Problematic Internet Use in adolescents: New psychometric evidence for the Spanish short form of the Compulsive Internet Use Scale

Abstract Objectives Problematic Internet Use (PIU) has become a worldwide problem in recent years. Among screening instruments for PIU, the Compulsive Internet Use Scale (CIUS) is perhaps the most widely used. Psychometric properties of the full CIUS are not convincing, however, and the short form (CIUS‐S) has shown promising results, albeit limited to the English version, with little evidence in Spanish. Therefore, the aim of the present work was to study the psychometric properties of the CIUS‐S scores in a large sample of Spanish adolescents. Method The sample consisted of 1790 participants, 816 male (45.6%), 961 female (53.7%), and 13 other (0.7%) . Mean age was 15.70 years (SD = 1.26). Results The five‐items one‐dimensional model displayed appropriate goodness‐of‐fit indices. Strong measurement invariance model across age and partial across gender was found. Furthermore, the CIUS‐S was positively associated with several indicators of poor well‐being and mental health, and negatively associated with prosocial behavior, self‐esteem, and feeling of belonging. Conclusion Overall, the CIUS‐S scores appear reliable and valid in its Spanish version for adolescent populations, supporting its aim to detect and prevent a problem that has become a major worldwide issue in the last years.

PIU is defined as a generalized and compulsive use of the Internet associated with a loss of control and negative consequences for the individual (Caplan, 2002(Caplan, , 2010. It occurs when individuals remain connected for long periods of time, mostly for activities not related to work or study, causing difficulties at school, family, and social relationships (Sarmiento et al., 2021). Numerous studies have examined the psychopathological comorbidity between different mental health problems and PIU (Anderson et al., 2017;Shapira et al., 2003;Tsitsika et al., 2014;Werling & Grünblatt, 2022). Depression, anxiety, attention deficit hyperactivity disorder, and substance use disorders, among others, have shown a relationship with PIU Rial et al., 2014;Sussman et al., 2018). Some studies also suggest that PUI may be associated with increased rates of suicidal behaviors and/or self-harm (Herruzo et al., 2023) and psychotic experiences (Lee et al., 2019). PIU has received increasing research and clinical attention, but it has not yet been recognized by diagnostic classification systems. Indeed, it is not clear whether it should be considered as a mental disorder or it just reflects other underlying mental health disorders Vink et al., 2016). The 11th revision of the International Classification of Diseases (ICD-11; World Health Organization, 2019) includes two behaviors (gambling disorder and gaming disorder) that have been classified as "disorders due to addictive behaviors." While there is growing evidence that addictions to sex, pornography, social network sites, exercise, work, and online shopping may be genuine disorders among a minority of individuals, none of these behaviors is likely to be included in formal psychiatric manuals in the near future until there is more high-quality data on all research fronts (e.g., epidemiological, neurobiological, psychological, and clinical) (Griffiths, 2022).
The increasing research interest in this phenomenon has led to the development of numerous scales. Laconi et al. (2014) identify 45 measuring instruments (23 languages) and report on their psychometric properties. Among the different tools, literature suggests the Internet Addiction Test (IAT) (Young, 1998), the Generalized Problematic Internet Use Scale (GPIUS) (Caplan, 2002) and its second version (GPIUS-2) (Caplan, 2010), the Problematic Internet Use Questionnaire (PIUQ) (Demetrovics et al., 2008), and the Compulsive Internet Use Scale (CIUS) (Meerkerk et al., 2009). Specifically, in Spanish populations the CIUS, the GPIUS-2, the Internet-Related Experiences Questionnaire (IREQ) (Beranuy et al., 2009), and the PIUS-a (Rial, Gómez, Sorna, et al., 2015) are the most used instruments. Compared to the GPIUS-2 (Caplan, 2010;Gámez-Guadix et al., 2013;Marzo et al., 2022), the CIUS (Meerkerk et al., 2009;Ortuño-Sierra et al., 2022) comprised one item less and has been used in long epidemiological surveys conducted in Spain (e.g., Drug Use in Secondary Education Survey [Encuesta sobre uso de drogas en Enseñanzas secundarias en España, ESUDES]). In addition, the response format is a 5-point Likert-type scale instead a 6-point Likerttype scale. CIUS (Meerkerk et al., 2009) is one of the most frequently internationally adapted psychometric instrument developed to assess PIU. It is a well-validated tool and demonstrated adequate psychometric properties in terms of reliability and validity evidences in various countries and across cultures (Dhir et al., 2016;López-Fernández et al., 2019;Sarmiento et al., 2021). This scale originally included 14 items related to aspects such as loss of control (items 1, 2, 5, and 9), preoccupa-tion (items 4, 6, and 7), withdrawal symptoms (item 14), coping (items 12 and 13), and conflict (items 3, 8, 10, and 11) with a 5-point Likert-type scale ranging from "never" to "very often." It was constructed based on criteria of substance abuse dependency and pathological gambling described in the Dianostic and Statistical Manual of Mental Disorder IV (DSM-IV) along with literature on behavioral addictions (Meerkerk et al., 2009).
Instead of elaborating new instruments, translating and adapting existing validated scales can be especially useful. Laconi et al. (2014) suggested that the number of items for detecting PIU varies between 7 and 91 depending on the instrument. López-Fernández et al. (2019) tested the psychometric properties of four CIUS versions (i.e., CIUS-14, CIUS-9, CIUS-7, and CIUS-5) across eight languages (i.e., German, French, English, Finnish, Spanish, Italian, Polish, and Hungarian) to examine their psychometric properties. Findings suggested that the short CIUS-5 was robust for screening CIU in adults. Besser et al.

Participants
Data came from a population of 15,000 students in La Rioja (northern Spain

Strengths and Difficulties Questionnaire
The Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) has a Likert-type response format with three options (0 = not true,

Procedure
The questionnaires were administered collectively by computers, in groups of 10-30 students, during normal school hours. With the aim to standardize the administration process, all the researchers had a protocol that they had to follow before, during, and after conducting the administration of the questionnaires. Participants were informed of the confidentiality of their responses and the voluntary nature of the study. No incentive was provided for their participation. All parents or legal guardian were asked to provide a written informed consent for their child to participate in the study. The present study was approved by the Ethical Committee of Clinical Research of La Rioja.

Data analysis
First, we calculated the descriptive statistics and the percentage distri- Mean Square Residual (SRMR). In order to obtain an adequate model fit, Hu and Bentler (1999) (Byrne, 2008). Considering that the ∆χ 2 has shown several limitations as it is very sensitive to sample size, Cheung and Rensvold (2002) proposed the increase or decrease in CFI values (∆CFI) to determine if nested models are practically equivalent. We, in addition, calculated latent mean differences across gender and age. To this end, we fixed the latent mean values to zero in the male and in the lower age group. With the aim to stablish comparisons among the studied variables, statistical significance was based on the z-statistic. We considered the reference group the one in which the latent mean was fixed to zero. Fifth, we analyzed the internal consistency of the scores. We

Descriptive statistics for the CIUS-S and prevalence rates
Descriptive statistics for the five-item unidimensional model are depicted in Table 1. In addition, percentage of the different answers' options for the CIUS-S is included. For example, 10.1% of adolescents answered "very often" to Item 12 (How often do you go to the Internet when you are feeling down?). In addition, the distribution of the scores was studied. With this regard, the CIUS total score was nonnormally distributed, according to the Kolmogorov-Smirnov test = .076 (p < .01).

Measurement invariance of the CIUS-S' scores by gender and age
We studied the MI across gender and age in this model. With the aim to study MI by gender, only male and female were considered, due to the limited amount of people not identifying a gender (n = 13). In order to study the MI by age, we divided the sample into two different groupsfirst group (adolescents from 14 to 16 years old) and second group (adolescents from 17 to 18 years old)-attending to the Spanish educational system (compulsory/post compulsory). As can be seen in Table 2, the ΔCFI lower than .01 permits confirming strong MI both by gender and age.

Analysis of latent mean scores
The comparison across groups in latent means revealed statistically significant differences in the CIUS Total Score. Therefore, the comparison across groups in latent means indicated that, on average, men scored 0.183 units under women in the CIUS Total Score (p < .01). With regard to the age, comparison across groups in latent means revealed no statistically significant differences between the younger and the older group.

Study of the reliability of the CIUS-S' scores
With the aim to study the internal consistency of the CIUS-S scores, we calculated the McDonald's Omega coefficient. The total score displayed a coefficient of 0.73. All the discrimination indices were over .30.

Relation of the CIUS-S' with well-being and mental health variables
Finally, we analyzed the correlation between the CIUS-S' scores and variables related to socioemotional adjustment. Results are depicted TA B L E 1 Prevalence rates for the Compulsive Internet Use Scale-Short (CIUS-S) and descriptive statistics for the different items in the total sample.

Prevalence (%)
Descriptive statistics in Table 4. CIUS-S' scores were positively associated with suicidal behavior, depressive symptoms, emotional and behavioral problems, and psychotic like experiences. Moreover, CIUS-S' scores were negatively correlated with self-esteem, prosocial behavior, and the feeling of belonging. In addition, the CIUS-S' scores were very strongly associated with the CIUS original form (r = .927).

DISCUSSION
Today, Internet access and use for children and adolescents in European countries is almost ubiquitous (Anderson et al., 2017;Smahel et al., 2020). Time spent online has increased during and since the world pandemic COVID-19, not only for adults but also for adolescents. Thus, PIU is receiving an increased amount of attention, and more research is devoted to understand and screen for a phenomenon that has been associated with different problems including, among others, sleeping and eating disorders, social skills deficits, sedentary lifestyles, family conflicts, or poor school performance (Golpe-Ferreiro et al., 2017;Vila et al., 2018). Bearing that in mind, screening at early stages seems increasingly relevant and assessment tools seem highly necessary. Therefore, the main goal of the present study was to analyze the psychometric adequacy of a short form of the CIUS in its Spanish version.
Results with regard to the internal structure suggest that the tries. In addition, Dhir et al. (2015) found that an eight-item version had adequate psychometric properties. The study of Besser et al. (2017) indicated that both the five-and the seven-item versions were adequate. Nonetheless, they retained the five-item version as better discrimination indices were found.
Another critical aspect of a measurement instrument is the fact that its dimensional structure could be replicated by attending to the different variables of the object of study (e.g., gender, age, culture, etc.). With this regard, we studied the MI of the five-item version by gender and age. The results indicated that partial MI was tenable for both gender and age. Previous studies, similarly to our results, revealed MI of the CIUS by gender, as well as Internet use (Meerkerk et al., 2009). However, few studies have analyzed the MI of the instrument in it shorter form. For instance, Gmel et al. (2019) revealed that the short form of eight items was invariant across gender, age, and region. It is worth noting that the study was conducted with a population of young adult participants, so results may not apply to adolescent populations. Studying MI isalso relevant as it contributes essential evidence of construct validity for the shorter form of the CIUS scores in adolescents. We also studied the comparison in the latent means across gender and age.
Results found indicated that men scored statistically significantly lower than women overall. This is consistent with previous studies indicating higher prevalence of PIU among women (Andrade et al., 2021;Gómez et al., 2017;Rial et al., 2018). For instance, in Spain, recent research indicated a higher percentage among women (36.1%) compared to men (29.8%) (Andrade et al., 2021). These results are relevant in order to understand the manifestation of CIUS during adolescence. This stage of life is a critical developmental period with different biopsychological changes that have, potentially, a different impact with regard to the gender and the age of the person (Liu et al., 2022).
With regard to validity evidences based on relations with other variables, the CIUS-S revealed positive and statistically significant associations with variables related to mental health such as emotional difficulties, behavioral problems, psychotic experiences, depressive symptoms, and suicidal behaviors. The association was negative and statistically significant with prosocial behavior and feeling of belonging.
In addition, the CIUS-S and the CIUS long form were also correlated.
Other studies have also revealed correlations between PIU and mental health problems like depressive symptoms and insomnia (Jain et al., TA B L E 2 Goodness-of-fit indices for the hypothetical model tested and measurement invariance across gender and age. 2020). In addition, problematic internet use is correlated with mood symptoms (Gao et al., 2020), as well as substance abuse and suicidal behavior (Serrano et al., 2017). Considering the prevalence of PIU behaviors during adolescence and the associations between PIU and different mental health problems, it may be important to screening both in school and clinical settings.

Model
The present study has some limitations. First, it is based on selfreported measures, and there are well-known problems related to this kind of measures. Therefore, experimental data or studies including parents, teachers, or relatives could add valuable information to this area of study. In addition, we conducted a cross-sectional study, which limits establishing cause-effect associations. Thus, longitudinal studies analyzing PIU are still needed.
Notwithstanding the mentioned limitations, the present work provides valuable information for the screening of PIU in adolescent populations. This is, to the best of our knowledge, the first study providing evidence of the psychometric adequacy of a short form of the CIUS in its Spanish version and in adolescents' populations. To sum up, the CIUS-S is a brief and easy-to-use screening instrument that allows study of this increasing phenomenon.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data presented in this study are available on request from the corresponding author.