Color coded: racial descriptors in television coverage of intercollegiate sports.

African American athletes, once excluded from participating in professional and intercollegiate sports because of institutionalized discrimination, now participate in many sports at a rate that equals or greatly exceeds their representation in the population. This is especially true in the case

At the same time that African American participation in professional and collegiate sports was growing, so too was television's (overage of sports. Thanks to the development of satellite technology and cable television, and the growing popularity and marketability of sports, televised sports coverage has proliferated. McCarthy and Jones (1997) noted that the marriage of sports and television has produced one of the more mutually beneficial relationships in the marketplace. Sports coverage has delivered a larger viewing audience to the networks, and the sporting entities have used that coverage to increase their visibility and revenue.

However, at the intersection of these "growth spurts" lies a potential dilemma. Throughout the history of television, African Americans and the African American community have been underrepresented in, and often entirely excluded from, television coverage. When African Americans have appeared, they have often been pigeonholed into demeaning, stereotype-ridden portrayals showing them as bestial, brutish, buffoonish, comical, criminal, dependent on government entitlements or support, ignorant, lazy, menacing, oversexed, and prone to out-of wedlock births (see Bogle, 1994, 2001; Dates & Barlow, 1990; Davis & Harris, 1998; Dixon & Linz, 2000; Entman & Rojecki, 2001; Lule, 1995; MacDonald, 1992; Romer, Jamieson, & de Coteau, 1998; Staples & Jones, 1985; United States Commission for the Study of Civil Rights, 1977; United States Riot Commission Report, 1968).

It is these two contradictory, yet overlapping phenomena that prompted this research project. Specifically, the investigation explored whether the equality that African Americans have gained on the field or court is undermined by television coverage that perpetuates racial bias.

Sports, Television, and Race

The competitive component of sports, especially when viewed in the context of its unscripted action and unknown outcome, creates a fertile field for investigation. As Birrell (1989) noted:

   Sport provides a particularly public display of relations of
   dominance and subordination.... The point of sport is to display
   publicly the processes of challenge and struggle between two sides
   alleged to begin in equal terms but determined to produce and
   sustain relations of dominance vis-a-vis one another. Moreover,
   sport as a meritocracy based on skill quietly reaffirms our
   national common sense; individuals who work hard and possess the
   right stuff will always prevail. Turned on its head, this lesson
   becomes even more insidious: those who are at the top must have
   risen to the top through fair means and thus deserve their
   position. In contrast, those not at the top do not possess the
   requisite talent for such privilege. Even the runner up is a loser.

In addition to their competitive nature, sports possess another aspect that makes it worthy of study. By their nature, sports provide objective measures to evaluate the performance of a player and/or a team. Statistics such as win loss percentage, shooting percentage, and baiting average provide a baseline by which observers can establish a commonly held interpretation of success and failure.

The investigation of sports gains added value when television coverage is factored into the equation. McCarthy and Jones (1997) argued that television has proven to be the perfect medium for covering sports. As the action moves from the field or court via television, the combination of the visual and the aural elements of coverage fulfills many emotional needs of audience members. In addition to these emotional elements, televised sports coverage also carries a personalized, dramatized immediacy that provides the opportunity to create ideological reproductions (Whannel, 2000). These reproductions might not the themselves be reflections of the world at large. Instead, they might be versions constructed by those responsible for producing the coverage. It is at this point where race, the third element of the equation, comes into play.

As Hoberman (1997) noted, sportscasters serve as the unofficial representative of the predominantly White power structure of sports. As such, they have the opportunity to "frame issues and interpret behavior immediately to enormous audiences" (p. 38). Staples and Jones (1985) took Hoberman's assertion one step further by saying that when announcers frame issues and behaviors, they do so in accordance with race-based misperceptions. Birrell (1989) concurred, noting that sports' message of domination and subordination often breaks down along lines of race, class, and gender.

The combination of a predominantly White reporting and announcing corps and an often overwhelmingly large number of African American athletes, coupled with the live, unscripted nature of televised coverage, (an create an environment that is ripe for the reproduction of racialized representations. The literature on stereotype activation and expression has shown that when a person is primed with the presence of the object of stereotypical beliefs, and placed in a situation that creates high anxiety and a need to respond quickly, those stereotypical beliefs rise to the surface and are readily expressed (Devine, 1989, 2001; Kawakami & Dovidio, 2001; Lambert et al., 2003; Lowery, Hardin, & Sinclair, 2001).

In live coverage of sporting events, announcers are placed in just such an environment. Having events unfold before them at a breakneck pace puts announcers under enormous pressure to generate commentary to fill the time and keep the program moving. Such pressure might cause announcers to utter stereotypical beliefs that, in a less frantic environment, might be more controlled.

The question of bias in sports coverage is neither new nor limited to race. To date, several research endeavors have uncovered bias across race, gender, and ethnicity (Davis & Harris, 1998; Duncan & Messner, 1998; Eastman & Billings, 2001; Jackson, 1989; Kane & Lenskyj; 1998; Lule, 1995; McCarthy & Jones, 1997; Rada, 1996; Rainville & McCormick, 1977; Sabo, Jansen, Tate, Duncan, & Leggett, 1996; Tuggle, 1997; Tuggle & Owen, 1999; Whannel, 2000). The presence of bias has not been limited by venue either. Research has found bias across a wide range of sporting endeavors ranging from professional and intercollegiate sports in the United States to international events such as the Olympics. Research has also demonstrated that bias can take many forms, from what is heard (the spoken commentary of the on-air talent) to what is seen (the production practices of those covering the games).

This study focused on previous research investigating racial bias found in announcers' commentary (luring the televised coverage of professional football and intercollegiate men's basketball in the United States. One form of racial bias that researchers have consistently uncovered is the "brawn versus brains" descriptions directed toward the players (Davis & Harris, 1998; Eastman & Billings, 2001; Jackson, 1989; McCarthy & Jones, 1997; Rada, 1996; Rainville & McCormick, 1977). On the surface, complimenting an athlete for his athletic ability and physical attributes would seem to be a positive reflection on that person; however, bias reveals itself when such commentary is viewed through the prism of race. Researchers have found that announcers are more likely to confine their descriptions, and praise, of African American athletes to statements regarding their athletic abilities and physical attributes.

The assumed physicality of African Americans has its roots in the antebellum-era stereotype of the Black Buck, or Brute--an image that is still present in modern-day portrayals of African Americans (Bogle, 1994; Lule, 1995). When that image carries over to the athletic arena, two more elements are added to the portrayal: the natural athleticism and the animal-like physical attributes of African American athletes (Davis & Harris, 1998; McCarthy & Jones, 1997; Rada, 1996; Rainville & McCormick, 1977). Portraying African Americans as naturally athletic or endowed with God-given athleticism exacerbates the stereotype by creating the impression of a lazy athlete, one who does not have to work at his craft (McCarthy & Jones, 1997).

Along with the overt associations that come with describing African American athletes in animal terms, such descriptions also create the impression that they are closer to nature and, thus, further away from civilization (Hargreaves, 1986). These stereotype-ridden expressions came to the fore in 1989 when Jimmy "the Greek" Snyder, then an on-air personality with CBS Sports, openly expressed the belief that the success of African American athletes was the result of selective and effective breeding on the part of slaveowners (Almond, 1989; Wilbon, 1988).

The problem becomes magnified when compared with the portrayal of White athletes. Announcers seem to have ceded the physical element of athletic competition to African American athletes. When announcers do refer to the White athletes' physical accomplishments and/or attributes, they often highlight two factors: the players' intellectual and cognitive prowess, and a strong, blue collar-like work ethic (Davis & Harris, 1998; McCarthy & Jones, 1997; Rada, 1996; Rainville & McCormick, 1977).

Another form of racial bias has been shown to exist through descriptions that create a more overtly negative image of African Americans, especially when announcers endeavor to describe other aspects of a player, such as intellect and character (Davis & Harris, 1998; Rada, 1996; Rainville & McCormick, 1977). Once again, the disparity grows when it is held up against the image announcers craft for White athletes. White players receive an increasingly disproportionate amount of positive commentary in these categories, and negative commentary toward White players is scarce, often to the point of being entirely absent.

Considering the well-established body of research into racial bias in televised sports coverage, we do not intend for this study to be strictly a replication of prior studies. Instead, we seek to use the previous findings as a stepping-stone for a heightened investigation into the distinction between the player-as-athlete and player-as-person. We assert that it is through these categories that announcers take attention away from the player and his or her on-field achievements--an area where the player has much more control over what transpires--and redirect that attention to other aspects of the player's life and character. Through this transformation from the player-as-athlete to player-as-person, the announcers are also able to shift the commentary--and thus the portrayal--from positive to negative.

The Message of Sports

The relevance of this research comes, in part, from the realization that stereotype-ridden descriptions are not just playful banter among the announcers. Consider that for most sports fans, watching the games on television is a ritual endeavor, one repeated several times over the course of a season (Gantz & Wenner, 1991). These consistent consumers are by no means passive. Instead, they spend more time beforehand "preparing" to watch, are emotional]y aroused while watching, and communicate with others more while watching (Gantz & Wenner, 1991). In such an environment, the opportunity exists for a viewer to receive messages that, through repetition, become incorporated into what Squire, Knowlton, and Musen (1993) referred to as semantic memory, "a general knowledge about the world" (p. 459). If this general knowledge includes racial misperceptions, then the possibility exists that exposure to these messages might affect impressions of, and attitudes toward, African American athletes in particular and African Americans in general (Eastman & Billings, 2001).

Researchers have demonstrated a causal link between messages and audience response in sports coverage (for a review, see Bryant & Zillmann, 1991) and television news coverage including portrayals of African Americans (see Pan & Kosicki, 1996; Rada, 2000). Several studies have demonstrated that a stereotype-based general knowledge about African Americans has indeed taken hold in society (Entman & Rojecki, 2001; Lapchick, 2000). Other research has revealed a direct link between people's perceptions and the aforementioned race-based stereotypes of athletes (Sailes, 1996; Stone, Lynch, Sjomeling, & Darley, 1999).

Hypotheses

Many have noted that whereas the role of the announcers during sporting events is not clearly defined, it is clear that on air talent do not simply report objectively on games. Instead, they comment subjectively and seem to go out of their way to offer praise (Hoberman, 1997; McCallum & O'Brien, 1998; Rada, 1996; Stewart, 1990; Wolff & Stone, 1995). This serves as the basis for our first hypothesis:

[H.sub.1]: Announcers will make more positive statements than negative statements.

The results of Hypothesis 1 will provide a baseline for further comparison. If the announcers are overly positive in their commentary, on those occasions when they do choose to criticize, one aspect of bias will be revealed by analyzing to whom this criticism is directed.

[H.sub.2]: African American players will receive more negative comments than will White players.

Once the frequency of who receives the positive and negative commentary has been determined, the next step is to turn to the what: What does the comment refer to? Consistent with previous research, we expect to find that announcers will not hesitate to praise the athleticism of African American athletes. For it is through their overly effusive praise of the African American player-as-athlete that the announcers have confined the imagery of African Americans. In addition, the overrepresentation of comments in this category directed toward African Americans serves to explain the success of African American athletes solely as a result of their physical attributes and/or athletic ability. We do not expect to find such confinement of commentary directed toward the White athletes.

[H.sub.3]: African American players will receive more comments pertaining to physical attributes than will White players.

[H.sub.3a]: African American players will receive more positive comments pertaining to their physical attributes than will White players.

Recall that one of the primary distinctions previous research has demonstrated is the disparity in how announcers credit, or criticize, the cognitive capabilities of African American and White players. Building on these findings, we expect to find similar results in this study. However, to better explicate the distinction between player-as-athlete and player-as-person commentary, we have broken the cognitive category into on-field intellect (player-as-athlete) and off field intellect (player-as-person).

[H.sub.4]: White players will receive more positive comments pertaining to on field intellect than will African American players.

[H.sub.4a]: African American players will receive more negative comments pertaining to on-field intellect than will White players.

[H.sub.5]: White players will receive more positive comments pertaining to off-field intellect than will African American players.

[H.sub.5a]: African American players will receive more negative comments pertaining to off-field intellect than will White players.

In seeking to further investigate the distinctions between player-as-athlete and player-as-person descriptions, our final hypotheses focus on announcers' commentary directed to the personal side of the players' persona.

[H.sub.6]: White players will receive more positive statements about their character than will African American players.

[H.sub.6a]: African American players will receive more negative statements about their character than will White players.

[H.sub.7]: White players will receive more positive personal interest stories than will African American players.

[H.sub.7a]: African American players will receive more negative personal interest stories than will White players.

Method

This research examined televised coverage of men's NCAA Division I-A intercollegiate football and basketball games. The football portion of the sample is comprised of college football games that aired during the 1998 regular season. The sample of basketball games is comprised of CBS's coverage of the 1999 Men's Division I-A Championship Tournament, an event often referred to as "March Madness."

The reasons for choosing televised coverage of intercollegiate sports as the sample for analysis are fourfold. First, the sample selected encompasses live, unscripted, and unedited coverage--the environment in which stereotype priming has been shown to exist. Second, unlike the Olympics, which takes place during a 2- to 3-week span every 2 years, our sample is available annually approximately 6 months out of the year. This provides the repetition that is needed for messages to become part of the general knowledge referred to earlier (Squire et al., 1993). Third, the demographics of intercollegiate football and basketball provide a comparable baseline for analysis: African Americans account for 46% of intercollegiate football players and 56% of intercollegiate basketball players (Center for the Study of Sport in Society, 2001). The final reason intercollegiate sports were chosen comes from examining time results of previous research. Recall that one of the more significant aspects of racial bias has been revealed through commentary focusing on the intellectual or cognitive capabilities. We assert that the coverage of intercollegiate athletics, where players assume the role of student-athletes, will provide more opportunities for announcers to discuss a player's intellect both on and off the field or court.

To gather a sample of college football games, Rada's (1996) method for collecting a composite game was used. Each week, one quarter of a game from each of three networks was taped. (1) The first week, the first quarter was taped; the second week, time second quarter, and so on. Thus, instead of taking one entire game, which would involve only two teams and one announcing crew, our method of sampling composite games presents the opportunity for including eight different teams and four different announcing crews within the same time frame. The sample was drawn from games that were played during a 6-week window from the middle of September through the end of October. We reasoned that taking games from this time frame would allow for the inclusion of a broader range of teams, or individual players, that were having successful seasons. A total of 20 hr of football coverage was taped. This sampling method resulted in a total of 19 quarters, (2) and 25 different teams. Because the stations choose to air games involving teams that are performing well, some teams appeared in the sample more than once. So, too, some announcers appeared more than once. Overall, 18 different announcers were included in the sample, all male, 2 of whom were African American. However, as we note in the Results section, there were no significant differences between sports, the networks, regional coverage, or individual announcers.

A total of 63 games are played in the Men's Division I-A Championship Basketball Tournament. Because multiple games take place simultaneously during the first two rounds, we were not able to obtain all 63 games in their entirety. During the first three rounds, coverage is offered on a regional basis, and during coverage of a game, there are frequent "break-ins" that provide bonus coverage of other games. Thus, for these two rounds, we were able to tape 20 games in their entirety as well as selected segments from other games. For the regional finals, national semifinals, and national final, nationwide coverage is offered. Thus, we were able to tape all 7 of these games in their entirety.

Unlike football, the entire tournament is covered by one network. Because the tournament is broken into four different regions, the coverage still provides the opportunity to analyze different announcing crews. For this sample, each region was covered by two different announcing crews, each consisting of three announcers (or reporters). The sample yielded a total of 24 different on-air talents, 6 from each region, 4 of whom were then used to cover the national semifinal and final. (3) Of the announcers, 2 were African American and 2 were female. A total of 55 hr of basketball coverage was taped for inclusion in the sample.

This research examined only actual game coverage, because of previous findings that reported a significantly higher proportion of commentary fell during the actual game coverage, as opposed to pregame, halftime, or postgame (Rada, 1996). In addition, as mentioned previously, it is during the live, unedited, and unscripted coverage of the game that the stereotype-priming environment is present. The sample used in this research, as measured in number of games and/or hours of coverage, is equal to--and in some cases greater than--those used in previous research (Rada, 1996; Rainville & McCormick, 1977; Wonsek, 1992).

Coding Instrument

Negative comments alone are not necessarily indicative of racial bias. Wonsek (1992) notes that racial stereotypes can manifest themselves both through the presence of specific messages, or the absence thereof, depending on the context in which those messages do, or do not, occur. With this in mind, our analysis centered around what was said about whom and how frequently.

This research employed many of the categories used in previous research (Davis & Harris, 1998; Eastman & Billings, 2001; Jackson, 1989; McCarthy & Jones, 1997; Rada, 1996; Rainville & McCormick, 1977). The first five categories consisted of the network, talent's role (play-by-play, color, or sideline), talent's race, the player's race, and the player's position.

The next portion of the coding instrument focused on the player as an athlete. In contrast to previous studies, this research did not include the category of "play-related praise/criticism." Because statements in this category were usually just recitations of events unfolding on the field or court, this research sought to focus its attention on non-play-related comments by announcers. The first of these categories is physical attributes. This category includes those statements relating to the physical characteristics of a player's body (e.g., height and weight) as well as physical abilities (e.g., agility, leaping ability, strength, speed, etc.). It is important to note that because size is relevant in both football and basketball, statements that provide only a player's "dimensions" were not coded; however, when the description included more than a player's dimensions, the statement was coded. For example, in reading the starting lineups for each team, the announcers will usually give the height of basketball players and the height and weight of football players. Thus, saying, "Smith is six-five, 330 pounds," would not have been coded. However, if the announcer elaborated by saying, "At six-five, 330 pounds, Smith is huge," that would be coded into this category (for an example of statements in each category, see Appendixes A and B).

As mentioned previously, in looking at the categories referring to the intellectual or cognitive descriptions of athletes, this research deviated slightly from that which has been done before. We broke the "cognitive/intellectual" category into two subcategories: on-field intellect and off-field intellect. On-field intellect refers to statements that describe an athlete's intellect in relation, or as a reaction, to the events unfolding on the field or court. This category still falls under the player-as-athlete designation.

Off-field intellect is the first of three player-as-person categories. This category included statements describing a player's academic or other off-field accomplishments. The next category is character reference. Any statements referring to a player's morality, integrity, or values were coded into this category. The third category of player as-person comments is personal interest. At times during the broadcast, the announcers fill time with personal interest stories about a player.

The last category analyzed is the comment attribute, that is, whether the statement was positive or negative. Any statement that was seen as criticizing some aspect of a player or casting a negative light on a player was coded as negative. Statements that were seen as praising a player or presenting a favorable impression were coded as positive. If a statement did not appear to present a definitive positive or negative angle, it was coded as neutral.

All games were taped and coded by the two authors. To reduce the likelihood of intercoder bias, after an initial conference to clarify the coding instrument, each author independently coded a portion of the sample. Using Holsti's (1969) formula, intercoder reliability was calculated at 92%. Two categories that presented a disparity between coders were those of character and personal interest. Of the 62 comments that fell into these categories, there was a small number of comments (n = 5) that coders could not independently determine which of two categories would be more appropriate. In cases such as this, the two coders discussed the statement to determine which of the categories best represented the dominant theme of the statement.

It is important to note here that in trying to determine a player's race, there always exists the possibility of misidentification. It is well beyond the scope of this research to engage in a discussion of the anthropological, biological, cultural, political, and sociological ramifications of race. Using a framework similar to that of McCarthy and Jones (1997), coders limited themselves to the visual information presented before them to determine whether the player was White or non-White (see also Birrell, 1989).

There are certainly many categories that can fall under the umbrella of "non-White." As shown in the previously mentioned demographic breakdown of intercollegiate sports, if a player is not White, he is most likely African American (Center for the Study of Sport in Society, 2001). The coders sought to determine the player's race using the same amount of information available to the viewer at home--namely, a visual identification of the player as he appears on the screen. If an instance arose when a coder was unsure as to the race of the player being described at that moment, the coder sought to find other instances d u ring the game w here more detailed footage of the player was employed. (4)

Results

A total of 486 comments was coded: 151 statements came from football games and 335 from basketball games. Of the comments coded, 330 fell into the categories of physical attributes, cognitive attributes, and character and personal interest. Of the remaining comments, 125 were coded as leadership or credit (not included in this analysis), 23 did not fall into any of the established categories, and 8 were directed toward players who were neither African American nor White, or whose race could not be determined. Because this number was too small to allow for adequate comparison, these cases were not included in the overall results.

About 77% of the players who received an announcer's comments were African American and 23% were White. These percentages are consistent with those found in previous research (Eastman & Billings, 2001; Rada, 1996). It should be noted that although our sample size and demographic breakdown are consistent with other studies, fewer comments were coded in this research due to the more narrow focus of the coding categories. However, when play-related commentary is removed from the samples of some of the other studies, the number of coded comments is consistent with previous studies (McCarthy & Jones, 1997; Rada, 1996).

Ninety-one percent of the comments were made by White announcers, and all but one coded comment was made by male announcers. Seventy-two percent of the coded comments were made by color commentators, 27% by play-by-play announcers, and 1% by other announcers.

No significant relationship was found between the network and the comment attribute, or the player's position and the comment attribute. Nor was there a significant relationship between the announcer's race and the comment attribute. In addition, although the coverage of basketball games provided more than twice as many coded comments than the football games, there was no significant difference between categories across sports. However, a significant relationship was uncovered when looking at the talent, the player's race, and the comment attribute.

About 79% of the comments made by play-by-play announcers were positive, and 89% of the comments made by color commentators were positive. Of the 47 negative statements, play-by-play announcers made 20 (43%), a significantly higher than expected percentage, [chi square] (1) = 6.2, p < . 01, [PHI] = .19. All of the 20 negative statements made by play-by-play announcers were directed toward African American athletes, [chi square] (1) = 8.7, p < .01, Contingency Coefficient = .18.

One other significant relationship was discovered. Of the comments made by color commentators, a significantly higher percentage fell into the player-as-athlete categories of physical attributes (60%) and on-field intellect (29%). In contrast, of the comments made by play-by-play announcers, a significantly higher percentage fell into the player-as-person categories of off-field intellect (16%), character (13%), and personal interest (32%), [chi square](5) = 95.6, p < .01, [PHI] = .53. This relationship held true when controlling for positive and negative comments (see Table 1).

Hypothesis 1 stated that announcers would be more likely to provide positive commentary. Hypothesis 1 was supported. Of the 328 comments for which coders were able to determine a positive or negative attribute, 281 (86%) were positive, t(327) = 59.6, p < .001.

Hypothesis 2 stated that African American players would receive a significantly higher proportion of the negative comments than would White players. Hypothesis 2 was supported. As shown in Table 2, of the 49 comments coded as negative, African American players were the recipients of 92%, [chi square] (1) = 8.4, p < .01, [PHI] = .15. In anticipating the results expected when testing Hypothesis 3, data were submitted to another chi-square, this time controlling for statements relating to physical attributes. The results showed an increase in both the significance and the strength of relationship, [chi square] (1) = 29.9, p < .001, [PHI] = .42.

The results of our tests of Hypotheses 3 through 7 are shown in Table 3. Hypothesis 3 stated that African American players would receive more comments pertaining to physical attributes. Hypothesis 3 was supported. One hundred fifty-one (92%) of the comments referring to an athlete's physical attributes were directed toward African American players, [chi square] (1) = 36.6, p < .001, [PHI] = .32. This category accounted for 57% of the total comments that were directed toward African American players. In contrast, this category accounted for only 19% of the total comments directed toward White players.

Hypothesis 3a staled that African American players would receive more positive comments pertaining to their physical attributes than would White players. Hypothesis 3a was supported. Of the 161 comments relating to physical attributes that were coded as positive, African Americans received a significantly higher percentage (92%) than White players, [chi square] (1) = 6.01, p < .05, [PHI] = .19.

Although attribution or explanation for an athlete's athleticism was not hypothesized, we sought to investigate whether there were any differences by race in how announcers accounted for a player's athletic abilities. There were only seven cases of ability attribution. Of those, all four that attributed the players' ability to natural or God-given talent were directed toward African Americans. All three that credited a player's willingness to work hard were directed toward White players.

Hypothesis 4 stated that White players would receive more positive comments pertaining to on-field intellect. Hypothesis 4 was supported. Overall, African Americans received 71% of the comments related to on-field intelligence. This number is in line with African American representation in the sample. However, when controlling for positive and negative comments, in relation to their representation in the sample, White players received a significantly higher percentage of the positive commentary referring to on-field intellect, [chi square] (1) = 5.0, p < .05, [PHI] = .12.

Hypothesis 4a stated that African American players would receive more negative comments pertaining to on-field intellect that would White players. Hypothesis 4a was supported. Of the 24 negative comments pertaining to on-field intellect, African Americans received 96%, [chi square] (1) = 10.1, p < .001, [PHI] = .36.

Hypothesis 5 stated that White players would receive more positive comments pertaining to off-field intellect than would African American players. Hypothesis 5 was not supported. There was no significant difference for positive comments referring to off-field-intellect.

Hypothesis 5a stated that African American players would receive more negative comments pertaining to their off-field intellect than would White players. Hypothesis 5a was supported. In looking at negative comments pertaining to off-field intellect, African Americans were the recipients of all the negative comments, [chi square] (1) = 3.67, p < .05, [PHI] = .44.

Hypotheses 6 stated that White players would receive more positive comments pertaining to their character. Hypothesis 6 was supported. Of the 24 positive comments relating to a player's character, White players received 58%, [chi square] (1) = 13.9, p < .001, [PHI] = .22.

Hypothesis 6a stated that African American players would receive more negative comments pertaining to their character than would White players. Hypothesis 6a was not supported. There were just two negative comments made in reference to a player's character. Although both comments were directed toward African Americans, the numbers are too small to allow for comparison.

Hypothesis 7 stated that White players would receive more positive comments relating to personal interest stories. Hypothesis 7 was supported. Of the 27 positive personal interest stories, White players received 56%, [chi square](1) = 7.78, p < .005, [PHI] = .47.

Hypothesis 7a stated that African Americans would receive more negative comments related to personal interest stories than would White players. Hypothesis 7a was supported. Of the eight negative comments associated with personal interest stories, all were directed toward African American players, [chi square] (1) = 7.78, p < .005, [PHI] = .47.

During the course of this analysis, a pattern worth mentioning was discovered. Recall that Hypothesis 3 stated that African American players would receive more comments pertaining to their physical attributes. This hypothesis focused on comments directed toward individual players. During the course of this analysis, a similar trend developed on a team-related scale. Eighteen times during the basketball coverage, the announcers made reference to the overall athletic ability of the team as a whole (e.g., "[School] really has an athletic group on the floor right now"). Every time this happened, the team referred to had five African American players on the court, [chi square](1) = 6.3, p < .05, [PHI] = 1.00.

A selection taken from the sample of college football games provided the opportunity to analyze comments directed at an almost equal number of players from the same position. In the 19 quarters of college football games included in the sample, there were a total of 38 quarterbacks: 20 African American and 18 White. There were not enough comments to test for differences across all categories; however, there was ample commentary to test for differences in how the announcers described the quarterbacks' physical and cognitive attributes.

Consistent with the larger sample, African American quarterbacks received a significantly higher proportion (90%) of the overall comments pertaining to physical attributes, [chi square] (4) = 12.82, p <. 01, Contingency Coefficient = .47. The relationship held true when controlling for positive comments relating to physical attributes (n = 18, or 95%), [chi square] (4) = 11.8, p < . 02, Contingency Coefficient = .58.

Although there were no significant differences in positive commentary related to a player's intellectual or cognitive attributes, a significant relationship was found for negative commentary within this category. African Americans received all of the negative comments (n = 10) pertaining to on field intellect, [chi square] (4) = 11.78, p < .02, Contingency Coefficient = .70.

Building on the results of our hypothesis testing, we present the findings from another angle. Table 3 provides a breakdown of the results as a percentage of components within the coded categories; however, we also sought to investigate the commentary in specific categories as a percentage of the comments directed toward players of different races. As shown in Figures 1 and 2, as announcer comments about African Americans move further away from a player as an athlete, and more toward a player as a person, the frequency of negative comments increases. For White players, the exact opposite is the case.

[FIGURES 1-2 OMITTED]

Predicting Race by the Announcer's Comment

Our final analysis was an attempt to determine whether it was possible to predict the race of a player based on the type of comments made about him. Data were submitted to a discriminant analysis. The player's race was the dependent variable, and the type of comment and the comment attribute were entered as predictor variables.

The analysis yielded four significant relationships. The strongest relationship came from comments relating to a player's physical attributes. The positive correlation between physical attributes and a player's race shows that if a statement was made that referred to a player's physical attributes, statistically speaking, we can predict that the player was African American (r = .69, p < .001).

The second significant predictor is comments relating to a player's character. If the statement referred to time player's character and the statement was negative, we can predict that the player was African American (r = -.40, p < .001).

The third significant predictor was the comment attribute. Overall, if time statement was negative, we can predict that the player was African American(r = -.32, p<.01).

The final significant predictor was personal interest stories about the player. If the statement was a personal interest story about the player and the statement was negative, then we can predict that the player was African American (r = .30, p < .001).

Looking at the group centroids provided by the discriminant analysis also revealed one other significant finding. Variables were dummy coded so that they would fall within a continuum where a positive statement equaled 1.0. The group centroids for White players (.95) shows the proximity between this group and the positive end of the continuum. In contrast, the group centroids for African American players (-.30) reveals a significant difference between this group and both the group centroids for White players and the positive end of the continuum. These results show that, statistically speaking, if a statement is negative, it is almost always directed at an African American player.

Discussion

This research sought to determine if the strides that African American athletes have made on the field or court have been undermined by television coverage that perpetuates racial bias. More than 25 years after Rainville and McCormick (1977) first found racial bias in television's coverage of professional football, our research has found that these disparities still persist.

Consistent with previous research, we found that announcers continue to paint a picture wherein African American athletes are portrayed as physical specimens using their God-given, natural ability, whereas White athletes are hard working and intellectually endowed. More troubling is the disparity in how the announcers craft positive and negative descriptions of the players. The door to positive imagery appeared to open slightly when we found that there were no significant differences across race in positive descriptions of off-field intelligence.

However, that door slammed shut when we looked at the negative commentary announcers directed toward the athletes. Of particular interest to this study was the disparity in how announcers describe the player-as-athlete versus descriptions of the player-as-person. When it came time to describe the African American players as people, the announcers' criticism knew no bounds. Of the 18 negative statements made about a player's off field intelligence, character, or personal interest stories, all 18 were directed toward African Americans. White players, in contrast, received only positive comments across these categories.

One more illuminating finding comes from the ability to predict the race of the player based on the type of comment. When Rainville and McCormick (1977) conducted their research into racial bias, it was noted that one of the authors, who was blind, was able to distinguish the race of a player based strictly on the announcer's description. The discriminant analysis employed by this research provided a similar scenario. The results showed that when a statement is critical of the player in any way, without using any visual information, we can predict that the player is African American.

In sum, the results of this research show that although the doors of membership in the once exclusive club of positive imagery have opened a bit for African American athletes, the opening is slight, and the privileges are restricted. In contrast, African Americans hold almost exclusive membership in the club of negative imagery. As has been discussed throughout, this dichotomy between the player-as-athlete and player-as-person is a troubling one. By shifting the focus from the athlete to the person, the announcers are also able to shift the venue from the athletic arena to the society at large. Through their negative descriptions, the announcers portray African American athletes as being at odds with society.

In effect, announcers have painted African American players into a corner wherein they can be accepted as athletes but only as athletes. Should African Americans endeavor to journey beyond the field or court, they are then typecast into the same stereotype-ridden portrayals that have been found in other venues of television programming. This raises the question, If African Americans cannot receive equal treatment in one of the venues of society where they have achieved equal or majority representation based on merit, where can they hope to receive it?

Suggestions for Further Research

With a well-established body of research consistently showing a disparity in the messages that are directed toward players of different races, we suggest two avenues for future investigation. The first is the messenger. As mentioned earlier, the role of on-air talent during the course of coverage has never been clearly defined. There are those who opine that the people charged with covering the game do not adhere to any standard of journalistic objectivity but instead fall into the role of cheerleader. These assertions gain validity when we look at the results of this study. Considering the announcers' overebullience, it is especially troubling that, in those rare instances when they do criticize, their criticism is overwhelmingly directed toward African Americans.

We would suggest further investigation into the process of message construction, specifically, the training announcers receive before taking to the air. Particularly enlightening would be an investigation of the decision making processes that are used to gather and disseminate personal information about the athlete. In addition, it would be beneficial to examine the kind of training that potential on-air announcers receive before they venture into the booth. This avenue of exploration would be especially worthwhile considering that during live coverage, the announcing booth becomes an environment that is conducive to priming racial stereotypes.

A second avenue for future researchers to explore is the potential effect(s) that racially biased messages found in televised sports coverage might have on a viewer. Previous research has demonstrated a link between media portrayals and audience perceptions. It would be beneficial to further explore how direct that link might be, especially in light of the negative off-field imagery of African American athletes found in this study. If such portrayals contribute to the development of audience perceptions, the negativity directed to African Americans on the screen might translate into perceptions imposed on the hearts and minds of society. If those perceptions translate into behavior, then the messages found in televised sports coverage and in other mediated communications could very well serve as a barrier to realizing a level playing field in society for African Americans.

Appendix A:

Representative Positive Comments by Category

Physical:

* "[Player] is a load--260 pounds."

* "He has great legs, great quickness, and tremendous leaping ability."

On-Field Intellect:

* "Smart play--he knew the defender was coming, so he threw it away."

* "[Player's] so smart--the ability, to find holes in the defense."

Off-Field Intellect:

* "[Player] the punter--he's also an excellent student."

* "[Player] has one degree and is working on another one ... extremely bright young man."

Character:

* "He's pleasant, affable, a sort-spoken athlete."

* "[Player's] unselfish--willing to sit the bench and let others start for the benefit of the team."

Personal Interest:

* "[Player] growing up on the family farm in North Dakota--family needs him on the farm, but they don't want to stand in the way of his dream of playing ball."

* "He's an art and computer graphics major ... he's a piano player--he also works with the computer graphics team that run the scoreboards at [school's arena] ... and also at [school's stadium]."

Appendix B: Representative Negative Comments by Category

Physical:

* "[Player's] quick enough to grab the rebound, but he doesn't have the strength to put it back up."

* "He is not really that strong physically."

On-Field Intellect:

* "Look at [player]--he's looking at his wristband and he's looking at the route, looking at the options and obviously he forgot what he did--too much going on."

* "He's a freshman, what a time for a breakdown in concentration."

Off-Field Intellect:

* "After successful first two years on court, [player] had academic difficulties and was suspended by [coach]."

* "[Player] a junior out of Baltimore--he was recruited by [school] but his grades weren't good enough."

Character:

* "Some guys have a hard time getting it all together they have a hard time marching with everyone else--[player's] one of those guys."

* "He's had an interesting off court session at [school]--two DUIs, suspended, but now he's back."

Personal Interest:

* "[Player] who did have some severe suspensions laid on him this year--an altercation in the [school] game this year."

* "[Player] is in the ballgame after sitting out the first quarter for his unexcused absence on Thursday."

Table 1 Comment by Announcer Role

                                 Play-by-     Color Co-
                       Total       Play       mmentator

                       n    %    n    % (a)    n     %

Physical attributes   166   50   28   28      138   60 *
On-field intellect     76   23   10   10       66   29 *
Off-field intellect    19    6   16   16 *      3    1
Character              27    8   13   13 *     14    6
Personal interest      43   13   32   32 *     10    4
Total                 330        99           231

(a) Percentages do not equal 100 due to rounding.

* p < .01.

Table 2 Total Comments by Comment Attribute by Race

                                              African American
                            Total                Players

                       Posi-      Nega-     Posi-      Nega-
                        tive      tive       tive      tive

                       n    %    n    %    n     %     n     %

Physical              161   58    5   11   148   70    3      7
On-field intellect     56   20   24   51    34   16   23     52
Off-field intellect    11    4    8   17     7    3    8     18
Character              24    9    2    4    10    5    2      5
Personal interest      27    9    8   17    12    6    8     18
Total                 279        47        211        44 *

                       White Players

                       Posi-    Nega-
                        tive    tive

                      n    %    n   %

Physical              13   19   2   67
On-field intellect    22   32   1   33
Off-field intellect    4    6   0
Character             14   21   0
Personal interest     15   22   0
Total                 68        3

Note: N = 326. Two statements were coded as neutral, and two negative
statements (one directed toward an African American athlete and one
directed toward a White athlete) fell into other categories.

* p < .01.

Table 3 Specific Comment by Comment Attribute by Player's Race

                                   On-Field    Off-Field
                      Physical    Intellect    Intellect

                       n     %      n     %     n     %
Positive
  African American   148 *   92   34 *    60   7      64
  White               13 *    8   22 *    40   4      36
Negative
  African American     3     60   23 **   96   8 *   100
  White                2     40    1 **    4   0 *     0

                                    Personal
                      Character     Interest

                       n      %      n      %
Positive
  African American   10 **    42   12 **    44
  White              14 **    58   15 **    56
Negative
  African American    2      100    8 **   100
  White               0        0    0 **     0

Note: Two statements were coded as neutral, and two negative statements
(one directed toward an African American athlete and one directed
toward a White athlete) fell into other categories.

* p < .05. ** p < .01.

Notes

(1) One of the two broadcast networks that cover collegiate football during the regular season, ABC, offers regional coverage. For this sample, two regions were taped: the "Southeast" region, which aired on the networks' Atlanta affiliate; and the "Far West" region, which aired on the network's San Diego affiliate. The only national cable network that covered college football during the 1998 season, ESPN, was included in this study.

(2) The sample yielded an actual total of 20 quarters. One quarter was removed from the results of the study and used for testing intercoder reliability.

(3) During regional coverage, the network used one sideline reporter. For the Final Four, the network used two sideline reporters.

(4) At the beginning of each game, when announcing the starting lineups for both teams, stations usually superimpose close-up photos of each individual player. In addition, at various points throughout the game, stations will use those same photos when drawing attention to an individual player whom they wish to highlight.

References

Almond, E. (1989, April 28). Question of race; debate over whether Black athletes are superior to Whites is not a new one. Los Angeles Times, p. C3.

Birrell, S. (1989). Racial relations theories and sport: Suggestions for a more critical analysis. Sociology of Sport Journal, 6, 212-227.

Bogle, D. (1994). Toms, coons, mulattoes, mammies, and bucks: An interpretive history of Blacks in American films. New York: Continuum.

Bogle, D. (2001). Primetime blues: African Americans on network television. New York: Farrar, Straus & Giroux.

Bryant, I., & Zillmann, D. (Eds.). (1991). Responding to the screen: Reception and reaction processes. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

The Center for the Study of Sport in Society: Northeastern University. (2001). Racial and gender report card. Retrieved June 1, 2002, from http://www.sportinsociety.org

Dates, J., & Barlow, W. (1990). Split intake: African Americans in the mass media. Washington, DC: Howard University Press.

Davis, L. R., & Harris, O. (1998). Race and ethnicity in U.S. sports media. In L. Wenner (Ed.), Mediasport (pp. 154-169). London: Routledge.

Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18.

Devine, P. G. (2001). Implicit prejudice and stereotyping: How automatic are they? Journal of Personality and Social Psychology, 81, 757-759.

Dixon, T., & Linz, D. (2000). Race and the misrepresentation of victimization on local television news. Communication Research, 27, 547-573.

Duncan, M. C., & Messner, M. A. (1998). The media image of sport and gender. In L. Wenner (Ed.), Mediasport (pp. 170-186). London: Routledge.

Eastman, S. T., & Billings, A. C. (2001). Biased voices of sports: Racial and gender stereotyping in college basketball announcing. The Howard Journal of Communications, 12, 183-201.

Entman, R. M., & Rojecki, A. (2001). The Black image in the White mind: Media and race in America. Chicago: University of Chicago Press.

Gantz, W., & Wenner, L. A. (1991). Men, women and sports: Audience experiences and effects. Journal of Broadcasting & Electronic Media, 35, 233-243.

Hargreaves, J. (1986). Sport, power and culture. Oxford, England: Polity.

Hoberman, J. (1997). Darwin's athletes: How sport has damaged Black America and preserved the myth of race. New York: Mariner.

Holsti, O. (1969). Content analysis for the social sciences and humanities. Reading, MA: Addison Wesley.

Jackson, D. (1989, January 22). Calling the plays in Black and White. Boston Globe, pp. A25, A28.

Kane, M. J., & Lenskyj, H. J. (1998). Media treatment of female athletes: Issues of gender and sexualities. In L. Wenner (Ed.), Mediasport (pp. 186-201). London: Routledge.

Kawakami, K., & Dovidio, J. (2001). The reliability of implicit stereotyping. Personality and Social Psychology Bulletin, 27, 212-225.

Lambert, A., Payne, B., Jacoby, L., Shaffer, L., Chasteen, A., & Khan, S. (2003). Stereotypes as dominant responses: On the "social facilitation" of prejudice in anticipated public contexts. Journal of Personality and Social Psychology, 84, 277-295.

Lapchick, R. E. (2000). Crime and athletes: New racial stereotypes. Society, 37(3), 14-20.

Lule, J. (1995). The rape of Mike Tyson: Race, the press and symbolic types. Critical Studies in Mass Communication, 12, 176-195.

Lowery, B., Hardin, C., & Sinclair, S. (2001). Social influence effects on automatic racial prejudice. Journal of Personality and Social Psychology, 81, 842-855.

MacDonald, J. (1992). Blacks and White TV? African Americans in television since 1948 (2nd ed.). Belmont, CA: Wadsworth.

McCallum, J., & O'Brien, R. (1998, March 30). More bull from Billy. Sports Illustrated, 88(13), 29.

McCarthy, D., & Jones, R. L. (1997). Speed, aggression, strength, and tactical naivete. Journal of Sport & Social Issues, 21, 348-362.

Pan, Z., & Kosicki, G. (1996). Assessing news media influences on the formation of Whites' racial policy preferences. Communication Research, 23, 147-178.

Rada, J. A. (1996). Color blind-sided: Racial bias in network television's coverage of professional football games. The Howard Journal of Communications, 7, 231-240.

Rada, J. A. (2000). A new piece to the puzzle: Examining effects of television portrayals of African Americans. Journal of Broadcasting & Electronic Media, 44, 704-715.

Rainville, R., & McCormick, E. (1977). Extent of covert racial prejudice in pro football announcers' speech. Journalism Quarterly, 54, 20-26.

Romer, D., Jamieson, K., & de Coteau, N. (1998). The treatment of persons of color in local television news: Ethnic blame discourse or realistic group conflict? Communication Research, 25, 286-306.

Sabo, D., Jansen, S., Tate, D., Duncan, M., & Leggett, S. (1996). Televising international sport: Race, ethnicity, and nationalistic bias. Journal of Sport & Social Issues, 21, 7-21.

Sailes, G. A. (1996). An investigation of campus stereotypes: The myth of Black athletic superiority and the dumb jock stereotype. In R. Lapchick (Ed.), Sport in society: Equal opportunity or business as usual? (pp. 193-202). Thousand Oaks, CA: Sage.

Squire, L. R., Knowlton, B., & Musen, G. (1993). The structure and organization of memory. Annual Review of Psychology, 44, 453-495.

Staples, R., & Jones, T. (1985, May/June). Culture, ideology and African American television images. The Black Scholar, pp. 10-20.

Stewart, L. (1990, March 30) Packer, Tompkins both get caught up in all the madness. Los Angeles Times, p. C3.

Stone, J., Lynch, C., Sjomeling, M., & Darley, J. (1999). Stereotype threat effects on Black and White athletic performance. Journal of Personality and Social Psychology, 77, 1213-1227.

Tuggle, C. A. (1997). Differences in television sports reporting of men's and women's athletics: ESPN Sportscenter and CNN Sports Tonight. Journal of Broadcasting & Electronic Media, 41, 14-24.

Tuggle, C. A. & Owen, A. (1999). A descriptive analysis of NBC's coverage of the centennial Olympics: The "games of the women." Journal of Sport & Social Issues, 23, 171-182.

United States Commission for the Study of Civil Rights. (1977). Window dressing on the set: Women and minorities in television. Washington, DC: U.S. Government Printing Office.

United States Riot Commission Report. (1968). Report of the National Advisory Commission on Civil Disorders. New York: Bantam.

Whannel, G. (2000). Sport and the media. In J. Coakely & E. Dunnings (Eds.), Handbook of sports studies (pp. 291-308). London: Sage.

Wolff, A., & Stone, C. (1995, April 3). Backscratchers. Sports Illustrated, 82(13), 22.

Wilbon, M. (1988, January 17). The Greek's apology is difficult to accept. Washington Post, p. D1.

Wonsek, P. L. (1992). College basketball on television; a study of racism in the media. Media, Culture and Society, 14, 449-461.

James A. Rada (Ph.D., The University of Georgia) is an Assistant Professor in the John H. Johnson School of Communications at Howard University. His research interests include mediated portrayals of African Americans and the potential effects those portrayals may have on an audience.

K. Tim Wulfemeyer (Ed.D., UCLA) is a Professor in the School of Communication at San Diego State University. His research interests included the content of radio-television newscasts, audience preferences, ethics in journalism, and media literacy.