Original Article

Validation of the Internet Gaming Disorder Scale–Short‑Form and the

Gaming Disorder Test in Singapore

Abstract Peter K. H. Chew1,

Introduction: Given that 90% of younger Singaporeans play video games, there is an urgent need Kuhanesan N. C.
for psychometrically valid instruments as a screening tool for problematic gaming. The Internet Naidu2,3,
Gaming Disorder Scale–Short‑Form (IGDS9‑SF) and the Gaming Disorder Test (GDT) have not
been validated for use among Singaporeans. The current study aimed to examine the psychometric Jing Shi4,
properties of both instruments among Singaporeans. Methods: Participants were a representative Melvyn W. B.
sample of 1001 adults (aged 18–40 years) in Singapore recruited using stratified sampling. The study Zhang5
used a correlational design. Participants completed instruments that assess Internet gaming disorder 1
School of Social and Health
(IGD), gaming disorder (GD), and negative emotional states. Data collection was conducted and Sciences, James Cook
completed in August 2023. The data were analyzed using (multigroup) confirmatory factor analyses, University, 2Department of
reliability analyses, Pearson’s correlation coefficient, and t‑tests. Results: The results provided Psychological Medicine,
support for the one‑factor model of the IGDS9‑SF (Comparative Fit Index [CFI] = 0.972) and National University Health
System, 3Saw Swee Hock School
GDT (CFI = 0.996). Both instruments also had configural, metric, scalar, and strict invariances across
of Public Health, National
gender (∆CFI was <0.010). The instruments were also reliable, with a Cronbach’s alpha of 0.93 and University of Singapore, 4Health
0.90, respectively. Finally, the instruments were significantly correlated with each other (r = 0.83) and Social Sciences, Singapore
and with depression, anxiety, and stress (r ranged from 0.65 to 0.71). Conclusion: Overall, both Institute of Technology,
instruments are reliable and valid among Singaporeans. Future research could recruit samples with 5
Nanyang Technological
a wider age range to include children and adolescents and those over the age of 40 years and to University, Singapore, Singapore
compare scores of the instruments against clinician‑administered assessments.

Keywords: Behavioral addiction, Internet gaming disorder, psychometrics, reliability, Singapore,

validity

Introduction activities because of gaming, (6) continued

gaming despite problems, (7) deceiving
The Diagnostic and Statistical Manual of
family members or others about amount of
Mental Disorders, 5th edition (DSM‑5)
gaming, (8) gaming to escape or to relive
added Internet gaming disorder (IGD) to
negative moods, and (9) risk or loss of a
Section III as a condition that warrants
relationship, job, or educational or career Received: 25 September, 2024.
further studies and defined it as “a pattern Revised: 18 December, 2024.
opportunity because of gaming. Individuals
of excessive and prolonged Internet gaming Accepted: 03 February, 2025.
who meet five or more criteria during the Published: 30 May, 2025.
that results in a cluster of cognitive and
past 12 months would meet the diagnostic
behavioral symptoms, including progressive ORCID:
criteria for IGD. Meta‑analyses found a Peter K. H. Chew:
loss of control over gaming, tolerance, and
prevalence rate of 3.05% for IGD,[2] with a https://orcid.org/0000-0002-5243-
withdrawal symptoms, analogous to the 1481
higher prevalence rate of 10.10% in Asian
symptoms of substance use disorders.”[1] It
countries.[3]
should be noted that despite its name, IGD Address for correspondence:
encompasses both online and offline gaming Since IGD’s inclusion in the DSM‑5, the Prof. Peter K. H. Chew,
behavior. The nine criteria of IGD are (1) correlates of IGD have been well documented. James Cook University,
preoccupation with gaming, (2) withdrawal For example, the correlates of IGD include Singapore 387380, Singapore.
E‑mail: [email protected]
symptoms like irritability or anxiety when age (i.e. younger participants are at higher
unable to play games, (3) tolerance – the risk),[2] gender (i.e. males are at higher risk),[4]
need to increase time spent on games, (4) and personality (i.e. conscientiousness, Access this article online
unsuccessful attempts to reduce or stop extraversion, agreeableness, and
Website: www.
gaming, (5) loss of interest in other neuroticism).[5] IGD is also correlated with healthandbehavior.com
a range of negative consequences such as DOI: 10.4103/shb.shb_327_24
depression and anxiety[6] and poor sleep Quick Response Code:
This is an open access journal, and articles are
distributed under the terms of the Creative Commons quality.[7]
Attribution‑NonCommercial‑ShareAlike 4.0 License, which
allows others to remix, tweak, and build upon the work
non‑commercially, as long as appropriate credit is given and How to cite this article: Chew PK, Naidu KN, Shi J,
the new creations are licensed under the identical terms. Zhang MW. Validation of the internet gaming disorder
scale–short-form and the gaming disorder test in
For reprints contact: [email protected] Singapore. Asian J Soc Health Behav 2025;8:125-32.

© 2025 Asian Journal of Social Health and Behavior | Published by Wolters Kluwer - Medknow 125
 Chew, et al.: Internet gaming disorder

New instruments designed to assess the nine they should have an acceptable Cronbach’s alpha
criteria of IGD based on the DSM‑5 have been and McDonald’s omega of more than 0.70, and the
reviewed.[8] These instruments include the IGD standard error of measurement should be ≤standard
Scale‑Short‑Form (IGDS9‑SF),[9] the Ten‑Item Internet deviation (SD)/2 (H2). Third, it was expected that the
Gaming Disorder Test,[10] and the IGD Scale.[11] The instruments would demonstrate convergent validity by
current study focused on the IGDS9‑SF because it has sharing a strong positive correlation with each other (H3a).
been validated in both Asian[12] and Western[13] samples. Furthermore, the known correlates of IGD and GD will be
Overall, a systematic review of 21 studies encompassing used to examine the concurrent validity and known‑group
15 language versions concluded that the IGDS9‑SF has validity of the instruments. Specifically, it was expected
excellent psychometric properties.[14] that the instruments would be significantly correlated
with depression (H3b), anxiety (H3c), and stress (H3d).[6]
In addition to the DSM‑5, the International Classifications
Finally, it was expected that males would have higher IGD
of Diseases, 11th edition (ICD‑11) officially recognized
and GD scores than females among Singaporeans (H3e).[4]
gaming disorder (GD) and defined it as a pattern of
persistent gaming behavior, either online or offline, that Methods
includes the following features: (1) impaired control
over gaming, (2) increased priority given to gaming over Participants
other activities, (3) continued gaming despite problems, Participants were a representative sample of adults
and (4) impairment in various life domains (e.g. family, recruited by a survey panel based on the inclusion and
educational, and occupational).[15] Individuals who meet exclusion criteria. Specifically, participants should be (a)
all of the criteria during the past 12 months would meet Singaporeans or permanent residents, (b) played at least
the diagnostic criteria for GD. One study involving 560 one game in the past 12 months, and (c) between 18 and
participants from China and the UK found a prevalence 40 years of age. A total of 1560 participants were recruited.
rate of 1.80% for GD.[16] However, 559 (35.83%) cases were removed because they
Due to the recency of the ICD‑11 criteria, there are did not meet the inclusion criteria, resulting in a total
relatively fewer instruments designed to assess the four sample of 1001 participants.
criteria of GD. These instruments include the Gaming Instruments
Disorder Test (GDT),[16] the GAMing Engagement
Screener test,[17] and the Assessment of Criteria for Specific The Background Information Form
Internet‑use Disorders.[18] The current study focused on The Background Information Form was developed for the
the GDT because it has been validated in both Asian[19] purposes of the current study to collect demographic and
and Western[20] samples. Overall, a meta‑analysis of 17 gaming‑related information. Demographic variables included
studies encompassing 14 language versions concluded that nationality, age, gender, ethnicity, housing type, occupation,
the GDT has excellent internal consistency reliabilities.[21] and current/highest education level, whereas gaming‑related
More important, a review found that the GDT was the only variables included the average amount of time (in hours)
instrument to adequately assess all ICD‑11 GD criteria.[22] spent playing games in a typical weekday and weekend.
Overall, multiple studies have provided support for the The Internet Gaming Disorder Scale–Short‑Form
psychometric properties of the IGDS9‑SF[14] and the
GDT.[16] However, both instruments have not been validated The IGDS9‑SF is a 9‑item instrument designed to assess the
for use among Singaporeans. A recent survey found that nine criteria of IGD in the DSM‑5.[9] Participants are asked
76% of Singaporeans play games, with the percentage to report their gaming activity during the past 12 months.
increasing to 90% among younger Singaporeans (aged Responses were made on a 5‑point Likert scale that ranged
18–24 years).[23] Given the high percentage of gamers, from 1 = never to 5 = very often. The item scores were
it is likely that some might play games excessively. This summed, with higher scores indicating high levels of IGD.
emphasizes the importance of having psychometrically The Gaming Disorder Test
valid instruments as a screening tool for IGD and GD in
The GDT is a 4‑item instrument designed to assess the four
Singapore.
criteria of GD in the ICD‑11.[16] Participants were also asked
The current study aimed to examine the psychometric to report their gaming activity during the past 12 months.
properties of the IGDS9‑SF and GDT among Singaporeans. Responses were made on a 5‑point Likert scale that ranges
First, it was expected that the instruments would be from 1 = never to 5 = very often. The item scores were
valid.[14,16] Specifically, it was hypothesized that the summed, with higher scores indicating high levels of GD.
IGDS9‑SF and GDT would have a unidimensional
The Depression Anxiety Stress Scale
structure (H1a), and there would be measurement
invariance across gender (H1b). Second, it was expected The Depression Anxiety Stress Scale is a 12‑item instrument
that the instruments would be reliable.[14,16] Specifically, designed to assess depression, anxiety, and stress.[24]

126 Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025
 Chew, et al.: Internet gaming disorder

Responses were made on a 4‑point Likert scale that ranges Finally, Pearson’s correlation coefficient and t‑tests were
from 1 = did not apply to me at all to 4 = applied to me conducted to examine the validity of the instruments.
very much or most of the time. Appropriate item scores Convergent validity was examined by correlating scores
were summed for each factor, with higher scores indicating between the IGDS9‑SF and GDT. Concurrent validity was
higher levels of the respective negative emotional state. examined by correlating scores of the IGDS9‑SF and GDT
with scores of the Depression Anxiety Stress Scale. Finally,
Procedure
known‑group validity was examined by comparing scores
The current study is part of a larger study to examine of the IGDS9‑SF and GDT across gender.
gaming behavior among young adults in Singapore.
Participants completed the study online via Qualtrics. Results
Upon providing informed consent, participants completed The sample consisted of 1001 participants (50.15%
a screener question to ensure they played games in the females, 74.43% Chinese, 13.29% Malays, 9.29%
past 12 months and the Background Information Form. Indians, and 3.00% others). Their age ranged from 18
Subsequently, participants completed the IGDS9‑SF,[9] the to 40 years (mean = 28.47, SD = 6.21). The gender
GDT,[16] the Depression Anxiety Stress Scale,[24] and two and ethnic distribution is similar to the Singaporean
instruments unrelated to the aims of the current study. population (51.10% females, 74.30% Chinese, 13.50%
All instruments are in English and were administered in a Malays, 9.00% Indians, and 3.20% others).[32] The males
randomized order to control for fatigue and order effects. consisted of 499 participants (74.55% Chinese, 13.23%
Data collection was conducted and completed in August Malays, 9.22% Indians, and 3.01% others). Their age
2023. This procedure was approved by the first author’s ranged from 18 to 40 years (mean = 29.12, SD = 5.76).
institutional Human Research Ethics Committee (Approval The females consisted of 502 participants (74.30% Chinese,
number: H9100). 13.35% Malays, 9.36% Indians, and 2.99% others). Their
Statistical analyses age ranged from 18 to 40 years (mean = 27.84, SD = 6.57).
The demographic and gaming‑related information of the
The results were analyzed using JASP Version 0.18.1. First, samples are presented in Table 1.
confirmatory factor analyses (CFAs) were conducted to test
the one‑factor model of the IGDS9‑SF and GDT. Given the Confirmatory factor analysis of the Internet Gaming
tendency of the Chi‑square test to be significant in large Disorder Scale–Short‑Form
sample sizes, alternative fit indices like the Comparative A CFA was conducted to test the one‑factor model of
Fit Index (CFI)[25] and the Root Mean Square Error of the IGDS9‑SF (Model 1). While Model 1 provided a
Approximation (RMSEA) were used to evaluate model satisfactory fit to the data, an inspection of the Modification
fit. Specifically, a CFI ≥0.950 and a RMSEA <0.080 were Indices suggested that the model can be improved by
considered indicators of good fit.[26] specifying one error covariance between Items 7 and 9
Second, multigroup CFAs (MGCFA) were conducted (Modification Index = 104.997; Model 1a). This covariance
for the IGDS9‑SF and GDT across gender based on is justified since the items belong to the same factor. The
current best practices.[27‑29] The one‑factor model of results showed that Model 1a represented a better fit to the
the instruments was first tested separately for males data than Model 1. All factor loadings were significant, and
and females. If required, the model was modified they ranged from 0.67 to 0.84 [Figure 1]. The fit indices of
accordingly before being used to test for equivalence the CFA are presented in Table 2.
across gender. Subsequently, the model was tested Multigroup confirmatory factor analyses of the Internet
for measurement invariance in the following order: Gaming Disorder Scale–Short‑Form
configural, metric, scalar, and strict invariance. Similarly,
Two CFAs were conducted to test the one‑factor model of
alternative fit indices like the CFI and RMSEA were
the IGDS9‑SF separately for males and females (Models
used. A ∆CFI <0.010 and a ∆RMSEA <0.015 indicated a
2 and 3, respectively). While Models 2 and 3 provided
nonsignificant change in model fit and provided evidence
a satisfactory fit to the data, an inspection of the
of invariance.[30]
Modification Indices suggested that the models can be
Third, Cronbach’s alpha, McDonald’s omega, and the improved by specifying one error covariance between
standard error of measurement were calculated to examine Items 7 and 9 (Modification Index = 40.193 for males and
the reliability of the instruments. A Cronbach’s alpha >0.70 58.588 for females; Models 2a and 3a). This covariance
and a McDonald’s omega >0.70 were considered is justified since the items belong to the same factor.
acceptable levels of internal consistency. The standard The results showed that Models 2a and 3a represented a
error of measurement was calculated by multiplying better fit to the data than Models 2 and 3, respectively.
the SD by √(1‑Cronbach’s alpha). A standard error of Consequently, this model with one error covariance was
measurement ≤SD/2 was indicative of an acceptable level used to test for equivalence across gender. MGCFA was
of measurement error.[31] conducted to test for measurement invariance in the

Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025 127
 Chew, et al.: Internet gaming disorder

Figure 1: Standardized estimates for the modified one‑factor model of the Internet Gaming Disorder Scale–Short‑Form with one error correlation
(Model 1a)

Table 1: Demographic and gaming‑related information of samples, n (%)

Variables Total sample (n=966–1001) Male sample (n=478–499) Female sample (n=488–502)
Nationality
Singaporean 880 (87.91) 446 (89.38) 434 (86.45)
Permanent resident 121 (12.09) 53 (10.62) 68 (13.55)
Age, mean (SD) 28.47 (6.21) 29.12 (5.76) 27.84 (6.57)
Gender
Male 499 (49.85) ‑ ‑
Female 502 (50.15) ‑ ‑
Ethnicity
Chinese 745 (74.43) 372 (74.55) 373 (74.30)
Malay 133 (13.29) 66 (13.23) 67 (13.35)
Indian 93 (9.29) 46 (9.22) 47 (9.36)
Others 30 (3.00) 15 (3.01) 15 (2.99)
Housing type
1‑Room HDB flat 12 (1.20) 8 (1.60) 4 (0.80)
2‑Room HDB flat 31 (3.10) 15 (2.81) 17 (3.39)
3‑Room HDB flat 275 (27.50) 129 (25.85) 146 (29.14)
4‑Room HDB flat 312 (31.20) 162 (32.47) 150 (29.94)
5‑Room HDB flat 180 (18.00) 87 (17.44) 93 (18.56)
Condominium 159 (15.90) 82 (16.43) 77 (15.37)
Landed properties 28 (2.80) 14 (2.81) 14 (2.79)
Others 3 (0.30) 3 (0.60) 0 (0.00)
Occupation status
Student 177 (17.72) 49 (9.82) 128 (25.60)
Employed 768 (76.88) 431 (86.37) 337 (67.40)
Unemployed 34 (3.40) 10 (2.00) 24 (4.8)
Others 20 (2.00) 9 (1.80) 11 (2.20)
Education level
Below secondary 1 (0.10) 0 1 (0.20)
Secondary 37 (3.76) 19 (3.87) 18 (3.66)
Postsecondary (nontertiary) 59 (6.00) 30 (6.11) 29 (5.89)
Diploma and professional qualification 211 (21.47) 103 (20.98) 108 (21.95)
University 675 (68.67) 339 (69.04) 336 (68.29)
Gaming time, mean (SD)
Weekday 2.98 (2.73) 3.31 (3.15) 2.66 (2.21)
Weekend 4.29 (3.18) 4.80 (3.44) 3.78 (2.81)
The sample sizes varied due to missing data on some variables. SD: Standard deviation

128 Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025
 Chew, et al.: Internet gaming disorder

following order: configural, metric, scalar, and strict scalar, and strict invariances. The fit indices of the MGCFA
invariance. For all levels, the ∆CFI was <0.010 and are presented in Table 3.
the ∆RMSEA was <0.015, providing evidence of metric,
Reliability of the Internet Gaming Disorder Scale–
scalar, and strict invariances. The fit indices of the
Short‑Form and Gaming Disorder Test
MGCFA are presented in Table 2.
The IGDS9‑SF had a Cronbach’s alpha of 0.93, 95%
Confirmatory factor analysis of the Gaming Disorder Test
confidence interval [CI] (0.92–0.93] and a McDonald’s
A CFA was conducted to test the one‑factor model of the omega of 0.93, 95% CI [0.92–0.93]. The GDT had a
GDT. The results showed that the model represented a good Cronbach’s alpha of 0.90, 95% CI [0.89–0.91] and a
fit to the data. All factor loadings were significant, and they McDonald’s omega of 0.90, 95% CI [0.89–0.91]. These
ranged from 0.79 to 0.87[Figure 2]. The fit indices of the statistics were >0.70, providing evidence of internal
CFA are presented in Table 3. consistency. In addition, the standard error of measurement
of both instruments was ≤SD/2, indicating an acceptable
Multigroup confirmatory factor analyses of the Gaming
level of measurement error. The descriptives of the
Disorder Test
variables are presented in Table 4.
Two CFAs were conducted to test the one‑factor model of
Validity of the Internet Gaming Disorder Scale–Short‑Form
the GDT separately for males and females. Both models
and Gaming Disorder Test
represented a good fit to the data. Consequently, this
model was used to test for equivalence across gender. The IGDS9‑SF and GDT share a strong positive
MGCFA was conducted to test for measurement invariance correlation, r (999) = 0.83, P < 0.001. In addition, both
in the following order: configural, metric, scalar, and the IGDS9‑SF and GDT had strong positive correlations
strict invariance. For all levels, the ∆CFI was <0.010 and that ranged from 0.65 to 0.71 with depression, anxiety,
the ∆RMSEA was <0.015, providing evidence of metric, and stress, all P < 0.001. Finally, males had higher

Table 2: Confirmatory factor analyses and multigroup confirmatory factor analyses for the Internet Gaming Disorder
Scale–Short‑Form
Model Baseline Difference
χ2 df P CFI RMSEA (90% CI) ∆χ2 ∆df P ∆CFI ∆RMSEA
Total sample
Model 1 290.662 27 <0.001 0.953 0.099 (0.089–0.109) ‑ ‑ ‑ ‑ ‑
Model 1a 186.308 26 <0.001 00.972 0.078 (0.068–0.089) ‑ ‑ ‑ ‑ ‑
Males
Model 2 163.589 27 <0.001 0.956 0.101 (0.086–0.116) ‑ ‑ ‑ ‑ ‑
Model 2a 123.749 26 <0.001 0.968 0.087 (0.072–0.102) ‑ ‑ ‑ ‑ ‑
Females
Model 3 150.265 27 <0.001 0.947 0.095 (0.081–0.111) ‑ ‑ ‑ ‑ ‑
Model 3a 91.472 26 <0.001 0.972 0.071 (0.055–0.087) ‑ ‑ ‑ ‑ ‑
Configural 215.221 51 <0.001 0.970 0.080 (0.069–0.091) ‑ ‑ ‑ ‑ ‑
Metric 234.338 59 <0.001 0.968 0.077 (0.067–0.088) 19.117 8 0.014 0.002 0.003
Scalar 260.976 68 <0.001 0.964 0.075 (0.066–0.085) 26.638 9 0.002 0.004 0.002
Strict 286.356 78 <0.001 0.962 0.073 (0.064–0.082) 25.38 10 0.005 0.002 0.002
Models 1, 2, and 3: One‑factor model; Models 1a, 2a, and 3a: One‑factor model with one error covariance between items 7 and 9.
CFI: Comparative Fit Index, RMSEA: Root Mean Square Error of Approximation, CI: Confidence interval

Table 3: Confirmatory factor analyses and multigroup confirmatory factor analyses for the Gaming Disorder Test
Model Baseline Difference
χ2 df P CFI RMSEA (90% CI) ∆χ2 ∆df P ∆CFI ∆RMSEA
Total sample 12.289 2 0.002 0.996 0.072 (0.037–0.112) ‑ ‑ ‑ ‑ ‑
Males 4.665 2 0.097 0.998 0.052 (0.000–0.115) ‑ ‑ ‑ ‑ ‑
Females 8.427 2 0.015 0.994 0.080 (0.030–0.139) ‑ ‑ ‑ ‑ ‑
Configural 13.092 3 0.004 0.996 0.082 (0.040–0.130) ‑ ‑ ‑ ‑ ‑
Metric 25.469 6 <0.001 0.992 0.081 (0.050–0.114) 12.377 3 0.006 0.004 0.001
Scalar 34.442 10 <0.001 0.989 0.070 (0.045–0.096) 8.973 4 0.062 0.003 0.011
Strict 50.438 14 <0.001 0.984 0.072 (0.051–0.094) 15.996 4 0.003 0.005 0.002
CFI: Comparative Fit Index, RMSEA: Root Mean Square Error of Approximation, CI: Confidence interval

Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025 129
 Chew, et al.: Internet gaming disorder

Figure 2: Standardized estimates for the one‑factor model of the Gaming Disorder Test

Table 4: Descriptives and intercorrelations of the across gender. This was consistent with previous studies
variables that found measurement invariance across gender for the
Variable 1 2 3 4 5 IGDS9‑SF.[14] In addition, the results extended on previous
1. IGD ‑ studies[16] by providing the first evidence of measurement
2. GD 0.83*** ‑ invariance across gender for the GDT.
3. Depression 0.68*** 0.65*** ‑ Second, the results supported the hypothesis that the
4. Anxiety 0.71*** 0.67*** 0.79*** ‑ IGDS9‑SF and GDT are reliable (H2). The IGDS9‑SF
5. Stress 0.71*** 0.68*** 0.84*** 0.81*** ‑ and GDT had Cronbach’s alpha and McDonald’s omega
Mean 20.67 8.96 8.00 7.78 8.19 of 0.93 and 0.90, respectively. These statistics were >0.70
SD 8.24 4.01 3.39 3.07 3.12 and consistent with previous studies that found a
Cronbach’s alpha 0.93 0.90 0.90 0.83 0.84 Cronbach’s alpha of 0.87 for the IGDS9‑SF[9] and 0.84
McDonald’s omega 0.93 0.90 0.90 0.84 0.85 for the GDT.[16] Furthermore, both instruments had an
SEM 2.18 1.27 ‑ ‑ ‑ acceptable level of measurement error since their standard
SD/2 4.12 2.01 ‑ ‑ ‑ error of measurement were ≤SD/2.[31] Taken together, both
Number of items 9 4 4 4 4
instruments are reliable.
Actual range 9–45 4–20 4–16 4–16 4–16
Potential range 9–45 4–20 4–16 4–16 4–16 Finally, the results supported the hypotheses that the
***P<0.001. GD: Gaming disorder, IGD: Internet GD, SEM: IGDS9‑SF and GDT shared a strong positive correlation
Standard error of measurement, SD: Standard deviation with each other (H3a), were significantly correlated with
depression (H3b), anxiety (H3c), and stress (H3d), and were
scores on the IGDS9‑SF (mean = 22.47, SD = 8.69) higher among males than females (H3e). The IGDS9‑SF
than females (mean = 18.88, SD = 7.35), t (999) = and GDT were strongly and positively correlated with each
7.06, P < 0.001. Males also had higher scores on the other (r = 0.83). This was consistent with previous studies
GDT (mean = 9.79, SD = 4.19) than females (mean = 8.13, that found a correlation of r = 0.83.[16] Both instruments were
SD = 3.65), t (999) = 6.67, P < 0.001. positively correlated with depression, anxiety, and stress
(r ranges from 0.65 to 0.71). This was consistent with previous
Discussion studies on the negative consequences of IGD (e.g. depression
and anxiety).[6] Finally, the finding that males had higher IGD
First, the results of this study supported the hypotheses
and GD than females was consistent with previous studies
that the IGDS9‑SF and GDT would have a unidimensional
that found males at higher risk for IGD than females.[4]
structure (H1a), and there would be measurement invariance
Overall, consistent with psychometric research among Asian
across gender (H1b). Our CFAs showed that the modified samples,[12,19] both instruments demonstrated excellent
one‑factor model of the IGDS9‑SF and the one‑factor psychometric properties in our Singaporean population.
model of the GDT provided a satisfactory fit to the data.
This was consistent with previous studies that found a Limitations
unidimensional structure for the IGDS9‑SF[9,14] and GDT.[16] Limitations of the study should be noted. First, the current
Furthermore, our MGCFAs showed that both instruments sample consisted of adults between 18 and 40 years of
had configural, metric, scalar, and strict invariances age. The methodology is justified since they are a group

130 Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025
 Chew, et al.: Internet gaming disorder

at risk for IGD.[2] However, this precluded an examination Chew PK, et al. Prevalence of internet addiction and gaming
of the structure of the instruments among adolescents and disorders in Southeast Asia: A meta‑analysis. Int J Environ Res
Public Health 2020;17:2582.
older adults. Second, the current study did not examine
4. Chew PK, Wong CM. Internet gaming disorder in the DSM-5:
the clinical validity of the instruments. Consequently, it
Personality and individual differences. Journal of Technology in
is unclear if the instruments, using the diagnostic criteria Behavioral Science. 2022;7:516-23.
for IGD[1] and GD,[15] are able to distinguish between 5. Chew PK. A meta‑analytic review of Internet gaming
problematic and nonproblematic gamers. In the future, disorder and the big five personality factors. Addict Behav
these limitations might be controlled by recruiting samples 2022;126:107193.
with a wider age range and comparing scores on the 6. Wong HY, Mo HY, Potenza MN, Chan MN, Lau WM, Chui TK,
instruments against clinician‑administered assessments et al. Relationships between severity of internet gaming
disorder, severity of problematic social media use, sleep quality
such as a structured clinical interview with people who
and psychological distress. Int J Environ Res Public Health
experience problems with gaming.[33] 2020;17:1879.
7. Krishnan A, Chew PK. Impact of social media addiction and
Conclusion internet gaming disorder on sleep quality: Serial mediation
This study conducted a comprehensive evaluation of analyses. Psychiatr Q 2024;95:185‑202.
the psychometric properties of the IGDS9‑SF and the 8. King DL, Chamberlain SR, Carragher N, Billieux J, Stein D,
Mueller K, et al. Screening and assessment tools for gaming
GDT among Singaporeans, demonstrating its robustness
disorder: A comprehensive systematic review. Clin Psychol Rev
in assessing IGD and GD. Our findings affirm the 2020;77:101831.
unidimensional structure, measurement invariance across 9. Pontes HM, Griffiths MD. Measuring DSM‑5 internet gaming
gender, and reliability for both instruments. Our findings disorder: Development and validation of a short psychometric
also attested to the instruments’ convergent validity and scale. Comput Hum Behav 2015;45:137‑43.
concurrent validity. The results add to the growing body 10. Király O, Sleczka P, Pontes HM, Urbán R, Griffiths MD,
of psychometric evidence of the IGDS9‑SF[14] and provide Demetrovics Z. Validation of the ten‑item internet gaming
disorder test (IGDT‑10) and evaluation of the nine DSM‑5
support for the relatively newer GDT.[16] Given the high
internet gaming disorder criteria. Addict Behav 2017;64:253‑60.
percentage of gamers among Singaporeans,[23] it is timely
11. Lemmens JS, Valkenburg PM, Gentile DA. The internet gaming
and important to have psychometrically valid instruments disorder scale. Psychol Assess 2015;27:567‑82.
to screen for IGD and GD. 12. Leung H, Pakpour AH, Strong C, Lin YC, Tsai MC,
Griffiths MD, et al. Measurement invariance across young
Author Contributions
adults from Hong Kong and Taiwan among three internet‑related
• Peter K. H. Chew: Conceptualization, methodology, addiction scales: Bergen social media addiction scale (BSMAS),
formal analysis, writing ‑ original draft, writing ‑ review smartphone application‑based addiction scale (SABAS), and
internet gaming disorder scale‑short form (IGDS‑SF9) (Study
and editing
Part A). Addict Behav 2020;101:105969.
• Kuhanesan N. C. Naidu: conceptualization, 13. Beranuy M, Machimbarrena JM, Vega‑Osés MA, Carbonell X,
writing ‑ original draft, writing ‑ review and editing Griffiths MD, Pontes HM, et al. Spanish validation of the internet
• Jing Shi: Conceptualization, writing ‑ review and editing gaming disorder scale‑short form (IGDS9‑SF): Prevalence
• Melvyn W. B. Zhang: Conceptualization, and relationship with online gambling and quality of life. Int J
writing ‑ review and editing. Environ Res Public Health 2020;17:1562.
14. Poon LY, Tsang HW, Chan TY, Man SW, Ng LY, Wong YL,
Data availability Statement et al. Psychometric properties of the internet gaming disorder
scale‑short‑form (IGDS9‑SF): Systematic review. J Med Internet
The data are available upon request. Res 2021;23:e26821.
Financial support and sponsorship 15. World Health Organization. ICD‑11. International Classification
of Diseases 11th Revision; 2019. Available from: https://icd.who.
The study was funded by James Cook University’s Internal int/en. [Last accessed on 2021 Apr 21].
Research Grant Scheme (Number: IRG20230005). 16. Pontes HM, Schivinski B, Sindermann C, Li M, Becker B,
Zhou M, et al. Measurement and conceptualization of gaming
Conflicts of interest disorder according to the World Health Organization framework:
There are no conflicts of interest. The development of the gaming disorder test. Int J Ment Health
Addict 2021;19:508‑28.
References 17. Higuchi S, Osaki Y, Kinjo A, Mihara S, Maezono M,
Kitayuguchi T, et al. Development and validation of a nine‑item
1. American Psychiatric Association. Diagnostic and Statistical short screening test for ICD‑11 gaming disorder (GAMES test)
Manual of Mental Disorders. 5th ed. Arlington, VA: American and estimation of the prevalence in the general young population.
Psychiatric Publishing; 2013. J Behav Addict 2021;10:263‑80.
2. Stevens MW, Dorstyn D, Delfabbro PH, King DL. Global 18. Müller SM, Wegmann E, Oelker A, Stark R, Müller A,
prevalence of gaming disorder: A systematic review and Montag C, et al. Assessment of criteria for specific internet‑use
meta‑analysis. Aust N Z J Psychiatry 2021;55:553‑68. disorders (ACSID‑11): Introduction of a new screening
3. Chia DX, Ng CW, Kandasami G, Seow MY, Choo CC, instrument capturing ICD‑11 criteria for gaming disorder

Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025 131
 Chew, et al.: Internet gaming disorder

and other potential internet‑use disorders. J Behav Addict 26. Schreiber JB, Nora A, Stage FK, Barlow EA, King J. Reporting
2022;11:427‑50. structural equation modeling and confirmatory factor analysis
19. Ghazi FR, Gan WY, Tung SE, Chen IH, Poon WC, Siaw YL, results: A review. J Educ Res 2006;99:323‑38.
et al. Problematic gaming in Malaysian university students: 27. Byrne BM. Structural Equation Modeling with AMOS: Basic
Translation and psychometric evaluation of the Malay Concepts, Applications, and Programming. 2nd ed. New York:
language versions of gaming disorder test and gaming disorder Routledge; 2010. p. 416.
scale for young adults. Eval Health Prof 2024;47:93‑104. 28. Milfont TL, Fischer R. Testing measurement invariance across
20. Cudo A, Montag C, Pontes HM. Psychometric assessment and groups: Applications in cross‑cultural research. Int J Psychol Res
gender invariance of the Polish version of the gaming disorder 2010;3:111‑30.
test. Int J Ment Health Addict 2024;22:1333‑56.
29. Vandenberg RJ, Lance CE. A review and synthesis of the
21. Jahrami H, Husain W, Lin CY, Björling G, Potenza MN, measurement invariance literature: Suggestions, practices,
Pakpour A. Reliability generalization meta‑analysis and and recommendations for or ganizational research. Organ Res
psychometric review of the gaming disorder test (GDT):
Methods 2000;3:4‑70.
Evaluating internal consistency. Addict Behav Rep
2024;20:100563. 30. Chen FF. Sensitivity of goodness of fit indexes to lack of
measurement invariance. Struct Equ Modeling Multidiscip J
22. Karhulahti VM, Martončik M, Adamkovič M. Measuring
2007;14:464‑504.
internet gaming disorder and gaming disorder: A qualitative
content validity analysis of validated scales. Assessment 31. Morrow JR, Mood D, Disch J, Kang M. Measurement and
2023;30:402‑13. Evaluation in Human Performance. 5th ed. Illinois: Human
23. YouGov. Half of Singaporean Gamers Gaming More in a COVID Kinetics; 2015. p. 480.
World; 2020. Available from: https://sg.yougov.com/en-sg/ 32. Singapore Department of Statistics. Singapore Census
news/2020/10/29/half-singaporean-gamers-gaming-more-covid- of Population 2020, Statistical Release 1: Demographic
world/. [Last accessed on 2023 Jun 13]. Characteristics, Education, Language and Religion. Singapore:
24. Ali AM, Hori H, Kim Y, Kunugi H. The depression anxiety Author; 2020. Available from: https://www.singstat.gov.sg/
stress scale 8‑items expresses robust psychometric properties as publications/reference/cop2020/cop2020‑sr1/census20_stat_
an ideal shorter version of the depression anxiety stress scale 21 release1. [Last accessed on 2021 Aug 05].
among healthy respondents from three continents. Front Psychol 33. Mihara S, Osaki Y, Kinjo A, Matsuzaki T, Nakayama H,
2022;13:799769. Kitayuguchi T, et al. Validation of the ten‑item internet gaming
25. Bentler PM. Comparative fit indexes in structural models. disorder test (IGDT‑10) based on the clinical diagnosis of IGD in
Psychol Bull 1990;107:238‑46. Japan. J Behav Addict 2022;11:1024‑34.

132 Asian Journal of Social Health and Behavior | Volume 8 | Issue 3 | July-September 2025