Quoted By:
properties—but not much is known about it. Recently, there has
been an interest by researchers to better understand this plant
and the substances within it. We would like to collect your
knowledge of Salvia if you have ever heard of it or come across it.
Salvia use was measured by a dichotomous variable which asked
respondents if they had ever tried salvia and was coded a 0 (no) or 1
(yes). Salvia use has a high level of desistance (Khey et al., 2008), thus
the current study was more concerned with salvia experimentation
rather than continual use. Indeed, given the high cost of salvia and the
short-lived experience, it was not surprising that very few habitual
users were found. The current study was mainly concerned with
predicting experimentation with salvia and not continued use.
Independent variables
Demographic
Several demographic variables were included: (1) age, (2) gender,
(3) race, (4) socioeconomic status, and (5) cumulative grade point
average. Age was a continuous variable which measured the age of a
subject on their most recent birthday. Gender was coded as 0 (female)
or 1 (male). Due to limited variation, race was recoded into a
dichotomous measure, coded as 0 (non-White)12 or 1 (White). Socio-
economic status was measured by student reports of their parents'
combined incomes, which was collapsed into four categories: low-
income (less than $50,000) lower middle-income ($50,000–$79.999),
upper middle-income ($80,000–$99,999), and high-income (over
$100,000).13 Grade point average was measured by asking respondents
to report their college grade point average.
Marijuana and alcohol use
Marijuana and alcohol use were each measured by one self-report
item asking about recent frequency of use. Each respondent was asked
in the past thirty days, how many days did you use marijuana (and
drink alcohol). Both measures ranged from 0 to 30, with mean
marijuana use in the past thirty days being 1.72 days; whereas, mean
alcohol use was 4.90 days in the past thirty days.
Self-control
Self-control was measured using Grasmick, Tittle, Bursik, and
Arneklev's (1993) twenty-four-item self-report scale. Several studies
had found this measure to be both reliable and valid (Nagin &
Paternoster, 1993; Piquero & Tibbetts, 1996), although others had
questioned its unidimensional properties (Gibson, 2005; Piquero
et al., 2000). Nonetheless, staying consistent with past research, the
current study summed responses across items to achieve a composite
score for each subject (α = .84). Responses for each item ranged on a
four-point Likert scale from 1 (disagree strongly) to 4 (agree strongly).
Higher scores on this measure indicated lower self-control. In
agreement with other studies (Brownfield & Sorenson, 1993; Nagin
& Paternoster, 1993), a principal component analysis indicated that all
items loaded on one factor (see also Arneklev, Grasmick, Tittle, &
Bursik, 1993; LaGrange & Silverman, 1999; Piquero & Rosay, 1998).
Analytic plan
The analytic plan unfolds using three general steps. First,
correlations were estimated to assess the bivariate relationships
between the independent variables and salvia use. Furthermore,
bivariate analyses are insightful for assessing relationships between
several of the independent variables, such as self-control and gender.
Second, a series of logistic regression models were calculated to arrive
at a final model predicting lifetime salvia use and also assess whether
low self-control can explain the effects of other independent variables
previously discussed. Finally, given that estimated coefficients from
logistic regression are not readily interpretable, a series of graphs are
displayed that show differences in predicted probabilities across
various combinations of independent variables to more clearly portray
the likelihood of salvia use, and more specifically to discern who has
the highest likelihood of having experimented with salvia.
Results
Table 2 shows bivariate correlations between the independent and
dependent variables. Of particular interest, several demographic and
other individual variables are significantly (p b .05) correlated with
lifetime salvia use. As for demographic characteristics, gender, race and
socioeconomic status are positively and significantly correlated with
salvia use. That is, males, White students, and students whose parents
have more income are more likely to have used salvia compared to
women, non-White students, and students whose parents have less
income. Furthermore, alcohol use, marijuana use, and self-control were
all positively and significantly correlated with salvia use. Students who
drink more often, smoke marijuana more often, and have lower self-
control are all more likely to report ever using salvia. Of all the bivariate
correlations, recent marijuana use exhibited the strongest correlation
with salvia use. These patterns of bivariate correlations may be
expected and are similar to the bivariate relationships Lange et al.
(2008) found between drug and salvia use among their sample, as well
as consistent with self-control theory's expectations.
Table 3 shows a series of logistic regression models predicting the
prevalence of lifetime salvia use. Four models are shown in a stepwise
fashion. Model 1 includes only demographic characteristics and shows
that while controlling for each, a similar picture emerges as seen in the
bivariate correlations. Gender, race, and socioeconomic status have
positive and statistically significant effects on life-time salvia use,
indicating that males are more likely than females to have used salvia;
Whites are more likely than non-Whites to have used salvia; and
students whose parents have more income are more likely to have
used salvia than students whose parents have less income. Model 2
shows only the effects of other drug use on lifetime salvia use. As
shown, when both recent alcohol and marijuana use are simulta-
neously considered, only marijuana use has a positive and statistically
significant effect on salvia use. Students who have used marijuana
increasingly more in the past thirty days are more likely to have used
salvia in their lifetime than those who have used less in the past thirty
days.
399
B.L. Miller et al. / Journal of Criminal Justice 37 (2009) 396–403