To find evidence of a racialized outdoor leisure identity, I performed a content analysis of magazine advertisements in which I recorded the geographical settings occupied by Black and White models, as well as the activities in which they participated and the image they conveyed. To ensure a wide variety of print advertising, I analyzed three different magazines that are targeted to reach three different, though non-mutually exclusive, audiences. First, I examined Time, a mainstream newsmagazine with a very high circulation, to see what images are presented by advertisers to the general society at large. Secondly, I examined Ebony, a magazine marketed primarily to Blacks, to determine the images to which Black Americans are exposed. Lastly, I examined Outside, a magazine marketed specifically to outdoor recreation participants, to learn what images are being presented to outdoor enthusiasts. Time and Ebony are both among the largest selling magazines in their respective classifications, and each has been used in content analyses by other researchers (Bowen & Schmid, 1997; Cox, 1969; Feasley & Stuart, 1987; Humphrey & Schuman, 1984; Shuey, 1953). Outside is a niche magazine, and while it has a smaller readership, it is the principal magazine devoted specifically to outdoor recreation.

My study spans the years 1985 to 2000. Besides the obvious advantage of being contemporary, the chosen time period also has other benefits. The advances of the Civil Rights Movement led to a decrease in the socioeconomic differences between Blacks and Whites in several important areas as evidenced by the emergence of an expanded Black middle class (Allen & Farley, 1986; Landry, 1987). Larger disposable incomes and enhanced purchasing power for Blacks, in combination with refinements in the targeted marketing strategies employed by advertisers, should provide for a greater variety of media images of Blacks than in previous decades. There was also an increase in the overall popularity of outdoor recreation activities during this time period. The combination of these factors makes the late 1980s and the 1990s an ideal era for this study.

I decided to analyze two issues, six months apart, from each year over the study period, and I wanted to examine what I considered to be standard editions (i.e., no seasonal, holiday, or special year-ending issues). With that in mind, I selected the March and September issues, and there is no reason to suspect that the results of the study would be substantially different had I chosen otherwise. In the case of Time, which is published weekly, I chose to examine the second issue of each applicable month. I coded all ads within those issues, and the unit of measurement, or the recording unit, was an ad page: defined simply as any page of the magazine that contains advertising. Because, however, I am interested in the individual identities of the characters depicted in the ads, I also made a separate record for each person appearing in the ad, up to four people per ad, beginning with the most prominent person. Whenever I am comparing the magazines to one another as a whole, I use ad pages as the unit of analysis, where each ad page counts only once regardless of the number of people who appear in the ad. Whenever I am comparing the people who appear in the magazines, I use individual people as the unit of analysis, where each person counts once regardless of in which ad they may appear.

In order to prevent measurement problems resulting from discrepancies in the size of the advertisements, only full page or multi-page ads will be included in the sample. Multi-page, or multi-spread ads, were counted and analyzed once for each page in the ad spread. Inserts (e.g. record or book clubs ads, etc.) were not included in the sample and were not counted to determine the number of ads per issue. As I am interested in the visual representation of a specific leisure identity, I only examined advertisements that contain pictures rather than include all other kinds of advertising (Feasley & Stuart 1987). Advertisements that contain only text will be counted to determine the total number of ads in each issue, but they will be coded as text-only and be excluded from the later analysis.

Ads with pictures were coded into two mutually exclusive categories based on the content of the ad: outdoor advertisements, and non-outdoor advertisements. The first category, outdoor advertisements, is defined as any ad that is either set in the Great Outdoors or that uses the Great Outdoors as a major theme. Also included in this category are ads in which a major character in the ad displays an outdoor leisure identity. The term outdoor leisure identity is operationally defined as a person who is depicted engaging in a wildland leisure activity (e.g, hiking, camping, rock climbing, or mountain biking). The second category, non-outdoor advertisements, includes all other ads not classified as outdoor ads. This category contains ads set in rural, suburban, or urban environments, as well as ads that are set indoors. Although the precise location along a rural/urban continuum is often impossible to determine for advertisements with indoor settings, such ads, by definition, do not exemplify a wilderness identity and are thus included in the non-outdoor group. Theoretically, it is possible for an individual to present an outdoor leisure identity even in an indoor setting, for instance an ad set inside a mountain cabin, but I did not encounter any such ads. Advertisements without a setting, or in which the setting is indeterminate, are coded into the above groups according to the image or activity depicted by the product or the actor. Some activities were coded as urban (baseball, basketball), and some activities are coded as outdoor (mountain biking, rock climbing), regardless of the indeterminacy of the setting.

The larger categories, outdoor ads and non-outdoor ads, were also then sub-divided into two mutually exclusive classes: ads that are not populated (that do not feature people), and ads that are populated (that do feature people). The first sub-class, non-populated ads, includes all ads where no people are present, or where the identities of visible persons are impossible to distinguish. Such ambiguity is a common occurrence in advertisements for automobiles where the outline or silhouette of the person driving is readily apparent, but their racial or sexual identity is impossible to differentiate. Also, in many of the outdoor ads depicting hikers, mountain climbers, or skiers, the models were often so bundled up with clothing that their individual identities were indistinguishable. The second sub-class, populated ads, includes any ad that contains at least one person, except as noted below. Advertisements that feature cartoon characters (e.g., Joe Camel, Mickey Mouse), and ads with drawings of people that are intentionally unrealistic (e.g., caricatures), are coded as ads without people. This was done because such ads are obviously fictionalized and they often do not project a distinct or consistent racial or sexual personality with which readers can identify.

The models themselves were coded according to their race, sex, and child/adult status. The variables sex and child/adult status were coded dichotomously, while race was coded trichotomously (Black, White, and other), and determinations were made subjectively based on visual cues such as skin color, hair type, facial features, and style of dress (for sex). Skin color is perhaps the most salient external feature for distinguishing one person from another and researchers often classify individuals visually during observation studies (Hughes & Hertel, 1990; Hutchison, 1987). All coding used for this study was done by the author. However, to check the reliability of the coding scheme, two colleagues each rated one entire issue of each of the three magazines (a total of 223 records). Agreement among the coders was determined using Cohen’s Kappa, a conservative measure of intercoder reliability, and the results were above acceptable levels for all variables coded (Lombard, Snyder-Duch, & Bracken, 2002). The average agreement between the three coders was 82.1% for race, and 91.2% for outdoor image.

Because there were so few depictions of models other than Blacks or Whites (Asians or Hispanics for instance), the racial identity of such models was coded as “other” and they were excluded from the analysis. Black and White models were then coded as to whether or not they displayed an outdoor leisure identity. This was usually determined by the setting, but in the case where there was no setting or the setting was ambiguous, the activity in which the model was engaged was often telling, as was the text of the ad in some instances. Because I am interested primarily in leisure identities in the U.S., tourism ads for foreign countries were excluded from the analysis (0.6%). Ads for other types of media (e.g., television programs, movies, or books) were also excluded (2.5%).

A special category, collage ads, was created for ads in which there is a myriad of different images, or for ads that feature a large group of diverse people where no individual identity is dominant (< 1.0%). In the preliminary analysis, a typical collage advertisement featured a schoolteacher and a classroom of students, with several students of both sexes and at least one representative from several racial and ethnic groups. In such cases, the identity of the main character (i.e., the teacher) was noted, and the rest of the advertisement was coded as a collage. If there is no main character, then the entire ad was coded as a collage and was excluded from the analysis.

A few further rules were also observed when coding the ads. Advertisements that show multiple photos of the same person counted as only one person. Also, as mentioned above, only four people per ad were coded. In ads that featured large groups of people from a single racial group (4+), regardless of their sex, the entire group was counted as four people, with the sexes coded to best reflect the true composition of the group. For example, one issue of Time magazine featured the Zion Baptist Gospel Choir (four Black males and four Black females), however it was the only advertisement in that particular issue that featured Black models. Counting each member of the choir separately would lead to the false impression that blacks were more heavily represented throughout the issue than they actually were. Following the above scheme, the ad was coded as displaying two Black males and two Black females. Thus, no single ad page can contain more than four people, although a two-page spread could contain eight people, subject to the above constraint that no person appears more than once in any ad, regardless of the number of additional pages.