Hispanic Adolescents Research Paper

Journal of Child & Adolescent Substance Abuse, 21:273–291, 2012
Copyright # Taylor & Francis Group, LLC
ISSN: 1067-828X print=1547-0652 online
DOI: 10.1080/1067828X.2012.700851

Screening for Alcohol Risk in Predominantly
Hispanic Youths: Positive Rates and
Behavioral Consequences
JOE TOMAKA, REBEKAH A. SALAIZ, STORMY MORALES-MONKS, and SHARON THOMPSON
University of Texas at El Paso, El Paso, TX, USA

SARAH MCKINNON
University of Texas at Austin, Austin, TX, USA

KATHLEEN O’ROURKE
University of South Florida, Tampa, FL, USA

The present study examined relationships between CAGE alcohol risk scores and predisposing factors for alcohol use, current alcohol use, and behavioral consequences in a large sample of secondary students. Students completed the
CAGE, measures of demographics, potential predisposing factors, and consequences of alcohol use.
More than 18 % of students screened positive for potential alcohol risk using traditional CAGE criteria, and another 23 % scored moderate risk using a more liberal criterion. CAGE scores were related to a variety of predisposing factors and were strongly related to current drinking and alcohol-related behavioral consequences. It was recommended that investigators examine multiple options for appropriate alcohol screening instruments.
KEYWORDS alcohol, CAGE, Hispanic, youth

The project was supported by two grants from the Center for Border Health Research
(Paso del Norte Health Foundation of El Paso, Texas) and the Substance Abuse and Mental
Health Services Administration (TI 11715).
Address correspondence to Joe Tomaka, Department of Public Health Sciences, College of Health Sciences, 500 W. University Avenue, University of Texas at El Paso, El Paso, TX
79968, USA. E-mail: [e-mail]
273

274

J. Tomaka et al.

INTRODUCTION
Alcohol abuse among youths remains a public health concern. Indeed, surveys indicate more than 131.3 million Americans ages 12 and older use alcohol (Substance Abuse and Mental Health Services Administration, 2011).
Alcohol use and abuse among secondary school students is widespread, with about 30% to 45% of all secondary school students (Chen, Yi, Williams, &
Faden, 2009) and 43% of all seniors (Monitoring the Future, 2008) having reported alcohol use within the past 30 days. Even higher rates have been reported in a predominantly Hispanic sample whereby 84% of secondary school students from all grade levels reported consuming alcohol at least once during the past 30 days (McKinnon, O’Rourke, Thompson, & Berumen,
2004).
Although college-based intervention programs have successfully used screening instruments to determine eligibility for intervention programs, the present investigation examined if such screening procedures generalize to secondary school samples. Specifically, the present study was conducted to examine the validity of the CAGE in a high school sample by assessing the relationship between CAGE risk scores and multiple factors associated with alcohol use in high school (i.e., predisposing factors, current drinking behaviors, alcohol-related behavioral consequences).
Heavy alcohol consumption, or ‘‘binge drinking,’’ is defined as the consumption of five or more drinks in one sitting for men; four or more drinks for women (National Institute on Alcohol Abuse and Alcoholism [NIAAA],
2002). Estimates of binge drinking in secondary school vary. For example,
Chen and colleagues (2009) found that 18% of males and 15% of females ages 16 to 17 years reported binge drinking during the past 30 days. These numbers rise to 41% and 30%, respectively, for men and women ages 18 to 20 years. In a sample of secondary school students along the border between the United States and Mexico, McKinnon and colleagues (2004) found 44% of secondary school students reported binge drinking over the same time period. These rates were higher than state (i.e., Texas) and national averages (31% and 30%, respectively).
Alcohol use by individuals ages 12 to 20 years is illegal, and contributes to short- and long-term health and social problems including increased accidents, poor academic performance, increased aggression and violence, and unprotected sexual activity (Hingson, Heeren, Winter, & Wechsler, 2005). Binge drinking is particularly risky, having been associated with a range of negative personal, health, and financial consequences (Hingson et al., 2005; McCarty et al., 2004). Motor vehicle accidents associated with alcohol use are the most serious consequence of adolescent drinking, accounting for nearly 3,000 deaths per year (National Highway Traffic Safety Administration [NHTSA],
2009) and remaining the number one cause of death among teens (Centers

Screening for Alcohol Risk

275

for Disease Control and Prevention, 2010). Unintentional injuries are a related consequence of adolescent drinking as well, with one study showing that 42% of all emergency department visits by adolescents were alcohol-related (Drug
Abuse Warning Network, 2006). Personal and social consequences include diminished academic performance and greater truancy rates among drinkers
(Almodovar, Tomaka, Thompson, McKinnon, & O’Rourke, 2006; Arata,
Stafford, & Tims, 2003). Binge drinking also increases the likelihood of other risky and deviant behaviors including fighting, acquaintance rape, suicides, and homicides (Bailey, Baker, Webster, & Lewin, 2004).
The consequences of adolescent drinking carry an enormous cost not only for the individuals themselves, but also for the community and society as a whole. For example, across all ages, the National Institute on Alcohol
Abuse and Alcoholism estimates that alcohol and substance abuse contribute to more than 100,000 deaths (Mersy, 2003) and that alcohol-related crashes in the United States cost more than $230.6 billion per year (Blincoe et al., 2002).
Early alcohol use can also have lasting consequences. For example, individuals who used alcohol before 15 years of age were four times more likely to become dependent on alcohol, or abuse alcohol, than those who started later (Grant & Dawson, 1998), and age at first consumption of an alcoholic drink has been shown a reliable predictor of subsequent consumption behaviors and drinking-related problems (Almodovar et al., 2006). Early alcohol use among adolescents may also serve a gateway function, leading to the use of other substances such as cigarettes and marijuana (Pacula, 1998).
Most approaches to intervention with those who are already engaging in risky drinking and experiencing alcohol-related problems rely on screening instruments to efficiently assess risk and determine appropriateness for intervention. Brief and frequently self-administered instruments typically include questions assessing the frequency and quantity of alcohol consumption, binge drinking, and experience of alcohol-related problems. Common instruments include the Conjoint Screening Test, or CAGE, developed to detect alcohol use disorders among young adults; the TWEAK, developed for young women of childbearing ages; or the Alcohol Use Disorders Identification
Test, or AUDIT, a self-administered questionnaire used for individuals with heavy drinking problems (Ewing, 1984; Russell, 1994; Saunders, Aasland,
Babor, De La Fuente, & Grant, 1993; Stewart & Connors, 2004=2005).
One frequently used scale, the CAGE, consists of four screening questions which have been described as nonjudgmental and nonconfrontational (Harwood, 2005). Used with both adults and adolescents over the age of 16 years, and administered either verbally or in writing, the instrument evaluates four areas related to lifetime alcohol use. Original items include the following: 1. Have you ever felt the need to Cut down on your drinking?
2. Have you ever felt Annoyed by someone’s criticizing your drinking?

276

J. Tomaka et al.

3. Have you ever felt bad or Guilty about your drinking?
4. Have you ever needed a drink first thing in the morning to steady your nerves and get rid of a hangover (Eye-opener)?
Two or more affirmative (i.e., yes) responses suggest the possibility of alcohol abuse disorder in an adult population and the need for further assessment and possible intervention.
Shields and Caruso (2004) comprehensively reviewed the literature regarding the reliability of the CAGE across multiple studies. Although they report that authors frequently do not report reliability estimates, their review of 13 studies of 22 distinct samples found that the CAGE had a mean reliability coefficient of .74 ranging from .52 to .90. They suggested that this estimate puts the CAGE on par with other screening instruments, such as the AUDIT, especially when relative length (i.e., redundant items increase reliability) is taken into account. Similarly, Laux, Salyers, and Kotova (2005) found the
CAGE to have a reliability coefficient of .74 in a recent college sample.
The CAGE has also been shown valid in multiple samples including the elderly (Jones, Lindsey, Young, Soltys, & Farani-Enayat, 1993), attendees at a medical clinic (King, 1986), and via computer administration (Barry &
Fleming, 1990). Among adolescent samples, Couwenbergh, van der Gaag,
Koeter, de Ruiter, and van den Brink (2009) recently showed both self-report and parent-report versions of the CAGE to have good diagnostic accuracy for predicting substance use disorder assessed via DSM-IV clinical diagnoses among adolescents ages 12 to 18 years. Knight, Sherritt, Harris, Gates, and
Chang (2003) also examined the sensitivity and specificity of the CAGE in relation to the Adolescent Diagnostic Interview in an adolescent sample ages
14 to 18 years. Although it performed less well than some other instruments, the CAGE demonstrated good specificity but limited sensitivity (sensitivity was improved by using a cutoff score of one affirmative response, rather than the typical two affirmative responses).
The CAGE has been widely used in studies of college students. Foote,
Wilkens, and Vavagiakis (2004), for example, examined alcohol screening procedures at 249 university student health centers and found that nearly
12% of the centers reported using the CAGE. To date, however, studies have not assessed its use in predominantly Hispanic samples.

Purpose and Hypotheses
The present investigation examined the validity of the CAGE screening instrument for identifying alcohol risk in a predominantly Hispanic secondary school sample. Specifically, the present study examined relationships between CAGE risk scores and multiple factors associated with alcohol use in high school, including predisposing risk factors, current drinking behaviors, and alcohol-related behavioral consequences.

Screening for Alcohol Risk

277

METHOD
This study reflects a secondary analysis of data previously collected by
McKinnon and colleagues (2004). The purpose of the original study was to determine the rates of alcohol consumption and alcohol-related risk behaviors among a sample of U.S. secondary school students living on the U.S. and Mexico border. The McKinnon et al. study did not report data relating to CAGE scores.

Participants
A total 1,366 students from 16 different secondary schools throughout El Paso
County, Texas, participated in an anonymous self-administered survey assessing alcohol use and alcohol-related behaviors. The study sample included public, private, and parochial schools from all geographic areas of the county. The final sample had relatively similar numbers of boys (N ¼ 639) and girls (N ¼ 732). The participants’ ages ranged from 13 to 19 years, with an average age of 15.8 (SD ¼ 1.17). The final sample consisted of 389
(28%) freshmen, 495 (36.4%) sophomores, 280 (20.7%) juniors, and 191
(14.1%) seniors (11 did not report grade level). Most students identified themselves as Hispanic (70.1%), followed by white, non-Hispanic (20.5%), and 10% self-identifying as ‘‘other’’ regarding racial category.

Measures
PREDISPOSING FACTORS
Although the data are cross-sectional, six retrospective items reflect influences that logically would have preceded any regular drinking by students and were classified as predisposing factors. These included levels of maternal and paternal education, eligibility to receive free or reduced school lunch (as a gross estimation of socioeconomic status), estimated age at consumption of first alcoholic drink, and participation in extra-curricular activities. Levels of parental education were coded as 1 ¼ elementary and=or some high school,
2 ¼ high school graduate and=or some college, or 3 ¼ graduated college or better. Eligibility for free or reduced lunch was coded 0 or 1 for not eligible and eligible, respectively. Finally, specific extracurricular activities included athletics, band=orchestra, choir, drama=speech, drill team=cheerleader, student government, student newspaper or yearbook, academic clubs or societies, work study, and other clubs, each coded 0 or 1 for relative participation.
Total activities were summed to produce a single estimate of extent of participation in extracurricular activities. Individuals were classified into one of four participation levels (none, 1 activity, 2 activities, and 3 or more activities). Finally, one question assessed parental approval of alcohol

278

J. Tomaka et al.

consumption: ‘‘How do your parents feel about kids your age drinking alcohol?’’ Five response options ranged from ‘‘strongly disapprove’’ to ‘‘strongly approve.’’ CURRENT DRINKING BEHAVIORS
Four items assessed drinking behaviors within the past 30 days. The first question assessed frequency of drinking by asking, ‘‘During the past 30 days, on how many days did you have at least 1 drink of alcohol?’’ The second question assessed the number of average drinks on drinking days by asking,
‘‘On the days you drank, about how many drinks did you drink on average?’’
The third question assessed the occurrence of drinking five or more drinks:
‘‘During the past 30 days, on how many days did you have 5 or more drinks of alcohol in a row, that is, within a couple of hours?’’ The fourth question assessed drinking to intoxication by asking, ‘‘In the past 30 days, how often did you drink enough to get drunk?’’
BEHAVIORAL CONSEQUENCES
Twelve items assessed behavioral consequences of alcohol consumption.
Three items assessed drinking and driving including the frequency of riding with a driver who had been drinking (any alcohol), personal drinking and driving (any alcohol), and driving after consuming five or more drinks. All were assessed on a 5-point scale of increasing behavioral frequency with
0 ¼ none=never, 1 ¼ one time, 2 ¼ two or three times, 3 ¼ four or five times, and 4 ¼ six or more times.
Students also completed nine items addressing various consequences of drinking. All came from standardized surveys including the Youth Risk
Behavior Surveillance (Grunbaum et al., 2002) and the Texas Commission on Alcohol and Drug Abuse (1988). All items used a standard time frame and explicitly linked the outcome to drinking (e.g., ‘‘Since the beginning of the school year have you ever missed class because you had been drinking?’’). Each was assessed along 4-point scales of increasing behavioral frequency including 0 ¼ never, 1 ¼ one or two times, 2 ¼ three or more times, and 3 ¼ five or more times. Based on a previous factor analysis of these data
(Almodovar et al., 2006), we calculated three coherent subscales including after-effects of drinking (Cronbach’s alpha ¼.75), academic difficulties (alpha
¼.78), and unplanned or unprotected sex (alpha ¼.72). After-effects included having a hangover, doing something you regretted, forgetting where you were or what you did, hurting oneself, and arguing with friends. Academic problems included behaviors such as missing a class, performing poorly on a test or project, and falling behind in schoolwork. Unplanned or unprotected sex included engaging in unplanned sexual activity or not using protection during sexual intercourse.

Screening for Alcohol Risk

279

Procedure
Students completed questionnaires during the Fall and Spring semesters of 2001. The surveys were conducted in El Paso, Texas. As noted, the study included schools from all areas of the city. A total of 1,920 surveys were distributed, with a total of 1,366 completed for a response rate of
71%. To enhance reliability and validity of responding, questions were based on nationwide and standardized surveys including the YBRSS and the Texas School Survey of Substance Abuse (see McKinnon et al.,
2004). The questionnaire was administered in a classroom setting, by graduate research assistants. Teachers were not present in the classroom during the survey. Informed consents were obtained through written notices from parents. Anonymity of participation was maintained by requesting students not to place their names or the name of their school on the survey.

RESULTS
Descriptive Statistics and Overall CAGE Scores
The first column of Table 1 contains descriptive statistics for all study variables. The table also contains comparisons between girls and boys (using analysis of variance or chi-square tests as appropriate to the level of measurement). As shown, the sample averaged just less than 16 years of age with girls significantly older than boys. Most students were sophomores or freshmen.
Analysis of CAGE scores showed an average score of less than 1.00.
However, almost one-fifth (18.3%, N ¼ 243) of the secondary school students surveyed scored positive for alcohol risk using the traditional CAGE cutoff score of two or more positive responses (Ewing, 1984).
An additional
22.9% (N ¼ 303) scored one. Mean CAGE scores did not differ significantly between boys and girls.
Among the predisposing factors, the percentages for father’s and mother’s education showed that the majority of parents had high school or greater education levels, but that a significant number had less than a high school education. Reflecting the relatively low overall socioeconomic status of the sample, more than half of the students qualified for free or reduced lunch. Students on average participated in about two extracurricular activities, with girls having slightly higher participation rates than boys. Among those students who had ever consumed alcohol, the average age at first drink was approximately 13 years of age, with girls reporting significantly later alcohol initiation than boys. Finally, students reported a mean of 1.72 (on a 5-point Likert-type scale) for parental approval of alcohol consumption, a number that indicates that most students felt their parents disapproved of kids their age drinking alcohol.

280

J. Tomaka et al.

TABLE 1 Descriptive Statistics for Study Variables and Comparisons Between Boys and Girls
Variable

Overall
Boys
Girls
Mean (SD)=% Mean (SD)=% Mean (SD)=%

Demographic variables
Age
15.8 (1.17)
Grade level
Freshman (N ¼ 392)
28.8%
Sophomore (N ¼ 495)
36.4%
Junior (N ¼ 281)
20.7%
Senior (N ¼ 192)
14.1%
CAGE score (0–4)
.67 (.94)
Score of 0 (N ¼ 779)
58.7%
Score of 1 (N ¼ 303)
22.9%
Score of 2 or greater (N ¼ 242)
18.3%
Predisposing factors
Father’s education
Elementary and=or some
25.7%
high school
Graduated high school=
38.4%
some college
Graduated college or greater
35.9%
Mother’s education
Elementary and=or some
28.5%
high school
Graduated high school=
42.9%
some college
Graduated college or greater
28.6%
.54 (.50)
Qualify for free=reduced luncha
Number of extracurricular
1.83 (1.67) activities Age of first drink (years)
12.9 (2.53)
Parental approval
1.72 (.91)
Current drinking behaviors
Frequency of drinking (past 30
2.38 (4.40) days) Number of drinks on drinking
2.73 (4.24) days 5 or more drinks (past 30 days) 1.83 (4.29)
Drinking to intoxication (past
.70 (1.33)
30 days)
Behavioral consequences
Ride with someone who was
1.17 (1.73) drinking Drive after any drinking
3.51 (.94)
Drive after 5 or more drinks
.18 (.76)
Total alcohol-related problems
1.91 (2.57)
Aftereffects of drinking
1.20 (1.56)
Academic problems
.37 (.81)
Unplanned=unprotected sex
.27 (.61)

F or v2
F ¼ 4.72Ã

15.7 (1.15)

15.9 (1.18)

42.1%
53.1%
46.4%
41.8%
.62 (.93)
62.5%
19.2%
18.4%

57.8%
46.8%
53.5%
58.1%
.70 (.94)
55.4%
26.2%
18.5%

v2 ¼ 13.17ÃÃ

23.1%

28.0%

v2 ¼ 5.94Ã

37.6%

39.2%

39.4%

32.8%

23.1%

33.1%

45.4%

40.8%

31.5%
.51 (.50)
1.71 (1.45)
12.6 (2.75)
1.73 (.89)

F ¼ 2.55 v2 ¼ 9.92ÃÃ

v2 ¼ 15.28ÃÃ

26.1%
.56 (.50) F ¼ .10
1.91 (1.82) F ¼ 3.16
13.2 (2.29)
1.72 (.93)

F ¼ 10.36ÃÃÃ
F ¼ .01

2.61 (4.90)

2.17 (3.91) F ¼ 3.34

2.99 (4.68)

2.49 (3.80) F ¼ 4.61Ã

2.21 (4.95)
.76 (1.41)

1.50 (3.58) F ¼ 7.07ÃÃ
.64 (1.23) F ¼ 1.73Ã

1.10 (1.73)

1.23 (1.73) F ¼ 3.01

.40
.23
1.93
1.18
.33
.35

.31
.13
1.89
1.22
.40
.21

(1.02)
(.86)
(2.67)
(1.57)
(.77)
(.67)

(.86)
(.60)
(2.49)
(1.56)
(.85)
(.54)

F ¼ 2.71
F ¼ 5.13Ã
F ¼ .10
F ¼ .18
F ¼ 2.28
F ¼ 17.70ÃÃÃ

Notes. N ranges from 1,069 to 1,351 due to missing data for the above analyses, unless specified. a 0 ¼ no, 1 ¼ yes.
Ã
p < .05. ÃÃ p < .01. ÃÃÃ p < .001.

Screening for Alcohol Risk

281

Regarding current alcohol consumption, on average, students reported drinking on 2.38 days during the past 30 days and consuming 2.73 drinks on the days they drank alcohol. Students reported consuming 5 or more alcoholic drinks on 1.83 days during the past 30 days and drinking to intoxication on .70 days. Boys reported significantly greater consumption than girls did on all four measures. In frequency terms (data not shown in table), 1,135 students (84%) reported having had an alcoholic drink at least once in their lifetimes, and 781
(68%) of them reported having had a drink in the past 30 days.
Regarding behavioral consequences, overall means for riding with someone who had been drinking, driving after any drinking, and driving after consuming 5 or more drinks, all assessed in relation to the past 30 days, were fairly low. Perhaps reflecting gender roles, girls were more likely to report riding with someone who had been drinking than boys, but were less likely to drive after drinking or drive after consuming 5 or more drinks. In frequency terms (data not shown in table), 625 (45.9%) of all students reported riding with someone who had been drinking at least one time in the past 30 days, whereas 254 (19.0%) reported driving after drinking and
113 (8.4%) reported driving after drinking in the same time period.
Because only a minority of secondary school students are eligible to drive, particularly without supervision, we repeated these analyses among the 362 students who were age 17 years or older and who were, hence, the most likely to be eligible to drive without supervision. Among this group, the frequencies of riding with someone who had been drinking, driving after any drinking, and driving after 5 or more drinks increased to 194 (54.6%), 107
(29.8%), and 58 (15.6%), respectively (data not shown in table).
Finally, students experienced on average almost two alcohol-related problems, most of these being aftereffects of drinking, but also significant levels of academic problems and unplanned and=or unprotected sex. Girls and boys reported similar levels of total problems, aftereffects of drinking, and academic problems; however, boys reported significantly higher rates of unplanned=unprotected sex, with 23.8% (N ¼ 150) of all boys reporting participation in unplanned or unprotected sex when they were drinking.
In terms of frequencies, 706 (52.7%) of all students reported experiencing at least one alcohol-related problem. Regarding subtypes, 664 (49.2%) experienced at least one aftereffect from drinking, 279 (20.8%) experienced at least one academic problem, and 254 (18.9%) reported participating in unplanned or unprotected sex because of drinking.

Non-Hispanic Students Versus Hispanic Students
Table 2 contains comparisons between non-Hispanics and Hispanics using analysis of variance (ANOVA) or chi-square (v2) tests as appropriate.
Although the ANOVA showed no mean difference in CAGE scores between
Hispanics and non-Hispanics, the v2 analysis showed that a greater

282

J. Tomaka et al.

TABLE 2 Descriptive Statistics for Study Variables and Comparisons Between Non-Hispanics and Hispanics
Variable
Demographic variables
Age
Grade level
Freshman (N ¼ 381)
Sophomore (N ¼ 479)
Junior (N ¼ 273)
Senior (N ¼ 188)
CAGE score (0–4)
Score of 0 (N ¼ 756)
Score of 1 (N ¼ 295)
Score of 2 or greater (N ¼ 238)
Predisposing factors
Father’s education
Elementary and=or some high school
Graduated high school= some college
Graduated college or greater
Mother’s education
Elementary and=or some high school
Graduated high school= some college
Graduated college or greater
Qualify for free=reduced luncha
Number of extracurricular activities Age of first drink (years)
Parental approval
Current drinking behaviors
Frequency of drinking (past 30 days) Number of drinks on drinking days 5 or more drinks (past 30 days)
Drinking to intoxication (past
30 days)
Behavioral consequences
Ride with someone who was drinking Drive after any drinking
Drive after 5 or more drinks
Total alcohol-related problems
Aftereffects of drinking
Academic problems
Unplanned=unprotected sex

Non-Hispanic
Mean (SD)=%

Hispanic
Mean (SD)=%

15.8 (1.14)

15.8 (1.18)

31.2%
32.9%
20.3%
15.6%
.62 (.92)
64.9%
20.1%
15.0%

28.0%
37.4%
20.8%
13.8%
.70 (.93)
56.5%
23.8%
19.7%

7.5%

33.1%

38.9%

38.3%

53.6%

28.6%

6.5%

36.5%

42.1%

43.2%

51.4%
.23 (.42)
2.09 (1.85)

20.2%
.66 (.47)
1.72 (1.56)

12.8 (2.60)
1.64 (.91)

13.0 (2.49)
1.65 (.88)

F or v2
F ¼ .12 v2 ¼ 2.87

F ¼ 2.58 v2 ¼ 7.35Ã

v2 ¼ 92.82ÃÃÃ

v2 ¼ 153.52ÃÃÃ

F ¼ 189.03ÃÃÃ
F ¼ 12.43ÃÃÃ
F ¼ .97
F ¼ .07

2.24 (4.29)

2.47 (4.50)

F ¼ .63

2.12 (2.86)

2.96 (4.65)

F ¼ 9.52ÃÃ

1.65 (4.04)
.62 (1.23)

1.92 (4.42)
.73 (1.37)

F ¼ .99
F ¼ 1.61

1.01 (1.65)

1.23 (1.76)

F ¼ 4.25Ã

.29
.12
1.73
1.14
.26
.26

.37
.20
1.96
1.21
.41
.28

F ¼ 1.85
F ¼ 2.73
F ¼ 2.12
F ¼ .51
F ¼ 7.40ÃÃ
F ¼ .18

(.87)
(.62)
(2.45)
(1.56)
(.70)
(.60)

(.97)
(.80)
(2.58)
(1.55)
(.85)
(.61)

Notes. N ranges from 1,042 to 1,323 due to missing data for the above analyses, unless specified. a 0 ¼ no, 1 ¼ yes.
Ã
p < .05. ÃÃ p < .01. ÃÃÃ p < .001.

Screening for Alcohol Risk

283

proportion of Hispanic students scored one or two, compared with the non-Hispanic students where a greater proportion scored 0.
Among predisposing factors, there were highly significant differences in parental education and qualifying for free=reduced lunch, as well as a more modest difference in participation in extracurricular activities. Specifically, the parents of non-Hispanic students had significantly more education than did Hispanic students and Hispanic students were much more likely to report being eligible for free=reduced lunch than were non-Hispanic students. Hispanic students also reported participating in fewer extracurricular activities than did non-Hispanic students. In contrast, there were no significant ethnic differences for age of first alcoholic drink and parental approval.
Regarding current drinking behaviors, non-Hispanic and Hispanic students significantly differed only regarding the typical number of drinks consumed on a typical drinking day with Hispanic students reporting significantly more drinks on those days than did non-Hispanic students.
The groups did not differ on any other consumption measure.
Finally, regarding behavioral consequences, Hispanics were significantly more likely to report riding in a car with someone who had been drinking and to report experiencing academic problems related to drinking than were non-Hispanic students. The groups did not differ on any other consequences measure.

Relation to Predisposing Factors, Current Drinking Behavior, and
Behavioral Consequences
The primary aim of this study was to assess whether CAGE alcohol risk scores were related to predisposing factors, current drinking behaviors, and consequences of alcohol use in a secondary school sample. The study also assessed whether gender moderated these relationships and whether CAGE scores of
1, which do not normally indicate risk in adult samples, indicate risk in a younger sample. The primary analytic approach was to conduct 2 Â 3 analyses of variance (ANOVAs) on various dependent measures using gender (M=F) and
CAGE risk (i.e., low, moderate, high) as the independent variables. In this design, significant interaction terms would suggest moderation by gender (see
Baron & Kenny, 1986). Table 3, Figure 1, and Figure 2 summarize the results of these analyses. Since Table 1 presented differences between boys and girls
(i.e., the main effects for gender), we only address main effects for CAGE risk and gender x risk interactions in the following discussion. Post hoc comparisons within risk group were conducted using the Tukey-A method.
PREDISPOSING FACTORS
As Table 3 shows, CAGE alcohol risk scores were significantly related to parental education, including both maternal and paternal educational

284

J. Tomaka et al.

TABLE 3 Relationship of CAGE Risk Scores to Predisposing Factors, Current Drinking
Behaviors, and Behavioral Consequences of Alcohol Use
CAGE Risk (CAGE Score)
0
1
2þ
% or M (SD) % or M (SD) % or M (SD)

Variable
Predisposing factors
Father’s education
Elementary and=or some high school
Graduated high school= some college
Graduated college or greater
Mother’s education
Elementary and=or some high school
Graduated high school= some college
Graduated college or greater
Qualify for free=reduced luncha
Current drinking behaviors
Frequency of drinking (past 30 days) Number of drinks on drinking days 5 or more drinks (past
30 days)
Drinking to intoxication (past
30 days)
Behavioral consequences
Ride with someone who was drinking Drive after any drinking
Drive after 5 or more drinks
Total alcohol-related problems
Aftereffects of drinking
Academic problems
Unplanned=unprotected sex

F or v2

v2 ¼ 8.03

23.2%

28.1%

29.3%

37.6%

39.4%

40.7%

39.2%

32.5%

29.8%

27.2%

27.0%

33.6%

41.9%

42.4%

45.5%

31.0%
.55b (.50)

30.6%
.51b (.50)

20.9%
.63c (.48)

1.82b (3.83)

3.07c (4.44)

3.37c (5.27) F ¼ 38.62ÃÃÃ

1.97b (3.43)

3.80c (5.08)

4.12c (4.99) F ¼ 36.99ÃÃÃ

1.31b (3.65)

2.46c (4.66)

2.96c (5.50) F ¼ 34.77ÃÃÃ

.46b (1.12)

.95c (1.44)

1.23c (1.67) F ¼ 47.89ÃÃÃ

.92b (1.57)

1.58c (1.95)

1.59c (1.86) F ¼ 28.19ÃÃÃ

.58c
.31c
2.61c
1.36c
.49c
.40c

.53c (1.18)
.32c (1.03)
3.31d (3.02)
1.75d (1.43)
.67d (1.04)
.42c (.72)

.22b
.09b
1.25b
.79b
.24b
.18b

(.71)
(.50)
(2.21)
(1.32)
(.69)
(.51)

(1.19)
(1.01)
(2.51)
(1.23)
(.87)
(.70)

Notes. a0 ¼ no, 1 ¼ yes. b,c,d Values not sharing a common superscript differ at p <.05.
Ã
p < .05. ÃÃ p < .01. ÃÃÃ p < .001.

FIGURE 1 Gender by CAGE risk interactions for age at first drink.

v2 ¼ 9.31Ã

F ¼ 3.88ÃÃ

F ¼ 25.93ÃÃÃ
F ¼ 16.17ÃÃÃ
F ¼ 79.34ÃÃÃ
F ¼ 86.42ÃÃÃ
F ¼ 28.08ÃÃÃ
F ¼ 26.73ÃÃÃ

Screening for Alcohol Risk

285

FIGURE 2 Gender by CAGE risk interactions for academic problems.

attainment, and an additional indicator of socioeconomic status, qualifying for free lunches. High-risk students reported lower levels of parental educational attainment and lower socioeconomic status than low-risk students.
Unlike most of the analyses that will follow, moderate-risk students resembled low-risk students in terms of predisposing factors. There was also a significant gender by CAGE risk interaction for age at first drink, F(2,
1059) ¼ 3.13, p < .05. As shown in Figure 1, low- and moderate-risk girls had their first drinks later than boys of the same risk level (both p < .05).
High-risk girls, in contrast, reported having their first drink at a younger age than the other girls (p < .05), and at an age similar to boys of the same level.
CURRENT DRINKING BEHAVIOR
There were strong, consistent main effects for CAGE risk on all four measures of current drinking behavior. Moreover, for frequency of drinking, number of drinks on drinking days, consuming five or more drinks, and drinking to intoxication, moderate-risk students showed similar behavior patterns as high-risk students. Specifically, relative to low-risk students, moderate- and high-risk students reported greater frequency of drinking, greater quantity of drinking on the days they drank, greater frequency of consuming five or more drinks, and greater frequency of drinking to intoxication. Although all four of these variables also showed main effects for gender (see Table 1), with girls having lower values than boys, none of the gender by risk interactions reached significance.
BEHAVIORAL CONSEQUENCES
Like current drinking behaviors, there were strong consistent main effects for behavioral consequences and moderate-risk students again resembled high-risk students in terms of problematic behaviors. Specifically, relative

286

J. Tomaka et al.

to low-risk students, moderate- and high-risk students were more likely to ride in a car with someone who had been drinking, to drive after any drinking, and to drive after consuming five or more drinks. Moderate- and high-risk students also reported experiencing more alcohol-related problems overall, including aftereffects of drinking, academic problems, and unplanned or unprotected sex. In addition, high-risk students reported significantly more total problems, aftereffects, and academic problems than moderate-risk students. Finally, there was only one significant gender by risk interaction, F(2, 1088) ¼ 4.36, p < .05, for academic problems. As Figure 2 shows, high-risk girls had more academic problems than any other group
(p < .05).

DISCUSSION
This study examined the prevalence of alcohol risk in a large sample of predominantly Hispanic secondary school students using the Conjoint Screening
Test—a brief and widely used measure known as the CAGE. In addition to estimating the number of students who screen positive for alcohol use disorder, it also examined the validity of the CAGE instrument in this population by examining its relation to predisposing factors, current drinking behaviors, and consequences of use. The study also examined gender and ethnic differences in drinking risks and outcomes, and examined Knight and colleagues’
(2003) contention that scores of one, which do not indicate risk in adult samples, are associated with risks in younger samples.
Regarding screening for potential alcohol risk, the results indicated that nearly one-fifth of secondary school students screened positive for alcohol use disorder using the traditional cutoff score of two or more positive responses to CAGE items. Moreover, an additional 23% screen positive for alcohol risk disorder if we follow Knight and colleagues’ suggested criterion of one or more positive responses on the CAGE for younger samples. Combined, the present data suggest that 41% of this large secondary school sample were putting themselves at significant risk because of alcohol consumption, a number that is consistent with rates of risky drinking in college students (e.g., Boyd, McCabe, & D’Arcy, 2003). Analyses of the predisposing factors, consumption variables, and behavioral consequences provided additional evidence for the validity of the CAGE in this population.
Regarding predisposing factors, CAGE scores were associated with socioeconomic disadvantage, specifically lower levels of parental education and greater likelihood of qualifying for a free lunch (a proxy for family income).
Analyses of current drinking behaviors showed that CAGE scores were consistently associated with reports of overall drinking frequency, consumption during drinking episodes, drinking to self-reported intoxication, and binge drinking. These latter associations provide evidence for the concurrent

Screening for Alcohol Risk

287

validity of the CAGE in this sample. CAGE scores were not associated with permissive parental attitudes as assessed from perceptions of these youths regarding drinking and not associated with participation in extracurricular activities. Finally, analyses of behavioral consequences showed that CAGE scores were strongly associated with driving outcomes, including riding in a car with someone who had been drinking, driving after any drinking, and driving after five or more drinks. CAGE scores were also strongly associated with total alcohol-related problems and all three subscales: aftereffects of drinking, academic problems, and participating in unplanned or unprotected sex.
The results also showed that students classified as moderate risk, scoring one on the CAGE, resembled low-risk students only in relation to predisposing factors. For all other indicators, however, including most current drinking behaviors and behavioral consequences, moderate-risk students were virtually indistinguishable from high-risk students. As such, the data support a lower CAGE criterion score (i.e., a score or 1 or more) to indicate alcohol risk in secondary school samples. This finding is consistent with Knight and colleagues (2003), who found the CAGE to have greater sensitivity with one as the cutoff score.
Gender was a significant, but relatively minor, factor in the overall data.
Small gender differences emerged on many variables, including parental education levels (girls less than boys), qualifying for free lunch (girls more than boys), and most drinking variables and behavioral consequences (girls less than boys). The strongest gender differences were for age at first drink, binge drinking frequency, and participation in unplanned or unprotected sex. In addition to initiating drinking later than did boys, girls participated less in binge drinking and unplanned=unprotected sex than did boys.
The study also examined gender as a moderator of relationships between CAGE scores and alcohol-related outcomes. Despite strong statistical power afforded by the large sample, gender interactions were only statistically significant with two variables: age at first drink and academic problems. Both interactions involved high-risk girls who (1) initiated drinking at earlier ages than low- and moderate-risk girls (and at an age similar to boys) and (2) reported more academic problems. These interactions might suggest the necessity of special intervention with high-risk girls.
Analyses of ethnicity suggested modest differences between Hispanic and non-Hispanic students. Perhaps the most notable of these was that
8.5% more Hispanic students scored positive on the CAGE compared with non-Hispanic students, including 3.7% more who scored 1 positive response and 4.7% more who scored 2 or greater. The results for the other study variables, however, did not indicate consistent differences by ethnicity. Specifically, despite vast disparities in sociodemographics (e.g., education and likely income), Hispanics differed from non-Hispanics only on a couple of variables, including the number of drinks on drinking days, riding in a car with someone who had been drinking, and academic problems from

288

J. Tomaka et al.

drinking, with no differences emerging for the majority of consumption and consequence variables. Overall, the data suggest that Hispanic students may drink more on the days they choose to drink, perhaps making them more likely to report suffering academic consequences, perhaps both contributing to their greater likelihood of screening positive on the CAGE.
Given that the CAGE instrument asks about rather serious signs of alcohol risk, and that the students in the present study were all well below the legal drinking age, the observed rates of positive screening are disconcerting.
That nearly one-fifth of students screened positive for potential alcohol risk using traditional criteria, and that 41% screened positive using more liberal criteria, indicates that alcohol risk appears to be relatively high in this region and=or population.
Moreover, the absolute levels of current drinking behaviors (e.g., 68% reported having had a drink in the past 30 days), driving after drinking
(e.g., 29.8% of those old enough to drive on their own), and experience of alcohol-related problems (e.g., 53% reported experiencing at least one problem), suggests the need for primary and secondary alcohol prevention programs. Although studies on secondary school students are relatively rare, our results are consistent with other examinations of the consequences of alcohol use in secondary school samples. For example, Bailey and colleagues (2004) found that risky drinking among secondary school students was related to missing school or work and engaging in unplanned sexual behaviors. Similarly, Arata and colleagues (2003) reported that the most common problems among secondary school students in relation to alcohol consumption included hangovers, behaving in ways they later regretted, getting into arguments, and being unable to remember things.
Our results are also consistent with a number of studies on college students. For example, our observed proportion of positive CAGE screeners, using traditional scoring criteria, parallels studies in college student samples. Boyd and colleagues (2003), for example, found that 22.7% of their college student sample scored positive using the CAGE. Like the present study, Boyd and colleagues also found strong associations between
CAGE scores and negative consequences, such as having a hangover, trouble with police, driving drunk, memory loss, performing poorly on test, and getting into a fight or an argument. The consistency of these results with the present data suggests the utility of using the CAGE instrument in relatively young samples and generalization from the college experience to secondary school.

Limitations
In addition to the absence of formal probability sampling just discussed, limitations to the present study include the cross-sectional nature of the data and

Screening for Alcohol Risk

289

reliance on the self-report by the students themselves, without collaboration from other sources such as the use of collaterals or the utilization of a biological measure. Anonymity of the survey and consistency and intricacy of some of the results (e.g., relation of CAGE scores to parental approval) suggest that self-report biases were minimal. Finally, this study was a secondary analysis of an existing data set and, as such, did not have the liberty to choose among available screening measures. Due to the secondary analysis nature of this study, other analyses comparing the CAGE to other screening instruments were not possible. Although the CAGE appeared to perform well in this context, we are aware of its limitations and recommend that investigators examine multiple options for appropriate screening instruments (Knight et al., 2003).

REFERENCES
Almodovar, A., Tomaka, J., Thompson, S., McKinnon, S., & O’Rourke, K. (2006). Risk and protective factors among high school students on the US=Mexico border.
American Journal of Health Behavior, 30, 745–752.
Arata, C. M., Stafford, J., & Tims, M. (2003). High school drinking and its consequences. Journal of Adolescence, 38, 567–579.
Bailey, K. A., Baker, A. L., Webster, R. A., & Lewin, T. J. (2004). Pilot randomized controlled trial of a brief alcohol intervention group for adolescents. Drug and Alcohol Review, 23, 157–166.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.
Barry, K., & Fleming, M. (1990). Computerized administration of alcoholism screening tests in a primary care setting. Journal of the American Board of Family
Practice, 3, 93–98.
Blincoe, L., Seay, A., Zaloshnja, E., Miller, T., Romano, E., Luchter, S., & Spicer, R.
(2002, May). The economic impact of motor vehicle crashes, 2000 (Technical report DOT HS 809 446). Washington, DC: U.S. Department of Transportation,
National Highway Traffic Safety Administration, National Center for Statistics and Analysis.
Boyd, C. J., McCabe, S. E., & D’Arcy, H. (2003). A modified version of the CAGE as an indicator of alcohol abuse and its consequences among undergraduate drinkers.
Substance Abuse, 24, 221–232.
Centers for Disease Control and Prevention. (2010). Web-Based Injury Statistics
Query and Reporting System (WISQARS). Retrieved from www.cdc.gov/injury/ wisqars/leading_causes_death.html Chen, C. M., Yi, H., Williams, G. D., & Faden, V. B. (2009). Trends in underage drinking in the United States, 1991–2007 (Surveillance report #86). National
Institute on Alcohol Abuse and Alcoholism. Retrieved from http://pubs.niaaa. nih.gov/publications/surveillance86/Underage07.htm Couwenbergh, C., van der Gaag, R. J., Koeter, M., de Ruiter, C., & van den Brink, W.
(2009). Screening for substance abuse among adolescents: Validity of the

290

J. Tomaka et al.

CAGE-AID in youth mental health care. Substance Use and Misuse, 44(6), 823–
834.
Drug Abuse Warning Network. (2006). Emergency department visits involving underage drinking. The New DAWN Report. Retrieved from http://www.samhsa. gov/data/2k6/underage/underage.htm Ewing, J. A. (1984). Detecting alcoholism: The CAGE questionnaire. Journal of the
American Medical Association, 252, 1905–1907.
Foote, J., Wilkens, C., & Vavagiakis, P. (2004). A national survey of alcohol and referral in college health centers. Journal of American College Health, 52, 149–157.
Grant, B. F., & Dawson, D. A. (1998). Age at onset of drug use and its association with DSM-IV drug abuse and dependence: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of Substance Abuse, 10, 163–173.
Grunbaum, J. A., Kann, L., Kinchen, S., Williams, B., Ross, J. G., Lowry, R., & Kolbe,
L. (2002). Youth Risk Behavior Surveillance—United States, 2001. MMWR CDC
Surveillance Summary, 51, 1–64.
Harwood, G. (2005). Alcohol abuse screening in primary care. Nurse Practitioner,
30, 56–61.
Hingson, R., Heeren, T., Winter, M., & Wechsler, H. (2005). Magnitude of alcoholrelated mortality and morbidity among U.S. college students ages 18–24:
Changes from 1998 to 2001. Annual Review of Public Health, 26, 259–279.
Jones, T., Lindsey, B., Young, P., Soltys, R., & Farani-Enayat, B. (1993). Alcoholism screening questionnaires: Are they valid in elderly medical outpatients? Journal of General Internal Medicine, 8, 359–369.
King, M. (1986). At risk drinking among general practice attenders: Validation of the
CAGE questionnaire. Psychological Medicine, 16, 213–217.
Knight, J. R., Sherritt, L., Harris, S. K., Gates, E. C., & Chang, G. (2003). Validity of brief alcohol screening tests among adolescents: A comparison of the AUDIT,
POSIT, CAGE, and CRAFFT. Alcoholism: Clinical and Experimental Research,
27, 67–73.
Laux, J. M., Salyers, K. M., & Kotova, E. (2005). A psychometric evaluation of the
SASSI-3 in a college sample. Journal of College Counseling, 8, 41–51.
McCarty, C. A., Ebel, B. E., Garrison, M. M., DiGiuseppe, D. L., Christakis, D. A., &
Rivera, F. P. (2004). Continuity of binge and harmful drinking from late adolescence to early adulthood. Pediatrics, 114, 714–719.
McKinnon, S. A., O’Rourke, K. M., Thompson, S. E., & Berumen, J. H. (2004). Alcohol use and abuse by adolescents: The impact of living in a border community.
Journal of Adolescent Health, 34, 88–93.
Mersy, D. (2003). Recognition of alcohol and substance abuse. American Family
Physician, 67, 1529–1532.
Monitoring the Future. (2008). Bureau of Justice Statistics. Retrieved from http:// www.ojp.usdoj.gov/bjs/dcf/du.htm National Highway Traffic Safety Administration (NHTSA). (2009). Fatality Analysis
Reporting System (FARS). Washington, DC: U.S. Department of Transportation,
National Highway Traffic Safety Administration, National Center for Statistics and Analysis.
National Institute on Alcohol Abuse and Alcoholism [NIAAA]. (2002, April). A call to action: Changing the culture of drinking at U.S. colleges. Task Force of the

Screening for Alcohol Risk

291

National Advisory Council on Alcohol Abuse and Alcoholism. Retrieved from www.collegedrinkingprevention.gov Pacula, R. (1998). Adolescent alcohol and marijuana consumption: Is there really a gateway effect? (NBER Working Paper No. W6348). Available at SSRN: http:// ssrn.com/abstract=226099 Russell, M. (1994). New assessment tools for drinking in pregnancy: T-ACE, TWEAK, and others. Alcohol Health and Research World, 18, 55–61.
Saunders, J., Aasland, O., Babor, T., De La Fuente, J., & Grant, M. (1993). Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption II.
Addiction, 88, 791–804.
Shields, A. L., & Caruso, J. C. (2004). A reliability induction and reliability generalization study of the CAGE questionnaire. Educational and Psychological Measurement, 64, 254–270.
Stewart, S. H., & Connors, G. J. (2004=2005). Screening for alcohol problems. Alcohol
Research and Health, 28, 5–15.
Substance Abuse and Mental Health Services Administration. (2011). Results from the
2010 National Survey on Drug Use and Health: Summary of national findings
(NSDUH series H-41, HHS publication no. (SMA) 11-4658). Rockville, MD:
Author.
Texas Commission on Alcohol & Drug Abuse. (1998). Texas school survey of substance abuse: Secondary survey instrument, 1998. Retrieved from www.tcada. state.tx.us/research/grades7-12/1998/98form1.pdf Copyright of Journal of Child & Adolescent Substance Abuse is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder 's express written permission. However, users may print, download, or email articles for individual use.