs proposed in the CUTIC-28, which had already been studied in another context. This study was able to verify the factorial structure in a sample of Peruvian university students, affirming coherence and consistency with the study of [51] and confirming the initially proposed 28 items. These items are related to the dimensions of frequency of ICT use, behavior/emotion generated by ICT, and perceived usefulness of ICT, by medium or support (computer/laptop or tablet versus cell phone).

Group Analysis

Table 12. Gender of respondents.

Gender	Number	%
Male	132	41.5%
Female	186	58.5%
Total	318	100%

Source: Data collected and IBM SPSS.

,

Table 13. University type.

Type	Number	%
Public university	70	22%
Private university	248	78%
Total	318	100%

Source: Data collected and IBM SPSS.

and

Table 14. Age of respondents.

Age Range	Number	%
From 18 to 21 years old	146	45.9%
From 22 to 25 years old	82	25.8%
From 26 to 29 years old	25	7.9%
From 30 to 33 years old	21	6.6%
From 34 to 37 years old	17	5.3%
From 38 to over	27	8.5%
Total	318	100%

Source: Data collected and IBM SPSS.

show the sample size of the control variables. This inspection of the sample size is necessary for the confidence of the model parameter estimates. It should be noted that, for samples that are not so large, the fit deteriorates with respect to the fit of the model with the whole sample. In the case of age ranges, where small samples (less than 30) were observed, contiguous ranges were regrouped to achieve greater sample representativeness. For the analysis, this was performed to form the groups from 18 to 21 and from 22 to higher ages.

In the female group, as previously mentioned, the measures of fit deteriorate, but within expectations, which allows us to affirm that a good fit is maintained (CMIN/DF < 2, TLI and CFI > 0.9, and RMSEA < 0.08). In the male group, the measures of fit also deteriorate, but within expectations (CMIN/DF < 2, TLI and CFI very close to 0.9, and RMSEA < 0.08); this can be seen in

Table 15. Adjustment indicators of the structural model to determine the frequency and extent of use in university students: female and male groups.

Indicator	Favorable Reference Values	Female Group	Male Group
Ratio Chi-square/gl or CMIN/DF	<2	1.57	1.67
Goodness of fit index (GFI)	≥0.90	0.82	0.78
Residual square root (RMR)	≤0.05	0.08	0.08
Adjusted Goodness of Fit Index (AGFI)	>0.90	0.78	0.74
Parsimony Goodness-of-Fit Index (PGFI)	0.50–0.70	0.67	0.64
Normalized fit index (NFI)	≥0.90	0.81	0.79
Comparative Fit Index (CFI)	≥0.90	0.92	0.90
Tucker–Lewis index (TLI)	≥0.90	0.91	0.89
Error cuadrático medio de aproximación (RMSEA)	0.05–0.08	0.04	0.05

Source: Benchmark favorable values taken from [52,53].

.

In both groups (male and female), it can be seen that all latent factors of the model are confirmed significantly at a level of 5%, and the standardized regression weights represent the factor loadings. The first six rows measure the contribution of the dimensions to their subdimensions at the group level by sex, and the subsequent rows represent the contribution of the dimensions to the items. The regression weights are mostly greater than 0.50 (with the exception of some items, which do not condition the result) and even reach higher than 0.80, which highlights the robustness of the model conditional on the effect or separation of group by sex.

Table 16. Estimated parameters of covariances and correlations of the factorial model: Female group.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.063	0.025	2.47	0.01	0.631
F3	<-->	F2	0.023	0.034	0.674	0.5	0.063
F1	<-->	F2	0.023	0.02	1.311	0.19	0.163

Source: Data collected and IBM SPSS.

shows that in the group of women, a positive and significant correlation at 5% was found between the frequency of ICT use for gaming, messaging, and social networking (RRSS) and the use of ICTs in the educational environment.

Table 17. Estimated parameters of covariances and correlations of the factorial model: Male group.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.126	0.041	3.065	0.002	0.47
F3	<-->	F2	0.211	0.069	3.041	0.002	0.379
F1	<-->	F2	0.176	0.049	3.584	***	0.556

*** Values are highly significant at 0.01. Source: Data collected and IBM SPSS.

shows that in the group of men, positive and significant correlations at 5% were verified between all dimensions and subdimensions. Differences in the confirmatory model by sex were observed.

Table 18. Indicators of structural model fit to determine frequency and extent of use in university students: public university and private university groups.

Indicator	Favorable Reference Values	Public University Group	Private University Group
Ratio Chi-square/gl or CMIN/DF	<2	1.82	1.67
Goodness of fit index (GFI)	≥0.90	0.78	0.83
Residual square root (RMR)	≤0.05	0.09	0.08
Adjusted Goodness of Fit Index (AGFI)	>0.90	0.73	0.80
Parsimony Goodness-of-Fit Index (PGFI)	0.50–0.70	0.64	0.68
Normalized fit index (NFI)	≥0.90	0.76	0.83
Comparative Fit Index (CFI)	≥0.90	0.87	0.92
Tucker-Lewis Index (TLI)	≥0.90	0.86	0.91
Root mean square error of approximation (RMSEA)	0.05–0.08	0.06	0.04

Source: Benchmark favorable values taken from [52,53].

shows that the group, Public University, was represented by a sample size of 70, so it was expected that the fit measures would be more deteriorated, as well as the significance of the latent structures, but the fit would still be good. It could even be affirmed that, if the sample were larger, the fit would improve, as well as the significance of the latent factors (CMIN/DF < 2, TLI and CFI close to 0.9, and MSEA < 0.08). On the other hand, in the group, Private University, the fit measures deteriorated, but within what was expected; thus, it can be affirmed that a good fit is maintained (CMIN/DF < 2, TLI and CFI > 0.9, and RMSEA < 0.08).

In both groups (public and private universities), it can be seen that all latent factors of the model are confirmed significantly at a level of 5% (with the exception of the model for the public university in the first dimension), and the standardized regression weights represent the factor loadings. The first six rows measure the contribution of the dimensions to their subdimensions at the group level by university type, and the subsequent rows represent the contribution of the dimensions to the items. The regression weights are mostly greater than 0.50 (with the exception of some items, which do not condition the result) and even reach higher than 0.80, which highlights the robustness of the model conditional on the effect or separation of group by university type.

Table 19. Estimated covariance and correlation parameters of the factor model: Public University Group.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.053	0.024	2.175	0.03	0.607
F3	<-->	F2	0.026	0.052	0.495	0.62	0.056
F1	<-->	F2	0.028	0.023	1.224	0.22	0.199

Source: Data collected and IBM SPSS.

shows that in this group of public university students, a positive and significant correlation at 5% was verified between the frequency of ICT use for gaming, messaging, and social networking (RRSS) and the use of ICTs in the educational environment.

Table 20. Estimated covariance and correlation parameters of the factor model: Private University Group.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.055	0.018	3.000	0.003	0.588
F3	<-->	F2	0.099	0.045	2.217	0.027	0.203
F1	<-->	F2	0.062	0.021	3.039	0.002	0.482

Source: Data collected and IBM SPSS.

shows that in this group of students from private universities, positive and significant correlations at 5% were verified between all dimensions and subdimensions. Differences in the confirmatory model by type of university were observed.

For the group aged 18 to 21 years, as previously mentioned, the fit measures deteriorated, but within expectations, allowing us to state that a good fit is maintained (CMIN/DF < 2, TLI and CFI very close to 0.9, and RMSEA < 0.08), while in the group aged 22 years and older, a good fit is maintained and even improved (CMIN/DF < 2, TLI and CFI > 0.9, and RMSEA < 0.08); as can be seen in

Table 21. Indicators of structural model fit to determine frequency and extent of use in college students: Age group.

Indicator	Favorable Reference Values	Group from 18 to 21 Years Old	Age Group 22 and Over
Ratio Chi-square/gl or CMIN/DF	<2	1.681	1.576
Goodness of fit index (GFI)	≥0.90	0.78	0.81
Residual square root (RMR)	≤0.05	0.08	0.08
Adjusted Goodness of Fit Index (AGFI)	>0.90	0.74	0.77
Parsimony Goodness-of-Fit Index (PGFI)	0.50–0.70	0.64	0.67
Normalized fit index (NFI)	≥0.90	0.78	0.82
Comparative Fit Index (CFI)	≥0.90	0.89	0.92
Tucker–Lewis Index (TLI)	≥0.90	0.88	0.91
Root mean square error of approximation (RMSEA)	0.05–0.08	0.05	0.04

Source: Benchmark favorable values taken from [52,53].

.

In both groups (18 to 21 years and 22 years and older), it can be seen that all latent factors of the model are confirmed significantly at a level of 5% (except for the model for the 18 to 21 years group in the first dimension), and the standardized regression weights represent the factor loadings. The first six rows measure the contribution of the dimensions to their subdimensions at the age group level, and the subsequent rows represent the contribution of the dimensions to the items. The regression weights are mostly greater than 0.50 (with the exception of some items, which do not condition the result) and even reach higher than 0.80, which highlights the robustness of the model conditional on the effect or age-group separation.

Table 22. Estimated covariance and correlation parameters of the factorial model: Age group 18 to 21 years old.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.032	0.019	1.737	0.08	0.415
F3	<-->	F2	0.068	0.040	1.706	0.08	0.196
F1	<-->	F2	0.026	0.015	1.676	0.09	0.313

Source: Data collected and IBM SPSS.

shows that in the group of students between 18 and 21 years of age, no significant correlations at 5% were found among all dimensions. No relevance or robustness of the factorial model was observed for this age group.

Table 23. Estimated covariance and correlation parameters of the factorial model: Age group 22 years and older.

Dimensions			Covariances (Estimates)	S.E.	C.R.	P	Correlations
F1	<-->	F3	0.075	0.026	2.918	0.004	0.672
F3	<-->	F2	0.088	0.057	1.560	0.119	0.167
F1	<-->	F2	0.088	0.031	2.867	0.004	0.477

Source: Data collected and IBM SPSS.

shows that in this group of students aged 22 years and older, positive and significant correlations at 5% were verified between the frequency of ICT use for gaming, messaging, and social networking (RRSS) and the use of ICTs in the educational environment, and the frequency of ICT use for gaming, messaging, and social networking (RRSS) and the behavior/emotion generated by ICTs. Differences were then observed in the confirmatory model by age group.

4. Discussion

The study sought to validate by means of a confirmatory factor analysis the constructs proposed in the CUTIC-28, an instrument proposed by [51] to measure the frequency and scope of use of ICTs with internet connection. The referred study had a sample of 178 university students studying Social Work at the Complutense University of Madrid, and an exploratory factor analysis was applied, reducing the number of items from 48 items to 28 items that showed better properties; it was concluded that the instrument was a valid and reliable tool to estimate the frequency of use of technology. However, this instrument did not have a confirmation of its theoretical and methodological proposal, so it was imperative to demonstrate the interrelationships between its dimensions and sub-dimensions, and in a new context. In this study, Peruvian university students, of all careers, were recruited for a final sample of 318. Similar to the previous study, the robustness of the 28 proposed items and their dimensions and subdimensions (frequency of ICT use, behavior/emotion generated by ICT, and perceived usefulness of ICT, by medium or support in terms of computer/laptop or tablet versus cell phone) were confirmed, affirming coherence and consistency between the studies with the use of a confirmatory analysis; this implies the finding of consistency between the estimation methods (see Table 9) and the submitted questionnaire. Thus, the proposed theoretical model reproduces the relationships between variables (dimensions and sub-dimensions), representing, in a new context, the relationships between the full composition of the explored construct, but now with the highest possible fidelity.

The confirmation of the model opens the possibility of verifying its interrelationships with other studies. Specifically, for the dimension of ICT use in the educational setting, similarity is found with the findings of [9] who, when exploring the various types of ICT applications used and the level of knowledge of nursing students at a South African university, concluded that the use of technology in nursing education is essential when training future nurses, especially in an IT-rich workplace. This is an aspect that is confirmed in this study, even with the subdivisions of university type, gender, and age that present their due particularities, as detailed in the results section.

Similarity is also found with the results of Hu et al. (2018), who examined the relationship between ICT and students’ mathematical, reading, and scientific literacy and concluded that most of the factors related to ICT significantly influence students’ learning outcomes, with the recommendation that ICT should be employed with care in educational environments and that it is essential to adopt new measures to improve the quality of ICT integration in education. The latter echoes with what has been referred to by [7], who indicates that there are differences in the ease of internet use in rural areas, with the need for a change in the proportion of telecommunication infrastructure to allow greater coverage. This is also the case in the Peruvian context, highlighting the notorious ICT gaps, which have been especially revealed since the beginning of the COVID-19 pandemic.

The above findings also correspond with [12], who analyzed the factors affecting the perception of ICT policy implementation in education in the empirical context of higher education in Namibia and concluded that the implementation of ICT policy in higher education in Namibia is mainly affected by the lack of ICT literacy and limited access to learning and training content, a factor that should be mapped very carefully and improvement actions should be taken in the studied context.

The findings about the frequency and extent of use of ICTs in this research correspond with what have been reported by [18,19,20], highlighting that ICTs play a very notable role in how people communicate, work, study or spend their free time. These findings clearly reflect how people interact with their social environment, although it is also important to assess the behavior and emotions generated by the use of ICTs, as this could create dependence, isolation, deterioration of family relationships, etc. Even so, the benefits are greater since, undoubtedly, ICTs have the potential to positively influence students’ lives. As reported in [26], social networks can serve to foster interpersonal relationships with peers, providing important communication and interaction.

From the analysis of the fit indicators of the confirmatory model developed to determine the frequency and extent of ICT use in university students, it is observed that most of them are met (six out of a total of nine, see Table 9); however, some oscillation is observed in some indicators, This is in line with [56], who reveals that, overall, the adjustment indexes are not particularly stable, due to the conditioning factors that affect the sample size; thus, he proposes to take into account mainly the RMSEA which has a more stable quality, and, in the present model, it reaches this by being at 0.05.

5. Conclusions

In conclusion, the results, as shown in each metric of the confirmatory factor analysis approve and corroborate the empirical sustainability of a questionnaire to evaluate the frequency and extent of ICT use, the CUTIC-28, in Peruvian university students; the set of data reported offers the certainty that it is a defensible and sustainable factorial model. The covariances and correlations between the dimensions and subdimensions are highly significant and positive and, therefore, the factorial structure is confirmed by the sample data. This scale has adequate properties that allow it to be considered a valid and reliable measure in future research, even after adding other variables, such as sex, age, and type of university, among the variables of interest that show significant differences. The results also show that there is still a knowledge gap to be covered.

It is significant to mention that the global dimensional structure composed of the frequency of ICT use, the behavior/emotion generated by ICT, and the perceived usefulness of ICT, by medium or support (computer/laptop or tablet versus cell phone) is confirmed, so it is recommended to universities to develop action plans to understand how their students use ICTs and with what frequency and media, and how active they are for the development of educational or academic activities. This will also undoubtedly serve to communicate with them in the best way, appealing to behavioral and emotional factors and achieving greater engagement. In addition, based on the particular characteristics of the groups studied here, there are differences between men and women, differences by ages (in the younger, the interrelations are not observed), and differences by types of university. These characteristics need to be taken into account to make better decisions, not only about the use of ICTs, but also for the development of outreach strategies.