Applying the Health Belief Model to Examine News Coverage Regarding Steroids in Sports by ABC, CBS, and NBC Between March 1990 and May 2008

Abstract

The investigation described here examined ABC, CBS, and NBC news coverage of steroids in sports between March 1990 and May 2008. Employing a framing analysis guided by the health belief model (HBM), coverage of the barriers and benefits of using steroids is reported. Overall, the trend by these three news affiliates was to emphasize the illegality of using steroids, whereas considerably less coverage was devoted to the health costs, in terms of both severity and susceptibility, of using steroids. Furthermore, of the health costs reported, the specific consequences of steroid use varied considerably. The results are reported across four timeframes: 1990–2008, 1990–1996, 1997–2002, and 2003–2008.

“So our sports heroes taught our youth a few lessons this past season. Life lessons. Such as cheating, disobeying the law and lying,” writes Devine (2008).

The steroid scandal in sports during the past two decades has tarnished our perceptions of professional athletics and triggered a genuine mistrust, placing insurmountable doubt about the purity of sports in general (see Kruse, 2008). Although the media have extensively covered the steroid scandal, a comprehensive review of how steroids in sports have been framed remains understudied. How the media frame steroid use in sports could reasonably affect potential and current steroid users' perceptions, as well as kids' perceptions, of these performance-enhancing drugs. Certainly, the presence of steroid use by professional athletes contributes to the usage rates, which range between approximately 2% and 12% among adolescents (Bahrke, Yesalis, & Brower, 1998; Cafri, van den Berg, & Thompson, 2006; Field et al., 2005). Although this variation in usage among adolescents is likely due to poor measurement (Kanayama, Boynes, Hudson, Field, & Pope, 2007), nevertheless, few would disagree that steroid use among adolescents is a major health problem.

Although few reports have examined media coverage of steroids in sports, research on steroids exists. Previous research examines steroid use among male (e.g., Pope & Brower, 2000) and female (e.g., Elliot & Goldberg, 2000) athletes (Dickinson et al., 2005) as well as nonathletes (Yesalis, Barsukiewicz, Kopstein, & Bahrke, 1997) alike. With steroid use on the rise, it is not surprising that several investigations have identified the health costs associated with steroid use (Cafri et al., 2005; Dickinson et al., 2005) as well as the attributes and characteristics of steroid users (Dickinson et al., 2005; Elliot & Goldberg, 1996; Miller, Barnes, Sabo, Melnick, & Farrell, 2002; Yesalis et al., 1997). In response to this rising health concern, others have developed curriculum to curb the use of performance-enhancing drugs among male (Fritz et al., 2005; Goldberg, MacKinnon, Elliot, Moe, Clarke, & Cheong, 2000) and female (Elliot et al., 2006) athletes. Research also examines the influence of socializing agents such as peers, parents, and the media on steroid use (Cafri et al., 2006; Denham, 2004a, 2004b, 2006; Field et al., 2005; Smolak, Murnen, & Thompson, 2005). From these reports, in America, males are driven to increase their muscles in a way similar to females desiring to be thin (Cafri et al., 2006; Harrison & Bond, 2007). Considering steroid use offers a shortcut to achieving these goals for males and females, steroids affect everyone.

Media coverage about steroids in sports has received modest attention with most investigations concentrating on a particular event, e.g., Carl Lewis's positive test result (Denham, 2004a) or Ken Caminiti's steroid use confession (Denham, 2004b). The present study seeks to extend this research by examining television news coverage across ABC, CBS, and NBC between March 1990 and May 2008 using a framing framework (Bartlett, 1932). Although framing serves as the overarching framework for this investigation, the key themes analyzed are derived from the health belief model (HBM) (Janz & Becker, 1984). Prior to revealing the methodological specifics of this content analysis, a discussion of framing followed by a brief review of the HBM is presented.

FRAMING

The present investigation is interested in how steroid usage in sports is depicted through the privileging of certain angles or frames by the media when covering steroids in sports. A frame can refer to the manner (i.e., images, language choices) in which a speaker conveys a message to his or her audience (Gamson & Modigliani, 1989). McCombs and Ghanem (2001) refer to framing as the selection, emphasis, and elaboration of a specific issue. Similarly, Kahneman and Tversky (2000) state that a prevailing set of features identified by the news media regarding a certain issue is reported and thus influences how individuals view an issue. For instance, an emphasis on the illegality of steroid use could send the message that steroids use can tarnish one's reputation. Framing as a micro-level construct concentrates on how individuals form impressions based on their news exposure of issues (Scheufele & Tewksbury, 2007).

Within the context of health communication, framing has received much attention both within experimental settings (see Salovey, Schneider, & Apanovitch, 2002) and through content analyses. Within an experimental setting, the majority of this research examines the advantages and disadvantages of framing an argument in terms of potential gains or losses. For example, within the context of antidrug ads, Cho and Boster (2008) discovered that loss-frame messages were more persuasive than their gain-frame counterparts for adolescents who had friends that used drugs. To the contrary, Reinhart, Marshall, Feeley, and Tutzauer's (2007) three-experiment study about organ donation found that gain-frame messages were most influential for college students. Along with these mixed results, meta-analytic research reports null to modest effect sizes when comparing the persuasiveness of gain- and loss-frame messages (see O'Keefe & Jensen, 2006). Increasingly more and more studies examine media coverage of a particular issue (e.g., Denham, 2004b; Lapinski, 2006; Morgan, Harrison, Chewning, Davis, & DiCorcia, 2007). For example, Denham's (2004b) analysis of newspaper stories following a report published in Sports Illustrated about steroid use by Ken Caminiti, a former professional baseball player, played a key role in prompting lawmakers to encourage baseball to enact a drug-testing policy. In short, framing of an issue by the media can shape audience beliefs and subsequent attitudes.

Investigating the framing of steroids in sports by the television news media is not surprising, considering that television is a credible and popular source of information in Western industrial societies. The Pew Research Center for the People & the Press (2006) found that, on average, individuals most often received their news via local television (54%) and newspapers (40%). Similarly, when asked where their news came from yesterday, 57% identified the television compared to 40% who identified a newspaper. From these reports, television appears to be the preferred channel for receiving news. Thus far, an argument has been advanced for using framing as the theoretical foundation for the present investigation as well as the importance of television coverage in reaching individuals; however, to gain a richer understanding of how the television news media communicate about steroids in sports, the HBM was employed.

HEALTH BELIEF MODEL (HBM)

Arguably one of the most influential models in health communication is Janz and Becker's (1984) Health Belief Model (HBM). More than two decades after its inception, the HBM remains one of the most popular theoretical frameworks to guide public health interventions (e.g., Chew, Palmer, Slonska, & Subbiah, 2002; Kennedy, O'Leary, Beck, Pollard, & Simpson, 2004; Mayer, Housemann, & Piepenbrok, 1999). For the purposes of the present investigation, the HBM serves as a theoretical guide in identifying key variables to analyze, which will provide a greater understanding of how the television news media communicate about steroids in sports. Specifically, Janz and Becker's HBM identified five original factors: (a) cues to action, (b) threat susceptibility, (c) threat severity, (d) barriers, and (e) benefits (Witte & Lapinski, 1998). Each factor is discussed with an emphasis on how it applies to steroid coverage.

Of these five factors, cues to action are the least studied variable. In short, cues to action provide the catalyst to engage in a certain behavior (Witte & Lapinski, 1998). Cues to action are derived internally (i.e., side effects from using steroids) or externally (i.e., public service announcement, television news story, or a movie clip). In short, Janz and Becker (1984) predict that cues to action can increase perceptions of the severity and susceptibility of a threat, which in turn increases the likelihood of performing the advocated behavior. It is assumed that television news coverage of steroid usage in sports can serve as a cue to action among its viewers. For this reason, analyzing news coverage is an important step toward understanding public perceptions of steroids in sports.

Two elements commonly used to communicate the dangers of a health threat are threat susceptibility and severity. Threat susceptibility refers to the likelihood of experiencing the dangers associated with the admonished behavior (Mattson, 1999). Similarly, threat severity pertains to the perceived dangers of performing a certain behavior (Chew et al., 2002). Within the context of the present study, individuals perceiving steroids to be harmful to users' health illustrates the latter component, whereas people believing that they may experience these health consequences by using steroids exemplifies the former. According to the HBM, an individual cognizant of the dangers of performing unhealthy behaviors and having a belief that experimenting with these risky behaviors result in negative consequences is less likely to perform these behaviors (Greene & Brinn, 2003). Thus, from a public health standpoint, attention to the negative health costs of using steroids is an effective strategy to deter usage.

Barriers to performing a behavior are the fourth factor identified by the HBM. A barrier relates to the costs of performing an unhealthy behavior or not ascribing to a healthy act (Witte & Lapinski, 1998). To the contrary, benefits refer to the rewards of performing a healthy behavior or not engaging in an unhealthy act. For example, barriers to using steroids in sports include the health risks associated with using these illegal substances, penalties (i.e., suspensions, monetary fines), and drug-testing concerns (Denham, 2004b; Wilson & Schmidt, 2007). On the other hand, steroid use provides athletes with a competitive edge. Also, medical doctors can prescribe steroids to aide in rehabilitation. According to the HBM (Janz & Becker, 1984), individuals weigh the barriers with the benefits when determining the right course of action.

The HBM has been employed as a theoretical framework used to design and evaluate myriad health interventions. For example, Mattson (1999) discovered that participants' perceived severity of and susceptibility to HIV/AIDS along with the benefits of and barriers to engaging in safe sex practices were associated with performing safer sex activities. Similarly, Greene and Brinn's (2003) study found that messages featuring statistical evidence enhanced perceptions of susceptibility to experience the health consequences of tanning whereas narrative evidence increased perceived realism among college females. In addition, both statistical and narrative evidence prompted participants to have decreased intentions to tan. Finally, Chew and colleagues examined the effectiveness of a television series promoting healthy lifestyles in Poland (Chew et al., 2002; Chew, Palmer, & Soohong, 1998). Their results reveal that threat severity and susceptibility, efficacy, and perceived salience were more closely related to beliefs about the dangers of not exercising, losing weight, changing eating habits, and quitting smoking compared to their nonviewing counterparts (Chew et al., 2002). Moreover, Chew et al. (1998) discovered that health motivation and salience mediate the association between perceived susceptibility and efficacy with respect to eating a healthy diet. In general, research suggests that perceived barriers and susceptibility are the best predictors of a behavior whereas perceived severity is the weakest (Witte & Lapinski, 1998).

SUMMARY AND RESEARCH QUESTIONS

The present investigation examines ABC, CBS, and NBC television news coverage of steroids in sports. The framing research states that the media are influential in determining how we think about an issue (Kosicki, 1993; Price & Tewksbury, 1997, Scheufele & Tewksbury, 2007). Additionally, this literature is useful in illustrating the role of the media in shaping public perceptions and, at times, influencing policy. The HBM is utilized to identify key variables regarding steroid usage (Janz & Becker, 1984). Specifically, television news stories were analyzed for their reporting of the benefits and barriers of athletes using steroids. Thus, the following two research questions were advanced.

RQ1: How often are the barriers of using steroids communicated in the news from 1990–2008 (RQ1a), 1990–1996 (RQ1b), 1997–2002 (RQ1c), and 2003–2008 (RQ1d)?

RQ2: How often are the benefits of using steroids communicated in the news from 1990–2008 (RQ2a), 1990–1996 (RQ2b), 1997–2002 (RQ2c), and 2003–2008 (RQ2d)?

In addition to analyzing the barriers and benefits of using steroids, the HBM also places an emphasis on the perceived severity and susceptibility of experiencing the costs associated with a particular behavior (Janz & Becker, 1984). Following the HBM, research suggests that individuals aware of the severity and susceptibility of using an unhealthy substance are less likely to engage in the risky behavior (Witte & Lapinski, 1998). With respect to types of evidence, research suggests that statistical and narrative evidence are effective in communicating a threat; however, research is not consistent regarding which strategy is the most effective (Feeley, Marshall, & Reinhart, 2006; Greene & Brinn, 2003; Kopfman, Smith, Ah Yun, & Hodges, 1998). Despite these inconsistencies, individuals aware of the health costs of steroid use (e.g., acne, cancer, diabetes, heart disease, and infertility) along with a belief that using steroids can result in these aversive outcomes should be less likely to use these banned substances (Greene & Brinn, 2003; Mattson, 1999). Thus, it is important to discover how much attention is given by the media to the severity and susceptibility of using steroids. Thus,

RQ3:  How frequently did the media communicate the health severity associated with steroid use from 1990–2008 (RQ3a), 1990–1996 (RQ3b), 1997–2002 (RQ3c), and 2003–2008 (RQ3d)?

RQ4: How frequently did the media provide statistical and narrative evidence demonstrating the susceptibility that steroid users experience these health costs from 1990–2008 (RQ4a), 1990–1996 (RQ4b), 1997–2002 (RQ4c), and 2003–2008 (RQ4d)?

METHOD

Sample and Procedure

The current investigation examined television news coverage of steroid usage in sports from March 1, 1990, through May 1, 2008. In total, 2,124 news broadcasts were identified. From this list of news broadcasts, a systematic sampling procedure was performed with every other news transcript being analyzed (n = 1,062) (Krippendorf, 1980). Of the news transcripts analyzed, CBS (n = 429) contained the majority of stories, followed by ABC (n = 334), and NBC (n = 299). Because the starting point was selected at random, each news broadcast had an equal chance of being included in the analysis (Riffe, Lacy, & Fico, 1998). Each news broadcast served as the unit of analysis.

News broadcasts were retrieved using Lexis-Nexis. Specifically, available coverage of steroids in sports occurred between (a) May, 1990 and May, 2008 for ABC, (b) March, 1990 and May, 2008 for CBS, and (c) February, 1997 and May, 2008 for NBC. See the Appendix for a complete list of programs from which these news stories were aired. The present investigation was interested in discovering how steroids were discussed in these news transcripts. As a result, any broadcast that mentioned steroids in sports anywhere in the text was analyzed. To uncover a wide range of stories, the following syntax was employed, in syntax form: ‘steroid! or muscle building drug! or performance enhancing drug! or designer steroid! or human growth hormone! or anabolic steroid! and sport! or baseball or basketball or body build! or box! or cycl! or football or golf! or hockey or lacrosse or power lift! or rugby or run! or swim! or track or field or soccer or softball or tennis or volleyball or wrestl! or martial arts or gym! or olymp!’.

Two individuals were trained by the author. Following approximately 8 hours of training, both coders were assigned to code 10% of the sample broadcasts independently (Lacy & Riffe, 1996; Lombard, Snyder-Duch, & Bracken, 2002). The intercoder reliabilities were calculated before the disagreements were resolved among the coders. After the intercoder reliability was established, both coders independently coded the remainder of the news broadcasts. Reliability between the coders was determined using Brennan and Prediger's (1981) kappa. Brennan and Prediger's kappa was chosen over Cohen's (1960) kappa because with Cohen's kappa, coders are penalized for higher rates of simple agreement. Brennan and Prediger's (1981) procedure corrects for this by disregarding simple agreement altogether and instead assumes chance agreement is determined solely by the number of categories in the coding scheme. Simple agreement was good across all of the variables. For each of the categories in the following discussion, simple agreement (SA) among coders along with the Brennan and Prediger's kappa (K_{B & P}) are reported.

Categories

Barriers

Each news broadcast was coded to determine whether the news story addressed barriers of steroid use (SA = .93, K_{B & P} = .86). Barriers were grouped into six subcategories: health costs, issues of illegality, drug testing, financial expenses of purchasing steroids, fines for steroid usage, and suspension for steroid usage (Denham, 2004b; Wilson & Schmidt, 2007). Each of the subcategories obtained a respectable reliability: (a) health costs (SA = .95, K_{B & P} = .91), (b) issues of illegality (SA = .88, K_{B & P} =.76), (c) steroid testing policies (SA = .89, K_{B & P} = .77), (d) financial expenses of purchasing steroids (SA = 1.0, K_{B & P} = 1.0), (e) fines for steroid usage (SA = .99, K_{B & P} = .98), and (f) suspension for steroid usage (SA = .94, K_{B & P} = .88).

Benefits

Each news broadcast was coded for the presence or absence of communicating the benefits of steroid use (SA = .94, K_{B & P} = .88). Benefits were divided into two subcategories: (a) gain a competitive advantage (SA = .95, K_{B & P} = .90) and (b) rehabilitation (SA = .99, K_{B & P} = .98) (Rogerson et al., 2007; Strauss et al., 1985).

Severity

A primary purpose of this investigation was to determine how much attention is devoted to the health consequences (National Institute on Drug Abuse, 2008) of using steroids (SA = .85, K_{B & P} = .70). Specific health costs (e.g., aggressiveness, liver damage, and suicide) received a respectable reliability with SA ranging from .98 and 1.0 and K_{B & P} between .96 and 1.0.

Susceptibility

News coverage of the likelihood that a person will experience the aversive health costs linked to using steroids was examined (SA = .98, K_{B & P} = .96). Each type of evidence was reliable: factual-narrative evidence (SA = .98, K_{B & P} = .96), hypothetical-narrative evidence (SA = .98, K_{B & P} = .96), and statistical evidence (SA = 1.0, K_{B & P} = 1.0).

Data Analysis Strategy

To determine if the major categories differed in frequency, chi-square goodness-of-fit tests were conducted. Each of the coded subcategories represented nonindependent dichotomous nominal data. Each subcategory was nonindependent because it was entirely possible that each news article contained multiple categories (i.e., illegality of steroid usage, health costs). For this reason, Cochran's Q tests were run to determine if certain categories emerged more than others whenever there were more than two related variables. Following a statistically significant Cochran's Q, McNemar tests were performed to determine where statistical significant differences occurred. To minimize the likelihood of committing a type I error, Bonferroni corrections were employed by dividing the conventional alpha level (.05) by the number of pairwise comparisons.

RESULTS

RQ1a: Television News Coverage of Barriers to Steroid Usage 1990–2008

The first research question examined television news coverage of the barriers associated with steroid use. Results suggest that the majority of news transcripts communicate the barriers associated with steroid usage (n = 685) while others did not (n = 376), χ²(1, N = 1061) = 89.99, p < .001. Specifically, issues of illegality received the most coverage, followed by drug testing, health costs, suspensions for steroid use, fines for steroids usage, and the financial expenses of purchasing steroids. A Cochran's Q test indicated that statistically significant differences were found, Q(5, N = 1061) = 1960.91, p < .001. McNemar tests were performed to indicate where significant differences occurred across all pairwise comparisons. Results from RQ1 are presented in Table 1.

TABLE 1 Comparison of Coverage of the Barriers of Steroid Usage

Barrier	1990–1996	1997–2002	2003–2008	Total 1990–2008
Health costs	13b c	24b	70bc	107b
Issues of illegality	36a	101a	492a	629a
Drug testing	2cd	24b	107b	133b
Purchasing steroids expense	2d	3c	1d	2d
Fines for steroid usage	0d	1c	5d	6c
Suspension forsteroid usage	15b	16b	73c	104b
Note. Frequencies in the same column that do not share the same subscripts differ at p < .003.

RQ1b: Television News Coverage of Barriers to Steroid Usage 1990–1996

From 1990 to 1996, television news coverage of the barriers associated with steroid use was sporadic, with little difference between news stories communicating the barriers (n = 42) and those not communicating the barriers (n = 57), χ²(1, N = 99) = 2.27, p = .13. Coverage among the six barriers differed significantly, Q(5, N = 99) = 101.09, p < .001. Similar to RQ1a, issues of illegality received the most coverage, followed by suspensions for steroid use, health costs, drug testing, and financial expenses of purchasing steroids. Fines for steroids usage received no coverage.

RQ1c: Television News Coverage of Barriers to Steroid Usage 1997–2002

With respect to coverage between 1997 and 2002, attention to the barriers of using steroids received mixed attention, with no difference between stories communicating the barriers of steroid use (n = 116) and those not communicating the barriers (n = 131), χ²(1, N = 247) = .91, p = .34. Issues of illegality obtained the most coverage, followed by health costs, drug testing, suspensions from sports, fines for usage, and financial expenses of purchasing steroids, Q(5, N = 247) = 289.97, p < .001.

RQ1d: Television News Coverage of Barriers to Steroid Usage 2003–2008

In stark contrast to the preceding two time periods, television news coverage between 2003 and 2008 communicated the barriers associated with steroid use (n = 527) compared to those not communicating these barriers (n = 188), χ²(1, N = 715) = 160.73, p < .001. The Cochran Q test suggests that certain barriers received more coverage than others, Q(5, N = 715) = 1596.79, p < .001. Again, the lion's share of attention went to the illegality of steroid use, followed by testing policies, suspensions for steroid use, health costs, fines for steroid usage, and financial expenses of purchasing steroids.

RQ2a: Television News Coverage of the Benefits of Steroid Usage 1990–2008

RQ2a analyzed coverage of the benefits of using steroids between 1990 and 2008. The majority of television news stories did not draw attention to the benefits of using steroids (n = 813), χ²(1, N = 1061) = 300.87, p < .001. The McNemar test found that more coverage was devoted to the competitive edge (e.g., increased strength and speed) gained from steroid usage (n = 178) compared to the benefits of using steroids for rehabilitation purposes (n = 92), p < .001.

RQ2b: Television News Coverage of the Benefits of Steroid Usage 1990–1996

Examining coverage between 1990 and 1996, RQ2b found that the majority of television news stories did not draw attention to the benefits of using steroids (n = 64), χ²(1, N = 99) = 8.50, p < .01. The McNemar test found more coverage was given to using steroids for rehabilitation purposes (n = 25) than for gaining a competitive edge (n = 15), although this difference was not significant.

RQ2c: Television News Coverage of the Benefits of Steroid Usage 1997–2002

With respect to news coverage between 1997 and 2002, again, results found that the majority of television news stories did not draw attention to the benefits of using steroids (n = 178), χ²(1, N = 247) = 48.10, p < .001. Similar to coverage between 1990 and 1996, McNemar tests found no difference in coverage between using steroids for rehabilitation purposes (n = 44) and gaining a competitive edge (n = 38).

RQ2d: Television News Coverage of the Benefits of Steroid Usage 2003–2008

Interestingly, coverage about steroids between 2003 and 2008 continued to not draw attention to the benefits of using steroids (n = 571), χ²(1, N = 715) = 255.01, p < .001. However, unlike the previous two time frames, more coverage communicated the competitive edge gained from using steroids (n = 125) rather than for rehabilitation purposes (n = 23) according to the McNemar test, p < .001.

RQ3a: Television News Coverage of the Severity of Steroid Usage 1990–2008

The third research question examined television news coverage of the health costs associated with steroid use. Overall, results suggest that relatively little attention is given to the specific health consequences associated with using steroids, χ²(1, N = 1061) = 672.97, p < .001. That is, news reports mentioning the health costs of using steroids (n = 108) pale in comparison to reports not mentioning the health consequences (n = 953). Of those mentioned, aggressiveness was cited most frequently, followed by liver failure, mental illness (i.e., depression), cancer, acne, and a deepened voice. See Table 2 for a complete list of all the health costs covered between 1990 and 2008.

TABLE 2 Comparison of Coverage of the Health Severity of Steroid Usage

Health cost	1990–1996	1997–2002	2003–2008	Total 1990–2008
Acne	3	1	12	16
Aggressiveness	5	4	32	41
Baldness	2	1	3	10
Birth defects in children	0	0	0	1
Blood clots	0	1	0	1
Breast development	5	1	1	4
Breast shrinkage	0	1	0	1
Cancer	6	1	13	20
Cholesterol increases	3	0	5	8
Crippling	1	0	0	1
Deepened voice	4	6	2	12
Diabetes	0	0	4	4
Facial hair	0	5	3	8
Fluid retention	0	1	0	2
Headache	1	0	2	3
Heart attack	4	0	0	4
Heart complications	0	3	0	3
High blood pressure	3	0	8	11
Infertility	1	0	1	2
Insomnia	1	0	0	1
Kidney failure	3	1	0	4
Liver failure	7	7	8	22
Menstrual irregularity	3	2	1	6
Mental illness	3	3	16	22
Mood swings	2	0	0	2
Nauseau	0	0	2	2
Nosebleeds	4	0	0	4
Osteoporosis	0	0	1	1
Ovarian infections	0	1	0	1
Sex-drive alteration	0	1	0	1
Sperm-count reduction	2	1	0	3
Stroke	2	0	0	2
Stunt growth	0	2	0	2
Suicide/death	3	0	3	6
Taste alteration	0	1	0	1
Testicles shrinkage	5	0	1	6
Testosterone increase	0	1	0	1
Testosterone reduction	0	0	2	2

RQ3b: Television News Coverage of the Severity of Steroid Usage 1990–1996

With respect to television news coverage between 1990 and 1996, results suggest that attention to the health costs associated with using steroids was scarce, χ²(1, N = 99) = 50.92, p < .001. More coverage did not identify the health costs of using steroids (n = 85) than specifically mentioned the aversive effects associated with these substances (n = 14). Receiving the most coverage were liver failure, cancer, breast development, aggressiveness, testicle shrinkage, heart attack, nosebleeds, and a deepened voice.

RQ3c: Television News Coverage of the Severity of Steroid Usage 1997–2002

Coverage between 1997 and 2002 reported infrequent coverage of the health costs associated with steroid usage (n = 223) compared to reports in which potential health costs were mentioned (n = 24), χ²(1, N = 247) = 160.33, p < .001. Liver failure received the most coverage, followed by a deepened voice, facial hair, aggressiveness, mental illness, and heart damage.

RQ3d: Television News Coverage of the Severity of Steroid Usage 2003–2008

Once again the overall trend in coverage between 2003 and 2008 suggests that infrequent news reporting was given to the health costs associated with steroid usage (n = 645) in comparison to coverage of the potential health costs (n = 70), χ²(1, N = 715) = 462.41, p < .001. Of the health costs cited, aggressiveness received the most coverage, followed by mental illness, cancer, acne, liver failure, high blood pressure, increased cholesterol, and diabetes.

RQ4a: Television News Coverage of Health Susceptibility of Steroid Usage 1990–2008

RQ4a examined whether television news stories conveyed the likelihood of experiencing the health costs associated with steroid use through their presentation of statistical and narrative evidence. Results found that some news reports communicated health susceptibility (n = 81), while others did not (n = 980), χ²(1, N = 1061) = 761.74, p < .001. Cochran's Q test found that differences occurred in coverage of evidence, Q(2, N = 1061) = 89.45, p < .001. Factual narratives appeared the most, followed by hypothetical narratives, and statistical evidence. See Table 3 for results of the pairwise comparisons.

TABLE 3 Comparison of Evidence Type

Evidence type	1990–1996	1997–2002	2003–2008	Total 1990–2008
Factual narrative	12a	11a	46a	69a
Hypotheticalnarrative	3b	6ab	9b	18b
Statistics	0c	0b	1b	1c
Note. Frequencies in the same column that do not share the same subscripts differ at p < .0167.

RQ4b: Television News Coverage of Health Susceptibility of Steroid Usage 1990–1996

With respect to 1990 and 1996, results found that most news reports did not communicate health susceptibility (n = 86), χ²(1, N = 99) = 53.83, p < .001. Cochran's Q test found that differences occurred in coverage of the three types of evidence, Q(2, N = 99) = 18, p < .001. Again, factual narratives received the most coverage, followed by hypothetical narratives, and statistical evidence.

RQ4c: Television News Coverage of Health Susceptibility of Steroid Usage 1997–2002

During the time period of 1997 and 2002, the majority of news reports did not mention that steroid users were susceptible to experience the health costs (n = 232), whereas a mere 15 stories explicitly made this connection, χ²(1, N = 1061) = 761.74, p < .001. The Cochran's Q test discovered that significant differences occurred in coverage, Q(2, N = 247) = 10.71, p < .01. Again, factual narratives received the most coverage, followed by hypothetical narratives and statistics.

RQ4d: Television News Coverage of Health Susceptibility of Steroid Usage 2003–2008

From 2003 to 2008, television news stories continued their trend of not communicating the susceptibility of experiencing health costs associated with using steroids. In fact, 662 news reports did not communicate susceptibility, χ²(1, N = 715) = 518.72, p < .001. The Cochran's Q test discovered that significant differences occurred between coverage, Q(2, N = 715) = 64.04, p < .001. Consistent with the earlier time periods, factual narratives received the majority of the coverage, followed by hypothetical narratives, and statistical evidence.

DISCUSSION

The present investigation examined how ABC, CBS, and NBC news media communicate about steroids in sports between March 1990 and May 2008. Employing a framing analysis, it was understood that the news media perform a central role in affecting viewers' perceptions of steroids by concentrating on specific elements of an issue (McCombs & Ghanem, 2001) and, as a result, shape how individuals' think about an issue (Scheufele & Tewksbury, 2007). Results and implications are discussed next.

A goal of the present study was to determine how frequently the barriers of using steroids were presented. Overall, the majority of news stories (65%) addressed at least one barrier: issues of illegality (59%), drug testing policy (13%), health costs (10%), suspensions (10%), fines (1%), and financial expenses of purchasing steroids (<1%). However, a closer look at the trends tells a slightly different story. Between 1990 and 1996 (42%), as well as in 1997 through 2002 (47%), modest attention was given to these barriers. During both time periods, no statistical difference occurred between stories that did and did not explicitly cite one of the aforementioned barriers. All the while, steroid use was believed to be on the rise in professional athletics. Coverage of the barriers associated with steroid use in sports increased significantly from 2003 to 2008 (74%). Baseball received the most coverage, followed by track and field, bicycling, football, and wrestling. News stories about steroid usage by Lance Armstrong, Chris Benoit, Barry Bonds, Jose Canseco, Roger Clemens, Marion Jones, Mark McGwire, and Rafael Palmeiro abound. Perhaps this increased coverage was due to political pressure on professional sports to adopt a more stringent drug-testing policy or as speculation grew regarding the number of athletes using these illegal substances (see Denham, 2004b). Either way, by 2003 to 2008, the framing of steroids in sports had shifted by accentuating the barriers of using performance-enhancing drugs.

Across the three time periods analyzed in this study, the barrier receiving the most coverage was that nonmedical use of steroids is illegal. Several stories demonstrated how one's reputation can be seriously damaged following a positive test (e.g., Barry Bonds, Roger Clemens, and Marion Jones). Receiving considerably less coverage were stories citing the steroid testing policy, health costs, and suspensions for usage, with each appearing in approximately 10% of the analyzed stories. Coverage of how expensive steroids are to purchase and fines for usage appeared in less than 1% of the news stories.

Steroids are often used for rehabilitation purposes or to gain a competitive edge (Rogerson et al., 2007; Strauss et al., 1985). The former is under the direction of a medical doctor whereas the latter is not. Results suggest that for the most part, the media frame steroid use as a strategy to gain a competitive edge. However, examining this trend across three time periods revealed some shifts in framing. Between 1990 and 1996, the majority of coverage focused on steroid usage for rehabilitation purposes. From 1997 to 2002, a major focus was on using steroids for rehabilitation reasons, but attention on steroid usage to gain a competitive advantage was on the rise. By 2003 to 2008, the focus was largely on using steroids to gain a competitive edge, with modest exposure to rehabilitation purposes. A major turning point in the framing of steroid use followed Caminitti's confession in Sports Illustrated (Denham, 2004b) as well as President George Bush's (2004) reference to steroid use in sports in his State of the Union address. Both events shifted the focus on steroid use from rehabilitation purposes to increasing one's competitiveness. Consistent with extant research (e.g., Edy & Meirick, 2007), it appears that a news story can affect the selection and emphasis of certain elements of an issue.

From a health standpoint, underscoring the severity of steroid use received modest coverage. A mere 10% of the stories analyzed mentioned the health costs of using these banned substances. The specific health costs of using steroids identified varied greatly. During 1990 to 1996 and 1997 to 2002, liver failure was identified as the leading health cost of using steroids, as were cancer, aggressiveness, and more noticeable aversive side effects (e.g., breast development, deepened voice, facial hair, testicle shrinkage). More recently, between 2003 and 2008, aggressiveness was cited as an affect of steroid use. Frequent coverage of Chris Benoit, a former professional wrestler, when he murdered his wife and child, followed by his suicide, attributed his actions to this finding. Additional health costs cited liver failure and cancer along with other noticeable side effects (e.g., acne). Due to inconsistent mention of the health costs linked with steroid use, potential and current users are likely unsure of the physical and mental consequences of using these substances. Future research should examine societal perceptions of the side effects of steroid use. Certainly, health practitioners must identify the health costs of using steroids and make this information available to the public. Once this information is available, it is important for the news media to frame the health consequences associated with steroid use by telling a consistent story. Future research should use social cognitive theory (Bandura, 1977) as an explanatory framework to understand how individuals learn from what they see in these news reports about steroids.

The implications of these trends suggest a telling fact about television news coverage of steroid use by athletes. The fact that infrequent coverage was given to these barriers may shed light on why steroid use was on the rise during the past two decades. As increased attention was given to steroid use in the media from 2003 to 2008, coverage of the barriers also increased. The majority of this coverage framed steroid use as an illegal activity, rather than emphasizing the health costs of using these dangerous substances. From a public health standpoint this is a concern. Communicating the severity of a health threat is a key component of dissuading people from unhealthy lifestyles (Janz & Becker, 1984, Witte, 1992). Framing steroid use in this manner, the media are communicating to viewers that steroids are costly if you get caught. Unfortunately, the health consequences of using steroids are likely, regardless of whether a user gets caught or not. Framing steroids as an illegal substance should receive coverage; however, emphasizing the health costs of steroid use in a tenth of the stories is troubling.

Disappointingly, evidence supporting the likelihood of experiencing these aversive health effects from steroid use received sparse attention. Results suggest that coverage of narrative (both factual and hypothetical) as well as statistical evidence was thin. Overall, factual narratives received the most coverage, although this evidence type appeared in less than 7% of the television news stories; hypothetical narratives appeared in less than 2% of the stories. Factual narratives received the most coverage, although between 1997 and 2002 the difference between evidence types was not significant. Surprisingly, statistical evidence was present in only one story. Future research should examine what types of evidence are the most persuasive. Research suggests that statistical evidence prompts feelings of susceptibility and narratives make a health threat appear more realistic (Green & Brinn, 2003). Regarding organ donation, research suggests that statistical evidence is the most persuasive (Kopfman et al, 1998), while others argue that narrative evidence is more convincing (Feeley et al., 2006).

It is not uncommon for the media to cover an issue with a bias. Entman (2007) classifies framing bias into three categories: (a) Distortion bias occurs when news falsifies reality; (b) decision-making bias is when a journalist manufactures biased material; and (c) content bias happens when one element of a story receives more coverage than another. A content bias has occurred since 1990 with respect to coverage of steroids in sports. That is, the media privilege the illegality of steroid use while at the same time provide minimal coverage to the health costs. Research suggests that programs geared toward male (Fritz et al., 2004; Goldberg et al., 1991; Goldberg et al., 2000) and female (Elliot et al., 2006) athletes are effective in enhancing knowledge about the dangers of these illicit drugs as well as reducing intentions to use, and in decreasing steroids usage. The media can play an influential role in this domain. Consistent with past framing research, the media shape how we think about an issue (Dearing & Rogers, 1996; Scheufele & Tewksbury, 2007). Drawing attention to certain aspects of steroid use, while not addressing others, likely shapes audiences' perceptions. The message to audiences is dangerous. By framing the message as “taking steroids without a doctor's prescription is illegal,” a subtle message is that using performance-enhancing drugs is okay as long as you do not get caught. From a public health standpoint, the primary barrier should be the health consequences, both mental and physical, from using these performance-enhancing substances. However, the legal and economic costs of using steroids should also be addressed. Future research should analyze why this bias exists with respect to media coverage of steroids in sports.

In light of these findings, professional athletes' lives are in the spotlight and research suggests that their actions can exert an influence on society (Brown et al., 2003). Magic Johnson's influence on HIV prevention is well documented (Basil & Brown, 1994; Brown & Basil, 1995). After Magic announced he was HIV positive, heterosexual individuals reported an increase in HIV/AIDS awareness as well as changes in beliefs, attitudes, and sexual behaviors (Basil & Brown, 1994; Brown & Basil, 1995). More recently, Brown et al. (2003) found that participants identifying with Mark McGwire were aware that he used androstenedione, had an awareness of this supplement, and more importantly, had an intention to use this product. These studies suggest that professional athletes exert an influence on others, particularly individuals who identify with these athletes. Future research must look at how news coverage of steroid use among professional athletes affects viewers' beliefs, attitudes, and intentions to use these performance-enhancing drugs. Edy and Meirick's (2007) results revealed that understanding individuals' adoption of a frame following the events of September 11, 2001, is a complex process. In the same manner, how individuals adopt frames with respect to steroid usage is likely compounded by multiple factors, and understanding of adoption would benefit from future inquiry into these factors. Research should certainly examine how exposure affects individuals who identify with these athletes in terms of the aforementioned outcomes.

Future research should build on the limitations within this investigation. For starters, relying on news transcripts, which do not take visual communication into account, certainly restricts some of the meaning. A reliance on ABC, CBS, and NBC as opposed to ESPN limits the generalizability of this project. ABC, CBS, and NBC were chosen because Lexis-Nexis covers these three networks whereas ESPN is not included. Additionally, using Lexis-Nexis limited the search to transcripts that were available in this online database. Future research should analyze ESPN's coverage of steroids, as well as coverage by magazines commonly read by athletes, i.e., Sports Illustrated, ESPN the Magazine, Men's Health, and Women's Health. Although research has examined newspaper coverage of Carl Lewis's positive steroid test result (Denham, 2004a), it would be interesting to see whether newspaper coverage resembles television trends. Future research should investigate adolescents' perceptions of the health costs of using steroids. Formative research would aid in message design strategies for future campaigns.

ABC, CBS, and NBC networks between March 1990 and May 2008 have emphasized that steroids are illegal while devoting inadequate attention to the economic, health, and political costs. In sum, the increased presence of steroids in sports has left many unsure about the legitimacy of sports. Sadly, uncertainty about the prevalence of steroid use in sports is not the only uncertainty. Uncertainty about the health consequences of using steroids remain. How severe are steroids to users and how likely is it that steroid users will experience these adverse health effects? Uncertainty about the effects of news coverage about the illegality of steroids rather than the health costs of using these substances lingers. Despite all of these uncertainties, one certainty remains—the issue of steroids in sports deserves much more attention by health communication scholars.

ACKNOWLEDGMENTS

The author thanks Teri Thompson and three anonymous reviewers for their comments and suggestions on this article. Also, thanks to Brian Chan, Jonathan Ortiz, and Nick Smith for their assistance on this project and to Scott Althaus, Isabelle Bauman, and Dave Tewksbury for their recommendations regarding this project.

APPENDIX: TELEVISION NEWS PROGRAMS SAMPLED IN CONTENT ANALYSIS

ABC

• ABC Breaking News

• ABC News Nightline

• ABC News Now

• ABC News Special

• ABC This Week

• Barbara Walters Special

• Downtown

• Good Morning America

• Hopkins 24/7

• ICU

• ICU Arkansas

• Nightline

• Primetime Live

• Special Report

• The Health Show

• Turning Point

• 20/20

• World News Now

• World News Sunday

• World News This Morning

• World News Tonight

CBS

• America Tonight

• CBS This Morning

• Early Show

• Evening News

• Eye to Eye with Connie Chung

• Face the Nation

• 48 Hours

• 48 Hours II

• 48 Hours Investigates

• Live Coverage of O.J. Simpson Trial: Opening Statements

• Morning News

• Osgood File

• Public Eye

• Saturday Early Show

• Saturday Morning

• 60 Minutes

• 60 Minutes II

• Sunday Morning

• Sunday Night News

NBC

• Dateline

• Dateline 2000

• Meet the Press

• NBC Nightly News

• News at Sunrise

• Nightly News

• The News

• Special Report

• Today Show