2.1 Study design and sample
This cross-sectional study with questionnaires was conducted at the Faculty of Education, Palacky University in Olomouc (Czech Republic) and the Faculty of Educational Sciences, Chongqing Normal University in Sichuan (China).
Inclusive criteria:
• College/university students studying full-time study programs at the Faculty of Education.
• Minimal age of 18 years.
• Czech/Chinese nationality.
• Only students of teaching study programs (who passed the subject educational competence) were recruited.
• Gender, socioeconomic status, language, and academic results were not limited.
Excluded were students with nationalities other than Czech / Chinese, and part-time students, or students of study programs without qualification. To fill the data in the questionnaire, respondents had to give consent for voluntarily participation in the study and anonymous publication of data in the final publication outcome.
The target sample of respondents eligible towards the inclusion criteria at Palacky University Olomouc totalled more than 2000 students, and at Chongqing Normal University in Sichuan, approximately 1800 students. Participants were contacted via official university emails (Czech Republic) and via the WeChat social app (China). Data were collected from February to April 2022 and from February to May 2023 using an online questionnaire (Google Forms in Czech Republic and Wenjuanxing in China) in two identical forms in the Czech and Chinese language. No questionnaires were excluded.
The final sample consisted of 1075 Czech (mean age ±SD: 24.07 ±6.7, 913 (84.9%) females) and 710 Chinese (mean age ±SD: 21.0 ±2.3, 388 (54.6%) females) students. The estimated response rate in Czech students was approximately 50%, whereas in Chinese students it was approximately 40%. Czech students were most frequently in the second (N=524, 48.7%) and first (N=326, 30.3%) year of study, while Chinese students were most frequently in the second (N=407, 57.3%) and third (N=191, 26.9%) year of study.
As the study involved an anonymous questionnaire, there is no risk of exposing participants' personal information. Therefore, specific consent for publication of the data is not required.
2.2 Measurements
The outcomes of this study were technology addiction (internet and smartphone addiction), mental health (depression, anxiety, stress, and resilience), personality traits, and academic achievement:
Mobile phone addiction was measured using the short version of the Smartphone Addiction Scale (SAS-SV, [43]). The scale consists of 10 items (each with a 6-point scale from strongly disagree to strongly agree), with higher scores indicating a higher level of mobile phone addiction. The reliability in this study was ω=0.90. Past research has shown good internal consistency reliability in studies using this scale with adolescent groups in Italy (Cronbach’s alpha = .79) and adolescents in Brazil (α = 0.81; ω = 0.78) [44, 45].
Internet addiction was assessed using a six-item (5-point scale from strongly disagree to strongly agree) version of the Internet Addiction Test (IAT, [46]), a short unidimensional scale. Scores range from 0 to 30, with a higher score representing stronger addiction. This instrument has previously demonstrated good psychometric properties [47, 48] and demonstrated excellent reliability (ω=0.86 and α=0.96) in the present study.
Mental health was measured using the 4-item version of the Patient Health Questionnaire (PHQ-4) [49]. Each question is scored on a scale of 0 (not at all) to 3 (nearly every day), yielding a total possible score range from 0 to 12. The total score lies within one of the following categories: 0-2: Normal, 3-5: Mild, 6-8: Moderate, 9-12: Severe. This short tool is designed to screen for the presence of depressive and anxiety symptoms. Higher scores indicate greater presence of symptoms. Despite its shortness, it has shown good psychometric properties. Reliability in this study was ωdepr.=0.77, ωanx.=0.85.
Perceived stress was measured using the 10-item Perceived Stress Scale (PSS, [50, 51]). The scale is scored on a 5-point Likert scale ranging from 1 = never to 5 = always. A higher the score indicates increased perceived stress. Studies have shown good reliability in different cultural settings. According to previous studies, the PSS is an easy-to-use questionnaire with established acceptable psychometric properties [52, 53]. A higher the score indicates increased perceived stress. The reliability in this study was ω=0.81.
Resilience was assessed by the short version of the Connor-Davidson test (CD-RISC, Vaishnavi et al., 2007)[54]. It is a 2-item (5-point scale from not at all to almost always) unidimensional tool for the quick assessment of resilience level that shows similar psychometric properties as longer versions. Reliability in this study was ω=0.80.
Personality traits were assessed using the Ten Item Personality Measure (TIPI, Gosling et al., 2003)[55]. This short, 10-item questionnaire assesses five personality traits based on the Big Five: extraversion (item 1,6), emotional stability (neuroticism, item 4,9), agreeableness (item 2,7), conscientiousness (item 3,8), and openness to experience (item 5,10). Each consist of a pair of descriptors that were scored from 1 (strongly disagree) to 7 (strongly agree). Each of the five dimensions has two corresponding items: the former is a forward score, and the latter is a reverse score. Despite being short, the questionnaire has proven its usefulness and validity when compared to longer scales (Gosling et al., 2003; Vorkapić, 2016), although some psychometric indicators suffer from some limitations due to the TIPI design (Kline, 2000; Woods & Hampson, 2005).
Academic achievement was assessed using an Academic Achievement Questionnaire (AAQ, Křeménková & Novotný, 2020)[30]. This new 9-item scale measures overall academic achievement and its three dimensions: performance (6-point scale from A to F), coping with study demands, and social adaptation (5-point scale from not at all to without problem). Reliability (McDonald’s omega) proved to be good (ωtotal=0.77, ωst.perf.=0.76, ωst.dem.=0.79, ωsoc.adapt.=0.59).
2.3 Statistical analysis
The reliability of the data was examined using McDonald’s omega. A preliminary presence of relationships between variables was verified by a Spearman correlation analysis. The models of the effects and interactions of the variables were then analyzed with structural equation modelling (path analysis) using maximum likelihood estimation. Analyses were performed separately for samples from both countries to capture cross-cultural differences.
Several goodness-of-fit indicators were used to assess the quality of the models: Chi-square statistics, Comparative Fit Index [56], Tucker-Lewis Index [57], Root Mean Square Error of Approximation (RMSEA)[58], Akaike information criterion (AIC)[59], and Bayesian information criterion (BIC)[60]. Another parameter of interest was the explained variance of individual intermediate and dependent variables, expressed by adjusted R-squared.
Individual models were gradually built by adding individual variables in different patterns of interactions (paths) in the analysis following several rules, such as academic achievement dimensions could only appear as dependent and intermediate variables (performance dimension only as a dependent variable equals an outcome), personality traits and gender could only be independent variables. Variables were included in the model in the order: addictions->mental health->personality->gender, while continuously verifying the significant impact of the variables already present in the model. To limit the number of models examined, only paths supported by significant preliminary correlations between variables were tested for mental health and personality trait variables. The final model included all variables and paths showing a significant effect, while exhibiting the best goodness-of-fit.
All statistical analyses were performed as two-tailed and all the results with P<0.05 were considered statistically significant. Data analysis and visualizations were performed in RStudio (v.2022.07.2 with R environment v.4.2.1). The path analysis was performed using the Lavaan package (v.0.6-15).
2.4 Ethical Approval and Consent to Participate
Ethical approval for this study was obtained from the Ethical Committee of Palacky University Olomouc (IGA_028_doc), ensuring that all experiments were performed in accordance with relevant guidelines and regulations. The approval by this committee validates that all experimental protocols were approved by a recognized institutional and/or licensing committee. Furthermore, informed consent was obtained from all subjects involved in the study, or from their legal guardian(s) where applicable. This process was conducted in strict adherence to ethical standards, safeguarding participant rights and well-being.
2.5 Availability of Data and Materials
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to their containing information that could compromise the privacy of research participants. Access to the data will be granted in accordance with ethical guidelines and under the condition that the requester will use the data solely for scientific and non-commercial purposes. Any requests for data access will be reviewed by the research team to ensure compliance with ethical standards and the protection of participant confidentiality.