Skip to main content

A longitudinal analysis of the network structure of internet gaming disorder and its associations with distress

Behaviour & Information Technology, 2025

Deon Tullett-Prado, Bruno Schivinski, Dylan Poulus, Mark D. Griffiths, Vasileios Stavropoulos

Behaviour & Information Technology, volume 44, issue 20, pages 4981–5001, 2025 (online 16 May 2025)

  • Open access
  • 462 adult gamers
  • Two waves, 12 months apart
  • Network analysis
  • Internet gaming disorder
  • Depression, anxiety and stress
Read the article
Keywords
  • internet gaming disorder
  • behavioural addiction
  • psychological distress
  • network analysis
  • longitudinal
  • IGD symptoms
  • IGDS9-SF
  • DASS-21
  • depression
  • anxiety
  • stress
  • expected influence
  • bridge centrality
  • network comparison test
  • escapism
  • mood modification
  • tolerance
  • persistence
  • gaming disorder
  • secondary symptom debate

Internet gaming disorder holds together as its own set of symptoms rather than being a side effect of depression, anxiety or stress. Following 462 adult gamers over 12 months, we found that gaming symptoms connected mainly with each other, and the main bridge to distress ran from using games to escape bad moods to depression.

Key findings

  1. Gaming disorder symptoms form their own cluster. In networks of the nine IGD symptoms and depression, anxiety and stress, connections within the gaming and distress clusters were stronger than those between them, supporting IGD as a distinct construct (Tullett-Prado et al., 2025).
  2. Escapism is the bridge to depression. The only notable link between the two clusters ran from escapism/mood modification to depression (edge weights 0.17 at both waves); escapism had the highest bridge influence of the gaming symptoms (Tullett-Prado et al., 2025).
  3. Persistence is consistently central. At the first wave, inability to stop, persistence and tolerance (with stress) had the highest expected influence; at the second, withdrawal and persistence led, with tolerance and stress next; escapism had the lowest (Tullett-Prado et al., 2025).
  4. The network is stable over 12 months. Network comparison tests found no significant differences between the two waves in global structure (p = .27) or global strength (p = .52), and stability coefficients were .60 to .75 (Tullett-Prado et al., 2025).

At a glance

Design
Two-wave online survey of English-speaking adult gamers, 12 months apart (first wave August 2020 to August 2021, second wave August 2021 to August 2022), recruited through social media, Reddit and a university learning system
Participants
968 gamers at the first wave and 462 at the second (mean age 30.8; 69.3% male), aged 18 to 64
Measures
Internet Gaming Disorder Scale–Short Form (IGDS9-SF; α = .87–.88) and the Depression, Anxiety and Stress Scale (DASS-21; α = .85–.94)
Analysis
Regularised (graphical lasso) networks in R (qgraph, networktools, bootnet), expected influence and bridge expected influence, case-dropping bootstrap stability, and the Network Comparison Test across waves
Ethics
Victoria University Human Research Ethics Committee (HRE20-169)
Funding
Australian Research Council Discovery Early Career Researcher Award (DE210101107); Victoria University Early Career Researcher Fund

Figure 1. Networks of gaming disorder symptoms and distress at both waves

Time 1Preoccupation – Withdrawal: 0.218Preoccupation – Tolerance: 0.254Preoccupation – Loss of interest: 0.132Preoccupation – Persistence: 0.071Preoccupation – Escapism: 0.081Withdrawal – Tolerance: 0.141Withdrawal – Unable to stop: 0.347Withdrawal – Stress: 0.082Tolerance – Unable to stop: 0.140Tolerance – Loss of interest: 0.156Tolerance – Persistence: 0.065Tolerance – Escapism: 0.212Tolerance – Depression: 0.073Unable to stop – Loss of interest: 0.214Unable to stop – Persistence: 0.247Unable to stop – Deception: 0.088Unable to stop – Negative consequences: 0.060Loss of interest – Persistence: 0.080Loss of interest – Deception: 0.065Loss of interest – Negative consequences: 0.127Persistence – Deception: 0.351Persistence – Negative consequences: 0.266Deception – Negative consequences: 0.156Escapism – Depression: 0.170Escapism – Stress: 0.063Depression – Anxiety: 0.191Depression – Stress: 0.361Anxiety – Stress: 0.548123456789DASTime 2 (12 months later)Preoccupation – Withdrawal: 0.246Preoccupation – Tolerance: 0.122Preoccupation – Loss of interest: 0.080Preoccupation – Persistence: 0.103Preoccupation – Escapism: 0.105Withdrawal – Tolerance: 0.277Withdrawal – Unable to stop: 0.178Withdrawal – Loss of interest: 0.063Withdrawal – Persistence: 0.078Withdrawal – Deception: 0.119Withdrawal – Negative consequences: 0.087Tolerance – Unable to stop: 0.161Tolerance – Loss of interest: 0.082Tolerance – Persistence: 0.115Tolerance – Deception: 0.087Tolerance – Escapism: 0.059Tolerance – Anxiety: 0.063Unable to stop – Loss of interest: 0.163Unable to stop – Persistence: 0.173Unable to stop – Deception: 0.123Loss of interest – Persistence: 0.191Loss of interest – Escapism: 0.124Loss of interest – Negative consequences: 0.061Persistence – Deception: 0.164Persistence – Negative consequences: 0.217Deception – Negative consequences: 0.237Escapism – Depression: 0.174Escapism – Anxiety: 0.062Escapism – Stress: 0.067Negative consequences – Anxiety: 0.093Depression – Anxiety: 0.119Depression – Stress: 0.467Anxiety – Stress: 0.464123456789DAS1Preoccupation2Withdrawal3Tolerance4Unable to stop5Loss of interest6Persistence7Deception8Escapism / mood modification9Negative consequencesDDepressionAAnxietySStressRegularised partial-correlation networks; edge weights from Tables 4 and 5 of the article; edges of at least 0.05 shown (thicker = stronger). Layout drawn for this page.
Drawn for this page from the edge weights printed in Tables 4 and 5 of the article (edges of at least 0.05).

Tables

Table 2–3. Descriptive statistics at both waves

Node T1 mean T1 SD T2 mean T2 SD T2 N
1 Preoccupation 2.78 1.19 2.52 1.07 461
2 Withdrawal 1.89 1.05 1.82 0.974 461
3 Tolerance 2.15 1.12 1.85 1.04 461
4 Unable to stop 1.73 1.02 1.65 0.942 459
5 Loss of interest 2.11 1.17 1.96 1.06 460
6 Persistence 1.63 1.06 1.64 1.00 460
7 Deception 1.33 0.805 1.32 0.756 459
8 Escapism / mood modification 3.15 1.24 3.14 1.16 460
9 Negative consequences 1.33 0.781 1.30 0.723 460
Depression (DASS-21) 7.94 5.94 7.18 6.10 293
Anxiety (DASS-21) 4.25 4.07 4.23 4.22 292
Stress (DASS-21) 6.81 4.72 6.41 4.72 292

From Tables 2 and 3 of the article. IGD items rated 1 (never) to 5 (very often); DASS-21 subscale scores 0 to 21. T1 N = 452–460.

Table 4. Edge weights at the first wave (T1)
1 2 3 4 5 6 7 8 9 Dep Anx Str
1 Preoccupation 0.218 0.254 0.025 0.132 0.071 0 0.081 0 0 0 −0.015
2 Withdrawal 0.141 0.347 0.024 0.008 0.039 0 0 0 0 0.082
3 Tolerance 0.140 0.156 0.065 0 0.212 0 0.073 0 0
4 Unable to stop 0.214 0.247 0.088 0.047 0.06 0 0 0
5 Loss of interest 0.08 0.065 0.002 0.127 0.026 0.007 0
6 Persistence 0.351 0 0.266 0.022 0 0
7 Deception 0 0.156 0 0.035 0
8 Escapism 0 0.170 0.043 0.063
9 Negative consequences 0.021 0.018 0
Depression 0.191 0.361
Anxiety 0.548

Upper triangle of Table 4 of the article (the matrix is symmetric).

Table 5. Edge weights at the second wave (T2)
1 2 3 4 5 6 7 8 9 Dep Anx Str
1 Preoccupation 0.246 0.122 0.035 0.08 0.103 0 0.105 0.016 0 0 −0.013
2 Withdrawal 0.277 0.178 0.063 0.078 0.119 0 0.087 0 0.01 0
3 Tolerance 0.161 0.082 0.115 0.087 0.059 0.000 0 0.063 0
4 Unable to stop 0.163 0.173 0.123 0.027 0.012 0 0.039 0
5 Loss of interest 0.191 0.031 0.124 0.061 0 0 0.025
6 Persistence 0.164 0 0.217 0 0 0
7 Deception 0 0.237 0 0.023 0.013
8 Escapism 0 0.174 0.062 0.067
9 Negative consequences 0 0.093 −0.005
Depression 0.119 0.467
Anxiety 0.464

Upper triangle of Table 5 of the article (the matrix is symmetric; its last column is labelled “Scale” in the article and refers to stress).

The question

Internet gaming disorder (IGD) is only provisionally recognised, and critics argue that it may be a secondary symptom of depression or anxiety. We asked how the nine IGD symptoms relate to each other and to depression, anxiety and stress, which symptoms are most central, and whether this network holds over 12 months.

Implications

Practical implications

Treatment of IGD could focus on persistence and tolerance, gamers with problematic play should be screened for distress, and interventions should address the use of gaming to cope with negative moods; dialectical behaviour therapy or coping-focused cognitive behavioural therapy are suggested candidates. The results support the IGDS9-SF for assessing problematic gaming.

Theoretical implications

IGD behaved as a construct distinct from psychological distress, supporting the behavioural addiction model. Strong links between symptoms that mark addiction (salience, withdrawal, tolerance) and its consequences (persistence, negative life outcomes) suggest that the nine-symptom framework could be simplified.

Limitations and future research

  • The sample was mostly male (about 70%) and Caucasian, from Western countries; cultural differences were not examined.
  • Self-report measures are open to social desirability and memory biases.
  • The gaming and distress clusters were assumed rather than detected; exploratory graph analysis could test them more robustly.
  • Future work: the order in which symptoms arise, how persistence and tolerance affect other symptoms, and mood modification, depression and anxiety as pathways between distress and IGD.

Questions this paper answers

Is internet gaming disorder just a symptom of depression or anxiety?

In a two-wave network study of adult gamers, IGD symptoms clustered separately from depression, anxiety and stress, with one main link through escapism to depression, which the authors read as support for IGD as a distinct disorder (Tullett-Prado et al., 2025).

How is gaming disorder linked to depression?

Mainly through escapism or mood modification, that is, using games to escape negative moods, which connected the gaming symptoms to depression at both waves (Tullett-Prado et al., 2025).

Which gaming disorder symptoms are most central?

Persistence (continuing to game despite problems) was central at both waves, alongside inability to stop and tolerance at the first wave and withdrawal at the second (Tullett-Prado et al., 2025).

Does the network of gaming disorder symptoms change over a year?

No significant differences were found between the two waves, 12 months apart (Tullett-Prado et al., 2025).

Key concepts

Network analysis of psychopathology
An approach that treats a disorder as a network of interacting symptoms (nodes) linked by their associations (edges).
Expected influence
A centrality index: the sum of a node’s edge weights, keeping their signs, showing how strongly a symptom is connected to the rest of the network.
Bridge expected influence
The same index counting only edges to another cluster, here the links between gaming symptoms and distress.
Internet gaming disorder
A condition for further study in the DSM-5, defined by nine symptoms such as preoccupation, withdrawal, tolerance, loss of control and continued gaming despite problems.

Open science and materials

How to cite

Tullett-Prado, D., Schivinski, B., Poulus, D., Griffiths, M. D., & Stavropoulos, V. (2025). A longitudinal analysis of the network structure of internet gaming disorder and its associations with distress. Behaviour & Information Technology, 44(20), 4981–5001. https://doi.org/10.1080/0144929X.2025.2504510

Open access under a Creative Commons CC BY-NC-ND 4.0 licence.

Related publications

All publications