Introduction
The prevalence of marijuana use has doubled in the past decade. One in eight US adults
now reports use of marijuana in the past year.1 Despite increasing use, our understanding
of the risks and benefits of marijuana use is limited.1–3 The National Academy of
Sciences, editorials, and numerous systematic reviews have called for more research.1–3
Despite concerns within the scientific community, only 33% of the public perceive
“great risk” from weekly marijuana use compared to 50% in 2002.1 The public appears
to believe that recreational marijuana use is safe and even has health benefits.1
The media contributes greatly to the national perspective on health 4 but may provide
incomplete information.4 We examined a cross-section of news media coverage to better
understand how the media may influence the public’s view of marijuana.
Methods
We identified the top 10 major news outlets by print circulation using Pew’s 2014
ranking.5 We selected those with articles available on LexisNexis, including The New
York Times, The Daily News New York, The New York Post, The Denver Post, USA Today,
and the Los Angeles Times; these have a combined total of approximately 4.9 million
daily visitors.
We selected articles published between 1/1/12 (Colorado’s recreational legalization)
and 5/1/2016 (the study’s initiation date), with major subject “marijuana.” If more
than 100 articles were available from any publication, 100 were randomly selected.
We excluded articles < 100 words, editorials, and those without a marijuana focus.
One reviewer (AW or RA) categorized each article by focus health, business, public
policy, crime, and entertainment/other and flagged them for health mentions. Each
article with a mention was abstracted independently by two reviewers (AZ, RA) for
overall message content. Overall content was categorized as follows: (1) marijuana
benefits outweigh harms, (2) marijuana harms outweigh benefits, or (3) neutral impression
of harms vs. benefits. Disagreements were adjudicated by a third abstractor (AA).
We analyzed the relationship between the article’s overall slant and other characteristics
using descriptive statistics.
Data Availability
Dataset available from corresponding author on request.
Results
We identified 564 articles of which 477 met inclusion criteria. Articles’ focus were
categorized as public policy (n = 190, 39.8%), crime (n = 113, 23.7%), business (n = 74,
15.5%), entertainment/other (n = 69, 14.5%), and health (n = 31, 6.5%). Health mentions
were found in a third of articles (n = 145, 30.4%), of which 89 (61.3%) mentioned
health benefits, and 98 (67.5%) mentioned harms (Table 1). About half of articles
with health mentions (n = 72, 49.7%) communicated a favorable impression of marijuana
(net benefit), 46 (31.7%) communicated a negative impression (net harm), and 27 (18.6%)
were neutral. Articles that were not focused on health were more likely to communicate
a favorable impression of marijuana (p < .001). Of the 31 articles that were focused
on health and public health, 13 (41.9%) acknowledged weaknesses in the evidence base
surrounding marijuana while 18 (58.1%) did not. Articles discussing both benefits
and harms were more likely to communicate a favorable impression of marijuana (p < .001).
The proportion of articles with favorable mentions of marijuana increased from 11.1%
in 2012 to 65% in 2016.
Table 1
Characteristics of Articles That Mentioned Health Effects
Positive impression (benefits outweigh harms) (N = 72)
Negative impression (harms outweigh benefits) (N = 46)
Neutral impression (N = 27)
Total
p value
Article focus
Health and public health
6 (19.4)
21 (67.7)
4 (12.9)
31
< .001
Non-health*
66 (57.9)
25 (21.9)
23 (20.2)
114
Total
72 (49.7)
46 (31.7)
27 (18.6)
145
Publication
USA Today
11 (33.3)
16 (48.5)
6 (18.2)
33
< .001
New York Times
13 (39.4)
11 (33.3)
9 (27.3)
33
Los Angeles Times†
21 (77.8)
2 (7.4)
4 (14.8)
New York Post
9 (60.0)
4 (26.7)
2 (13.3)
15
The Denver Post
5 (26.3)
9 (47.4)
5 (26.3)
Daily News New York
13 (72.2)
4 (22.2)
1 (5.6)
18
Health effects listed in article‡
Described both benefits and harms
24 (54.5)
9 (20.5)
11 (25.0)
44
< .001
Described benefits only
42 (93.3)
0 (0)
3 (6.7)
45
Described harms only
5 (9.3)
37 (68.5)
12 (22.2)
54
Year of publication
2012
1 (11.1)
6 (66.7)
2 (22.2)
9
0.10
2013
8 (36.4)
8 (36.4)
6 (27.3)
22
2014
28 (48.3)
21 (36.2)
9 (15.5)
58
2015
22 (61.1)
7 (19.4)
7 (19.4)
36
2016
13 (65.0)
4 (20.0)
3 (15.0)
20
*Articles related to business, public policy, celebrity/entertainment/other, crime
†Only the past 6 months were available on LexisNexis
‡Two articles while focused on health did not mention explicit benefits or harms and
were therefore not included in this analysis
Discussion
Most articles on marijuana do not mention health effects. Nearly half of articles
in our sample that mentioned health effects communicated a favorable impression of
marijuana. Articles that focused exclusively on health-related issues of marijuana
were more likely to communicate a nuanced view on marijuana use; however, these articles
represented a small portion of all marijuana articles (6.5%, 31/477). Articles covering
public policy and business aspects of recreational legalization tended to communicate
more benefits than harms. Given that many reported therapeutic benefits have insufficient
evidence,2,3 the media’s coverage slant is concerning and may lead the public to underestimate
health risks associated with frequent marijuana use.1,6
Our study has limitations that deserve comment. Top print circulation newspapers are
dominated by coastal outlets whose favorable messaging may not be generalizable. On
the other hand, the newspapers included have an outsized influence on public discourse.
In addition, print has declined in relation to digital media. However, while our sampling
strategy relied on print articles, these articles are also available in digital format
on news websites and widely shared in social media.
In conclusion, major news outlets communicate a more positive message regarding marijuana
use than is warranted by current evidence. A concerted public health response is needed
to counter the inaccurate information provided to the public.