Content Analysis of Tobacco, Alcohol, and Other Drugs in Popular Music

Objective  To perform a comprehensive content analysis of substance use in contemporary popular music.

Design  We analyzed the 279 most popular songs of 2005 according to Billboard magazine. Two coders working independently used a standardized data collection instrument to code portrayals of substance use.

Outcome Measures  Presence and explicit use of substances and motivations for, associations with, and consequences of substance use.

Results  Of the 279 songs, 93 (33.3%) portrayed substance use, with an average of 35.2 substance references per song-hour. Portrayal of substance use varied significantly (P < .001) by genre, with 1 or more references in 3 of 35 pop songs (9%), 9 of 66 rock songs (14%), 11 of 55 R & B/hip-hop songs (20%), 22 of 61 country songs (36%), and 48 of 62 rap songs (77%). While only 2.9% of the 279 songs portrayed tobacco use, 23.7% depicted alcohol use, 13.6% depicted marijuana use, and 11.5% depicted other or unspecified substance use. In the 93 songs with substance use, it was most often motivated by peer/social pressure (45 [48%]) or sex (28 [30%]); use was commonly associated with partying (50 [54%]), sex (43 [46%]), violence (27 [29%]), and/or humor (22 [24%]). Only 4 songs (4%) contained explicit antiuse messages, and none portrayed substance refusal. Most songs with substance use (63 [68%]) portrayed more positive than negative consequences; these positive consequences were most commonly social, sexual, financial, or emotional.

Conclusions  The average adolescent is exposed to approximately 84 references to explicit substance use daily in popular songs, and this exposure varies widely by musical genre. The substance use depicted in popular music is frequently motivated by peer acceptance and sex, and it has highly positive associations and consequences.

If music be the food of love, play on; Give me excess of it, that, surfeiting, The appetite may sicken, and so die. That strain again! It had a dying fall: O, it came o’er my ear like the sweet sound That breathes upon a bank of violets, Stealing and giving odour! William Shakespeare, Twelfth Night

Although the influence of music on human beings was recognized even before the time of Shakespeare, current technological and social changes dramatically magnify that influence. While 15- to 18-year-old adolescents are forming health attitudes and behaviors that will last a lifetime, they are exposed to 2.4 hours of music per day, according to a large nationally representative study.¹ There are few limits to access; 98% of children and adolescents live in homes with radio and CD or MP3 players, and 86% of 8- to 18-year-old children and adolescents have CD or MP3 players in their bedrooms.¹ These figures have increased substantially even over the past decade.^1,2

There is convincing evidence that exposure to certain media messages increases substance use in adolescents.^3-10 For instance, viewing smoking in movies prospectively predicts a substantial proportion of adolescent smoking initiation.^4,11 Similarly, exposure to smoking-related media promotions is associated with smoking initiation.^5-8,12 Alcohol use in movies and promotions is also linked to actual alcohol use.^4,13-15

While the most frequently studied genres for this research include movies, television, and advertising, health behavior theory strongly supports a link between music exposure and substance use. According to the social learning model, human beings learn not only by direct experience but also by exposure to modeled behavior, such as that represented in popular music.^16-18 This theory further suggests that individuals exposed to representations of substance use would be more likely to perform those behaviors themselves if they are linked with (1) motivations that are relevant, (2) associations that are desirable and familiar, and (3) consequences that are positive.^16-18 Music is well-known to connect deeply with adolescents and to influence identity development, perhaps more than any other entertainment medium.^19-22

In addition, prior work suggests that references to substances of abuse in music are common. Several years ago, a content analysis published by the Office of National Drug Control Policy^23,24 showed that multiple messages related to substance use are present in music lyrics and music videos. Of the top 1000 popular songs they studied, 18% referenced illicit drugs and 17% referenced alcohol.²³ Another report²⁴ found that, of 258 popular music videos, 20% verbally referenced illicit drugs and 37% displayed alcohol. In every case, marijuana was the illicit drug most commonly represented.^23,24

However, to our knowledge, no comprehensive content analysis of substance use in popular music lyrics has been published in the peer-reviewed medical literature. Furthermore, popular music is rapidly changing, and it has been more than a decade since the previous data described were collected. This is a particularly important omission because popular music exposure is increasing among young people.¹ Finally, a more comprehensive and theoretically based content analysis may more completely capture relevant factors, such as the motivations, associations, and consequences associated with substance use in popular music.

The purpose of this study was to perform a comprehensive content analysis of contemporary popular music, focusing on the presence and use of substance use and on the motivations, associations, and consequences of substance use. Based on prior data,^23-25 our a priori hypothesis was that alcohol and marijuana would be more commonly represented than tobacco. We also hypothesized that representation of substance abuse would differ among genres, and that use would be commonly juxtaposed with motivations, associations, and consequences likely to be deemed positive by adolescents.

We used Billboard magazine to identify the most popular songs of 2005.²⁶Billboard annually uses an algorithm that integrates data from sales and airplay to determine the top songs according to exposure. Sales data for this algorithm are compiled by Nielsen SoundScan from merchants representing more than 90% of the US music market, including sales from music stores, direct-to-consumer transactions, and Internet sales and downloads. Billboard's airplay data use Nielsen Broadcast Data Systems, which electronically monitors radio stations in more than 120 representative markets across the United States. Integrating these data, Billboard reported the following youth-relevant lists of popular song titles in 2005: the “Pop 100” (n = 100), the “Billboard Hot 100” (n = 100), “Hot Country Tracks” (n = 60), “Hot R & B/Hip-Hop Songs” (n = 100), “Hot Rap Tracks” (n = 25), “Mainstream Rock Tracks” (n = 40), and “Modern Rock Tracks” (n = 40). The lists are closed out at year end, after which time the song rankings do not change. Because some songs were included on more than 1 chart, only 279 unique songs composed this sample. Additional popular charts, such as the “Adult Top 40,” were also available, but these songs were not included in this analysis because we were primarily interested in youth exposure to popular music.

For each of the 279 songs, we coded multiple elements related to substance use. These measures were selected based on a comprehensive search of prior relevant content analyses of media^10,23,24,27 and inclusion of other measures based on the social cognitive theory (such as motivations for, associations with, and consequences of substance use).^16-18 Two coders familiar with popular music then independently analyzed the lyrics of each song for content related to each of the measures. Before the content analysis, each coder was given lists of frequently used substance use slang terms.

We computed interrater agreement and κ statistics²⁸ for each of the data elements coded and found a minimum of 74% agreement for all variables. In all interrater disagreements, we used 2 new confirmatory coders (M.V.C. and A.A.A.) to independently code each of the items on which the previous coders did not agree (blinded to the prior responses). When the confirmatory coders both agreed with 1 of the original coders, the coding of that individual was recorded. However, when the confirmatory coders disagreed with each other or agreed with each other but not with 1 of the initial coders, the item was discussed by the coders and the principal investigator of the project (B.A.P.) to achieve a consensus.

We coded descriptive information related to each song from Billboard's records, including song title, artist, album, song length in minutes and seconds, sex of singer, and primary song genre. Because songs are often associated with more than 1 genre, we used the following standardized approach to identify the primary genre. First, we determined its highest position at any time on each of the Billboard charts we analyzed. All songs were assigned to a primary genre based on their highest ranking on a specialty chart, regardless of the ranking on the “pop” or “hot” charts. For example, Gwen Stefani's song “Hollaback Girl” peaked at No. 1 on the pop 100 chart, No. 1 on the hot 100 chart, and No. 8 on the R & B/hip-hop chart and it was, therefore, classified as an R & B/hip-hop song. Only songs that never reached any specialty chart but did reach the pop and/or hot charts were defined as pop. We combined “modern rock” and “mainstream rock” because the line between these charts has become less distinct over the past 2 decades. Using this approach, each song was clearly and uniquely defined as country (n = 61), pop (n = 35), R & B/hip-hop (n = 55), rap (n = 62), or rock (n = 66).

We coded any clear reference to substance use (Table 1, example 1). We also coded 3 types of references that were not necessarily associated with explicit use: figurative, place, and wallpaper (Table 1, examples 2-4).

For each song with explicit substance use, we recorded the specific substance(s) mentioned (ie, tobacco, alcohol, marijuana, cocaine and other stimulants, heroin and other opiates, hallucinogens, inhalants, prescription drugs, over-the-counter drugs, and nonspecific substances) and the number of references to each substance. We report tobacco, alcohol, and marijuana separately and combine all other substances because they were less commonly reported.

Each song with a reference to substance use was then coded on 3 domains that were informed by the social learning model: (1) motivations for use, (2) associations with use, and (3) consequences of use.^16-18 Based on pilot testing and the work of others, we used a dichotomous variable to indicate the presence or absence of the following motivations: peer/social pressure, sex, mood management, financial, and addiction/craving.²³ Similarly, dichotomous variables were used to indicate whether the substance use was associated with violence, sex, humor, partying, dealing/trafficking, a reference to a specific brand, operation of a vehicle, refusal to use, and limit setting. To be coded positively, the association had to be present in the song and directly associated with the substance use (Table 1, examples 12-15). We judged consequences of substance use across 7 dimensions: mental, emotional, physical, social, legal, financial, and sexual. We used a 5-point scale from − 2 to 2 to indicate the degree to which the consequences were positive or negative. For ease of display and interpretation, we ultimately collapsed these responses into 3 categories: negative (− 1 or − 2), neutral (0), and positive (1 or 2).

To quantify the density of substance use per time by song and music genre, we divided the number of references to each substance in a song by the duration of the song in hours. For songs that had multiple references (eg, in the song's chorus), we counted each mention as a separate reference (Table 1, examples 5-8).

We used χ² tests to determine if there were statistically significant differences between the proportion of songs that contained any mention of substance use among the 5 primary musical genres (country, pop, R & B/hip-hop, rap, and rock). When χ² tests were statistically significant, we used post hoc testing to determine if there were significant differences between specific genres. We then examined the songs with documented substance use (n = 93) to determine what proportion had each of the motivations and associations with use and the proportion of songs that manifested negative, neutral, or positive consequences of substance use. We used χ² tests and post hoc testing to determine if there were statistically significant differences between genres. We used a 2-tailed P value of < .05 to define statistical significance.

Overall, 116 of the 279 unique songs (41.6%) had a substance use reference of any kind (explicit, figurative, place, or “wallpaper”). Ninety-three songs (33.3%) contained explicit substance use references. Alcohol use was referenced most frequently, followed by marijuana use and use of other substances (illicit, prescription, or nonspecific substances) (Table 2). Tobacco use, mentioned in only 2.9% of songs, accounted for the least number of substance use references. References to explicit use of alcohol, marijuana, and other substances varied significantly by song genre, with rap songs containing the highest frequency of references to each of these substances.

The motivations for substance use represented most commonly among all songs were peer/social pressure, sexual, and financial (Table 3). Except for peer/social pressure, each of the motivations differed significantly by song genre. Sexual motivations were most common in R & B/hip-hop and rap songs. Mood management as a motivation was highest in rock, pop, and country songs. Financial motivations were highest in rap and R & B/hip-hop songs, and addiction/craving was most commonly portrayed in rock songs.

The most common elements associated with substance use were partying, sex, violence, dealing/trafficking, and humor (Table 3). Of the 93 songs, 17 associated substance use with a specific brand name and 15 associated substance use with use of a vehicle. Only 4 songs contained a specific antiuse message, and none portrayed refusal to use a substance. Most associations differed significantly by genre. Humor, for example, was much more commonly associated with substance use in country and pop songs than in rap and rock songs. Dealing and/or trafficking was common among rap songs but was not found in the other 4 genres. Finally, all 4 songs with antiuse messages were rock songs.

Overall, of the 93 songs with substance use, 15 (16%) portrayed more negative than positive consequences, whereas 63 (68%) contained more positive than negative consequences (P < .001). In almost half (45 [48%]) of the songs, the social consequences were positive, compared with only 7 (8%) in which consequences were negative (P < .001). Sexual (30 [32%] vs 2 [2%]; P < .001), emotional (14 [15%] vs 5 [5%]; P = .04), and financial (22 [24%] vs 0; P < .001) consequences were also more likely to be positive than negative (Figure). In contrast, legal (0 vs 8 [9%]; P = .005) consequences were more likely to be negative than positive. Mental consequences were no more likely to be positive than negative (6 [7%] vs 8 [9%]; P = .59), as were physical consequences (5 [5%] vs 10 [11%]; P = .20).

There were a mean of 35.2 references to explicit substance use per song-hour in our sample, ranging from 2.1 references per hour in pop music to 104.5 references per hour in rap music (P < .001) (Table 4). Alcohol references were most common in country and rap songs. References to marijuana and other substances were most common in rap songs.

This study demonstrates that explicit substance use is represented in about one-third of the most popular songs in the United States, with alcohol and marijuana referenced most frequently. Overall, explicit substance use is portrayed most frequently in rap music (48 of 62 songs [77%]) and least frequently in pop music (3 of 35 songs [9%]). Substance use is most commonly motivated by peer/social pressure and sex, and it is associated with partying and sex. The social, sexual, emotional, and financial consequences of use are more commonly depicted as positive than negative, whereas the legal and physical consequences of use are generally depicted as more negative than positive.

Because adolescents aged 15 to 18 years are exposed to an average of 2.4 hours of popular music per day,¹ our results suggest that the average adolescent is exposed to approximately 84 references to explicit substance use per day, 591 references per week, or 30 732 references per year. This represents a pervasive source of exposure to positive portrayals of substance use. Furthermore, exposure varies substantially by genre: the average adolescent listening wholly to pop would be exposed to 5 references per day, whereas the average adolescent listening wholly to rap would be exposed to 251 references per day.

Our findings were consistent with previous studies showing that alcohol and marijuana use are far more common than tobacco use in popular music. The frequency of tobacco references (2.9%) found in our study was nearly identical to that previously reported.²³ However, the reference rate to alcohol that we documented (23.7%) was higher than the 17% reported by Roberts et al.²³ There are several possible explanations for these differences in findings. First, portrayal of alcohol use in popular music may be increasing. Second, we only examined the top 279 popular songs, whereas Roberts et al examined 1000 songs. It is possible that the more popular songs contain more references to alcohol. Finally, it is possible that our rigorous method may have been more sensitive in identifying references to alcohol compared with other studies. Often, popular slang terms are used to represent alcohol (sauce, hooch, and juice) and marijuana (dro, chronic, and haze) that many youth understand well but with which some coders may not be familiar. By using a complex method involving 4 coders, we may have captured a more complete set of references to substance use.

Documentation of a growing exposure of adolescents to substance use in popular music suggests that further study of the actual effect of this exposure is warranted. Because recent data indicate that exposure to film smoking is one of the strongest risk factors for smoking initiation and progression,^4,29 it is reasonable to hypothesize that exposure to substance use in music is also a strong risk factor for substance use initiation and progression. Although music lacks the visual element of film, adolescent exposure to music is much more frequent, accounting for an average of 16 hours each week, compared with about 6 hours each week.³⁰ In addition, music is known to be highly related to personal identity; young people often model themselves in terms of dress, character, and behavior after musical figures.^21,31,32

In view of the heavy exposure young people have to substance use in popular music, health education related to substance use may need to be rethought. Currently accepted health education regarding substance use in ninth grade traditionally uses approximately 6 hours over a year.³³ However, this study would imply that during that same year the average young person would be exposed to an estimated 30 000 references to substance use in popular music. Health educators, health professionals, and curriculum designers may want to be familiar with the messages young people are receiving regarding substance use in their music so that they can best respond to those messages. Also, this large disparity between the exposure to substance use in popular music and substance abuse education suggests that simply “fighting fire with fire” is not likely to be feasible. Instead, we may need to find creative ways of generating doubt in the minds of young people as to the veracity of the positive substance use messages they receive in their media. One way of doing this might be to include in anti–substance abuse programming more “media literacy,” in which young people learn to analyze and evaluate the media to which they are frequently exposed.^34-36

Our results also show that substance use in popular music is commonly associated with some positive and some negative consequences. However, the negative consequences manifested (legal and physical) are ones that are generally not as concerning to adolescents as the positive consequences (social, sexual, emotional, and financial). Developmentally, many adolescents are not concerned about legal and physical ramifications of actions because they often consider themselves “invincible” with regard to these realms.³⁷ However, they are simultaneously concerned about social, sexual, emotional, and financial issues.^38-40 Thus, whereas current health education often emphasizes physical and legal ramifications of substance use, it might be preferable to instead focus on rebutting the positive consequences (social, sexual, emotional, and financial) of substance use portrayed in popular music.

Finally, our results show that different genres portray different (1) substances, (2) amounts of substances, and (3) motivations for, associations with, and consequences of use. This is likely to be because of a number of social, political, and economic factors. Research will be necessary to determine more specifically the reasons for these differences. Meanwhile, this information can be used to our advantage in developing health promotion materials and campaigns for young people. For instance, because rap music most frequently contains references to marijuana use, this population may need targeted education regarding the dangers of marijuana use. A rap artist might be the ideal spokesperson for this public health message.

Our study was limited in that it focused on 1 year of popular music. However, we had sufficient power to detect differences of interest with the available data. Still, it will be important to continue to follow popular music content longitudinally using rigorous methods. In addition, this study did not assess the effect of popular music messages on young people; rather, it focused on analysis of the content. Future studies should address more carefully the effects of these messages on their audiences, in terms of knowledge, attitudes, and practices. Also, the coding of elements, such as motivations, associations, and consequences, can be subjective. It is for this reason that we used a detailed and comprehensive coding method in which (1) both initial coders coded all songs, rather than overlapping only somewhat; (2) 2 confirmatory coders scrutinized each and every discrepancy, blinded to previous ratings and to each other; and (3) a committee involving coders and the principal investigator resolved any remaining discrepancies. Finally, we did not examine the visual elements of these songs, such as their music videos, CD covers, or Web sites. Although these are interesting areas to explore in future studies, the purpose of this study was to focus on the song lyrics.

In summary, children and adolescents are heavily exposed to substance use in popular music, and this exposure varies widely by genre. Substance use in music is frequently motivated by peer acceptance and sex, and it has highly positive associations and consequences. Research is needed to (1) determine the potency of exposure to substance use messages in music in adolescents and (2) determine the effects of various types of substance abuse messages, such as those with certain associations and consequences. If future studies determine an impact, we will need to consider the potential for media literacy and other educational interventions to reduce the impact of these messages on adolescent substance use.

Correspondence: Brian A. Primack, EdM, MD, Center for Research on Health Care, University of Pittsburgh School of Medicine, 230 McKee Pl, Ste 600, Pittsburgh, PA 15213 (bprimack@pitt.edu).

Accepted for Publication: August 14, 2007.

Author Contributions: Dr Primack had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Primack. Acquisition of data: Primack, Carroll, and Agarwal. Analysis and interpretation of data: Primack, Dalton, Carroll, Agarwal, and Fine. Drafting of the manuscript: Primack, Carroll, and Agarwal. Critical revision of the manuscript for important intellectual content: Primack, Dalton, and Fine. Statistical analysis: Primack. Obtained funding: Primack. Administrative, technical, and material support: Primack, Carroll, and Agarwal. Study supervision: Primack, Dalton, and Fine.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by a K-07 career development award (K07-CA114315) from the National Cancer Institute (Dr Primack); by a Physician Faculty Scholar Award from the Robert Wood Johnson Foundation (Dr Primack); by a grant from the Maurice Falk Foundation (Dr Primack); and in part by a K-24 career development award (K24-AI01769) from the National Institute of Allergy and Infectious Diseases (Dr Fine).

Role of the Sponsor: The funding bodies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.

Additional Contributions: Thomas Radomski, BS, and Supria Batra assisted with coding (compensation was provided for their services).