Exploring the Intrinsic Association Between Perceived Social Support and Depressive Symptoms and Problematic Phone Use Among College Students Based on Network Analysis

Introduction

In recent years, with the continuous development of electronic technology and the internet, mobile phones have become common items in people’s daily lives.¹ The portability of this tool has brought convenience to people’s lives, but it has also had some potential negative impacts, such as causing physical harm and negative impacts on mental health levels. This issue is referred to as problematic mobile phone use (PPU).^2,3 For adolescents, especially those in college campuses, these negative effects have become increasingly severe.^4–6 Numerous studies have pointed out that excessive use of mobile phones by adolescents may develop into problematic mobile phone use, which is accompanied by the emergence of various psychological and behavioral problems such as depression, anxiety, social barriers, and insomnia.⁷ A longitudinal study has indicated that problematic mobile phone use among college students can exacerbate anxiety symptoms from multiple dimensions.⁸ Depressive symptoms have also been repeatedly confirmed to be highly correlated with problematic mobile phone use.^9,10

Depression is currently one of the most common mental health disorders worldwide.¹¹ According to the latest Global Burden of Disease study, depression ranks first among mental disorders in terms of disease burden and is the 13th among all diseases globally.¹² Recent research indicates that approximately 21.48% of college students exhibit depressive symptoms, and the detection rate of depression among Chinese college students has been on the rise over the past decade.¹³ Depressive symptoms are often accompanied by various dangerous comorbidities, severely endangering the psychological and physical health of college students.^14,15 It is crucial to promptly reduce the overall level of depressive symptoms in the college student population. Notably, there is a close relationship between depressive symptoms and problematic mobile phone use among college students.¹⁶ Implementing intervention measures to alleviate both of these negative impacts on college students may yield unexpected results.

Social support refers to the support and assistance that an individual receives from social groups through interactions with others in social relationships, including family members, friends, teachers.^17,18 Social support can effectively alleviate tense mental states and adverse psychological reactions, and enhance social adaptability. Perceived social support is a type of subjective social support, referring to an individual’s subjective feeling of being supported and understood, which has a positive role.^19,20 As an important psychological resource for people to face setbacks and difficulties in life, perceived social support can significantly alleviate mental health problems.²¹ Current research has proven that perceived social support can effectively help reduce the occurrence and development of depressive symptoms among college students and also has a good effect on alleviating problematic mobile phone use.^22–24 However, past research has only stayed at the overall level and has not explored the complex internal relationships from the perspective of specific symptoms.

The network analysis model offers a fresh perspective for our research. It enables us to observe the interconnections among perceived social support, depressive symptoms, and problematic mobile phone use. Moreover, it allows for a more in-depth examination of the complex internal relationships between these three factors from the specific symptom level.^25,26 Network analysis is a practical approach to investigating complex, dynamic relationships between individual psychiatric and behavioral symptoms.^27,28 In network analysis, nodes reflect psychiatric and behavioral symptoms, and edges between nodes reflect relationships between symptoms, including the activation spread from one sign to another through the network.^27,29 Nodes can also function as bridge symptoms that transfer symptom activation from one disorder to another. Therefore, network analysis has utility in elucidating connection patterns between individual psychiatric symptoms or behaviors and psychiatric disorders.³⁰ Depressive symptoms and problematic mobile phone use among college students are highly correlated, and past research has also indicated that receiving social support plays a positive role in alleviating psychological problems among college students and reducing problematic mobile phone use. However, previous studies have not examined these three aspects within the same research model. This study will utilize a network analysis model to fill this research gap.

Therefore, this study aims to construct a network structure of problematic mobile phone use, depressive symptoms, and perceived social support among college students, explore the nodes that play a prominent role in the network, investigate how social support can alleviate the emergence of negative events from the perspective of specific symptoms, and propose targeted interventions to alleviate problematic mobile phone use and depressive symptoms among college students.

Methods

Participants

This study used convenience sampling to collect data from five different universities in five different provinces (Jiangsu, Shandong, Zhejiang, Guangdong, and Heilongjiang) and distributed and collected the questionnaires through the online questionnaire software “Questionnaire Star”. A total of 4186 college students participated in the survey and the questionnaires were collected through the online platform. Due to the content of the questionnaire, we deleted all the questionnaires that could be completed in less than two minutes because we felt that the validity of such questionnaires could not be guaranteed. In addition, we also set some validity entries, and if these validity entries in the recovered questionnaires contradict each other, then this questionnaire will also be excluded. By checking the time of filling out the questionnaires and the valid entries in the questionnaires and deleting some invalid questionnaires, 3869 valid questionnaires were finally obtained. The response rate was 92.4%.

Measures

Perceived Social Support

This research employed the Perceived Social Support Scale (PSSS-12) to measure the level of perceived social support among college students.^31,32 The scale evaluates individuals’ perception of social support and is divided into three sections: family support, friend support, and other forms of support. It utilizes a 7-point rating scale ranging from 1 to 7, where higher scores reflect a greater perception of social support. Factor analysis results confirm the good validity of this questionnaire. The Cronbach’s alpha coefficient for this scale of current study is 0.941.

Problematic Smartphone Use

The extent of problematic smartphone use was evaluated using the Smartphone Addiction Scale-Short Version (SAS-SV), which has been adapted to suit the Chinese context and demonstrates strong reliability and validity.^33,34 Unlike the original scale, the Chinese version of the SAS-SV has been locally modified in China to facilitate better comprehension and response from college students whose native language is Chinese. This scale has a unidimensional structure and consists of 10 items (eg, “Missing planned work due to smartphone use”, “People around me tell me that I use my smartphone too much”). Each item is rated on a scale from 1 (Strongly Disagree) to 6 (Strongly Agree), with higher scores indicating more severe problematic smartphone use among participants. Factor analysis results confirm the good validity of this questionnaire. The Cronbach’s alpha coefficient for this scale of current study is 0.913.

Depressive Symptoms

The present study used the PHQ-9 (Patient Health Questionnaire-9) scale to assess the depressive symptoms of college students, which is a common and popular scale for assessing depressive symptoms internationally and has very good reliability both at home and abroad.³⁵ The scale is divided into nine items, each of which is scored using the Likert-4 scoring mechanism, and the overall scale score ranges from 9 to 36, with higher scores indicating higher levels of depression. The reliabilities of our survey in this study were 0.901, suggesting reliability.

Analytical Strategies

Descriptive Analysis

Descriptive analysis was conducted using SPSS (26.0) software. The general demographic information distribution and specific scores for each research variable were described. Descriptive analysis was conducted using t-tests and ANOVAs of two independent samples to compare differences in ratings of perceived social support, depressive symptoms, and problematic phone use across different demographic profile contexts.

Network Estimation

All analyses and network model were computed using R-studio (version 4.3.1). Paired Spearman correlation analysis estimated the symptom network illustrating the relationship between symptoms. In the network analysis method, each symptom is treated as a node, and the pairwise correlation between these nodes are treated as edges.³⁶ We estimated the Graphical Gaussian Model (GGM), with the graphic least absolute shrinkage and selection operator (LASSO) and Extended Bayesian Information Criterion (EBIC) model using R package “graph”.²⁷

Strength was used as a centrality indicator to investigate the significance of individual symptoms in the network. Strength is the sum weight of all direct relationships between a particular sign and other symptoms. Centrality measures are reported as standardized values (z-scores). The bridge function in R-package “networktools” was also used to investigate network bridge symptoms essential in linking two or more mental diseases. Use bridge strength to assess the centrality of bridge symptoms. Based on previous investigations, bridge symptoms were chosen using an 80th-percentile bridge strength limit.^37,38

Estimation of Network Accuracy and Stability

The node and bridge strength stability were assessed using the R-package “bootnet” and a case-dropping bootstrap procedure. This procedure removed many cases from the dataset and recalculated the centrality indices. Suppose most samples can be removed from the dataset without significantly affecting the node’s centrality index. In that case, the network is said to be stable, and the stability is quantified by the Correlation Stability Coefficient (CS-C). CS-C refers to the maximum number of cases that can be excluded from the sample. The centrality indices from the subsamples are correlated with the index from the original selection at a value of r = 0.7.^25,27 In general, CS-C should be greater than 0.25, preferably greater than 0.5. The confidence interval (CI) for the correctness of an edge weight was computed using a non-parametric bootstrap procedure. The observations are randomly resampled to generate new data sets from which a 95% CI can be calculated, with a narrower CI indicating a more trustworthy network. To analyze variations in network characteristics, we performed a bootstrap differential test with 1000 permutations and ran the results.

Ethical Considerations

The study was approved by the Medical Ethics Committee of Xuzhou Medical University, and informed consent was obtained from the school, students, and parents. Our study complies with the Declaration of Helsinki.

Result

Descriptive Statistics

Table 1 presents the demographic characteristics of the participants and specific scores for problematic smartphone use, depressive symptoms and perceived social support. A total of 3869 study participants were included in this study, of which 1772 (45.80%) were males and 1908 (49.32%) were freshmen, 1710 (44.20%) were of urban origin. More than 90% of the population was Han Chinese. We found no significant differences in perceived social support scores, depressive symptoms scores and problematic phone use scores for the most part, as long as all three showed differences in scores across grades. For demographic information and specific scores for the three variables see Table 1. The specific symptoms represented by each node in the network are shown in Table 2.

Table 1 Description of General Demographic Information of Participants

Table 2 All Nodes in the Network and the Specific Symptoms They Represent

Network Estimation and Centrality Measure Analysis

Figure 1 depicts the problematic smartphone use, depressive symptoms and perceived social support network and the centrality measures of all the symptoms. The composite network structure shows that the edge PSS1 (There are people who are there for me when I have a problem) and PSS2 (I can share happiness and sorrow with some people) have the strongest correlations, followed by the edge PSS3 (My family can help me in a concrete way) and PSS4 (I can get emotional help and support from my family when I need it). The three different communities are very visible in the network model and the three show different interrelationships between them. A weighted adjacency matrix was used to examine the numerical interactions between these symptoms and the results are shown in Supplementary Table 1. Node PPU6 (Having my smartphone in my mind even when I am not using it) is the most influential in the network, followed by DEP4 (Feeling tired or having little energy), PPU 2 (Having a hard time concentrating in class while doing assignments, or while working due to smartphone use). The highest strength values indicate that these nodes are most likely to have an impact on other nodes in the network model.

Figure 1 The network structure of Perceived Social Support, Depressive Symptoms, and Problematic Phone Use among college students.

Based on previous research findings, the bridge strength of a node can effectively reveal the impact of this node on surrounding nodes. As shown in Figure 2, in terms of bridge symptoms, DEP4 (Feeling tired or having little energy), PSS10 (There are certain people in my life who care about my feelings), and PPU9 (Using my smartphone longer than I had intended) were the most prominent bridging symptoms in this study. These specific symptoms that are seen as bridge nodes are generally thought to be key in connecting the different communities, and the nodes mentioned above are thought to connect the three communities of perceived social support, depressive symptoms, and problematic phone use in this study.

Figure 2 Centrality and bridge centrality strength of network models.

Network Accuracy and Stability

As shown in Figure 3, we analyzed the stability of the network and found it to be very stable. The case-dropping bootstrap procedure resulted in CS-Cs of 0.75 for node strength, indicating that 75% of the samples could be removed without significant changes in the network structure compared to the original structure. The edge weights of the current model, especially those with larger weights, are consistent with the edge weights of the bootstrap sample, indicating that the existing network structure is stable (Supplementary Figure 1). The bootstrap 95% CIs for the estimated edge weights suggest that the network model is reliable and stable. Supplementary Figure 2 indicated that bootstrap difference tests showed that most edge-weight comparisons were statistically significant. The Bootstrapped 95% CI of the estimated edge weights indicates that the network model is reliable and stable (Supplementary Figure 3).

Figure 3 The Stability of centrality indices by case-dropping bootstrap. Notes: The stability of strength and bridge strength using case-dropping bootstrap. The x-axis indicates the percentage of cases of the original sample included at each step. The y-axis indicates the average correlations between the centrality indices from the original network and the centrality indices from the networks that were re-estimated after excluding increasing percentages of cases. Each line indicates the correlations of strength and bridge strength, while areas indicate 95% CI.

Discussion

This study is the first to explore the complex internal relationships between perceived social support, problematic mobile phone use, and depressive symptoms among college students using network analysis methods. Through the network analysis model, we can observe the specific interconnections between symptoms. The results also provide us with key bridge nodes and connection paths in the network model, offering clues for proposing targeted intervention measures in the future.

Observing the overall network model, we can find that the two nodes with the highest strength in the network are PPU6 from the problematic mobile phone use community and DEP4 (Feeling tired or having little energy) from the depressive symptoms’ community. At the same time, DEP4 is also the node with the highest bridge strength in the entire network, which means that this node plays a crucial role in connecting three different communities. DEP4 represents the most common depressive symptom among college students.³⁹ This may stem from the heavy academic pressure and intense competitive exam stress they face, as well as the feelings of discomfort and fatigue brought about by the campus environment.⁴⁰ By observing the edge weights of the network, we found that the node PSS7 (I can count on my friends in times of trouble) shows a negative association with DEP4 (Feeling tired or having little energy), revealing that having friends to rely on during difficult times is very important for alleviating depressive symptoms. The support from friends can provide emotional comfort and understanding, helping individuals to reduce psychological stress and feelings of loneliness.⁴¹ In addition, the companionship and encouragement from friends can also promote individuals to participate in more social activities, thereby improving their emotional state. The role of this social support has a significantly positive impact on alleviating depressive symptoms.^23,35 A nationwide longitudinal study has pointed out that high-quality friendships during childhood are of great significance in reducing depressive symptoms in old age and the physical damage caused by these symptoms.⁴²

Some studies have indicated that peer bullying, school dissatisfaction, and depressive symptoms all show significant negative correlations, and social support can act as a mediating variable to reduce the occurrence and development of depressive symptoms caused by peer bullying.^43–45 In this study, four nodes related to peer support are clustered together in the network, and most of their associations with the depressive community are also significantly negative. The interaction among peers is crucial for alleviating depressive symptoms and other negative emotions in adolescents, which can be explained through emotional and companionship support in social support theory.^22,46 The intimate relationships between peers can provide emotional support and comfort. When adolescents encounter psychological distress or stress, they can confide in their peers, gaining understanding and sympathy, which in turn reduces feelings of loneliness and depressive emotions, thereby decreasing the tendency to use mobile phones and other items to alleviate emotions. Additionally, the companionship of peers makes adolescents feel less alone, with someone willing to spend time with them. This companionship can shift their focus away from negative emotions, reducing excessive attention and indulgence in these emotions, thus alleviating depressive symptoms.⁴⁷ It is worth noting that compared to the negative impact of peer support on depressive symptoms, peer support does not effectively reduce people’s dependence on mobile phones. This may be due to the prevalence of mobile phones in the digital age, where social and entertainment activities among peers are often based on mobile phones and online platforms, thereby increasing phone usage and exacerbating dependence.^48,49

Family members’ support also shows a negative correlation with the research subjects’ problematic mobile phone use and depressive symptoms. Through the edge weight results, we can find that nodes PSS3 (My family can help me in a concrete way) and PSS8 (I was able to talk to my family about my problems) show a negative correlation with most results from the other two communities. Both of these nodes reflect that families can provide tangible help when facing practical difficulties. Family support has been proven to reduce depressive symptoms in adolescents.^24,50,51 Additionally, some studies have indicated that support from family can reduce adolescents’ feelings of loneliness, as well as the associated issues of problematic mobile phone use and internet addiction.^52,53

In the entire network model, the perceived social support community’s PSS10 (There are certain people in my life who care about my feelings) shows a negative correlation with the depressive symptoms community’s DEP9 (Thoughts that you would be better off dead or of hurting yourself in some way) and the problematic mobile phone use community’s PPU 7 (I will never give up using my smartphone even). This indicates that the more concern a person feels from others, their suicidal ideation and dependence on mobile phones will be weakened. This aligns with the social support theory proposed by previous studies, which highlights the impact of the external social environment on an individual’s internal factors.^54,55 A review and meta-analysis have pointed out that social support has significant practical importance in reducing suicidal ideation among psychiatric patients and related populations.⁵⁶ Whether it is support from family, peers, or other people around them, it is very important for the mental and behavioral health of adolescents.

Observing the overall network model, it can be seen that most nodes within the problematic mobile phone use community and the depressive symptoms community among college students exhibit positive correlations. Meanwhile, nodes in the perceived social support community generally show negative correlations with the other two communities, which aligns with previous research indicating that perceived social support can alleviate depressive symptoms and problematic mobile phone use.^24,57 However, there are still some counterintuitive relationships within the network model. For example, DEP9 (Thoughts that you would be better off dead or of hurting yourself in some way) and PPU9 (Using my smartphone longer than I had intended) show a negative correlation. The negative correlation between excessive mobile phone use and suicidal ideation can be explained from different perspectives. Excessive mobile phone use can serve as a way to divert attention, allowing individuals to temporarily detach from negative emotions and suicidal thoughts, reducing their focus on suicidal ideation. This short-term shift in attention may alleviate suicidal ideation to some extent.^58,59 Additionally, some studies have indicated that mobile phones provide individuals with a wealth of information and resources, including knowledge about mental health, suicide prevention, and coping strategies.⁶⁰ People with suicidal ideation can access relevant information through their phones to understand the dangers of suicide, seek help, and learn how to deal with psychological distress. This information and these resources can help them better understand their condition, enhance their confidence and ability to cope with problems, thereby reducing suicidal ideation.⁶¹ However, it is important to note that while excessive mobile phone use may have some positive effects on reducing suicidal ideation in certain situations, the negative impacts it brings cannot be overlooked. Interestingly, a study has pointed out that using smartphones can help alleviate depressive and anxiety symptoms in breast cancer patients, further emphasizing the importance of using mobile phones wisely in daily life.⁶²

The research results reveal the intrinsic connections between perceived social support, problematic smartphone use, and depressive symptoms among college students, providing new clues for professionals responsible for the psychological and behavioral health of college students on campus. First, schools need to provide more support to adolescents, including encouragement in daily life and mental aspects, strengthen mental health education, offer professional psychological counseling services, and establish a mental health early warning system. In addition, adolescents should establish positive connections with family members and peers to obtain sufficient emotional support. As the core bridge node PSS10 in the research results shows, feeling cared for by others is most likely to reduce the emergence of depressive symptoms and problematic smartphone use.

Limitations

There are still some limitations in this study. First, the questionnaire survey used in this study is based on self-report, which inevitably has some uncontrollable biases. Additionally, the study involves multiple different cities and schools, and the discipline at the questionnaire completion sites varies, making it difficult to ensure a uniform quality of responses. Furthermore, the study collected limited general demographic information about the subjects, which makes it challenging to adequately control for environmental influences on individuals. Third, this study is a cross-sectional one, and the stability of the results needs to be further validated through more repeated experiments and longitudinal studies, which is a specific direction for future research. In future studies, we plan to collect study participants from more areas and gather more comprehensive general demographic profile information. In addition, subsequent studies will be followed longitudinally to obtain more stable and accurate results using the longitudinal network analysis model test. It is important to expand the existing research and strive to overcome the current limitations, continuing to explore in depth the specific mechanisms by which perceived social support alleviates depressive symptoms and problematic mobile phone use among college students.

Conclusion

This study is the first to use network analysis to explore the intrinsic connections between perceived social support, problematic smartphone use, and depressive symptoms among Chinese college students. The results show that PSS10 (There are certain people in my life who care about my feelings) is a key bridge node connecting three communities in the network, while DEP4 (Feeling tired or having little energy) and PPU6 (Having my smartphone in my mind even when I am not using it) are the nodes with the highest strength in the network. Identifying these key and bridge nodes is of great theoretical significance for intervening among college students with problematic smartphone use and depressive symptoms. By providing precise social support interventions for these adolescent groups and targeting the intrinsic connections between problematic smartphone use and depressive symptoms, it is possible to combat the worsening states of both issues simultaneously, thereby achieving more effective and comprehensive treatment.