Professional Documents
Culture Documents
Internet Gambling Is Common in College Students and Associated With Poor Mental Health
Internet Gambling Is Common in College Students and Associated With Poor Mental Health
Internet Gambling Is Common in College Students and Associated With Poor Mental Health
This study evaluated prevalence of Internet gambling A growing literature indicates that pathological
and its association with pathological gambling and mental gambling, regardless of preferred form of gambling, is
health status in college students. The South Oaks Gambling
Screen and General Health Questionnaire were adminis- associated with financial difficulties, emotional distress,
tered to 1356 undergraduates, and 23% reported ever and psychiatric problems.1,11–13 In a survey of more than
gambling on the Internet, with 6.3% reporting Internet 43,000 respondents, pathological gambling was associated
gambling weekly. Almost two-thirds (61.6%) of regular with increased odds of every psychiatric disorder assessed,
Internet gamblers were pathological gamblers, compared even after controlling for demographics that may be
with 23.9% of infrequent Internet gamblers and 5.0% of
non-Internet gamblers. Internet gambling frequency was linked to those disorders.2 Mental health status and
significantly associated with poor mental health, after con- well-being may also be impacted by pathological gam-
trolling for demographics and pathological gambling. These bling in college students. Furthermore, Internet gambling,
data call for prevention and treatment efforts of Internet a sedentary and isolating activity, may place one at risk
gambling in students. (Am J Addict 2007;16:325–330) for mental health problems. However, an association
between Internet gambling participation and mental
health status has rarely been evaluated using standardized
Pathological gambling is characterized by continued gam- questionnaires.
bling in spite of negative consequences. Prevalence rates To our knowledge, this study is the first to report upon
of pathological gambling range from 0.4 to 1.9% in gen- rates of Internet gambling in a large sample of college stu-
eral population studies.1–3 Among college students, rates dents. Goals were to compare Internet and non-Internet
of pathological gambling may be even higher, with gamblers in terms of prevalence rates of pathological
Shaffer, Hall and Vander Bilt4 reporting 3–6% of gambling and indices of global mental health status.
students suffering from pathological gambling. Recent One hypothesis was that Internet gambling would be
studies report rates of pathological gambling to be as high fairly common among students, and Internet gamblers,
as 3% to 9% in college students,5–8 although some studies especially those with frequent Internet wagering, would
find rates of pathological gambling in students to be have high rates of pathological gambling. A second hypo-
consistent with general population rates.9 thesis was that Internet gambling would be associated
Gambling appears to have risen with legalization and with poor mental health.
availability of betting opportunities,4 and today’s youth
has grown up exposed to a variety of gambling options.
One type of gambling that is now widely accessible is METHOD
Internet gambling, with more than 2,500 gambling Web Participants
sites online.10 As college students regularly use the Inter- Participants (N ¼ 1,356) were students recruited in
net for schoolwork and recreation, Internet gambling classroom settings, near the cafeterias, and during general
may be high in this group. Little data exist about rates screenings at three university campuses between March
of Internet gambling in young adults, the largest users 2005 and May 2006. University Institution Review Boards
of the Internet. approved the study.
325
15210391, 2007, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1080/10550490701525673 by Indian Institution Of Mgmt - Indore, Wiley Online Library on [19/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
data on study refusal rates were not collected, RAs for variables putatively related to mental health, including
reported few refusals (<20%). Comparisons of survey pathological gambling itself.
response data revealed than <1% of demographic data To further isolate the effects of Internet gambling fre-
had all similar responses (eg, age, gender, year, ethnicity), quency, analyses were repeated with lifetime frequencies
and the bulk of the data were collected on nine screening of other types of wagering entered in the analyses: one
days (and in different classrooms). Thus, duplicative sub- that occurred at similar rates as Internet gambling (dice
jects are unlikely to be included in the database. betting) and one that is fairly common (lottery). In addi-
tion, overall frequency of past two-month gambling
Measures (including all forms) was entered into the analyses to
ascertain if general gambling frequency was associated
The questionnaire included items about basic demo-
with mental health status. Data were analyzed using
graphics and lifetime participation in several forms of
SPSS, and two-tailed alphas <.05 were considered statis-
gambling, including Internet gambling. The South Oaks
tically significant.
Gambling Screen (SOGS) was included, with scores ran-
ging from 0–20 and scores >5 indicative of lifetime prob-
able pathological gambling.14 Using a cutoff of 5, the hit
rate is 0.98, sensitivity is 0.95, specificity is 0.996, false RESULTS
positive rate is 0.004, and false negative rate is 0.05.15,16
Overall, 1043 (76.9%) of students had never wagered
Test-retest reliability is 0.71.14 Cronbach’s alpha internal
on the Internet; 141 (10.4%) tried Internet gambling 1
consistency in this sample was 0.85. Additional questions
to 10 times in their lives, 86 (6.3%) gambled on the
asked about recent gambling frequency and expenditures
Internet >10 times but never as often as weekly, 52
and desire for gambling treatment.
(3.8%) reported weekly or more frequent yet less than
The General Health Questionnaire was included in the
screening instrument. This 12-item scale assesses global daily Internet gambling, and 34 (2.5%) stated they
gambled on the Internet everyday. Participants endorsing
physical and mental health, with items asking how one
the second or third categories were combined into one
has been feeling in the past few weeks according to a
group (infrequent Internet gamblers, n ¼ 227), and those
four-point scale.17 By summing responses (and inversing
endorsing the last two categories into a frequent Internet
scores for reverse coded items), normally distributed total
gambling group (n ¼ 86).
scores of 0–36 are obtained, with higher scores reflecting
The three groups differed in terms of some demo-
poorer health. Internal consistency in this sample was 0.89.
graphics (see Table 1). Internet gamblers were more likely
to be male, Caucasian, and from one campus than non-
Data Analysis Internet gamblers. They also had lower grade point
The item inquiring about lifetime Internet gambling averages and higher incomes than non-Internet gamblers.
participation was rated on a five-point Likert scale: never, Internet gamblers were also more likely to endorse
1–10 times, more than 10 times but less than weekly, lifetime gambling problems, as noted by higher SOGS
weekly or more than once a week but never daily, or scores and greater proportions scoring in the probable
daily. Participants were classified into never Internet gam- pathological gambling range. Large proportions of Inter-
blers, infrequent Internet gamblers (‘‘1–10 times’’ and net gamblers, especially frequent Internet gamblers,
‘‘more than 10 times but never weekly’’), and frequent wagered more than $100 in the past two months in
Internet gamblers (combining the last two categories). contrast to fewer non-Internet gamblers. Frequency of
Differences among groups with respect to demographics gambling overall also differed, such that frequent Internet
were assessed using chi-square tests for categorical data, gamblers wagered more often than other groups in the past
ANOVA for continuous normalized data, and Kruskal- two months when all types of gambling were considered.
Wallis tests for data that could not be normalized. In the univariate analyses, age, gender, and pathologi-
General linear models (GLM) evaluated group differ- cal gambling status were all significantly associated with
ences on GHQ scores, with Internet gambling status GHQ scores, F(1,1257) ¼ 6.30, 11.05, and 10.56, respec-
(never, infrequent, frequent) as an independent variable. tively, p < .02. Women had higher GHQ scores than
Analyses controlled for age, gender, recruitment site, men, indicating worse functioning (means and 95% con-
ethnicity (Caucasian versus other), and pathological fidence intervals (CI) of 13.7 (12.5–15.2) and 11.8 (10.9–
gambling status, with age entered as a continuous variable 12.6), respectively), and older age was also associated
and all others as fixed. Two-way interactions (Internet with higher scores. Probable pathological gamblers had
gambling status by site, by gender, by pathological gam- significantly higher GHQ scores than non-pathological
bling status, and by ethnicity) were included to determine gamblers, with mean scores of 13.7 (12.6–14.8) and 12.0
if relationships varied by Internet gambling status within (10.9–13.0), respectively. Neither ethnicity ( p > .33) nor
groups. Thus, analyses examined the effects of Internet campus recruitment site ( p > .98) was significantly asso-
gambling frequency on mental health, after controlling ciated with GHQ scores.
Gamble on Gamble on
Never gambled Internet but Internet at least Statistic
on Internet <weekly weekly (df) p
N 1,043 227 86
Recruitment site, %(n) v2(2) ¼ 5.31 <.001
Campus A 77.9% (813) 69.2% (157) 38.4% (33)
Campus B 18.6% (194) 29.1% (66) 48.8% (42)
Campus C 3.5% (36) 1.8% (4) 12.8% (11)
Age 21.2 4.5 20.8 2.7 20.8 3.1 F(2,1352) ¼ 1.30 .27
Female, %(n) 66.3% (691) 38.3% (87) 11.6% (10) v2(2) ¼ 141.23 <.001
Ethnicity, %(n) v2(10) ¼ 29.39 <.001
Caucasian 65.4% (679) 64.8% (147) 77.9% (67)
African-American 16.7% (173) 15.9% (36) 7.0% (6)
Hispanic 8.3% (86) 8.4% (19) 2.3% (2)
Native American 0.1% (1) 0.9% (2) 2.3% (2)
Asian 3.5% (36) 4.0% (9) 8.1% (7)
Other 6.2% (64) 6.2% (14) 2.3% (2)
Years of education 13.9 1.5 14.1 1.3 14.0 1.4 F(2,1279) ¼ 1.64 .19
Grade point average v2(8) ¼ 17.77 <.05
3.5 to 4.0 23.6% (238) 15.0% (33) 24.7% (21)
3.0 to 3.4 37.0% (373) 40.9% (90) 30.6% (26)
2.5 to 2.9 29.2% (294) 36.8% (81) 34.1% (29)
2.0 to 2.4 8.7% (88) 4.5% (10) 9.4% (8)
Below 2.0 1.5% (15) 2.7% (6) 1.2% (1)
Single marital status, %(n) 91.5% (954) 92.5% (210) 93.0% (80) v2(2) ¼ 0.46 .79
Yearly personal income, %(n) v2(8) ¼ 23.93 <.01
Under $5,000 53.8% (548) 40.7% (90) 34.9% (30)
$5,000–10,000 28.2% (287) 33.9% (75) 40.7% (35)
$10,001–$25,000 14.0% (143) 20.8% (46) 18.6% (16)
$25,001–$50,000 3.0% (31) 3.2% (7) 3.5% (3)
More than $50,000 0.9% (9) 1.4% (3) 2.3% (2)
$ spent gambling past two months %(n) v2(8) ¼ 496.60 <.001
$0 51.0% (522) 16.0% (36) 1.2% (1)
$1–$10 23.9% (245) 19.1% (43) 0.0% (0)
$11–$49 12.8% (131) 27.1% (61) 13.1% (11)
$50–$100 7.2% (74) 20.9% (47) 13.1% (11)
$101–$500 3.9% (40) 12.4% (28) 48.8% (41)
More than $500 1.2% (12) 4.4% (10) 23.8% (20)
Times gambled in past two v2(8) ¼ 523.46 <.001
months in any form, %(n)
None 50.5% (519) 15.6% (35) 0.0% (0)
1–2 times 31.3% (322) 28.1% (63) 2.4% (2)
3–5 times 9.9% (10.2) 27.2% (61) 12.7% (10)
6–10 times 3.8% (39) 14.7% (33) 15.9% (13)
>10 times 4.5% (46) 14.3% (32) 69.5% (57)
SOGS score median (IQ range) 0.0 (1.0) 2.0 (4.0) 5.0 (5.0) v2(2) ¼ 323.70 <.001
Probable pathological gamblers %(n) 5.0% (52) 23.9% (54) 61.6% (53) v2(2) ¼ 282.50 <.001
General health questionnaire score 10.6 5.0 11.7 5.2 12.9 4.7 F(2,1277) ¼ 9.94 <.001
Note. Values represent means standard deviations unless otherwise stated. Numbers may not add to group sample size due to missing responses.
IQ ¼ Interquartile range.