exploring the psychometric properties of the english ... · pdf file5 mohsin raza1, spogmai...
TRANSCRIPT
Exploring the psychometric properties of the English versionof the Internet Addiction Test in the Pakistani population: across-sectional surveyAhmed Waqas, Faisal Farooq, Anum Bhatti, Saamia Javed, Mahrukh Elahi Ghumman, Mohsin Raza, Spogmai Khan, Waqas Ahmad
Introduction: Despite growing concerns over pathological use of the Internet, studiesbased on validated psychometric instruments are still lacking in Pakistan. The presentstudy aimed to examine the psychometric properties of Young’s Internet Addiction Test(IAT) in a sample of the Pakistani population. We examined the validity, internalconsistency, readability and floor and ceiling effects of IAT scores.Methods: This cross-sectional study was conducted at CMH Lahore Medical College and Institute of Dentistry,Lahore, Pakistan from 1 March 2015 to 30 May 2015. A total of 522 medical and dentalstudents completed the questionnaire, which consisted of three sections: (a) demographicsand percentage grades in annual examinations, (b) a categorical question to record theestimated number of hours spent on the Internet per day, and (c) the English version ofthe IAT. All data were analyzed in SPSS v. 20. Principal axis factor analysis was used tovalidate the factor structure of the IAT in our study sample. An alpha coefficient > .7 wassought in the reliability analysis. Histograms and the values of skewness and kurtosis wereanalyzed for floor and ceiling effects. In addition, readability of the IAT was assessed as theFlesch Reading Ease score and Flesch-Kincaid Grade level function. Results: A total of 522medical and dental students participated in the survey. Most respondents were femalemedical students enrolled in preclinical years of their degree program. Median age (min-max) of the respondents was 20 years (17-25 years). A single-factor model for IAT scoreexplained 33.71% of the variance, with a high alpha coefficient of .893. In addition, the IAThad good face and convergent validity and no floor and ceiling effects, and was judgedeasy to read by participants. Conclusion: The English version of the IAT showed goodpsychometric properties in a sample of Pakistani university students. A single-factor modelfor assessing internet addiction showed good reliability and was found suitable with ourstudy sample.
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Exploring the psychometric properties of the English version of the Internet Addiction Test
2 in the Pakistani population: a cross-sectional survey
3
4 Ahmed Waqas1, Faisal Farooq1, Anum Bhatti2, Saamia Tahir Javed1, Mahrukh Elahi Ghumman1,
5 Mohsin Raza1, Spogmai Khan1, Waqas Ahmad1
6
7 Affiliations: 1Undergraduate students, CMH Lahore Medical College and Institute of Dentistry,
8 Lahore Cantt, Pakistan
9 2Undergraduate student, Portland State University, Oregon, USA
10
11 Running title: Psychometric properties of Internet Addiction Test in the Pakistani population
12
13 Corresponding author:
14 Ahmed Waqas
15 Email: [email protected]
16 Contact #: +92-03434936117
17 Address: CMH Lahore Medical College and Institute of Dentistry, Lahore Cantt, Pakistan
18
19 Abstract word count: 319
20 Manuscript word count: 2569 (excluding tables and references)
21 Funding: none
22 Conflicts of interest: none
23
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
24 Abstract
25 Introduction:
26 Despite growing concerns over pathological use of the Internet, studies based on validated
27 psychometric instruments are still lacking in Pakistan. The present study aimed to examine the
28 psychometric properties of Young’s Internet Addiction Test (IAT) in a sample of the Pakistani
29 population. We examined the validity, internal consistency, readability and floor and ceiling
30 effects of IAT scores.
31 Methods:
32 This cross-sectional study was conducted at CMH Lahore Medical College and Institute of
33 Dentistry, Lahore, Pakistan from 1 March 2015 to 30 May 2015. A total of 522 medical and
34 dental students completed the questionnaire, which consisted of three sections: (a) demographics
35 and percentage grades in annual examinations, (b) a categorical question to record the estimated
36 number of hours spent on the Internet per day, and (c) the English version of the IAT. All data
37 were analyzed in SPSS v. 20. Principal axis factor analysis was used to validate the factor
38 structure of the IAT in our study sample. An alpha coefficient > .7 was sought in the reliability
39 analysis. Histograms and the values of skewness and kurtosis were analyzed for floor and ceiling
40 effects. In addition, readability of the IAT was assessed as the Flesch Reading Ease score and
41 Flesch-Kincaid Grade level function.
42 Results:
43 A total of 522 medical and dental students participated in the survey. Most respondents were
44 female medical students enrolled in preclinical years of their degree program. Median age (min-
45 max) of the respondents was 20 years (17-25 years). A single-factor model for IAT score
46 explained 33.71% of the variance, with a high alpha coefficient of .893. In addition, the IAT had
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
47 good face and convergent validity and no floor and ceiling effects, and was judged easy to read
48 by participants.
49 Conclusion:
50 The English version of the IAT showed good psychometric properties in a sample of Pakistani
51 university students. A single-factor model for assessing internet addiction showed good
52 reliability and was found suitable with our study sample.
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
70
71
72 Introduction
73 There has been tremendous growth in the number of Internet users in Pakistan, from 0.1% of the
74 total population in 2000 to 10.9% (over 20 million) in 2013 1. This sudden rise in Internet usage
75 has attracted much attention from mental health experts interested in exploring the potential
76 harms associated with its excessive or pathological use.
77 The problematic, uncontrollable and impulsive use of the Internet has been
78 conceptualized as a behavioral addiction comparable to pathological gambling 2. The etiology of
79 these behavioral addictions is rooted in biological processes such as conditioned learning and the
80 brain’s reward system. Hence, certain behaviors elicit short-term rewards, promote continuous
81 behavior and diminish control over behavior 3. Such behaviors are being recognized as a
82 compulsive-impulsive spectrum disorder 2 entailing at least three domains: gaming, pornography
83 and emailing/text messaging 4. According to Shapira et al., the diagnostic criteria for Internet
84 addiction include (a) maladaptive and excessive use of the Internet for longer times than planned,
85 (b) significant impairment in social, occupational and other domains of functioning, and (c)
86 excessive use that cannot be accounted for by any axis I disorders 5.
87 Previous studies have increasingly delineated the harmful effects of problematic use of
88 the Internet in different populations. Such use has been associated with significant psychosocial
89 impairment leading to low work performance, relationship problems, loneliness, self-destructive
90 behaviors and several psychiatric disorders such as depression, anxiety, social phobias and
91 Attention Deficit Hyperactivity Disorder5–11. Recently, a number of cross-sectional and
92 longitudinal studies have explored the association of Internet use with axis II disorders. An
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
93 interesting study by Floros and colleagues reported a high comorbidity of axis II disorders
94 including personality disorders and use of impaired ego defenses in a clinical sample of people
95 with Internet addiction disorder (IAD) 12. However, in several longitudinal studies, the evidence
96 for the direction of a causal relationship between IAD and axis I and II disorders is not yet clear
97 10–12, and more evidence is required.
98 Because of the potential harms of Internet addiction, internet gaming disorder has been
99 incorporated into section III of the Diagnostic and Statistical Manual of Mental Disorders fifth
100 edition (DSM-5) 13. A new diagnostic category of behavioral addictions including Internet
101 addiction is also being introduced in the International Classification of Diseases (ICD-11) 14.
102 At present, there are no estimates of the prevalence of IAD in the Pakistani population.
103 However, according to Jadoon et al., 44% of Pakistani medical students are regular Internet users
104 15. Despite growing concerns over pathological use of the Internet, studies based on validated
105 psychometric instruments are still lacking in Pakistan. Therefore, the paucity of validated tools to
106 study IAD in the Pakistani population warranted the present study. This study aimed to examine,
107 in a sample of the Pakistani population, the psychometric properties of the Internet Addiction
108 Test (IAT) devised by Young 9. We studied the validity, internal consistency, readability and
109 floor and ceiling effects of IAT scores.
110 The IAT continues to be one of the most extensively used validation tools to study IAD
111 in different populations, and in both clinical and nonclinical settings. This test was selected for
112 analysis because it has shown excellent psychometric properties in many languages, cultures,
113 ethnicities and countries including China 16, Germany 17, France 18, Sweden, Lebanon 19, Greece
114 20 and Bangladesh 21. However, validations in different settings have led to heterogeneous factor
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
115 structures of the IAT, ranging from a one-factor structure to as many as six factors 16–21,
116 sometimes giving rise to different constructs.
117
118 Methods
119 Study design
120 This cross-sectional study was conducted at CMH Lahore Medical College (CMH LMC) and
121 Institute of Dentistry, Lahore, Pakistan from 1 March, 2015 to 30 May, 2015. Ethical approval
122 was sought from and granted by the Ethical Review Committee of CMH LMC. A total of 550
123 questionnaires were distributed among medical and dental students enrolled in all years of the
124 medical or dental degree program. Respondents were selected with a random sampling approach
125 using computer software. All respondents read and signed a consent form, and were ensured
126 anonymity and that only group findings would be reported.
127 The questionnaire consisted of three sections: (a) demographics of the respondents, (b) a
128 categorical question to document the estimated number of hours spent on the Internet per day,
129 and (c) the IAT developed by Young 2,22. This instrument has shown excellent psychometric
130 properties 16–21 in a variety of settings. It consists of 20 items that investigate the respondent’s
131 potentially problematic use of the Internet and disruption in psychosocial functioning 22.
132 Responses are recorded on a 6-point Likert scale of frequencies, ranging from “does not apply”
133 (0) to “always” (5). For purposes of analysis, a global score is obtained by adding the scores for
134 the responses to each item.
135
136 Pilot survey
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
137 Before starting the survey, a pilot study was conducted at CMH LMC in a sample of 20 medical
138 students selected by convenience sampling. We received positive comments from the
139 participants that the English version of the IAT was easily comprehensible. Therefore, we did not
140 feel the need to translate it into Urdu (the official language of Pakistan).
141
142 Sample size calculation
143 Sample size calculations are generally based on expected effect sizes and variability in the study
144 sample. Both of these were unknown before we started the study; so we relied on “rules of
145 thumb” to calculate the sample size for factor analysis. Comrey and Lee suggested samples of
146 500 or more for factor analysis studies 23. According to their rating scale, a sample size of 100 is
147 considered poor, 200 fair, 300 good, 500 very good, and 1000 or more as excellent. Therefore,
148 we aimed for a sample size of at least 500 respondents as recommended by Comrey and Lee 23.
149
150 Data analysis
151 All data were analyzed in SPSS v. 21 (IBM Chicago, IL, USA). Frequencies and were calculated
152 for demographic variables, and descriptive statistics were obtained for total scores on the IAT.
153 Assumptions of normality and floor and ceiling effects of IAT scores were verified by plotting
154 histograms and Q-Q plots. The percentages of individuals with the lowest and highest possible
155 IAT scores were recorded to examine floor and ceiling effects, and values >20% were considered
156 significant.
157 Exploratory factors with principal axis factoring and quartimax rotation were studied to
158 analyze the factor structure of the IAT. The suitability of principal axis factoring for our
159 purposes was determined with the following criteria: correlation coefficient >0.3 for all
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
160 variables, Kaiser-Meyer-Olkin (KMO) value greater than 0.6, and a statistically significant
161 Barlett test of sphericity (P < 0.05). The maximum number of components to be retained was
162 determined with a Cattell’s scree plot, eigenvalues >1, an interpretability criterion, the amount of
163 variance explained, and reliability analysis. Factor loading values > 0.3 were sought. Internal
164 consistency of the IAT was evaluated with Cronbach’s alpha reliability analysis, and an alpha
165 coefficient of > .70 was considered acceptable 24. Item total correlations were analyzed with
166 Pearson’s product moment correlation coefficient, and values between 0.2 to 0.8 were sought 24.
167 Convergent validity of the IAT was evaluated by analyzing the association between the number
168 of hours spent on the Internet per day and the IAT score. Participants were asked, “How many
169 hours do you estimate you spend on the Internet each day?”. Responses were recorded on the
170 following scale; (a) <30 min to 1 hour, (b) >1 hour to 3 hours, (c) >3 to 6 hours or more. This
171 association was analyzed with one-way analysis of variance (ANOVA). A similar study that
172 analyzed the convergent validity of the IAT in a Greek population sample used the same method
173 to determine convergent validity 20.
174 Readability of the questionnaire was recorded as the Flesch Reading Ease score and
175 Flesch-Kincaid Grade level function 25. Associations between gender, residence, year of study,
176 degree program and respondents’ IAT scores were analyzed with independent sample t-tests.
177
178 Results
179 Of the 550 questionnaires distributed, 522 (94.90%) were returned. Most of the respondents were
180 female medical students enrolled in preclinical years of their degree program. Median age (min-
181 max) of the respondents was 20 years (17-25 years). Most students were average users of the
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
182 Internet (from >1 to 3 hours daily). Detailed results for demographic characteristics and Internet
183 use are shown in Table 1.
184 The independent sample t-test showed that male students scored significantly higher on
185 the IAT than females. According to one-way ANOVA and post-hoc least significant difference
186 (LSD) tests, IAT scores were positively associated with hours spent using the Internet. Detailed
187 results are given in Tables 1 and 2.
188 The IAT demonstrated easy readability, with a Flesch reading ease value of 71.1 (rated as
189 fairly easy) and a Flesch-Kincaid level of 7.1. Respondents indicated that items 1, 2, 12, 14, 16
190 and 17 were the easiest to understand, whereas items 3, 4, 19, 20 were the hardest.
191
192 Construct validity
193 The overall KMO value for sample adequacy was 0.920, which is classified as meritorious which
194 according to Kaiser’s criteria (1974). The Bartlett test of sphericity was significant (P < 0.001).
195 Therefore, the data in the present study were suitable for exploratory factor analysis. Inspection
196 of the correlation matrix showed that all variables had at least one correlation coefficient greater
197 than 0.3.
198 Initially, principal axis factor analysis with the quartimax method extracted three
199 components/factors with an eigenvalue >1 and a high degree of cross-loading of statements on
200 different factors. Most of the variance in IAT scores was explained by the first factor (33.71%),
201 while the second factor was associated with a modest variance of 7.68%, and the third with a
202 variance of 5.79%. However, the lack of theoretical underpinning for this three-factor structure,
203 visualization of the Cattell’s scree plot (Figure 1) together with the modest variance explained by
204 second and third factors favored a unidimensional model of the IAT. Factor loadings of
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
205 the rotated three-component solution for the 20-item IAT based on eigenvalues greater than 1 are
206 presented in Table 3. Factor loadings for the unrotated one-factor structure of the test are given
207 in Table 4.
208
209 Normality and floor and ceiling effects
210 Mean IAT score was 43.8. Inspection of skewness (−0.030, std. error = 0.107), kurtosis (−0.486,
211 std. error = 0.213) and the histogram (Figure 2) revealed that IAT scores did not deviate
212 significantly from normality, and no floor or ceiling effects were found in IAT scores for the
213 present sample.
214
215 Reliability analysis
216 The IAT consists of 20 items. Cronbach’s alpha value for the one-factor structure of the IAT was
217 .893, which indicates excellent reliability of this tool in the present study sample. Item total
218 statistics for the IAT are detailed in Table 5. All items had corrected item correlations greater
219 than 0.3, thus exhibiting the same construct. To further analyze internal consistency, item total
220 correlations adjusted for overlap were calculated for each item; these values ranged from 0.32 to
221 0.60, which reflect substantial and moderate correlation.
222
223 Discussion
224 The English version of the IAT was judged easy to read, and had good face, content and
225 convergent validity and high internal reliability in our sample of Pakistani medical and dental
226 students. In addition, our analysis detected no floor or ceiling effects in IAT scores.
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
227 Initial exploratory factor analysis revealed a three-factor structure based on Kaiser’s
228 criterion of eigenvalues >1. However, the factor structure of the IAT was not very clear cut, with
229 a few items cross-loading on different factors. The three-factor structure for the IAT lacks any
230 theoretical underpinning, therefore the Cattell’s scree plot together with the modest variance
231 explained by the second and third factors favor a unidimensional factor structure for the IAT.
232 These results are in consonance with the factor structures of the Arabic 19, French 26 and
233 Portuguese 27 versions of this test. However, studies of the psychometric properties of the IAT in
234 Germany 17, Korea 28, USA 29, Bangladesh 21, Malaysia 30, Italy 31, Greece 20 and China 16 have
235 reported heterogeneous factor structures for the IAT. Factor models ranging from two to six
236 components have been proposed in different studies with different factor loadings and constructs,
237 albeit with very high internal consistency for this instrument. These discrepancies may be due to
238 the use of different factor analysis techniques, and/or to variations in demographics, culture and
239 age groups of the respondents. As noted, sample sizes ranged from as low as 151 postgraduate
240 and undergraduate medical students in Greece 20 to as high as 1882 respondents in a validation
241 study of the IAT in Germany 17. The respondents in most studies were college and university
242 students; however, sampling strategies differed in some studies such as the German study 17, in
243 which respondents completed an online questionnaire as well as an offline sample, a study of
244 students in the USA who were contacted through Facebook29, and research in China16 and
245 Lebanon19, which involved adolescent male and female students. As noted in the literature, most
246 studies used principal components analysis, although others used exploratory factor analysis and
247 confirmatory factor analysis to validate the factor structure of the IAT in their respective
248 population samples. Regardless of the number of factors that have been extracted to date, all
249 studies reported high internal consistency with alpha values ranging from 0.89 21 to 0.93 26. Table
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
250 6 summarizes the factor models that have been proposed in different studies, highlighting the
251 population characteristics, methods and results of the different factor analysis techniques.
252 Our sample included medical and dental students from Pakistan, which is a
253 predominantly Muslim country. Therefore, we expected findings similar to those reported in
254 other Muslim populations. For example, as reported in an Arab population, our analysis revealed
255 that item 4 had a low inter-item correlation – a finding that highlights the cultural restrictions
256 regarding premarital relationships in Muslim cultures 19. Similarly, item 1 had the lowest inter-
257 item correlation (0.323), indicating that the number of hours spent online is not a strong
258 determinant of the IAT score, and other factors such as impairment in psychosocial life might
259 play a greater role than has been recognized thus far. A validation study of the IAT in a Bengali
260 population identified four sub-scales, i.e. ‘Neglect of duty’, ‘Online dependence’, ‘Virtual
261 fantasies’, and ‘Privacy and self- defense’21.
262 Male students scored higher on the IAT than their female counterparts, probably because
263 males are more likely to engage in cyber-sexual behavior, online gaming and gambling –
264 behaviors that are likely to affect their psychosocial health 32. Similar trends in the scores for
265 male and female respondents were observed in the Bengali population 21.
266
267 Limitations
268 This study evaluated the psychometric properties of the English version of the IAT in a sample
269 of Pakistani medical and dental students. The test demonstrated fairly easy readability and
270 comprehensibility in our study sample. But because most of the general population is not
271 proficient in the English language, further studies are advised to test the psychometric properties
272 of the translated Urdu version of the IAT. Our sample comprised randomly selected respondents
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
273 from a single medical school; therefore, future studies should include a more diverse study
274 sample. Moreover, comparator instruments were not included in the present study to establish the
275 criterion validity of the IAT; therefore further efforts to address this issue are needed.
276
277 Conclusion
278 The English version of the Internet Addiction Test demonstrated good psychometric properties in
279 a sample of the Pakistani population. However, future studies are encouraged to assess the
280 psychometric properties of the Urdu version of this instrument in the general Pakistani
281 population.
282
283 Acknowledgment
284 The authors thank K. Shashok (AuthorAID in the Eastern Mediterranean) for improving the use
285 of English in the manuscript.
286
287 References
2881. The World Bank. World Development Indicators: International Telecommunication Union,
289 World Telecommunication/ICT Development Report and database, and World Bank estimates.
290 [Internet]. 2013. [cited 2015 Jul 26]. Available from:
291 http://data.worldbank.org/indicator/IT.NET.USER.P2?cid=GPD_44
2922. Young K. Internet Addicition : Diagnosis and Treatment Considerations. Contemporary
293 Psychotheraphy. 2009. p. 241–6.
2943. Bergmark KH, Bergmark A, Findahl O. Extensive internet involvement-addiction or emerging
295 lifestyle? Int J Environ Res Public Health. 2011;8(12):4488–501.
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
2964. Block J. Pathological internet use in the USA. International symposium on counseling and
297 treatment of youth internet addiction; Seoul, Korea. 2007. p. 433.
2985. Shapira N, Lessig M, Goldsmith T, Szabo S, Lazoritz M, Gold M, et al. Problematic internet use:
299 proposed classification and diagnostic criteria. Depress Anxiety. 2003;17(4):207–16.
3006. Chen Y-F, Peng S. University students’ internet use and its relationships with academic
301 performance, interpersonal relationships, psychosocial adjustment, and self-evaluation.
302 CyberPsychology Behav. 2008;11:467–9.
3037. Leung L. Net-generation attributes and seductive properties of the Internet as predictors of online
304 activities and Internet addiction. CyberPsychology Behav. 2004;7:333–48.
3058. Skinner H, Biscope S, Poland B. Quality of internet access: barrier behind internet use statistics.
306 Soc Sci Med. 2003;57(5):875–80.
3079. Young K. Internet addiction: a new clinical phenomenon and its consequences. Am Behav Sci.
308 48:402–15.
30910. Gentile DA, Choo H, Liau A, Sim T, Li D, Fung D. Pathological video game use among youths:
310 A two-year longitudinal study. Pediatrics. 2011;127(2):e319–29.
31111. Ko C-H, Yen J-Y, Chen C-S, Yeh Y-C, Yen C-F. Predictive values of psychiatric symptoms for
312 internet addiction in adolescents: A 2-year prospective study. Arch Pediatr Adolesc Med.
313 2009;163(10):937.
31412. Floros G, Siomos K, Stogiannidou A, Giouzepas I, Garyfallos G. Comorbidity of Psychiatric
315 disorders with Internet addiction in a clinical sample: the effect of personality, defense style and
316 psychopathology. Addict Behav.
31713. American Psychiatric Association (APA). Diagnostic and statistical manual of mental disorders
318 (DSM-5). 2013;
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
31914. Mann K, Fauth-Bühler M, Seiferth N, Heinz A; Expertengruppe Verhaltenssüchte der DGPPN.
320 The concept of behavioral addiction and limits of the term addiction. Nervenarzt.
321 2013;84(5):548–56.
32215. Jadoon NA, Zahid MF, Mansoorulhaq H, Ullah S, Jadoon BA, Raza A, et al. Evaluation of
323 internet access and utilization by medical students in Lahore, Pakistan. BMC Med Inform Decis
324 Mak [Internet]. BioMed Central Ltd; 2011;11(1):37. Available from:
325 http://www.biomedcentral.com/1472-6947/11/37
32616. Lai CM, Mak KK, Watanabe H, Ang RP, Pang JS, Ho RCM. Psychometric properties of the
327 internet addiction test in chinese adolescents. J Pediatr Psychol. 2013;38(7):794–807.
32817. Barke A, Nyenhuis N, Kröner-Herwig B. The German Version of the Internet Addiction Test: A
329 Validation Study. Cyberpsychology, Behav Soc Netw. 2012;15(10):534–42.
33018. Barrault S, Varescon I. Psychopathologie des joueurs pathologiques en ligne: Une ??tude
331 pr??liminaire. Encephale [Internet]. L’Enc?®phale; 2012;38(2):156–63. Available from:
332 http://dx.doi.org/10.1016/j.encep.2011.01.009
33319. Hawi NS. Arabic validation of the Internet addiction test. Cyberpsychol Behav Soc Netw
334 [Internet]. 2013;16(3):200–4. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23441685
33520. Tsimtsiou Z, Haidich AB, Kokkali S, Dardavesis T, Young KS, Arvanitidou M. Greek version of
336 the Internet Addiction Test: A validation study. Psychiatr Q. 2014;85(2):187–95.
33721. Karim A. KMR, Nigar N. The internet addiction test: Assessing its psychometric properties in
338 Bangladeshi culture. Asian J Psychiatr [Internet]. Elsevier B.V.; 2014;10:75–83. Available from:
339 http://dx.doi.org/10.1016/j.ajp.2013.10.011
34022. Young KS. Caught in the net: how to recognize the signs of Internet addiction–and a winning
341 strategy for recovery. Hoboken: Wiley.;
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
34223. Comrey AL, Lee H. A first course in factor analysis. Hillsdale, NJ: Erlbaum. 1992.
34324. Streiner D. Health measurement scales: a practical guide to their development and use. 4th ed.
344 Oxford University Press; 2008.
34525. Flesch R. A new readability yardstick. J Appl Psychol. 1948;32:221–33.
34626. Khazaal Y, Billieux J, Thorens G, Khan R, Louati Y, Scarlatti E, et al. French validation of the
347 internet addiction test. Cyberpsychol Behav. 2008;11(6):703–6.
34827. Pontes HM, Patrão IM, Griffiths MD. Portuguese validation of the Internet Addiction Test : An
349 empirical study. 2014;3(2):107–14.
35028. Lee K, Gyeong H, Yu B, Song YM, Lee HK, Kim D. Reliability and validity of the Korean
351 version of the internet addiction test among college students. J Korean Med Sci. 2013;28(5):763–
352 8.
35329. Jelenchick L a., Becker T, Moreno M a. Assessing the psychometric properties of the Internet
354 Addiction Test (IAT) in US college students. Psychiatry Res. 2012;196(2-3):296–301.
35530. Chong Guan N, Isa SM, Hashim a. H, Pillai SK, Harbajan Singh MK. Validity of the Malay
356 Version of the Internet Addiction Test: A Study on a Group of Medical Students in Malaysia.
357 Asia-Pacific J Public Heal. 2012;
35831. Faraci P, Craparo G, Messina R, Severino S. Internet Addiction Test (IAT): Which is the best
359 factorial solution? J Med Internet Res. 2013;15(10).
36032. Cao H, Sun Y, Wan Y, Hao J, Tao F. Problematic Internet use in Chinese adolescents and its
361 relation to psychosomatic symptoms and life satisfaction. BMC Public Health [Internet]. BioMed
362 Central Ltd; 2011;11(1):802. Available from: http://www.biomedcentral.com/1471-2458/11/802
363
364
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 1(on next page)
Demographic characteristics of respondents and mean Internet Addiction Test scores (n= 522)
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Table 1. Demographic characteristics of respondents and mean Internet Addiction Test scores (n
2 = 522)
Variable Frequency
(n)
Median (min-
max)
Mean IAT
score (SD)
Statistical
value
Gender Male 194 (37.2%) - 48.96 (15.9) t value =
5.41
Female 328 (62.8%) - 40.77 (17.2)
Median age (min-max) 20 (17-25) 43.82 (4-95) -
Study year Preclinical 296 (56.7%) - 43.28 (16.7) t value =.8
Clinical 226 (43.3%) - 44.52 (17.8)
Degree MBBS 414 (79.3%) - 45.02 (17.4) t value = 3.1
BDS 108 (20.7%) - 39.19 (15.35)
Residence* Off-campus 267 (51.1%) - 44.27 (16.6) t value =.54
On-campus 253 (48.5%) - 43.46 (17.5)
Hours spent surfing
Internet per day*
<30 min – 1 hour 142 (27.2%) - 37.07 (17.3) F value=
25.691
>1 hour – 3 hours 220 (42.1%) - 43.40 (15.4)
>3 – 6 or more
hours
157 (30.1%) - 50.67 (16.9)
Level of addiction Average user 320 (61.3%) 20-49 Χ2 = 2881
Frequent problems 193 (37%) 50-79
Significant
problems
6 (1.1%) 80-100
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
3 *Missing values 2 or more, 1 denotes P < 0.001
4
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 2(on next page)
Post-hoc least significant difference test for hours spent surfing the Internet.(Dependent variable: Internet Addiction Test scores)
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Table 2. Post-hoc least significant difference test for hours spent surfing the Internet. (Dependent
2 variable: Internet Addiction Test scores)
3
4
95% Confidence interval(I) hours spent (J) hours spent Mean
difference (I-J)
Std. error Sig.
Lower bound Upper bound
>1 – 3 hours -6.32803* 1.76834 0.000 -9.8021 -2.8540<30 min – 1 hour
>3 – 6 hours or more -13.59694* 1.90243 0.000 -17.3344 -9.8595
<30 min – 1 hour 6.32803* 1.76834 0.000 2.8540 9.8021>1 – 3 hours
>3 – 6 hours or more -7.26891* 1.71624 0.000 -10.6406 -3.8972
<30 min – 1 hour 13.59694* 1.90243 0.000 9.8595 17.3344>3 – 6 hours or more
>1 –3 hours 7.26891* 1.71624 0.000 3.8972 10.6406
*Mean difference significant at the 0.05 level.
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 3(on next page)
Factor matrix for Internet Addiction Test scores in a sample of Pakistani medical anddental school students (n = 522)
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 3. Factor matrix for Internet Addiction Test scores in a sample of Pakistani medical and
dental school students (n = 522)
Factor
1 2 3
Item 1: How often do you find that you stay on-line longer than you
intended?0.349 0.461
Item 2: How often do you neglect household chores to spend more time
on-line?0.498 0.348
Item 3: How often do you prefer the excitement of the Internet to
intimacy with your partner?0.437
Item 4: How often do you form new relationships with fellow on-line
users?0.416 −0.366
Item 5: How often do others in your life complain to you about the
amount of time you spend on-line?0.549
Item 6: How often do your grades or school work suffers because of the
amount of time you spend on-line?0.609
Item 7: How often do you check your email before something else that
you need to do?0.418
Item 8: How often does your job performance or productivity suffer
because of the Internet?0.542 0.393
Item 9: How often do you become defensive or secretive when anyone
asks you what you do on-line?0.574
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Item 10 : How often do you block out disturbing thoughts about your life
with soothing thoughts of the Internet?0.583
Item 11 : How often do you find yourself anticipating when you will go
on-line again?0.563
Item 12: How often do you fear that life without the Internet would be
boring, empty, and joyless?0.483
Item 13: How often do you snap, yell, or act annoyed if someone bothers
you while you are on-line?0.641
Item 14: How often do you lose sleep due to late-night log-ins? 0.605
Item 15: How often do you feel preoccupied with the Internet when off-
line, or fantasize about being on-line?0.637
Item 16: How often do you find yourself saying “just a few more minutes”
when on-line?0.602
Item 17: How often do you try to cut down the amount of time you spend
on-line and fail?0.565
Item 18: How often do you try to hide how long you’ve been on-line? 0.628
Item 19: How often do you choose to spend more time on-line over going
out with others?0.611
Item 20: How often do you feel depressed, moody or nervous when you
are off-line, which goes away once you are back on-line?0.640
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 4(on next page)
Unrotated factor solution for Internet Addiction Test scores (n = 522)
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Table 4. Unrotated factor solution for Internet Addiction Test scores (n = 522)
Factor
Item 1 0.338
Item 2 0.487
Item 3 0.431
Item 4 0.405
Item 5 0.549
Item 6 0.601
Item 7 0.418
Item 8 0.530
Item 9 0.576
Item 10 0.586
Item11 0.565
Item 12 0.484
Item 13 0.641
Item 14 0.605
Item 15 0.634
Item 16 0.596
Item 17 0.565
Item 18 0.629
Item 19 0.606
Item 20 0.637
Extraction method: Principal axis factor analysis
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
2
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 5(on next page)
Item total statistics for statements in the Internet Addiction Test (n = 522)
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Table 5. Item total statistics for statements in the Internet Addiction Test (n = 522)
Item Scale mean if item
deleted
Scale variance if
item deleted
Corrected item
total correlation
Squared multiple
correlation
Cronbach’s alpha if
item deleted
Item 1 39.41 270.579 0.323 0.272 .894
Item 2 39.77 266.092 0.471 0.344 .889
Item 3 41.08 266.434 0.406 0.286 .891
Item 4 41.06 270.416 0.375 0.291 .892
Item 5 40.33 262.256 0.522 0.311 .888
Item 6 40.39 262.554 0.573 0.425 .886
Item 7 40.36 268.442 0.401 0.215 .891
Item 8 40.54 264.702 0.501 0.386 .888
Item 9 40.57 262.794 0.547 0.370 .887
Item 10 40.20 259.364 0.556 0.371 .887
Item 11 40.46 262.782 0.531 0.346 .887
Item 12 39.66 263.897 0.457 0.278 .890
Item 13 40.46 260.269 0.598 0.426 .885
Item 14 40.08 257.881 0.575 0.398 .886
Item 15 40.63 261.108 0.589 0.418 .886
Item 16 39.62 259.275 0.562 0.399 .886
Item 17 40.12 262.735 0.531 0.353 .887
Item 18 40.45 259.193 0.586 0.418 .886
Item 19 40.67 261.338 0.564 0.425 .886
Item 20 40.68 259.375 0.593 0.468 .885
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
2
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
Table 6(on next page)
Comparison of psychometric properties of different versions of the Internet AddictionTest
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1 Table 6. Comparison of psychometric properties of different versions of the Internet Addiction
2 Test
Version Characteristics of
study sample
Factor analysis method Criteria for
number of
factors
Names of factors, % variance
and reliability
Arabic19 817 intermediate
and secondary
school students,
mean age 15 (2.12)
years
PCA with oblimin
rotation
Parallel analysis and
Velicer’s minimum
average partial (MAP)
test
Confirmatory factor
analysis
Two
components
with eigenvalue
>1 but only one
of these
retained
One factor, 40.64%, α= .921
Malay30 162 undergraduate
medical students,
mean age 19 (0.19)
years
Principal component
analysis with varimax
rotation method
Five factors
with eigenvalue
>1
Lack of control, Neglect of duty,
Social relationship disruption,
Problematic use, email primacy,
63.84%, Factor-wise alpha
values: .55 – .89, For all items:
.91
Banglades
h21
177 internet users,
mean age 22.33
(2.01) years
Principal component
(PC) with varimax
rotation
Four based on
Cartell scree
plot
Neglect of duty, Online
dependence, Virtual fantasies,
and Privacy and
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
self-defense
55.68%, Cronbach’s alpha = .89
for the IAT, and .60 – .84 for the
factors
Italian31 485 college
students, mean age
24.05 (SD 7.3)
years
Principal axis factoring
with oblique rotation
(promax criterion)
Parallel
analysis, scree
plot, eigenvalue
>1
Emotional and cognitive
preoccupation with the Internet
and
Loss of control and interference
with daily life,
One factor model = 36.18%
Two factor model = 42.15%,
Alpha values for one-factor
solution (Cronbach’s alpha =
.91), and the two-factor
solution
(Cronbach’s alpha = .88 and
Cronbach’s alpha = .79)
Chinese16 844 Hong Kong
Chinese
adolescents
(37.7% boys),
mean age 15.9
(standard deviation
Confirmatory factor
analysis results
indicated 18-item
second-order
three-factor model
- Withdrawal
and social problems, time
management and performance,
and reality substitute. A total of
80.5%, 94.7% and 95.9% of the
total variances
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
3.5) years of the second-order factor were
accounted for.
Cronbach’s alpha = .93).
Factor wise = .87, .86 and
.70
Korean28 279 college
students at a
national university
Principal components
analysis with varimax
rotation
Eigenvalue >1 Excessive use, Dependence,
Withdrawal, and Avoidance of
reality, 58.91%, α= .91
Portugese
27
593 Portuguese
students, average
age 19.9 (SD =
2.7) years
Confirmatory factor
analysis, with robust
maximum-likelihood
estimates (MLR)
- One factor, α = .90
Greek20 151 postgraduate
and undergraduate
medical students
Exploratory factor
analysis with varimax
rotation
Visual
examination of
a scree plot and
eigenvalues >1
‘Psychological/Emotional
conflict, Time management and
Neglect work, 55.3%, α= .91
USA29 215 Undergraduate
students selected
through Facebook
Exploratory factor
analysis with varimax
rotation
Scree plot and
eigenvalues >1
Dependent use and Excessive
use, 91%, α= .90 – .93
German17 Online (ON)
sample (n= 1041,
age 24.2 – 7.2
EFA with varimax
rotation
Horn’s parallel
analysis
Emotional and cognitive
preoccupation with the Internet
and Loss of control and
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
years) and Offline
[OF] sample, n =
841, age: 23.5 –3.0
years
interference with daily life,
46.7% (ON) and 42.0% (OF), α
= .91 (ON) and α = .89 (OF)
French26 246 adults, age:
mean 24.11,
standard deviation
9, range 18–54
years)
Exploratory factor
analysis, confirmatory
factor analysis
Velicer’s
minimum
average partial
(MAP) test
One factor, 45%, α = .93
3
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
1Cattell's scree plot for Internet Addiction Test scores in a sample of Pakistani medicaland dental students
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015
2Histogram of the distribution of Internet Addiction Test scores in a sample of Pakistanimedical and dental students
PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1531v1 | CC-BY 4.0 Open Access | rec: 24 Nov 2015, publ: 24 Nov 2015