Social Networks and Internet Emotional Relations on Student Mental Health and Quality of Life: Structural Equation Modeling | BMC Psychiatry


Study design and participants

The present cross-sectional study was conducted in 2021 among students of the Alborz University of Medical Sciences in the province of Alborz and its six schools: medicine, dentistry, pharmacy, health, nursing and paramedicine.

Sample size and sampling method

According to a study conducted by Veisani et al. [19] and considering correlation= 0.15,β= 0.2,α= 0.05the sample size was determined to be 350 using the following formula

$$boldsymbol{n}={left{left({boldsymbol{Z}}_{boldsymbol{alpha}}+{boldsymbol{Z}}_{boldsymbol{beta)} right)/mathbf{C}right}}^{mathbf{2}}+mathbf{3}$$


The students were selected by the multistage random sampling method. The students in each school were selected using the method proportional to size (number of students in each school), then in each school the students were selected using the method of systematic stratified random sampling. The stratification was based on the number of semesters completed by the students.

All students aged 18 to 29 who had successfully completed at least one semester of college, had previous physical and mental health assessments according to their school records and self-assessment, and had a mobile phone allowing them to use virtual networks on their mobile phone were included in the study.

Students using psychotropics, narcotics, and antidepressants, who withdrew from the study for any reason, who discontinued their studies, experienced adverse events and trauma during this study (e.g., the death of their parents) and returned incomplete questionnaires were excluded.


Demographic, socioeconomic status, mental health status, internet use, and quality of life were assessed using validated questionnaires.

Demographic characteristics such as age, gender, nationality, marital status, education, field of study, academic semester, and occupation were asked of the students. Students were also asked to indicate the use of virtual networks on their mobile phone, the fact of being a virtual network user, the type of virtual network used and the average duration of use of social networks per day (in hours).

Socio-economic status questionnaire

The socio-economic status questionnaire consists of five main items and six demographic items and was developed by Ghodratnama in 2013 to assess the four dimensions of socio-economic status, namely income level, economic class, education and housing status. Items were scored on a five-point scale ranging from 1: very low to 5: very high (the score ranges from 5 to 25). Eslami et al. confirmed the face and content validity of this questionnaire and the internal consistency reliability in Iran [20]. Cronbach’s alpha in this study was 0.83.

Internet Emotional Relationships Questionnaire

The questionnaire developed by Barghi-Irani et al. [21] was used to assess online emotional relationships in this study. This 28-point questionnaire has five items, including trust, honesty, enjoyment, sexual desire and preference for virtual relationships, and is scored on a five-point Likert scale (from totally disagree to completely disagree). agreement), the score ranges from 28 to 140. The validity and reliability of the questionnaire was confirmed with Cronbach’s alpha coefficients of 0.73 for trust, 0.70 for honesty, 0.71 for pleasure , 0.79 for sexual desire, 0.84 for preferring virtual communication and 0.90 for the whole scale. [21]. In the present study, the Cronbach’s alpha of the questionnaire was 0.85.

Stress, Anxiety, Depression Scale (DASS-21)

The validated DASS-21 was used to assess the mental health status of the subjects [22]. It has 21 items in the three dimensions of stress, anxiety, and depression, each with seven items, and the final score for each subscale and the total score are found by summing the item scores for each subscale. ladder. Each item receives a score between zero (did not apply to me at all) to three (applied to me a lot); thus, the total score for each subscale ranges from 0 to 21, and the DASS-21 total score ranges from 0 to 63. Lower total and subscale scores indicate a lower level of depressive and anxiety symptoms and a lower level of lower stress. [22]. The validity and reliability of this scale have been confirmed in Iran by Sahebi et al. within a range of 0.77 to 0.79 [23]. The internal consistency of this questionnaire was validated (Cronbach’s alpha: 0.89) in this study.

36-Item Abbreviated Questionnaire (SF-36)

In this study, an abbreviated form of health survey called SF-36 was used to assess quality of life. This questionnaire was designed in the United States by Ware and Sherbourne (1992), and its validity and reliability were evaluated in different groups of patients. [24]. The survey contains 36 items across eight dimensions, including physical functioning, role limitations due to physical issues, role limitations due to emotional issues, energy/fatigue, emotional well-being, social functioning, pain and general health. In addition, two general subscales called physical health and mental health are obtained by combining these subscales. A lower total score in this questionnaire indicates a lower quality of life and vice versa. The score varies from 0 to 100. In Iran, Montazeri et al. (2005) confirmed the validity of this questionnaire at 0.58 to 0.95 and its reliability at 0.77 to 0.90 [25]. In the present study, Cronbach’s alpha was 0.78.

Social media quiz

We used the Iranian questionnaire developed by Jahanbani et al. [26] for evaluation. It comprises 19 items containing three dimensions: rate of use, type of use and user confidence in the networks. Scoring is based on a five-point Likert scale ranging from very little to a lot, the score ranges from 19 to 95. Jahanbani et al. confirmed the reliability of this questionnaire with a Spearman correlation coefficient of 0.90. The internal consistency of this questionnaire was also confirmed with the Cronbach’s alpha coefficient of 0.85 [26].


The study began after obtaining the necessary permissions from the university and a code of ethics from the university’s ethics committee. Due to the Covid-19 situation and the impossibility of physical presence of the students, a consent form to participate in the study was first sent to the students via related online networks, such as the student assistant, the student research committee and student groups. . Eligible students willing to participate were selected by convenience sampling. Then, the online questionnaires were sent to the students via these networks, and they were asked to complete them within the specified time frame (maximum of 2 weeks). The researcher’s telephone number was communicated to the students to answer any ambiguities.

The students were assured of the confidentiality of all their data and that they were under no obligation to participate in the study or to continue their cooperation, and that they would not encounter any problems or restrictions if they decided not to participate in the study.

statistical analyzes

This study assessed the suitability of a conceptual model to examine the simultaneous effect of social networks and emotional relationships on the Internet on students’ mental health and quality of life (Fig. 1). First, the normal distribution of quantitative variables was assessed using the Kolmogorov-Smirnov test, then the data was analyzed in SPSS-25 [27]PLS3 [28]and Lisrel-8.8 [29].

Fig. 1

A conceptual model of communication between social networks and Internet emotional relationships on mental health and quality of life

In order to test the model, the aforementioned questionnaires were first evaluated in the model by PLS. The factor loadings of the items of each questionnaire as well as the validity and reliability of the tools used in the model were evaluated in PLS. According to the results, the factor loadings of all the items of the questionnaire were higher than 0.4 and all the items were retained after the final test of the model. To determine convergent and divergent validity, indices including composite reliability (CR), mean extracted variance (AVE), maximum shared variance (MSV), and mean shared variance (ASV) were used.

Discriminant validity, meaning that latent variables that represent different theoretical constructs are statistically different, was assessed by the heterotrait-monotrait ratio of correlations (HTMT); to this extent, values ​​less than 0.90 are considered to have discriminant validity [30].

After evaluating the questionnaires in the model, we test the model by path analysis. Path analysis is considered a causal modeling technique; it can be done with cross-sectional or longitudinal data and is an extension of the usual regression that shows the direct effects as well as the indirect effects and the impact of each variable on the dependent variables. All variables in a trajectory model can be described as endogenous or exogenous. Endogenous variables are schematized as being influenced by other variables in the model. The variables schematized as independent of any influence are the exogenous variables. Dependent variables are always endogenous, but some independent (or predictor) variables may be endogenous if they are influenced by other independent variables in the model [31]. In this study, the exogenous variables were education, average Internet time use, socioeconomic status (SES), emotional relationships on the Internet, and the endogenous variable was mental health and quality of life. . Social networks were exogenous for mental health and quality of life, but endogenous for the other variables.

Correlation results were presented as Pearson’s correlation coefficient and trajectory analysis results as regression coefficient, beta standardized with significance level of T value > 1.96.


Comments are closed.