Addictive Behaviors, Vol. I, pp. 363-371, 1982
Printed in the USA. All rights reserved.
Copyright B 1983 Pergamon Press Ltd
THE DRUG ABUSE SCREENING TEST
HARVEY A. SKINNER
Addiction Research Foundation and University of Toronto
Abstract-The Drug Abuse Screening Test (DAST) was designed to provide a brief instrument
for clinical screening and treatment evaluation research. The 28 self-report items tap various
consequences that are combined in a total DAST score to yield a quantitative index of prob-
lems related to drug misuse. Measurement properties of the DAST were evaluated using a
clinical sample of 256 drug/alcohol abuse clients. The internal consistency reliability estimate
was substantial at .92, and a factor analysis of item intercorrelations suggested an unidimen-
sional scale. With respect to response style biases, the DAST was only moderately correlated
with social desirability and denial. Concurrent validity was examined by correlating the DAST
with background variables, frequency of drug use during the past 12 months, and indices of
psychopathology. Although these findings support the usefulness of the DAST for quantifying
the extent of drug involvement within a help-seeking population, further validation work is
needed in other populations and settings.
Much of our information about the nonmedical use of drugs has been collected by
surveys, especially of high school and college students. These surveys have been criti-
cized, however, since they are often based on nonstandardized questionnaires with
unknown measurement properties (Hochhauser, 1979; Stanton, 1977). A need has been
voiced for standardized instruments that have been carefully evaluated in both clinical
and nonclinical populations.
Several recent studies have provided evidence on the reliability and validity of drug
use scales. For instance, Single et al. (1975) examined the consistency of self-reporting
among high school students in a longitudinal survey. Self-reports of illicit drug use
were fairly consistent at one point in time but less consistent over time (5-6 month in-
terval). However, Single et al. (1975) concluded that inconsistencies were more the
result of poor recall than an active attempt by respondents to conceal drug use. Smart
and Blair (1978) found good test-retest reliability (r = .88) with a drug use scale admin-
istered to high school students on two occasions separated by 8 weeks. Furthermore, a
nine-item lie scale administered on the two occasions suggested little defensiveness in
self-reported drug use. From the perspective of attitudes toward drug use, Goodstadt et
al. (1978) found a median internal consistency reliability of .85 for 10 brief six-item
scales. These scales achieved a median correlation of .39 with reported use of specific
drugs and a median correlation of .48 with the intention to use certain drugs in the next
year. In clinical populations, investigators have reported favorable measurement prop-
erties for drug use assessment techniques (e.g., Cohen et al., 1977; Joe, 1974).
Overall, research to date suggests that drug use and related problems in student
populations can be reliably assessed by self-report methods. Nonetheless, since the ma-
jority of studies have been conducted with students in nonclinical contexts, whether or
not the results generalize to clinical settings remains an open question. In clinical
This research was supported by the Addiction Research Foundation. The author wishes to thank Romilla
Chhabra for her help in the preparation of this manuscript, and Simone Shrimpton for her assistance in
developing the original version of the DAST.
Reprint requests should be directed to Dr. Harvey Skinner, Addiction Research Foundation, 33 Russell
Street, Toronto, Canada M5S 2Sl. Copies of a revised version of the DAST and scoring instructions may be
obtained by writing the author.
364 HARVEY A. SKINNER
populations a Ijoor correlation has been found between self-report and objective
evidence (urinalysis) of drug use (Chambers & Taylor, 1973; Orrego et al., 1979). Cer-
tainly, one must be cautious about the reliability of drug use information with patients
who are presently using drugs or are experiencing a drug withdrawal reaction. In addi-
tion, the validity of self-reports should be questioned in situations where the individual
has incentives to conceal drug use behavior (e.g., criminal justice or employment con-
The present study describes an empirical evaluation of the Drug Abuse Screening
Test (DAST) in a clinical sample of persons who voluntarily sought help for problems
related to drug/alcohol abuse. The 28-item DAST is a self-report scale that yields a
quantitative index of the range of problems associated with drug abuse. The specific
aims of this study are fourfold: (1) to determine item characteristics and scale reliabil-
ity; (2) to examine the multidimensionality of DAST items using factor analysis; (3) to
evaluate the degree to which the DAST is influenced by social desirability, denial and
carelessness response styles; and (4) to examine correlations of the DAST with demo-
graphic characteristics, frequency of illicit drug use, and indices of psychopathology.
The sample consisted of 223 individuals (72% males, 28% females) who had volun-
tarily sought help at the Clinical Institute of the Addiction Research Foundation. All
subjects were tested as part of a comprehensive assessment program. Of the total sam-
ple, 58.6% were referred for alcohol problems, 25.4% for drug abuse, and 16.0% for
both alcohol and drug problems. The mean age for the sample was 32.47 with a stan-
dard deviation of 11.17. Of the subjects, 39.5% were single, 18.4% married while 28.5%
were either divorced or separated. With regard to education level, 46.5% had some
high school while 17.2% had completed high school. Of the sample, 47.6% were unem-
ployed at the time of assessment. The most heavily represented occupational categories
were labourers (35.7%), clerical workers (18.8%) and skilled workers (16.9%). Of the
subjects referred for either drug use or both alcohol and drug related problems, the
mean age at first steady use of drugs was 20.75 (Range of 1 l-59). Finally, of the 193 in-
dividuals who reported having tried drugs, cannabis was the most frequent drug first
used (43.5%) followed by barbiturates, sedatives, and tranquillizers (20.2%), halluci-
nogens (13.5%), and amphetamines (11.9Vo).
The 28 items in the self-administered DAST (Table 1) parallel items on the Michigan
Alcoholism Screening Test (MAST), which is a widely used assessment device for alco-
holism (Selzer, 1971). The MAST has proven to be a reliable and valid screening in-
strument for clinical and non-clinical settings (Jacobson, 1976; Skinner, 1979). The fre-
quently noted similarities among alcohol and drug abusers (Freed, 1975) suggest that a
variation of the MAST might be useful for the assessment of problems related to the
non-medical use of drugs. Recently, Cannel1 and Farazza (1978) administered a modi-
fied version of the MAST to screen for drug-related problems among college students.
However, no attempt was made to evaluate the psychometric properties of their scale.
The DAST was administered by trained assessment workers as part of a comprehensive
assessment program. Specific instructions were:
The following questions concern information about your involvement and abuse of
drugs. Drug abuse refers to (1) the use of prescribed or “over the counter” drugs in
The drug abuse screening test 365
excess of the directions and (2) any non-medical use of drugs. Carefully read each
statement and decide whether your answer is yes or no. Then circle the appropriate
response on the separate answer sheet.
The DAST total score is computed by summing all items that are endorsed in the di-
rection of increased drug use problems. Thus, the total score can range from 0 to 28.
The age and sex of each client was recorded during a structured interview. As well, a
composite index of social stability was computed from information on present accom-
modation, family contact, employment and legal status (Skinner, 1981). Also, a three
item social class index was derived from education level, occupational status and living
accommodation (Skinner, 1981). Also, clients completed the 25-item Michigan Alco-
holism Screening Test (Selzer, 1971; Skinner, 1979), as well as a measure of stressful
life event changes during the past 12 months (Holmes & Rahe, 1967). Frequency of
drug use during the previous 12 months was assessed by a structured interview using a
6-point scale (1 = none, 2 = less than weekly, 3 = weekly, 4 = twice weekly,
5 = daily, 6 = several times per day). Psychopathology was assessed by the Basic Per-
sonality Inventory (Jackson, 1976). This 240-item instrument contains 12 scales that
were developed using a construct validation framework (Jackson, 1971). Finally, three
response style measures were given including: (1) denial which taps an individual’s
tendency to be defensive and minimize problems (Jackson, 1976); (2) social desirability
which assesses the extent to which one presents an overly favourable picture of oneself
(Jackson, 1974); and (3) infrequency which identifies individuals who either did not un-
derstand the items or were responding carelessly (Jackson, 1974).
The results of an item analysis are summarized in Table 1. The item mean is the pro-
portion of subjects who answered the question “yes.” The item-scale correlation (cor-
Table 1. Item Analysis Summary
Item Standard Item-Total
Mean Deviation Correlation
1. Have you used drugs other than those required
for medical reasons?
2. Have you abused prescription drugs?
3. Do you abuse more than one drug at a time?
4.*Can you get through the week without using
drugs (other than those required for medical
S.*Are you always able to stop using drugs when
you want to?
6. Do you abuse drugs on a continuous basis?
7.*Do you try to limit your drug use to certain
.33 .47 .55
.48 .50 .61
.59 .49 .76
8. Have you had “blackouts” or “flashbacks” as
a result of drug use?
9. Do you ever feel bad about drug your abuse?
10. Does your spouse (or parents) ever complain
about your involvement with drugs?
1 I. Do your friends or relatives know or suspect
you abuse drugs?
12. Has drug abuse ever created problems between
you and your spouse?
.29 .46 .24
.64 .48 .74
.47 .50 .67
.58 .49 .72
.50 .50 .76
.65 .48 .62
366 HARVEY A. SKINNER
Table 1 (continued)
13. Has any family member ever sought help for
problems related to your drug ;se?
14. Have you ever lost friends because of your
use of drugs?
15. Have you ever neglected your family or missed
work because of your use of drugs?
16. Have you ever been in trouble at work because
of drug abuse?
17. Have you ever lost a job because of drug abuse?
18. Have you gotten into fights when under the
influence of drugs?
19. Have you ever been arrested because of unusual
behaviour while under the influence of drugs?
20. Have you ever been arrested for driving while
under the influence of drugs?
21. Have you engaged in illegal activities in order to
22. Have you ever been arrested for possession of
23. Have you ever experienced withdrawal symptoms
as a result of heavy drug intake?
24. Have you had medical problems as a result of
your drug use (e.g., memory loss, hepatitis,
convulsions, bleeding, etc.)?
25. Have you ever gone to anyone for help for a
26. Have you ever been in hospital for medical
problems related to your drug use?
27. Have you ever been involved in a treatment
programme specifically related to drug use?
28. Have you been treated as an out-patient for
problems related to drug abuse?
*Items 4, 5 and 7 are scored in the “no” or false direction.
rected for part-whole overlap) provides an index of the discriminating power of each
item. “Good” items are ones that correlate highly with the total DAST score (Nunnally,
1978). Except for item 7, all DAST items have moderate to substantial item-total scale
correlations. The three best items were:
23. “Have you ever experienced withdrawal symptoms as a result of heavy drug in-
take?” (r = .78)
15. “Have you ever neglected your family or missed work because of your use of
drugs?” (r = .77)
3. “Do you abuse more than one drug at a time?” (r = .76)
Finally, the internal consistency reliability (coefficient alpha) of .92 was substantial
which indicates that subjects were quite consistent when responding to all DAST items
In Figure 1, the total DAST score is depicted for subjects according to their reason
for seeking help: Drug Problems (n = 51), Alcohol Problems (n = 137), or mixed
Drug/Alcohol Problems (n = 35). A one-way analysis of variance (Table 2) revealed a
statistically significant difference among the three group means, F(2,220) = 112.9,
p < .OOl. The DAST total score clearly differentiated the group with primarily
Alcohol-related problems from the other two groups with Drug and mixed
Drug/Alcohol problems. Multiple range tests (Duncan, Student-Newman-Keuls)
The drug abuse screening test 367
Alcohol Drug and Drug
Problems Alcohol Problems
Only Problems Only
Fig. 1. DAST total score by reason for seeking treatment. The top of the box denotes the
group mean, whereas the vertical line represents * one standard deviation from the mean.
revealed statistically significant differences (p < .05) between pair wise comparisons
for all three groups.
Point-biserial correlations were computed among the 28 DAST items, and a prin-
cipal components model of factor analysis (Gorsuch, 1974) was used to explore impor-
tant dimensions that underlie the correlation matrix. The first principal component ac-
Table 2. Distribution of DAST Total Scores
Reason for seeking treatment
1. Alcohol Abuse 2. Drug/Alcohol Abuse 3. Drug Abuse
n = 137 n’= 35 n = 51
0 25% 0% 0%
l-5 49% 9% 0%
6-10 1% 14% 8%
11-15 11% 26% 19%
16-20 6% 25% 42%
21-25 1% 23% 29%
26-28 1% 3% 2%
1.5 15.7 18.0
4.5 15.2 17.8
6.1 6.5 5.0
ANOVA F(2,220) = 112.9, /J < .OOl
Multiple Range Tests (Duncan, Student-Newman-Keuls)
Group 1 significantly different from Group 2 (pc .05)
Group I significantly different from Group 3 (pe .05)
Group 2 significantly different from Group 3 (pc .OS)
368 HARVEY A. SKINNER
counted for 45.4% of the total variance, which suggests an unidimensional scale
among the DAST items. Indeed, the weights on this first factor are proportional to the
item-scale correlations of Table 1 (Henrysson, 1962). The distribution of eigenvalues
for the first seven factors was 12.7, 1 S, 1.3, 1.2, 1 .l, 1 .O, 0.8. The substantial differ-
ence between the first (45.4% total variance) and second factors (5.4% total variance)
indicates that the DAST measures a dominant single dimension of problems related to
drug abuse. The remaining factors largely reflect variation that is specific to each item.
Although a varimax rotation was conducted using 4, 5 and 6 factors, the single drug
abuse dimension predominated each solution.
A common clinical impression of individuals who abuse drugs and alcohol is their
tendency to minimize or deny problems. Accordingly, the DAST was correlated with
three measures of response bias (Table 3) including: Denial, Social Desirability, and In-
frequency (carelessness). Correlations were computed for the total sample (n = 223),
as well as for a subsample (n = 86) that excluded individuals referred only for alcohol
related problems. The total sample allows one to generalize results to a clinical setting
that deals with alcohol, drug or combined alcohol/drug abuse, whereas the subsample
results are relevant to a more restrictive population of drug abusers. Because the total
Table 3. Correlations with the DAST Total Score
Total Sample exluding Ss
Sample with only alcohol problems
Sex (I = M, 2 = F)
Stressful Life Events
- .28** -.13
- .38** -.31*
- .42** -.19*
-.14* - .23*
- .27*’ - .33*
- .31** - .30**
Frequency of drug use in past 12 months
Heroin .19* .I7
Other Opiates .35** .26*
Amphetamines .36** .lO
Barbiturates .47+* .30*
Hallucinogens .27** .08
Cannabis .55** .I0
Glue, Solvents .21* .I0
l p <.05
The drug abuse screening test 369
sample is more heterogeneous, one would expect these correlations to be greater in
The largest correlation was with Social Desirability, which suggests that individuals
who present an overly favorable or socially desirable picture of themselves also tend to
score lower on the DAST. However, the magnitude of this relationship was relatively
small: 14% common variance in the total sample, 9% common variance in the sub-
sample. Similarly, there was a modest relationship between Denial and the DAST,
which suggests that individuals who are higher on Denial tend to report fewer drug
related problems. In the total sample, there was a slight tendency for high scorers on
the DAST to be less careful when completing the assessment instruments. However,
since the internal consistency reliability estimate for the DAST was quite high (.92 in
total sample, .86 in subsample), the carelessness tendency had a negligible influence on
the DAST total score.
Correlations with background, drug use and psychopathology
Table 3 also contains correlations between the DAST and various background and
drug use measures. Again, correlations are given for both the total sample and the sub-
sample which excludes individuals with only alcohol problems. In general, the pattern
of correlations is similar for both samples, although the magnitude of relationships is
greater in the more heterogeneous total sample.
With respect to demographic characteristics, younger individuals, especially males,
tended to score higher on the DAST. This suggests age cohort differences in the clinical
population since older individuals tended to misuse alcohol more than drugs. High
scorers on the DAST were inversely related to indices of social stability and social class.
As one would expect, drug abuse tended to produce problems in maintaining stable ac-
commodation, work record and family contact. A lack of social stability was further
reflected by a higher prevalence of recent life event changes among high scorers on the
DAST. For comparison purposes, the DAST was correlated with the original Michigan
Alcoholism Screening Test. There was a slight inverse relationship (r = - .21) between
the DAST and MAST in the total sample. Next, the DAST was correlated with the fre-
quency of drug use (6 point scale ranging from none to daily) over the past 12 months.
A greater range of problems associated with drug abuse (DAST) was related to the
more frequent use of cannabis, barbiturates and opiates other than heroin. These find-
ings support the concurrent validity of the DAST. Finally, relationships were explored
between the DAST and measures of psychopathology. The largest correlations were
with the sociopathic scales of Impulse Expression and Social Deviation. High scorers
on the DAST tended to engage in reckless actions and express attitudes that are mark-
edly different from common social codes. Furthermore, the DAST was positively
related to interpersonal problems, suspiciousness, depressive symptoms and a preoc-
cupation with bodily dysfunctions. Thus, drug abuse tended to be manifest in, or
covary with, other psychopathological characteristics.
The Drug Abuse Screening Test was developed to provide a convenient instrument
for assessing the extent of problems related to drug misuse. The total DAST score
yields a quantitative index of problem severity. In this clinical population, the DAST
score was highly reliable and only minimally influenced by the response style biases of
denial and social desirability. However, caution must be exercised when generalizing
the results to other contexts. For instance, one could expect more defensiveness about
admitting to drug related problems in an employment or criminal justice setting.
370 HARVEY A. SKINNER
The finding that a dominant factor underlies the DAST items indicates that indi-
viduals in this clinical population tend to be ordered along a single dimension or axis.
Since this dimension provides quantitative information about the degree of problems
related to drug abuse, emphasis on a diagnostic cut-off point would be somewhat ar-
bitrary. As the DAST total score increases, one may interpret that a given individual
has accrued an increasingly diverse range of consequences. Similarly, Skinner (1979)
has argued that the Michigan Alcoholism Test is most appropriately interpreted as
classifying individuals along a continuum according to the degree of alcohol misuse.
On the other hand, it is interesting to note in Table 2 that 75% of individuals seeking
treatment for alcohol problems scored in the O-5 range, whereas only 9% of clients
with both drug/alcohol problems and none of the clients seeking treatment for drug
problems scored less than 6 on the DAST. Hence, one could evaluate the usefulness for
case finding purposes of a DAST score exceeding 5.
With respect to further research, a shortened version of the 28-item DAST was ex-
plored. The following 20 items had consistently high item-total scale correlations in
both the total sample and the subsample excluding clients with only alcohol problems
(items 1, 2, 3, 4, 5, 8, 9, 10, 12, 14, 15, 16, 17, 18, 21, 22, 23, 24, 25, 27). The 20-item
DAST correlated almost perfectly (r = .99) with the original 28-item version. More-
over, the internal consistency reliability of the 20-item DAST was extremely high (.95
for the total sample, .86 for the subsample excluding clients with only alcohol prob-
lems). Thus, the briefer 20-item version has excellent psychometric properties com-
parable to the original DAST. Another research line would be to differentiate among
various drug classes and their specific consequences. At present, the DAST considers
drugs as a generic group. In certain contexts one might profitably assess the conse-
quences related to each particular class of drug used. Further validation work is also
needed using objective criteria of drug use (e.g., urinalysis) and using populations
other than clients who already acknowledge having a drug problem. Hopefully, this
research will demonstrate that the DAST has diagnostic value for screening and assess-
ment programs in a variety of clinical and nonclinical settings.
Cannell, M.B., & Favazza, A.R. Screening for drug abuse among college students: Modification of the
Michigan Alcoholism Screening Test. Journal of Drug Issues, 1978, 8$ 119- 123.
Chambers, C.D., & Taylor, W.J. The incidence and patterns of drug abuse during maintenance therapy.
In CD. Chambers and L. Brill (Eds.), Methadone: experiences and issues. New York: Behavioral
Cohen, M., Karras, A., & Hughes, R. The usefulness and reliability of a drug severity scale. International
Journal of the Addictions, 1977, 12, 417-422.
Freed, E.X. Drug abuse by alcoholics: A review. International Journal of the Addictions, 1973,8,451-473.
Goodstadt, M.S., Cook, G., Magid, S., & Ciruson, V. The Drug Attitudes Scale (DAS): Its development and
evaluation. International Journal of the Addictions, 1978, 13, 1307-1317.
Gorsuch, R.L. Factor Analysis. Philadelphia: Saunders, 1974.
Henrysson, S. The relation between factor loadings and biserial correlations in item analysis. Psychometrika,
1962, 21, 419-424.
Hochhauser, M. Bias in drug abuse survey research. International Journal of the Addictions, 1979, 14,
Holmes. T.H.. & Rahe. R.H. The social readiustment rating scale. Journal sf Psychosomatic Research,
1967, 11, 213-218.
Jackson, D.N. The dynamics of structured personality tests: 1971. Psychological Review, 1971,78,229-248.
Jackson, D.N. Personality Research Form Manual. Goshen, N.Y.: Research Psychologists Press, 1974.
Jackson, D.N. Basic Personality Inventory. London, Canada: Author, 1976.
Jacobson. G.R. The alcoholisms: Detection, assessment and diagnosis. New York: Human Sciences Press,
Joe, G.W. Patient background indices for a drug-abusing population. In S.B. Sells (Ed.), Research on
The drug abuse screening test 371
patients, treatments and outcomes: Studies of the effectiveness of treatments for drug abuse. Volume II.
Cambridge, Mass.: Ballinger, 1974.
Nunnally, J.C. Psychometric theory. New York: McGraw-Hill, 1978.
Orrego, H., Blendis, L., Blake, J.E., Kapur, B., & Israel, Y. Reliability of assessment of alcohol intake
based on personal interviews in a liver clinic. Lancet, December 1979, 1354-1356.
Seizer, M.L. The Michigan Alcoholism Screening Test: The quest for a new diagnostic instrument. American
Journal of Psychiatry, 1971, 127, 1653-1658.
Single, E., Kandel, D., & Johnson, B.D. The reliability and validity of drug use responses in a large scale
longitudinal survey. Journal of Drug Issues, 1975, Fall, 426-443.
Skinner, H.A. A multivariate evaluation of the MAST. Journal of Studies on Alcohol, 1979, 40, 831-844.
Skinner, H.A. Assessment of alcohol problems: Basic principles, critical issues and future trends. In Y. Israel
et al. (Eds.), Research advances in alcohol and drug problems, Vol. VI. New York: Plenum, 1981.
Smart, R.G., & Blair, N.L. Test-retest reliability and validity information for a high school drug use ques-
tionnaire. Drug and Alcohol Dependence, 1978, 3, 265-271.
Stanton, M.D. Drug use surveys: Method and madness. International Journal of the Addictions, 1977,
12, 95-l 19.