Gender Difference in Internet Use and Internet Problems among Quebec High School Students

Abstract

Objectives:

There are presently no data available concerning Internet addiction (IA) problems among adolescents in Canada and the province of Quebec. The goal of this study is thus to document and compare the influence of gender on Internet use and addiction.

Method:

The study data were collected from a larger research project on gambling among adolescents. Activities conducted online (applications used and time spent) as well as answers to the Internet Addiction Test (IAT) were collected from 3938 adolescents from grades 9 to 11. The two most often employed cut-off points for the IAT in the literature were documented: (40-69 and 70+) and (50+).

Results:

Boys spent significantly more time on the Internet than did girls. A greater proportion of the girls made intense use of social networks, whereas a greater proportion of the boys made intense use of massively multiplayer online role-playing games, online games, and adult sites. The proportion of adolescents with a potential IA problem varied according to the cut-off employed. When the cut-off was set at 70+, 1.3% of the adolescents were considered to have an IA, while 41.7% were seen to be at risk. At a 50+ cut-off, 18% of the adolescents were considered to have a problem. There was no significant difference between the genders concerning the proportion of adolescents considered to be at risk or presenting IA problems. Finally, analysis of the percentile ranks would seem to show that a cut-off of 50+ better describes the category of young people at risk.

Conclusions:

The results of this study make it possible to document Internet use and IA in a large number of Quebec adolescents.

Clinical Implications

• With the more severe cut-off point employed, between 1.16% and 1.45% adolescents have an Internet problem. It is therefore important that they have access to treatment.

• A significant proportion of adolescents are probably at risk of developing addiction problems. Careful thought must be given to prevention for these young people. This prevention should be addressed specifically according to gender.

• The percentile ranks seem to indicate that new cut-off points should be employed to determine at-risk categories. The score of 50 (instead of 40) seemed appropriate in this study.

Limitations

• Despite a sizable number of participants, the use of a convenience sample limits the generalization of the results.

• The lack of a gold standard to which we could compare the IAT performance is a serious weak point in this field of research.

The rapid development of software applications and the growing use of the Internet raise numerous questions about the repercussions of this technology in the life of adolescents.^1,2 While most adolescents make reasonable use of the Internet, some lose control and their use becomes so excessive that it is often qualified as Internet addiction (IA).^3,4

There is as of yet no consensus regarding the conceptualization or operationalization of IA.^5–10 Despite these debates, several studies have reported IA prevalences oscillating between 0.6% and 26.7%.^4,10–18 These sizable variations in the prevalences—depending on the studies, the countries, samples, and the evaluation criteria^10,19,20—are of note.

Numerous studies also highlight the differences between genders in terms of both application utilization and the prevalence of IA problems.^4,12,14 The prevalence of IA problems would seem to be higher in boys than in girls.^4,12,14

There are presently no data available concerning IA problems among adolescents in the province of Quebec.¹⁹ These data are of interest considering that the cultural differences associated with young Quebecers may have an impact on Internet use.^21,22 Furthermore, numerous health practitioners wonder whether preventive intervention is needed.^2,23,24 The goal of the present study is to document Internet use and the proportion of Internet addiction according to gender.

Method

Participants

The study data were collected in 2012 during a research project on gambling among adolescents.²⁵ A convenience sampling method was employed for schools selection. Fourteen public and private secondary schools were approached, and 11 agreed to participate. Participants in grades 9, 10, and 11 were invited to complete the questionnaires individually in their classroom for a duration of 50 minutes.

Instruments

For Internet activities, the time spent on 9 applications was documented: social networks (Facebook, Twitter, etc.), YouTube, chatroom (MSN, Myspace, other than Facebook), blogs or discussion forums, news sites (Cyberpresse, Canoe), massively multiplayer online role-playing games or MMORPGs (World of Warcraft, League of Legends), online games with other players (Call of Duty, Counter Strike), downloading music or films, and visiting adult sites (XXX). For each of the applications and excluding time spent for schoolwork, they had to evaluate how many hours they spent online using a 7-point Likert scale (1 = 0 minutes, 2 = less than an hour, 3 = 5 hours, 4 = 15 hours, 5 = 20 hours, 6 = 30 hours, 7 = 40+ hours). Considering that more than 20 hours of Internet per week is high-intensity use,²⁶ time spent online was put into 2 categories: high intensity (20 hours or more) and low intensity (19 hours or less).

Internet addiction was measured with the French version²⁷ of the Internet Addiction Test (IAT).²⁸ This 20-item questionnaire was evaluated using a 5-point Likert scale and the scores varied from 20 to 100. The IAT scale was shown to have an internal consistency coefficient of 0.93. Although this scale has been employed in many studies,²⁹ different cut-off points have been used to determine which people are at risk or addicted.^4,10,27,30 Given the lack of sufficient consensus, we employed the 2 most common ways of categorizing the IAT scores: 1) by using the 3 categories recommended by Young^3,13,28 (no risk [20-39], at risk [40-69] or probable Internet addict [70-100]) and 2) by using 2 categories^27,31–33 (average users [20-49] or problematic users [50 or more]).

Data Analysis

Pearson’s chi-square test and Fisher’s exact test were used to compare the 2 groups’ means and proportions. Percentile scores of the IAT results were calculated for each gender. Due to the sample size, the significance level at <0.0001 was used to reduce the possibility for type II errors.³⁴

Results

We surveyed a total of 3938 high school students, with a response rate of 98%. Participants were 15.3 years old on average (range, 14-21 years old), of whom 43.7% were boys and 56.3% were girls. Almost all participants (88.8%) spoke French at home, and 87.8% were Canadian. Thirty-one percent of the sample were in grade 9, 36.1% in grade 10, and 32.44% in grade 11.

Intensity of Time Spent on Each Internet Activity

On average, the boys spent significantly more time on the Internet each week (19.65 hours, SD = 17.70) than did the girls (16.68 hours, SD = 15.13) (<0.0001).

Comparing the intensity of the time spent on Internet activities, a greater proportion of the girls made intense use of social networks (P < 0.0001), whereas a greater proportion of the boys made intense use of MMORPGs (P < 0.0001), online games (P < 0.0001), and adult sites (P < 0.0001) (see Table 1).

Table 1.

Intensity of Time Spent on Internet Activities Each Week According to Gender.

Applications	Boys, % (n)	Girls, % (n)	P Value
Social networks (n = 3880)
≤19 hours	89.91 (1527)	83.11 (1812)	<0.0001
≥20 hours	10.09 (172)	16.89 (369)
YouTube (n = 3882)
≤19 hours	92.10 (1562)	93.85 (2052)	0.0326
≥20 hours	7.90 (135)	6.15 (133)
Chat (n = 3871)
≤19 hours	96.88 (1635)	95.17 (2077)	0.007
≥20 hours	3.125 (53)	4.83 (106)
Blogs (n = 3871)
≤19 hours	99.29 (1681)	98.95 (2155)	0.25
≥20 hours	0.70 (12)	1.05 (23)
News (n = 3870)
≤19 hours	99.65 (1684)	99.86 (2177)	0.16
≥20 hours	0.35 (6)	0.14 (3)
MMORPGs (n = 3883)
≤19 hours	90.23 (1529)	99.41 (2173)	<0.0001
≥20 hours	9.77 (168)	0.59 (13)
Online games (n = 3874)
≤19 hours	87.30 (1481)	98.85 (2156)	<0.0001
≥20 hours	12.70 (215)	1.05 (22)
Downloading music or films (n = 3876)
≤19 hours	97.13 (1648)	97.90 (2134)	0.1239
≥20 hours	2.87 (49)	2.10 (45)
Adult sites (n = 3873)
≤19 hours	97.22 (1635)	99.77 (2186)	<0.0001
≥20 hours	2.78 (47)	0.23 (5)

MMORPG, massively multiplayer online role-playing game.

Internet Addiction Problems

The mean IAT score was 39.43 (SD = 11.30) for the boys and 40.03 (SD = 11.70) for the girls, with no difference between genders (P = 0.1234). There was no significant difference between gender with regard to the proportion who were considered at risk or who had IA problems, no matter which cut-off point was employed (see Table 2).

Table 2.

Internet Addiction According to Gender and Cut-Off Point Employed.

Cut-Off	Boys		Girls
	% (95% CI)	n	% (95% CI)	n	Total N	P Valuea
IAT: 3 categoriesb
20 to 40: no risk	57.78 (55.43-60.09)	995	56.27 (57.19-58.33)	1243	2238
40 to 69: at risk	41.06 (38.76-43.40)	707	42.28 (40.24-44.35)	934	1641
70 to 100: probable Internet addict	1.16 (0.75-1.79)	20	1.45 (1.03-2.04)	32	52	0.511
IAT: 2 categoriesc
20 to 49: average user	81.53 (79.63-83.29)	1404	78.9 (77.15-80.55)	1743	3147
50 to 100: problematic user	18.47 (16.71-20.37)	318	21.1 (19.45-22.45)	466	784	0.040

CI, confidence interval; IAT, Internet Addiction Test.

^aChi-squared.

^bCut-off with 3 categories according to Young.²⁸

^cCut-off with 2 categories according Khazaal et al.²⁷

Percentile Ranks

The classification standards employed in IA studies^35–37 often set the 95th or 99th percentile as the problem user category, whereas the 75th or 80th percentile is often used to determine the at-risk category. As described in Table 3, the cut-off points corresponding to these percentiles were similar for both genders.

Table 3.

Proposed Cut-Off Criteria of Categories of Internet-Users Based on Percentile.

User Category	Percentile	IAT Score (Boys)	IAT Score (Girls)
At-risk users	75th to 95th	46 to 58	48 to 60
	80th to 95th	48 to 58	50 to 60
Problem users	95th or greater	59+	61
	99th or greater	68+	69

IAT, Internet Addiction Test.

Discussion

To our knowledge, this is the first study to document Internet use and IA problems in a large number of adolescents in Quebec. In comparison with adolescents in other countries, young Quebeckers used the Internet for more hours per week. On average, they spent 17.9 hours compared to weekly averages ranging from 8 to 15 hours elsewhere in the world.^38–40 Moreover, the number of weekly hours spent on the Internet has increased since 2007, when a similar study found that adolescents spent 11 hours per week.⁴¹ The constant arrival of new applications, notably Facebook, which came online in 2008, might explain this difference.

As expected, the boys spent significantly more time on the Internet and used different applications than did the girls. These results were similar to those observed in other European countries where the boys played online more often (MMORPGs and online games) and went to adult sites more often, while the girls were on social networks more often.^17,18,32

Whereas the score of 40 was set as the cut-off point for determining the at-risk group,^3,18,30 it was also the average result for the girls (40.03) and close to the average result for the boys (39.43). These results on the IAT scale call into question the cut-off point proposed by Young²⁸ for the category of at-risk adolescents. The percentile ranks would seem to indicate that a cut-off of 50 would probably better delimit the category of adolescents at risk. As for the cut-off of 70 regularly used in studies,^3,18,30 it approximately corresponds to the 99th percentile for both genders. It is thus quite likely that it detected the adolescents who were in the greatest difficulty due to their Internet use. A clinical validation of the cut-off points should confirm the relevance of setting the points at 50 or 70.

Contrary to some studies,^4,11,12 there was no difference in the proportion of IA among boys and girls. Furthermore, the prevalence in IA problems varies considerably according to the cut-off point employed. When the three categories recommended by Young²⁸ were used, 41.7% of the adolescents were at risk and 1.32% had IA problems. This level of IA prevalence was lower than that observed in most countries,^{4,11–15,17,20,42–45} but the prevalence of the at-risk category was higher than those generally observed.^4,11,12,18 When using the 50+ cut-off point, 19.94% of adolescents were considered problematic, with a prevalence higher than that observed in Italy (5.8%)¹⁶ and in China (8.1%-12.2%)^18,46,47 but lower than that observed among Italian adolescents (36.7%).⁴⁸ The variation in the results according to the cut-off point that was employed highlights the inherent difficulties in conceptualizing and comparing IA prevalence in studies.

Despite the large number of participants, this study was based on a convenience sample, which cannot be considered representative of all Quebec adolescents. Other studies with representative samples of the Quebec and Canadian population will have to be conducted to arrive at a more accurate portrait of IA problems.

Regardless of this study’s limitation, these results would seem to indicate that a small proportion of adolescents has an IA problem and should thus have access to treatment. Furthermore, a large number of adolescents would seem to be at risk of developing addiction problems. Longitudinal studies will be necessary to better understand these young people and determine whether they are truly at risk.^10,22 These observations raise questions about the prevention that should be provided for these young people. Given that their Internet use is different, gender-specific prevention activities should be developed.