The Gerontologist

cite as: Gerontologist, 2020, Vol. 60, No. 4, e232–e241

doi:10.1093/geront/gnz054
Advance Access publication May 14, 2019

Measurement Article

Measuring Attitudes Toward Medical and Recreational

Cannabis Among Older Adults in Colorado

Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023

Kanika Arora, PhD,1,* Sara  H. Qualls, PhD,2 Julie Bobitt, PhD,3 Hillary  D. Lum, MD,4
Gary Milavetz, PharmD,5 James Croker, MA,1 and Brian Kaskie, PhD1
1
Department of Health Management and Policy, College of Public Health, University of Iowa. 2Department of Psychology
and Gerontology Center, University of Colorado, Colorado Springs. 3Interdisciplinary Health Sciences, University of Illinois at
Urbana Champaign. 4Division of Geriatric Medicine, University of Colorado Anschutz Medical Campus, Aurora. 5Department
of Pharmacy Practice and Science, University of Iowa.
*Address correspondence to: Kanika Arora, PhD, Department of Health Management and Policy, College of Public Health, University of Iowa,
145 N. Riverside Drive, Room N238, Iowa City 52242-2007, IA. E-mail: kanika-arora@uiowa.edu

Received: November 11, 2018; Editorial Decision Date: March 25, 2019

Decision Editor: Rachel Pruchno, PhD

Abstract
Background and Objectives:  Cannabis use among older adults is on the rise. Despite growing interest in the topic, there
exists a paucity of standardized measures capturing cannabis-specific attitudes among older adults. Using data from a
survey of older Coloradans, we create two scales that separately measure medical and recreational cannabis attitudes. We
also examine how these two attitudes relate to individual-level characteristics.
Research Design and Methods: We assess reliability using Cronbach’s alpha and item-rest correlations and perform
confirmatory factor analyses to test the two attitude models. We conduct a seemingly unrelated regression estimation to
assess how individual characteristics predict medical and recreational cannabis attitude scores.
Results:  Twelve indicators combined into two valid and reliable scales. Both scales had a three-factor structure with affect,
cognition and social perception as latent dimensions. For both scales, fit indices for the three-factor model were statistically
superior when compared with other models. The three-factor structure for both scales was invariant across age groups. Age,
physical health, and being a caregiver differentially predicted medical and recreational cannabis attitude scores.
Discussion and Implications:  Medical and recreational cannabis attitude scales can inform the development and evaluation
of tailored interventions targeting older adult attitudes that aim to influence cannabis use behaviors. These scales also
enable researchers to measure cannabis-specific attitudes among older adults more accurately and parsimoniously, which in
turn can facilitate a better understanding of the complex interplay between cannabis policy, use, and attitudes.
Keywords:  Scale construction, Substance use, Factor analysis

The passage of medical and recreational marijuana legisla-
tions by several states has shifted the emphasis to examining
determinants of use, particularly the role of cannabis-
specific attitudes. Although a majority of this work has fo-
cused on adolescents and young adults (Bachman, Johnson,
& O’Malley, 1998; Keyes et  al., 2011), increasing rates
of cannabis use among older adults (Kaskie, Ayyagari,
Milavetz, Shane, & Arora, 2017; Stoner, 2016) has drawn
attention to the need to understand their attitudes as well.
Despite this growing interest, there is a surprising pau-
city of standardized cannabis attitude measures to guide this
line of investigation. Although some studies on cannabis
use among older adults have included various attitude-
related measures (Black & Joseph, 2014; Choi, Marti,

© The Author(s) 2019. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. e232
For permissions, please e-mail: journals.permissions@oup.com.
The Gerontologist, 2020, Vol. 60, No. 4
DiNitto, & Choi, 2018; Han et  al., 2017; Salas-Wright
et al., 2017), little evidence exists concerning the reliability
and validity of these items. Variability in content has also
resulted in a lack of clarity regarding the actual construct
being measured. Moreover, prior research on the topic has
largely operationalized the attitude construct from a uni-
dimensional perspective where it is not possible to simul-
taneously hold both positive and negative evaluations of
the substance depending on purpose of use. Although this
may not be as relevant for adolescents and young adults, it
is conceivable that older adults are ambivalent: They may
disapprove of recreational cannabis and regard it as risky,
but at the same time favor the medical application of can-
nabis for its symptomatic and palliative benefits (Ahmed
et al., 2014; Wang & Chen, 2006). However, no previous
study has examined attitudes toward medical and recrea-
tional cannabis separately.
The purpose of this study was to demonstrate the de-
velopment of scales capturing older adults’ attitudes to-
ward medical and recreational cannabis using survey data
collected in 2017 from individuals age 60 and older in
Colorado. Procedures of scale construction, including reli-
ability and confirmatory factor analyses (CFA), were used
to combine several indicators of attitudes into two scales,
one assessing perceptions regarding medical cannabis and
another assessing perceptions regarding recreational can-
nabis. In particular, we hypothesized that a three-factor (af-
fect, cognition and social perception) representation would
be replicated in the CFA for both attitude scales. Finally,
individual-level predictors of medical and recreational
cannabis attitudes were examined using a joint estimation
technique.
Medical and recreational cannabis attitude scales
can inform the development and evaluation of tailored
interventions targeting older adult attitudes that aim to
influence cannabis use behaviors. These scales also enable
researchers to measure cannabis-specific attitudes among
older adults more accurately and parsimoniously, which in
turn can facilitate a better understanding of the complex
interplay between cannabis policy, use, and attitudes.
Theoretical Background and Literature Review
Attitudes have historically been a chief focus of social
psychological research given their assumed importance
in guiding behavior (Hess, 2006). Even though several
studies, including those focused on cannabis, typically tend
to measure attitudes in the form of a respondent’s summary
evaluation of overall favorable or unfavorable reaction to-
ward an external stimulus (for instance, through questions
measuring attitudes on like/dislike, pleasant/unpleasant,
approve/disapprove, right/wrong dimensions), theoretical
research on the topic has characterized attitudes as a com-
plex phenomenon reflecting the integration of numerous
factors (Hansen & Hansen, 2016; Rosenberg, 1960).
e233
One commonly held perspective portrays attitudes
as a superordinate construct with affective and cognitive
factors as underlying dimensions (Breckler & Wiggins,
1993; Eagly & Chaiken, 1993; Van der Pligt, Zeelenberg,
van Dijk, de Vries, & Richard, 1997). The affective dimen-
sion has traditionally been postulated to include the set of
physiological responses, positive and negative feelings and
emotions an individual associates with an object (Breckler
and Wiggins, 1993; Crites, Fabrigar, & Petty, 1994). The
cognitive dimension is assumed to be composed of know-
ledge regarding an object’s attributes or traits (Breckler &
Wiggins, 1993; Crites et al., 1994). In empirical analyses of
Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023
structural components of attitudes toward a broad range
of objects (including risky behaviors such as alcohol and
cannabis use), Simons and Carey (1998) have concluded
that affective and cognitive factors constitute discrete
dimensions that can be distinguished from each other and
from overall attitude evaluations.
In addition to self-centric affective and cognitive factors,
theorists have also long recognized attitudes to have an ex-
ternal, social perception dimension. For example, Festinger
(1954) argued that people evaluate their attitudes in refer-
ence to perceived attitudes of those around them. In addi-
tion, in their theory of reasoned action, Ajzen and Fishbein
(1980) examine a number of factors underlying attitudes,
including concepts related to the perceptions of significant
others in an individual’s social network as well as perceived
social norms and pressures (for instance, evolving policy
environment and cultural factors). Essentially, this di-
mension focuses on the perception of social behavior (be-
yond the self) that contributes to attitude formation at the
individual level.
With respect to cannabis, existing research has empiri-
cally operationalized the attitude construct using a variety
of measures. Recent studies using large nationally repre-
sentative data sets appear to have employed either one or
both of the affective and cognitive dimensions to measure
attitudes through the following questions (or a close var-
iant of these): How do you feel about adults trying mari-
juana once or twice? (with response options on an approve/
disapprove scale) and How much do people risk harming
themselves physically and in other ways when they smoke
marijuana? (with response options on a no risk/great risk
scale; Black & Joseph, 2014; Choi et al., 2018; Han et al.,
2017; Keyes et al., 2011; Miech et al., 2015; Salas-Wright
et al., 2017).
Other studies have attempted to examine broader, so-
cial facets of individual attitudes by measuring beliefs
regarding parent/peer perceptions of cannabis-specific
attitudes among adolescents (e.g., How do you think your
close friends/parents would feel about you trying mari-
juana or hashish once or twice?; Hansen & Hansen, 2016;
Hohman, Crano, Siegel, & Alvaro, 2014; Keyes et  al.,
2011; Wu, Swartz, Brady, & Hoyle, 2015), and perceptions
regarding legalization of cannabis (e.g., Do you think the
e234
use of marijuana should be made legal or not?; Moeller &
Woods, 2015; Nielsen, 2010).
Methods
In this article, we develop two scales capturing medical
and recreational cannabis attitudes among older adults. In
light of the foregoing discussion, we hypothesize that these
scales will comprise three latent dimensions: affect, cogni-
tion, and social perception. Below, we describe the sample,
methods for scale construction and evaluation, as well as
methods for examining individual-level predictors of med-
ical and recreational cannabis attitudes.
Sample
The data for this study come from a statewide, cross-sec-
tional survey of 274 older Coloradans. The state of
Colorado provides an appropriate setting to examine med-
ical and recreation cannabis attitudes because both forms
of cannabis use are currently legal in the state. A  pur-
posive, nonrepresentative, sample of older adults was
recruited from senior centers, wellness and health clinics,
state-registered cannabis clubs, cannabis retail shops and
dispensaries located in each of Colorado’s geographically
defined 16 Planning and Service Areas served by the state’s
Area Agencies on Aging, as well as the Gerontology Center
at University of Colorado, Colorado Springs (UCCS).
Individuals were encouraged to participate regardless of
their history of cannabis use. The study was approved by
the UCCS Institutional Review Board.
Scale Construction
The survey consisted of 83 questions on a variety of topics,
including attitudes, prevalence, predictors, patterns, and
health outcomes associated with cannabis and its use. This
present study only focuses on attitude measurement—anal-
ysis of other survey questions is covered elsewhere.
To generate survey items on cannabis-specific attitudes,
we began by conducting a comprehensive literature review
on theoretical and empirical measurement of attitudes. An
expert panel consisting of a social psychologist, neurolo-
gist, and a clinical pharmacist drew on this review to con-
struct the first iteration of attitude questions. Next, eight
focus groups of older adults were convened across four
Midwestern states with varying levels of legal cannabis
access. The focus group participants were asked to review
the draft set of survey questions, with attitude questions
embedded in the larger instrument, and then discuss their
overall attitudes and perceptions about using cannabis.
The focus group discussions illuminated that older
adults potentially perceive medical and recreational can-
nabis differently, which aided in the refinement of the at-
titude questions. In particular, participants embraced the
The Gerontologist, 2020, Vol. 60, No. 4
use of cannabis for medical purposes as legitimate, while
recreational cannabis use was perceived more variably,
with many associating it with societal risks and others
comparing it favorably against alcohol. To develop a con-
cise set of items, a preliminary analysis of survey responses
from 80 focus group participants was conducted to help
with item-reduction and response scale adjustment. These
analyses led to a pool of 22 attitude questions under
the following headings: “attitudes on medical use (7
questions),” “attitudes on recreational use (6 questions),”
and “general attitudes about marijuana (9 questions).”
After screening all items independently, a total of 12 items
Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023
were retained (The remaining 10 items were discarded
for one of the three reasons: (a) despite being placed in
a section on “Attitudes,” some survey items appeared to
reflect a different latent construct altogether. For example,
the question: Is medical (recreational) marijuana legal in
your state? was not used in scale construction because we
assessed them as primarily capturing respondent’s know-
ledge on state cannabis policies; (b) the item focused on
cannabis use in general instead of medical or recreational
use in particular; and (c) the item response had a different
level of measurement when compared with that of the
retained items.) These included six items for the medical
cannabis attitudes scale and six items for the recreational
cannabis attitudes scale.
In each scale, to operationalize the affective attitude di-
mension, we used two items assessing current and prior
acceptability of cannabis. Specifically, the survey asked
respondents the degree to which they agreed with the fol-
lowing statements: I currently believe that use of marijuana
for a medical (recreational) purpose is acceptable and When
I  was 18  years old, I  believed that using marijuana for a
medical (recreational) purpose was acceptable. The cog-
nitive attitude subscale was operationalized by examining
perceived harm associated with cannabis. Specifically, the
survey asked respondents the degree to which they agreed
with the following statements: Medical (recreational) ma-
rijuana use is risky and Using medical (recreational) mari-
juana leads to the use of harder drugs. Finally, to measure the
social component, we used two items assessing perceptions
that went beyond the individual. Specifically, the survey
asked respondents the degree to which they agreed with the
following statements: The important people in my life have
positive attitudes toward using medical (recreational) mari-
juana and Regardless of my current state law, I am in favor
of legalized marijuana for medical benefits (recreational
use). Participants responded to each item on a five-point
Likert scale (1 = “do not agree,” 5 = “strongly agree”).
To ensure that all items on each scale were coded in
the same direction, the two cognitive attitude items were
reverse coded before the scales were scored or reliability
analyses were conducted. On each scale, participant
responses across the six items were summed to generate
global attitude scores for medical and recreational cannabis
respectively.
The Gerontologist, 2020, Vol. 60, No. 4
Reliability
We assess scale reliability—a measure of how strongly the
scale’s items are correlated with each other—using two
measures of internal consistency: (a) item-rest correlations,
or the correlations between each item and the set of other
items included in the scale (Sabers & Gochyyev, 2010),
and (b) the proportion of total variation in the scale that
is shared by the included items as indicated by Cronbach’s
(1951) alpha coefficient. Alpha coefficients greater than .70
typically connote acceptable overall scale reliability (De
Vellis, 2003; Nunnally, 1978). We also examine whether the
scales retain acceptable internal consistency reliability across
gender and age categories. Finally, we determine medical and
recreational attitude scores for each sample respondent by
summing across item responses in each respective scale.
Confirmatory Factor Analyses
We conduct CFA to test the factor structures expected to
underlie medical and recreational cannabis attitude scales.
For both scales, we hypothesize that a three-factor model
(Model 1), with affect, cognition and social perception as la-
tent factors, would be replicated in the analyses. Note even
though total score was used in later analyses, the three-factor
structure was tested to provide evidence of construct va-
lidity for both scales. We assess the standardized loadings for
each factor, which can be interpreted as the correlations be-
tween the observed variables and the unobserved variables.
We also consider the overall fit of each hypothesized model
to the sample using three goodness-of-fit indices. The χ2 sta-
tistic compares the tested model with a saturated model,
with the null hypothesis that there is no difference between
the observed data and our hypothesized model. We also ex-
amine the comparative fit index, which indicates how well
the specified model fits compared with a baseline model.
Higher values indicate a better fit, and values above 0.90 are
generally considered to indicate a good fit (Diamantopoulos
& Sigaw, 2000). The final goodness-of-fit measure includes
the root mean square error of approximation (RMSEA). This
indicates the closeness of the model fit in relation to its degrees
of freedom. Previous research has suggested that values less
than .05 indicate adequate fit (Browne & Cudeck, 1993).
For both scales, we compare model fit of the hypothesized
three-factor model with two alternative models. First, a one-
factor model (Model 2) is tested, in which all six items load on
one primary factor. Second, a two-factor model (Model 3) is
tested where a person-centered component (affect and cogni-
tion combined) and a social perception component are treated
as two latent factors. To examine generalizability of the scales,
we also test for measurement invariance across age groups.
Predictors of Medical and Recreational Cannabis
Attitudes
We examine individual-level predictors of medical and rec-
reational cannabis attitudes. Because the two dependent
e235
variables in this case—medical cannabis attitude score and
recreational cannabis attitude score—pertain to the same
individual, error terms are assumed to be correlated across
both equations. Thus, the medical and recreational cannabis
attitude scores are jointly modeled as a two-equation seem-
ingly unrelated regression (SUR) system (Zellner, 1962).
SUR employs generalized least squares estimation, which
takes account of the cross-equation disturbance covariance
to improve efficiency of estimates over those attained under
ordinary least squares (Greene, 2003). Predictors for both
equations include variables on which information is asked
of all survey respondents (irrespective of past cannabis use),
Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023
thus enabling us to utilize the largest possible sample size.
These variables include age, gender, ethnicity, education,
marital status, employment status, whether the respondent
is currently providing informal care to an adult over
60 years of age, Patient-Reported Outcomes Measurement
Information System (PROMIS) scores (The PROMIS was
developed by the National Institutes of Health [NIH] to
provide a standardized metric for measuring physical,
mental, and social health across chronic diseases (www.
nihpromis.org). PROMIS instruments are publicly avail-
able, were developed using item response theory, and have
been tested in more than 20,000 individuals drawn from
the general U.S. population [Cella and colleagues, 2010].)
for mental and physical health, whether the respondent
used prescription opioids in the past year, and whether the
respondent used benzodiazepines in the previous year.
Results
Table 1 presents frequency distributions of the 12 retained
items. Participant responses provided initial evidence of po-
tential incongruence between medical versus recreational
cannabis attitudes. In particular, although about 60%
of older adults strongly agreed with the statement that
the “use of medical marijuana is acceptable,” only about
30% strongly agreed with the parallel statement on the
acceptability of recreational cannabis. A  similar pattern
was observed for being “in favor of legalized marijuana.”
Although about 60% strongly agreed with the statement in
relation to medical cannabis, less than 30% did so for rec-
reational cannabis. In comparison to medical cannabis, a
greater proportion of older adults appeared to perceive rec-
reational cannabis as risky and as a potential gateway drug.
Scale Reliability and Dimensions
Estimates of internal consistency reliability are reported
in Table 2. Cronbach’s alpha coefficients for both scales
were well above the threshold of .70 (.86 for medical can-
nabis attitudes and .91 for recreational cannabis attitudes),
indicating strong overall internal consistency reliability.
Furthermore, Cronbach’s alpha coefficients remained re-
markably stable across age and gender subgroups for both
medical and recreational cannabis attitude scales. Each
e236 The Gerontologist, 2020, Vol. 60, No. 4

Table 1.  Attitudes on Medical and Recreational Cannabis

Do not Strongly
Medical cannabis use agree agree

[1] [2] [3] [4] [5]

I currently believe that use of marijuana for a medical purpose is 5% 3% 15% 18% 59%
­acceptable. (N = 265)
When I was 18, I believed that using marijuana for a medical purpose 53% 9% 21% 6% 12%
was acceptable. (N = 256)
Medical marijuana use is risky. (N = 263) 28% 23% 28% 11% 10%
Using medical marijuana leads to the use of harder drugs (N = 262) 47% 16% 16% 6% 15%
The important people in my life have positive attitudes toward using 18% 11% 20% 20% 31%

Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023

medical marijuana. (N = 261)
Regardless of my current state law, I am in favor of legalized marijuana 8% 2% 13% 20% 60%
for medical benefits (N = 256)
Recreational marijuana use is risky. (N = 266) 18% 17% 17% 15% 33%
Using recreational marijuana leads to the use of harder drugs (N = 264) 35% 16% 16% 12% 22%
The important people in my life have positive attitudes toward using 31% 14% 24% 15% 17%
recreational marijuana. (N = 264)
Regardless of my current state law, I am in favor of legalized marijuana 31% 12% 13% 16% 28%
for recreational use (N = 257)

of the item-rest correlations greatly exceeded the cutoff
of .20, demonstrating strong reliability (Kline, 1986).
With the exception of the item assessing acceptability of
cannabis at age 18, all item-rest correlations were above
0.50. Specifically, the smallest item-rest correlation (.34)
was for the item on acceptability of medical cannabis
when younger among respondents over age 80. (When we
removed the item on acceptability of cannabis at age 18,
Cronbach’s alpha improved only marginally [from .86 to
.87 for medical cannabis attitudes and from .91 to .92 for
recreational cannabis attitudes]. Furthermore, the retention
of this item is justified on theoretical grounds as it may shed
light on the evolution of cannabis-specific attitudes over
the life course. Individuals with positive evaluations of can-
nabis acceptability in both young and old ages are likely to
have greater overall affective strength when compared with
those with a positive evaluation of cannabis acceptability in
older age only. Although this is difficult to test empirically
in this article as we do not have other items to measure
“affect,” we do find evidence that different groupings of
individuals exist. For instance, 17% [15%] of respondents
held positive evaluations of medical (recreational) cannabis
acceptability both currently and at age 18. However, 26%
[8%] of respondents held a negative evaluation of medical
(recreational) cannabis acceptability at age 18 but, at the
same time, a positive evaluation of it at current time. It is
possible that overall affective strength varies depending on
group membership. Therefore, the item on cannabis accept-
ability at age 18 was retained.) Similar to Cronbach’s alpha
coefficients, differences in item-rest correlations across
subgroups were small within both scales.
To verify that the items on each of the two scales
tapped into three different attitude components, CFA was
conducted separately for each scale. For both medical
and recreational cannabis attitudes, the data supported a
three-factor structure (Model 1), with dimensions similar
to those described earlier. In both attitude scales, the two
acceptability items loaded on the affective dimension; the
riskiness and gateway drug items loaded on the cognitive
dimension; and the significant other attitudes and views on
legalization loaded on the social perception dimension. The
standardized loadings for all indicators were above 0.60
on the corresponding dimensions. One exception was the
item on acceptability of cannabis at age 18. In the medical
cannabis scale, this item had a standardized loading of 0.45
reflecting moderate strength. In the recreational cannabis
scale, the standardized loading for this item was 0.56.
Fit statistics generally indicated a good fit of the model
to the data for both scales (Table 3). The χ2 statistic was
insignificant in both cases (medical cannabis: χ2  =  9.48,
p = .12; recreational cannabis: χ2 = 5.34, p = .50). This in-
dicated that we cannot reject the null hypothesis that there
is no difference between the patterns observed in these data
and the theoretically specified model. The comparative fit
indices of 0.99 in the medical cannabis attitudes case and
1.00 in the recreational cannabis attitudes case were both
well above the threshold of 0.90 for indication of a good
fit compared with the baseline model (Diamantopoulos
& Sigaw, 2000). The RMSEA of 0.04 in the medical can-
nabis attitudes scale and 0.01 in the recreational cannabis
attitudes scale indicated that the model is an adequate fit to
the data because for both scales, the value was lower than
the typical cutoff of 0.05 (Browne & Cudeck, 1993).
Table 3 also presents fit indices of two alternative factor
analytical models. For both medical and recreational can-
nabis attitude scales, χ2 values for Models 2 and 3 were
The Gerontologist, 2020, Vol. 60, No. 4 e237

Table 2.  Reliability of Recreational and Medical Cannabis Attitudes Scales, by Age Group and Gender

Item-rest correlations

Age groups Gender

Overall 60–70 70–80 80+ Female Male

Medical cannabis use

  Acceptable now .76 .81 .72 .74 .77 .73
  Acceptable when younger .44 .41 .47 .34 .46 .41
 Risky .62 .65 .53 .64 .63 .56
  Leads to use of harder drugs .74 .75 .73 .69 .73 .75
  Important people in favor .60 .64 .57 .50 .58 .69

Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023

  In favor of legalization .76 .84 .68 .71 .79 .69
  Cronbach’s alpha, N .86, N = 237 .87, N = 95 .84, N = 108 .83, N = 34 .86, N = 147 .85, N = 79
Recreational cannabis use
  Acceptable now .89 .91 .84 .89 .90 .88
  Acceptable when younger .52 .50 .43 .41 .49 .56
 Risky .76 .72 .73 .78 .78 .72
  Leads to use of harder drugs .72 .80 .65 .55 .71 .78
  Important people in favor .72 .73 .70 .61 .68 .81
  In favor of legalization .87 .91 .84 .72 .89 .86
  Cronbach’s alpha, N .91, N = 244 .91, N = 95 .88, N = 111 .86, N = 38 .91, N = 151 .92, N = 82

Table 3.  Fit Indices for Confirmatory Factor Models for individuals in the 80+ age category with less than 40
individuals.
Chi-square
As displayed in Table 4, indices showed good model fit
(df; p value) RMSEA CFI
in the two age categories, indicating similar configural pat-
Medical cannabis attitudes scale tern across the two cohorts for both medical and recrea-
  Model 1 (three factor) 9.48 (6; .12) 0.04 0.99 tional cannabis attitudes. We then tested for measurement
  Model 2 (one factor) 50.44 (9; .00) 0.13 0.93 invariance across age groups by comparing a model with
  Model 3 (two factor) 46.65 (8; .00) 0.14 0.94 factor loadings freely estimated with another, where factor
  N = 237 loadings were constrained to be equal across the two age
Recreational cannabis attitudes scale categories. As given in Table 5, both models fit the data
  Model 1 (three factor) 5.34 (6; .50) 0.01 1.00 quite well. Comparison between these two models showed
  Model 2 (one factor) 42.86 (9; .00) 0.12 0.96 that constraints of factor loadings did not significantly
  Model 3 (two factor) 42.80 (8; .00) 0.13 0.96 worsen the goodness of fit (medical scale: Δχ2  =  5.16,
  N = 244 Δdf  =  3, p > .05; recreational scale: Δχ2  =  4.22, Δdf  =  3,
p > .05). The results suggested that the constrained model
Note: CFI = confirmatory fit index; RMSEA = root mean square error of ap-
was preferred over the freely estimated one for both scales,
proximation.
revealing cross-age measurement invariance.

large and statistically significant. In addition, the RMSEA

for these models was also greater than 0.05 for both scales.
From these results, Model 1 was selected as the best fit
for the data. Standardized factor loadings and multiple
correlations for all models are provided in Supplementary
Material.
A multigroup structural analysis was used to investi-
gate whether the three factors from the CFA were invariant
across age groups. Age was treated as a binary variable,
with overall sample divided into those above or below the
mean age of 72.74 in the sample. We created two groups
based on mean age instead of three age categories (60–70,
71–80, and 80+) used in other analyses in the article due
to computational issues experienced in conducting a CFA
Medical and Recreational Cannabis Attitudes
Among Older Adults
The scores on both attitude scales ranged from a min-
imum of six points to a maximum of 30 points. For both
scales, higher scores indicated more positive/favorable
attitudes toward the type of cannabis use under consid-
eration. If an individual had a missing response to any
one of medical or recreational scale items, they were not
included in the calculation of attitude scores for that par-
ticular scale. This results in 237 complete observations
for the medical cannabis attitudes scale and 244 complete
observations for the recreational cannabis attitudes scale.
(We compared means of independent variables across
e238 The Gerontologist, 2020, Vol. 60, No. 4

Table 4.  Fit Indices for the Three-Factor Model in Different potential incongruence in the two types of attitudes for
Age Groups some individuals.
Next, we conducted a SUR estimation predicting med-
Chi-Square
(df; p value) RMSEA CFI
ical and recreational cannabis scores as a function of
various individual-level characteristics. After removing
Medical cannabis attitudes scale observations with missing values on other covariates, a
  Age < 72.47 8.40 (6; .21) 0.05 0.99 sample size of 192 individuals was retained. Among these,
  Age > 72.47 4.03 (6; .67) 0.01 1.00 about 37% were between the ages of 60–70, 48% between
Recreational cannabis attitudes scale the ages of 71–80, and 15% over age 80. Furthermore, 64%
  Age < 72.47 7.46 (6; .28) 0.04 0.99 of the sample were female and about 7% Hispanic. A ma-
  Age > 72.47 2.81 (6; .83) 0.01 1.00 jority of this sample reported to be married/partnered and
retired. Approximately 13% of the sample was currently

Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023

Note: CFI = confirmatory fit index; RMSEA = root mean square error of ap-
proximation.
Table 5.  Cross-Age Invariance
Chi-square
(df; p value) RMSEA CFI
Medical Cannabis Attitudes Scale
  Factor Loadings Freely Estimated 22.12 (14; .07) 0.07 0.98
 Factor Loadings Constrained to 27.28 (17; .05) 0.07 0.98
be Equal
Recreational Cannabis Attitudes Scale
  Factor Loadings Freely Estimated 14.28 (14; .42) 0.01 1.00
 Factor Loadings Constrained to 18.50 (17; .35) 0.02 0.99
be Equal
Note: CFI = confirmatory fit index; RMSEA = root mean square error of ap-
proximation.
complete and missing observations. A  t-test of mean
differences between the groups revealed only two statisti-
cally significant results. Individuals over age 80 were more
likely to miss questions on both medical and recreational
cannabis attitudes. In addition, widowed respondents
were more likely to miss questions on medical cannabis
attitudes. No other systematic pattern was found between
the two groups.)
The mean (SD, skewness, kurtosis) score of the med-
ical cannabis attitudes scale was 21.22 (6.03, −0.42, 1.97)
and that of the recreational cannabis attitudes scale was
17.11 (7.73, 0.66, 2.28). This yielded a statistically signifi-
cant paired two-tailed t-test (t = 13.33; df = 229; p < .001)
leading us to reject the null hypothesis that the mean dif-
ference between medical and recreational cannabis attitude
scores is equal to zero. Cohen’s d for this paired samples
t-test was 0.87th, indicating a large effect. Mean (SD) med-
ical (25.30 [3.04]) and recreational (23.25 [5.96]) cannabis
attitudes scores were greater among those who reported
cannabis use in the past year. Although medical and rec-
reational cannabis attitude scores were strongly correlated
in the positive direction (r = .81, p < .01), approximately
12% of respondents simultaneously scored above average
on the medical cannabis attitudes scale and below average
on the recreational cannabis attitudes scale, suggesting
providing informal care to an individual over age 60. About
24% and 15% had used opioids and benzodiazepines in the
previous year. The average (SD) PROMIS Mental Health
and Physical Health Scores were 15.50 (3.10) and 15.42
(3.04), respectively, with a range of 6–20. Though this var-
iable is not included in the regressions given concerns of
reverse causality, about 42% of this sample reported using
cannabis is the previous year.
Table 6 provides results of the SUR estimation. In ge-
neral, age was negatively associated with cannabis attitudes.
However, when compared with being in the age 60–70 cat-
egory, being in older age categories had a much stronger
negative association with recreational cannabis attitude
scores relative to medical cannabis attitude scores. The
negative coefficients on older age categories were also sta-
tistically significant only for recreational cannabis attitude
scores. When compared with being a high school graduate,
being more educated consistently had a strong statistically
significant positive effect on medical and recreational can-
nabis attitudes. Although better physical health was posi-
tively associated with medical cannabis attitude scores, this
association was not statistically significant for recreational
cannabis attitudes. Conversely, better mental health was
negatively associated with both types of cannabis attitudes.
Interestingly, being a caregiver was positively associated
with attitudes toward recreational cannabis, but not to-
ward medical cannabis. Though statistically insignificant in
both cases, being female had contrasting associations with
attitudes for medical and recreational cannabis. In partic-
ular, it was positively associated with medical cannabis
attitudes, but negatively associated with recreational can-
nabis attitudes. The R2 for both equations was small (.23
for medical cannabis attitudes and .33 for recreational can-
nabis attitudes), indicating that the model explained lim-
ited variability in attitudes.
Discussion and Implications
By combining multiple indicators of attitudes, we have
developed two reliable and valid scales assessing medical
and recreational cannabis attitudes among older adults.
Our results support the hypothesized three-factor struc-
ture (i.e., affect, cognition, and social perception) of
both attitude scales. This representation was found to be
The Gerontologist, 2020, Vol. 60, No. 4 e239

Table 6.  Seemingly Unrelated Regression Estimation of heightened risk for adverse effects such as greater levels of
Medical and Recreational Cannabis Attitude Scores injury and emergency department visits, psychiatric dis-
tress, mental illness, and comorbid substance use (Choi
Recreational
Medical cannabis cannabis attitude
et al., 2018; Han et al., 2017). Attitude scales developed
attitude score score in this article can enable researchers to identify and target
health behavior interventions to “high risk” older adults.
Age 71–80 –0.64 (0.84) –2.91** (1.03) For example, interventions educating older adults on po-
Age 80+ –2.51 (1.54) –5.66**(1.87) tential risks of drug interactions and concurrent use of
Female 0.62 (0.88) –1.34 (1.07) multiple substances may be designed specifically for those
Hispanic 0.50 (1.58) 0.09 (1.92) with strong favorable attitudes toward both recreational
Some college 3.37* (1.48) 2.42 (1.80) and medical cannabis. Similarly, other interventions fo-
College graduate 5.32*** (1.51) 5.64** (1.83) cused on health benefits of cannabis may be designed for
Graduate degree 4.92** (1.43) 5.31** (1.73)

Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023

older adults currently using opioids and holding strong
Divorced/separated –0.67 (1.07) –0.68 (1.29)
unfavorable attitudes toward medical cannabis. In this
Widowed –1.85 (1.11) –2.04 (1.34)
way, tailored interventions can be designed to change use
Never married 1.27 (2.23) 4.01 (2.70)
behavior (including frequency and mode of cannabis use)
Retired –0.12 (1.30) –0.49 (1.58)
among heterogeneous groups of older adults.
PROMIS_MH –0.46** (0.17) –0.44* (0.21)
Second, the development of these scales can potentially
PROMIS_PH 0.38* (0.19) 0.40 (0.22)
enhance future research by enabling researchers to empir-
Opioid use 1.78 (1.08) 2.29 (1.31)
Benzo use 1.72 (1.26) 2.67 (1.53)
ically account for several attitudinal dimensions by using
Caregiver 1.91 (1.17) 3.88** (1.42) two parsimonious indicators. This is relevant for analyses
Constant 18.40*** (3.14) 16.45*** (3.82) examining cannabis-specific attitudes as an outcome of
R2 0.24 0.33 marijuana legalization policies as well as those that use
N 192 cannabis attitudes as a risk factor in modeling variation
in cannabis use. Third, in light of potential distinctions be-
Notes: SE in parentheses. tween medical and recreational cannabis attitudes held by
*p < .05, **p < .01, ***p < .001. the same person, large nationally representative data sets
should consider separating questions on cannabis attitudes
invariant across two age categories. Findings of this study and use by purpose of use. Finally, our results also sug-
also provide evidence for our hypothesis that medical gest that when considering changes in marijuana legaliza-
and recreational cannabis attitudes among older adults tion policies, policy makers should assess attitudes by both
are positively related but distinct. Results also show that demographic subgroups and type of use.
favorability of recreational cannabis decreases with age; The results of this study should be interpreted with cau-
however, the same is not case for medical cannabis. In tion given nonrepresentative sampling and the relatively
contrast, being a caregiver is positively associated with small sample size. The study also has limited generaliza-
recreational cannabis attitudes, but not with medical can- bility as the survey results pertain to older adults in only
nabis attitudes. Although physical health was positively one state. Finally, due to limited sample size, we were un-
associated with medical cannabis attitudes, we found no able to examine associations between cannabis attitudes
detectable relationship between physical health and rec- and health outcomes and patterns of use. Despite these
reational cannabis. Education and better mental health limitations, we expect these results to move us closer to
generally predicted both types of cannabis attitudes in a understanding the relationship between public policy, can-
consistent direction. nabis use, and attitudes among older adults. This is espe-
The results of this study have important public health cially pertinent in light of the rapidly aging baby boom
implications. First, programs designed to promote changes population and the related projection of continued increase
in health behavior frequently include components that at- in the prevalence of cannabis use among this population.
tempt to change attitudes (Stacy, Bentler, & Flay, 1994).
Development of effective attitude-change interventions
may emerge as a priority in the coming years given sub- Supplementary Material
stantial increase in cannabis use over the past decade Supplementary data are available at The Gerontologist
among adults 65 and older as well as mixed evidence on online.
outcomes of cannabis use among older adults. On the
one hand, cannabis has been shown to have benefits in
treating certain conditions that affect older adults, in- Funding
cluding chronic pain and nausea (Han et  al., 2017). This work was supported by funding from the Colorado Department
On the other hand, however, studies have also shown of Public Health and Environment.
e240
Acknowledgments
The authors acknowledge Emma Cole and Karrey Shannon for their
assistance with project management. K.  Arora planned the study,
helped in survey instrument development, conducted the data anal-
ysis, and wrote the paper. S. H. Qualls, J. Bobbit, H. D. Lum, G. M.
Milavetz helped with survey instrument development and imple-
mentation and provided feedback on the manuscript. J.  Croker
assisted with data analysis. B.  Kaskie conceptualized the survey,
helped in study planning and survey instrument development and
implementation, and provided feedback on study results and the
manuscript.
Conflict of Interest
None reported.
References
Ahmed, A. I. A., van den Elsen, G. A., van der Marck, & Rikkert, M. G.
(2014). Medicinal use of cannabis and cannabinoids in older
adults: Where is the evidence? Journal of the American Geriatrics
Society, 62, 410–411. doi:10.1111/jgs.12661
Ajzen,  I., & Fishbein,  M. (1980). Understanding attitudes and
predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.
Bachman, J. G., Johnson, L. D., & O’Malley, P. M. (1998). Explaining
recent increases in students’ marijuana use: Impacts of perceived
risks and disapproval, 1976 through 1996. American Journal of
Public Health, 88, 887–892. doi:10.2105/AJPH.88.6.887
Black, P., & Joseph, L. J. (2014). Still dazed and confused: Midlife
marijuana use by the baby boom generation. Deviant Behavior,
35, 822–841. doi:10.1080/01639625.2014.889994
Breckler, S. J., & Wiggins, E. C. (1993). Emotional responses and the
affective component of attitude. Journal of Social Behavior and
Personality, 8, 281.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing
model fit. In K.  A.  Bollen and J.  S.  Long (Eds.), Testing struc-
tural equation models (Vol. 154, pp. 136–162). Newbury Park,
CA: Sage.
Cella, D., Riley, W., Stone, A., Rothrock, N., Reeve, B., Yount, S., ...
Hays, R. (2010). The Patient-Reported Outcomes Measurement
Information System (PROMIS) developed and tested its
first wave of adult self-reported health outcome item banks:
2005–2008. Journal of Clinical Epidemiology, 63, 1179–1194.
doi:10.1016/j.jclinepi.2010.04.011
Choi,  N.  G., Marti,  C.  N., DiNitto,  D.  M., & Choi,  B.  Y. (2018).
Older adults’ marijuana use, injuries, and emergency department
visits. The American Journal of Drug and Alcohol Abuse, 44,
215–223. doi:10.1080/00952990.2017.1318891
Crites,  S.  L., Fabrigar,  L.  R., & Petty,  R.  E. (1994). Measuring
the affective and cognitive properties of attitudes: Conceptual
and methodological issues. Personality and Social Psychology
Bulletin, 20, 619–634. doi:10.1177/0146167294206001
Cronbach, L. J. (1951). Coefficient alpha and the internal structure
of tests. Psychometrika, 16, 297–334. doi:10.1007/BF02310555
De Vellis, R. F. (2003). Scale development: Theory and applications
(2nd ed., Vol. 26). Thousand Oaks, CA: Sage Publications.
Diamantopoulos,  A., & Sigaw,  J. (2000). Introducing LISREL.
London: Sage Publications. doi:10.4135/9781849209359
The Gerontologist, 2020, Vol. 60, No. 4
Eagly,  A.  H., & Chaiken,  S. (1993). The psychology of attitudes.
Orlando, FL: Harcourt Brace Jovanovich College Publishers.
Festinger, L. (1954). A theory of social comparison processes. Human
Relations, 7, 117–140. doi:10.1177/001872675400700202
Greene,  W.  H. (2003) Econometric analysis. Delhi, India: Pearson
Education India.
Han, B. H., Sherman, S., Mauro, P. M., Martins, S. S., Rotenberg, J.,
& Palamar, J. J. (2017). Demographic trends among older can-
nabis users in the United States, 2006–13. Addiction, 112, 516–
525. doi:10.1111/add.13670
Hansen,  W.  B., & Hansen,  J.  L. (2016). Using attitudes, age
and gender to estimate an adolescent’s substance use risk.
Journal of Children’s Services, 11, 244–260. doi:10.1108/
Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023
JCS-06-2015-0020
Hess,  T.  M. (2006). Seventeen - Attitudes toward aging and their
effects on behavior. In J. E. Birren, K. W. Schaie, R. P. Abeles,
M. Gatz, & T. A. Salthouse (Eds.), Handbook of the psy-
chology of aging (6th ed., pp. 379–406). Academic Press. ISBN
9780121012649
Hohman, Z. P., Crano, W. D., Siegel, J. T., & Alvaro, E. M. (2014).
Attitude ambivalence, friend norms, and adolescent drug.
Prevention Science, 15, 65. doi:10.1007/s11121-013-0368-8
Kaskie, B., Ayyagari, P., Milavetz, G., Shane, D., & Arora, K. (2017).
The increasing use of cannabis among older Americans: A public
health crisis or viable policy alternative? The Gerontologist, 57,
1166–1172. doi:10.1093/geront/gnw166
Keyes, K. M., Schulenberg, J. E., O’Malley, P. M., Johnston, L. D.,
Bachman,  J.  G., Li,  G., & Hasin,  D. (2011). The social
norms of birth cohorts and adolescent marijuana use in
the United States, 1976–2007. Addiction, 106, 1790–1800.
doi:10.1111/j.1360-0443.2011.03485.x
Kline, P. (1986). A handbook of test construction: Introduction to
psychometric design. New York: Methuen.
Miech,  R.  A., Johnston,  L., O’Malley,  P.  M., Bachman,  J.  G.,
Schulenberg, J., & Patrick, M. E. (2015). Trends in use of ma-
rijuana and attitudes toward marijuana among youth before
and after decriminalization: The case of California 2007–
2013. The International Journal on Drug Policy, 26, 336–344.
doi:10.1016/j.drugpo.2015.01.009
Moeller, K. E., & Woods, B. (2015). Pharmacy students’ knowledge
and attitudes regarding medical marijuana. American Journal of
Pharmaceutical Education, 79, 85. doi:10.5688/ajpe79685
Nielsen, A. L. (2010). Americans’ attitudes toward drug-related issues
from 1975–2006: The roles of period and cohort effects. Journal of
Drug Issues, 40, 461–493. doi:10.1177/002204261004000209
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). Hillsdale, NJ:
Mcgraw-Hill.
Rosenberg,  M.  J. (1960). Cognitive, affective, and behavioral
components of attitudes. In M. J. Rosenberg & C. I. Hovland
(Eds.), Attitude organization and change: An analysis of con-
sistency among attitude components. New Haven, CY: Yale
University Press.
Sabers,  D., & Gochyyev,  P. (2010). Item-test correlation. In
N. J. Salkind (Ed.), Encyclopedia of research design (pp. 651–
653). Thousand Oaks, CA: Sage.
Salas-Wright,  C.  P., Vaughn,  M.  G., Cummings-Vaughn,  L.  A.,
Holzer,  K.  J., Nelson,  E.  J., AbiNader,  M., & Oh,  S. (2017).
Trends and correlates of marijuana use among late middle-aged
The Gerontologist, 2020, Vol. 60, No. 4
and older adults in the United States, 2002–2014. Drug
and Alcohol Dependence, 171, 97–106. doi:10.1016/j.
drugalcdep.2016.11.031
Simons, J., & Carey, K. B. (1998). A structural analysis of attitudes to-
ward alcohol and marijuana use. Personality and Social Psychology
Bulletin, 24, 727–735. doi:10.1177/0146167298247005
Stacy,  A.  W., Bentler,  P.  M., & Flay,  B.  R. (1994). Attitudes and
health behavior in diverse populations: Drunk driving. Alcohol
use, binge eating, marijuana use, and cigarette use. Health
Psychology, 13, 73–85. doi:10.1037/0278-6133.13.1.73
Stoner, S. A. (2016) Learn about marijuana: Factsheets: Marijuana
use by older adults. Retrieved October 22, 2018, from http://
learnaboutmarijuanawa.org/factsheets/olderadults.htm
Van der Pligt, J., Zeelenberg, M., van Dijk, W. W., de Vries, N. K.,
& Richard,  R. (1997). Affect, attitudes and decisions: Let’s be
e241
more specific. European Review of Social Psychology, 8, 33–66.
doi:10.1080/14792779643000074
Wang,  M., & Chen,  Y. (2006). Age differences in attitude change:
Influences of cognitive resources and motivation on responses
to argument quantity. Psychology and Aging, 21, 581–589.
doi:10.1037/0882-7974.21.3.581
Wu, L. T., Swartz, M. S., Brady, K. T., & Hoyle, R. H.; NIDA AAPI
Workgroup. (2015). Perceived cannabis use norms and can-
nabis use among adolescents in the United States. Journal of
Psychiatric Research, 64, 79–87. doi:10.1016/j.jpsychires.2015.
02.022
Zellner, A. (1962). An efficient method of estimating seemingly un-
related regressions and tests for aggregation bias. Journal of the
Downloaded from https://academic.oup.com/gerontologist/article/60/4/e232/5489149 by guest on 07 April 2023
American Statistical Association, 57, 348–368. doi:10.1080/016
21459.1962.10480664