Env. Mgmt. 25/4 - Springer Link

DOI: 10.1007/s002679910032

Toward an Understanding of Norm Prevalence: A Comparative Analysis of 20 Years of Research1 MAUREEN P. DONNELLY JERRY J. VASKE DOUG WHITTAKER Colorado State University Fort Collins, Colorado 80523, USA BO SHELBY Oregon State University Corvallis, Oregon 97331, USA

ABSTRACT / Norms are defined as evaluative standards regarding individual behavior or conditions in a given context. They define what behavior should be, rather than actual behavior. Norm prevalence refers to the proportion of individuals in a population who can articulate a norm in a given evaluation context. This paper empirically examines the prevalence of encounter norms in 56 evaluation contexts.

As recreation research has matured, investigators have expanded the focus of study to include comparative analyses of data aggregated across activities and resources. For example, comparative analyses have been reported for concepts such as satisfaction (Vaske and others 1982), motivation (Kuentzel 1990, Manfredo and others 1996), crowding (Shelby and others 1989), and norms (Vaske and others 1986, Shelby and others 1996).2 By contrasting identical measures of the same concept across a number of activities, resources, and visitor characteristics, aggregated data sets can reveal patterns in the findings and identify causal factors that typically cannot be manipulated in a single study (e.g., effects from different response formats). The current debate in the normative literature (e.g., see Roggenbuck and others 1991, Shelby and Vaske 1991, Shelby and others 1996) highlights the imporKEY WORDS: Encounter norms; Norm prevalence 1This

is a revised version of a paper presented at the 6th International Symposium on Society and Resource Management at The Pennsylvania State University, University Park, Pennsylvania, 18–23 May 1996. *Author to whom correspondence should be addressed.

Data for this comparative analysis were obtained from 30 studies that used a single-item question asking recreationists to indicate the highest number of encounters they would tolerate before the experience changed. Four predictor variables were examined: (1) type of resource, (2) type of activity, (3) type of encounter, and (4) question response format. As anticipated, norm prevalence varied by type of resource (backcountry or frontcountry), type of encounter (no conflict versus conflict), and question response format (two-category implicit, two-category explicit, and three-category). These three independent variables explained 64% of the variance in norm prevalence. Also as hypothesized, there was no relationship between type of activity (consumptive or nonconsumptive) and norm prevalence. Implications for future research and management are discussed; it is argued that prevalence is an important characteristic of social norms.

tance of systematic examination of the available empirical evidence. This debate was stimulated by a study of boaters on the New River in West Virginia, a high-use, frontcountry river (Roggenbuck and others 1991). In that investigation, encounter norms were measured for three different river experiences: a wilderness whitewater trip, a scenic whitewater trip, and a social recreation trip. Depending on the type of experience, only 29%– 50% of the respondents specified an encounter norm.3 These findings led the authors to question the ‘‘existence’’ of norms and raise methodological questions about previous studies of norms. In response, Shelby and Vaske (1991) presented data from several western rivers where the percent of visitors specifying encounter norms ranged from 73% to 84%. With findings so different from those on the New River, Shelby and Vaske pointed out that situational and methodological factors may have accounted for the low numbers in the New River study. When few respondents answer norm questions, either the norms are not relevant in that particular context or measurement problems make responding difficult (Shelby and Vaske 1991). This exchange has generated lively debate at conferences and in the journals (see McDonald 1996).

2Norms

are defined here as evaluative standards regarding individual behavior or conditions in a given context (Shelby and others 1996, Vaske and others 1986). Similar to Homans (1950, p. 124) definition, ‘‘norms are not behavior itself, but rather what people think behavior ought to be.’’

Environmental Management Vol. 25, No. 4, pp. 403–414

3Encounter norms are standards defining the number of contacts with other people an individual will tolerate over the course of a day (or trip).

r 2000 Springer-Verlag New York Inc.

404

M. P. Donnelly and others

Models used for assessing encounter norms measure ‘‘personal norms’’ at the individual level, then aggregate the individual data to describe ‘‘social norms’’ (Vaske and others 1986, Shelby and others 1996). Characteristics of social norms typically derived from these data include the optimal number of encounters, the range of tolerable encounters, level of agreement (crystallization), and intensity. The percent of respondents giving a norm is another characteristic of social norms. This proportion is an important first step in describing the social characteristics of norms and may be referred to as norm prevalence. Prevalence can range from 0% to 100% and is interesting regardless of whether the proportion of respondents giving a norm is low or high. If norm prevalence is low, the issue may not be relevant to respondents, or the measurement technique may be confusing or difficult. If prevalence is high, the norm is probably relevant for respondents and measurement has been successful. This paper focuses on norm prevalence.4 More specifically, the objectives are to: (1) summarize studies that have examined the prevalence of encounter norms, (2) evaluate variables influencing norm prevalence, (3) develop a model that explains norm prevalence, and (4) assess the theoretical and applied importance of this variable. The existing literature identifies both experiential and methodological variables that may account for a respondent’s willingness to specify an encounter norm. Among the experiential variables, differences have been suggested for different: types of resources— frontcountry versus backcountry (e.g., Manning and others 1996, Vaske and others, 1996), types of activities— consumptive versus nonconsumptive (e.g., Vaske and others 1982), and types of encounters—conflict versus no conflict (Vaske and others 1995a). Methodologically, variations in question response format (Roggenbuck and others 1991, Hall and others 1996) have also been suggested to influence the number of people who specify a norm. This study examines these three experiential variables and the one methodological variable.

Type of Resource Encounter norms for two types of resources— backcountry and frontcountry—have been reported in the literature. Because meeting others can disrupt the solitude of an experience, encounters have been shown to be important impacts in backcountry settings. In higher density, frontcountry settings, where users ex-

pect others to be present, other impacts may be relatively more important. For example, on rivers in southcentral Alaska that receive considerable use, conflicts and resource competition assumed greater importance than encounters, while on the lower density rivers encounters were relatively more important (Whittaker 1992). Similarly, on the high-density New River, Roggenbuck and others (1991) show that waiting at rapids was the impact for which users were consistently most likely to give a specific norm. This may imply that waiting times for rapids were more relevant than encounters. Alternatively, specifying a norm for a waiting time may have been easier than for number of encounters in a situation where numbers of boats were extremely high and therefore difficult to count. In general, papers summarizing the findings from backcountry studies (e.g., Shelby and Vaske 1991, Vaske and others 1986, Vaske and others 1993) consistently make two general observations. First, norms for encounters during a backcountry experience tend to be quite low (about four or fewer encounters per day; Shelby and others 1996), compared to frontcountry settings where the tolerance limits can exceed 100 encounters (Vaske and others 1996). Second, the percent of people who can specify an encounter norm for wilderness experiences is fairly high. For example, 84% of Grand Canyon respondents, 90% of Rogue River visitors, and 90% of Illinois River boaters were able to specify an encounter norm when asked (Shelby 1981). In frontcountry settings (Manning and others 1996, Roggenbuck and others 1991, Vaske and others 1996), the available data suggest more variation in respondents’ ability or willingness to specify a numeric estimate for acceptable encounters with others. In a crosscultural comparison of visitors to the Columbia Icefield in Jasper National Park (Vaske and others 1996), for example, the percent of visitors giving a norm (i.e., specifying a number) ranged from a low of 50% for American tourists to a high of 90% for British visitors. Findings from frontcountry settings suggest that compared to backcountry areas, where most visitors consistently give a norm, the willingness of frontcountry visitors to provide a numerical tolerance limit is more variable and generally lower (Vaske and others 1996). This probably occurs because in frontcountry situations, individuals expect and can tolerate the presence of other visitors. We therefore hypothesize that norm prevalence will be greater in backcountry when compared to frontcountry resources (H1 ).

Type of Activity 4We

do not address other social norm characteristics such as range of tolerable contacts or agreement (see Vaske and others 1986, Shelby and others 1996 for reviews of these concepts).

A review paper by Vaske and others (1982) theorized that participants in consumptive activities (e.g., hunters

Norm Prevalence: Comparative Analysis

and anglers) differ from those in nonconsumptive activities in terms of the specificity of their recreation goals and their control in achieving those goals. The findings showed that those in consumptive activities reported significantly lower satisfaction scores. Other comparative analyses (Shelby and others 1989), however, found no statistical differences between consumptive and nonconsumptive users in terms of crowding. The difference in findings between the satisfaction (Vaske and others 1982) and crowding (Shelby and others 1989) comparative analyses can be partially explained in terms of the degree of measurement correspondence among the concepts. Social psychologists have repeatedly noted that for a relationship to be observed, the two concepts must be measured at the same level of specificity (Eagly and Chaiken 1993). When there is a direct correspondence between the two measures (general to general or specific to specific), a relatively strong correlation will be observed. When there is no correspondence, the magnitude of the relationship declines. In the satisfaction article (Vaske and others 1982), both activity type (consumptive versus nonconsumptive) and satisfaction (overall evaluation) were measured at a general level, and a relationship was observed. In the crowding paper (Shelby and others 1989), type of activity was measured at a general level, while crowding was specific to the number of people encountered, and a relationship was not observed. The lack of a relationship between type of activity (consumptive versus nonconsumptive) and perceived crowding may also be explained by the role encounters play in different activities. For example, some hunters view deer hunting as a solitary experience and thus prefer few encounters (Vaske and others 1993). Others believe that the presence of additional hunters helps move deer, thereby increasing the likelihood of seeing game. For this group, a large number of encounters may be preferable. The same logic applies to nonconsumptive recreation. Individuals who are motivated by solitude may find the presence of others disruptive. Alternatively, large numbers of people on a beach are often expected and may enhance the quality of the social experience. Given this diversity of desired experiences within both consumptive and nonconsumptive activities, and the lack of measurement correspondence between the concepts, the role of activity type (a general concept) in determining encounter norm prevalence (a specific concept) may be similar to the relationship between encounters and crowding, where no association has been observed (Shelby and others 1989). Thus, we hypothesize that there will be no relationship between norm prevalence and type of activity (consumptive and nonconsumptive) (H2 ).

405

Type of Encounter It is now well documented that respondents’ evaluations of others vary by the type of encounter. Most research on recreational conflict, for example, has focused on the asymmetrical relationships between different activity groups (Kuss and others 1990). Several studies conducted in the Boundary Waters Canoe Area have shown the presence of a one-way conflict between paddling canoeists and motorboaters (Lucas 1964a,b, Lime 1975, Adelman and others 1982). Paddling canoeists disliked motorboaters, but the people using motorpowered craft were not bothered by, and often enjoyed seeing, paddlers. This one-way type of conflict has also been shown between hikers and trailbikers (McCay and Moeller 1976, Watson and others 1991), oar-powered and motor-powered whitewater rafters (Nielsen and Shelby 1977, Schreyer and Nielsen 1978, Shelby 1980, Whittaker and others 1990), cross-country skiers and snowmobilers (Knopp and Tyger 1973, Jackson and Wong 1982), backpackers and horsepackers (Stankey 1973, Watson and others 1994), water skiers and anglers (Gramann and Burdge 1981), and hunters and wildlife viewers (Vaske and others 1995b). In general, these studies show that for those recreationists for whom the encounter has negative consequences (e.g., disrupts the solitude of the experience, or inhibits one’s ability to catch fish or hunt game), conflict increases. This implies that encounters are more important for these recreationists. Thus, we hypothesize that norm prevalence will be greater in evaluation contexts where the encounter represents a potential conflict, compared to those where the encounter makes no difference or even enhances the experience (H3 ).

Question Response Format In addition to the experiential variables described above, methodological considerations may influence norm prevalence. The attitude literature, for example, has noted that respondents’ reports of their own attitudes are influenced by a variety of variables beyond the attitudes themselves. In particular, question wording or the context in which questions are asked can systematically influence responses (Schuman and Kalton 1985). Variations based on such factors are called response effects (Krosnick and Schuman 1988). Research on norms in recreation settings has raised similar questions about the effects of question format. A typical question for measuring encounter norms asks respondents to give the highest number they would tolerate, or they can check a category that says ‘‘makes no difference to me.’’ Some investigations (Hall and Shelby 1996, Hall and others 1996, Manning and others

406


1996, Roggenbuck and others 1991, Vaske and others 1995) have included a third response category. Rather than forcing a choice between giving a number or indicating that the number of encounters makes no difference, respondents may check the category ‘‘makes a difference but can’t give a number.’’ Similar to the debate over the inclusion or omission of ‘‘neutral’’ or ‘‘don’t know’’ categories in Likert scales (Dawes and Smith 1985, Gilljam and Granberg 1993), however, the use of two versus three response choices in norms questions has both advantages and disadvantages (Hall and others 1996). The three-choice option provides individuals with a more exhaustive set of response categories, but the findings have less managerial relevance. For example, knowing that a substantial proportion of people care about the number of encounters but cannot specify a number does not provide direction for determining how many is too many; desirable information when attempting to establish carrying capacities and provide high-quality experiences. A recent study compared the two- versus three-choice format (Hall and others 1996). In the two-choice format, kayakers and rafters on the Clackamas River in Oregon were asked to give a norm or to indicate that the impact did not matter to them. In the three-choice format, respondents were allowed to choose either of the above two options or to indicate that they considered the impact important to their experience, but could not assign a specific norm. In the three-choice format, between 16% and 39% chose the intermediate option. The two-choice format resulted in larger percentages who specified a personal norm, but it also produced larger percentages who said a particular impact was not important. Based on these findings, we hypothesize that norm prevalence will be greater when response categories encourage respondents to specify a number (two-category response format) as opposed to checking one of the alternatives (three-category format) (H4 ).

Methods Encounter norms have been examined for a wide variety of activities and settings since the mid-1970s. The activities include rafting, kayaking, motor boating, climbing, sight seeing, wildlife viewing, hunting, and fishing. The areas studied also show diversity, ranging from remote backcountry resources to developed frontcountry settings. The investigators collecting the data included academic and public agency researchers. This diversity of applications provides a foundation for assessing variables that influence norm prevalence across a variety of evaluation contexts. The data for this paper were obtained from secondary analysis of articles, dissertations, theses, and pub-

lished and unpublished reports. Table 1 details the location, sample size, and response rates utilized by these studies. Analyses are based on responses from 9806 individuals. The average response rate across studies was 78%. The study areas were located in seven different US states (Alaska, Arizona, California, Colorado, Oregon, Washington, and West Virginia) and two Canadian provinces (Alberta and British Columbia). In total, 30 different studies, representing 56 norm evaluation contexts are examined. An evaluation context refers to a specific situation where one type of user encounters another. For example, one evaluation context at Gwaii Haanas National Park Reserve in British Columbia involves kayaker norms for meeting motorboaters. A second evaluation context for this same study involves kayaker norms for meeting other kayakers. Evaluation contexts may also represent norms for different types of experiences (e.g., a wilderness as opposed to a social recreation experience). The 56 evaluation contexts in this comparative analysis constitute the basic unit of analysis. Variables Measured Dependent variable. An indicator of norm prevalence was derived from the single item measure that asked respondents to specify the highest number of encounters they would tolerate for a given situation by completing the statement ‘‘OK to have as many as ___ encounters per day.’’ As noted above, some investigations allowed respondents to check a category ‘‘makes no difference to me,’’ while other studies included the third response option ‘‘makes a difference but can’t give a number.’’ Norm prevalence was operationalized as the percent of individuals who gave a number in response to this question (i.e., the percent of respondents in each evaluation context who specified a number). Independent variables. To determine the characteristics affecting norm prevalence, each evaluation context was coded in terms of three experiential variables (type of resource, type of activity, type of encounter) and one methodological variable (question response format). Evaluation contexts were classified as either backcountry or frontcountry resource types based on accessibility of the setting, amount of use, and level of development. The activity type variable grouped hunting and angling as consumptive activities; all other activities (climbing, kayaking, rafting, motorboating, wildlife viewing, and sightseeing) were considered nonconsumptive activities. Type of encounter was operationalized as a twocategory variable based on the existing conflict literature. For example, evaluation contexts involving similar encounters (e.g., kayakers evaluating other kayakers) or


Table 1.

Description of studies

State/ province

Resource

Alaska

Alberta Arizona British Columbia California Colorado

Oregon

Washington West Virginia

Activity studied

Goodnews River Little Susitna River Lower Deshka River Lower Kanektok River Lower Lake Creek Middle Kenai River

Rafting/boating Rafting/boating Rafting/boating Rafting/boating Rafting/angling Angling

Talachulitna River Talkeetna River Canyon Togiak River Upper Deshka River Upper Kanektok River Upper Kenai River

Rafting/boating Rafting Angling Rafting/angling Rafting Angling

Upper Lake Creek Columbia Icefield

Rafting Snocoach visitors Toe of the glacier visitors Rafting Motorboating/kayaking Rafting/angling Climbing Hunting Hunting/wildlife viewing Kayaking/rafting Rafting Rafting Rafting Rafting Kayaking/rafting Rafting Rafting Rafting

Grand Canyon Gwaii Haanas Klamath River Mt. Shasta Colorado Mt. Evans Poudre River Illinois River Rogue River (1977) Rogue River (1984) Rogue River (1991) Clackamas Deschutes River White Salmon New River

those traditionally considered nonconfrontational situations (e.g., motorboaters’ evaluations of kayakers) were labeled ‘‘no conflict encounters.’’ Potential conflict encounters were represented by evaluation contexts such as nonmotorized visitors evaluating motorized recreationists or wildlife viewers evaluating hunters. Finally, response format was defined in terms of the number of options presented to an individual. The three-level response format allowed individuals to give a number, indicate that the number of encounters makes no difference, or check the category ‘‘makes a difference but can’t give a number.’’ The two-level explicit response format included only the first two options listed above for each question. In the two-level implicit format, respondents were also asked to specify a number, but were simply told in the instructions that they could mark an ‘‘X’’ if they could not give a number (i.e., the ‘‘makes no difference’’ option was not repeated for each question). To provide an initial understanding of the data, bivariate relationships between each of the independent variables and the dependent variable are pre-

Citation Whittaker (1996) Whittaker and others (1990) Whittaker and others (1990) Whittaker (1996) Whittaker and others (1990) Alaska Division of Parks and Recreation (1993) Whittaker and others (1990) Whittaker and others (1990) Whittaker (1996) Whittaker and others (1990) Whittaker (1996) Alaska Division of Parks and Recreation (1993) Whittaker and others (1990) Vaske and others (1996) Vaske and others (1996) Shelby (1981) Vaske and others (1995a) Shelby & Stein (1984) Puttkammer (1994) Fulton and others (1995) Vaske and others (1995b) Vaske and Donnelly (1993) Shelby and Colvin (1981) Shelby and Colvin (1979) Johnson and others (1990) Shelby and Shindler (1992) Rolloff and others (1995) Shelby and others (1987) Shelby and Wing (1992) Roggenbuck and others (1991)

407

Sample size

Response rate (%)

84 203 258 106 64 166

87 79 73 79 76 95

104 54 42 258 106 242

71 77 58 73 79 95

27 501 409 434 257 50 310 612 790 1065 263 268 466 253 365 576 857 616

76 97 90 39 67 43 50 97 68 97 92 78 79 98 67 83 95 67

sented using analysis of variance and Pearson correlations. A multivariate regression model was then used to predict norm prevalence based on the three experiential variables (i.e., type of resource, type of activity, type of encounter) and the methodological variable (i.e., response format). For all analyses, a significance level of P ⬍ 0.05 was used.

Results Across all evaluation contexts, the percent of respondents who reported an encounter norm (i.e., norm prevalence) ranged from a high of 97% for the Colorado bow elk hunters’ evaluations of other hunters to a low of 29% for the West Virginia New River rafters’ evaluations of other rafters in a social experience context (Table 2). The average norm prevalence was 70%, the median was 71%, and the standard deviation was 18%. These descriptive statistics suggest that across all evaluation contexts, nearly three quarters of respondents, on average, reported a numerical encounter norm when asked. Table 2 ranks the evaluation contexts


408

Table 2.

Ranking of norm prevalence for different resources and evaluation contexts Evaluation context

Study site

Evaluations by

Evaluations for

% Reporting It matters but It does not matter an encounter cannot give a number (%) to me (%) norma

Colorado Colorado Talkeetna River Canyon Upper Kanektok River Colorado Lower Kanektok River Rogue River (1984) Goodnews River Upper Kanektok Illinois River Talachulitna River Upper Lake Creek Goodnews River Gwaii Haanas Upper Kenai River Grand Canyon Lower Lake Creek Rogue River (1977) Lower Kanektok Talkeetna River at Clear Creek Togiak River Upper Deshka River Middle Kenai River Rogue River (1991) Klamath River Poudre River Deschutes River Lower Deshka River Gwaii Haanas Columbia Icefield Klamath River Columbia Icefield Little Susitna River Middle Kenai River Mt. Evans Deschutes River Deschutes River Mt. Shasta Mt. Evans Poudre River Mt. Evans Poudre River Clackamas Klamath River Mt. Shasta White Salmon Mt. Shasta New River Gwaii Haanas Mt. Evans

Bow elk hunters Bow elk hunters Rafters Floaters Bow elk hunters Floaters Rafters Floaters Floaters Rafters Rafters Rafters Floaters Kayakers Bank anglers Rafters Boaters/rafters/anglers Rafters Floaters

Other hunters Other archers Other rafters Powerboats Recreationists Powerboats Other rafters Powerboats Other float groups Other rafters Other groups Other rafters Other float groups Motorboaters Other anglers at high use times Other rafters Other groups Other rafters Other float groups

97 96 96 94 92 92 90 90 90 90 89 88 87 86 86 84 84 82 82

Boaters/anglers Fly-in anglers Rafters/boaters Bank anglers Rafters Summer rafters Kayakers Rafters Rafters/boaters Motorboaters Snocoach visitors Anglers Toe of the glacier visitors Rafters/boaters Bank anglers Wildlife viewers Rafters Rafters Hikers Hunters Kayakers Hunters Rafters Rafters Anglers Hikers Rafters Hikers Rafters Kayakers On-site visitors

81 80 80 78 78 78 77 73 72 71 71 70 69 68 67 62 62 62 61 60 60 57 56 56 56 54 52 51 50 48 48

New River Mt. Evans Clackamas Poudre River Gwaii Haanas

Rafters Wildlife viewers Rafters Rafters Motorboaters

New River

Rafters

Other groups Other angling groups Other groups Other anglers at low use times Other rafters Other rafters Rafters Other rafters—segment 1 Other groups Motorboaters Other visitors Float groups Other visitors Other groups Other anglers at high use times Hunters Other rafters—segment 2 Other rafters—segment 3 Other hikers while camping Other hunters Other kayakers Wildlife viewers Other rafters Other rafters Other anglers Other hikers while climbing Other rafters Other hikers at summit Other rafters—as wilderness Other kayakers Other visitors Other rafters—as scenic experience Other viewers Other rafters Kayakers Kayakers Other rafters—as social experience

aThe

percent reporting an encounter norm represents norm prevalence.

11

29

28

34 33

34 44

25 47 39

3 4 4 6 8 10 10 10 10 10 11 12 13 3 14 16 16 18 18 19 20 20 22 22 22 23 27 28 0 29 30 31 32 33 10 38 38 39 6 40 10 44 44 44 46 48 49 16 8 52

45 43 39 39 34

41

30 10 23 61 25

29

18

53


Table 3.

409

Aggregate variables affecting norm prevalence Dependent variable: Norm prevalence

Independent variables Experiential variables Type of resource Backcountry Frontcountry Type of Activity Consumptive Nonconsumptive Type of encounter Conflict No conflict Methodological Variable Question response formata 2-category (implicit) 2-category (explicit) 3-category aMeans

Number of evaluation contexts

Mean

Standard deviation

26 30

81.2 60.7

15.9 13.8

11 45

76.3 68.7

15.3 18.4

8 48

82.9 68.1

11.9 18.0

17 27 12

86.8a 67.9b 52.0c

8.3 13.8 16.2

F

P

26.8

.001

1.6

.215

5.0

.030

26.1

.001

followed by different letters differ significantly at P ⬍ 0.05.

from high to low based on the percent of individuals reporting an encounter norm. Several trends emerge when viewed from this perspective. Type of Resource Studies conducted in backcountry settings tend to be grouped in the upper half of the table where norm prevalence is highest. Sixteen of the 28 evaluation contexts with norm prevalence greater than the median (71%) are from studies conducted in Alaska. Data for other evaluation contexts in this group come from locations with known low densities. For example, use on the Rogue River is limited to 120 people per day, and norm prevalence ranged from 78% (1991 data) to 90% (1984 data). Use on the Colorado River in Grand Canyon was limited to 150 people per day during the time of the study, and averaged about 75 per day during the season (norm prevalence ⫽ 84%). Use on the Illinois River is not recorded by the managing agency, but during the time of the study, averaged less than 10 people per day (norm prevalence ⫽ 90%). By comparison, evaluation contexts with norm prevalence scores lower than the median are more often higher density, frontcountry settings. The New River, for example, is a day-use area with use levels averaging over 1000 persons per day on summer weekends. Norm prevalence for the New River ranged from 29% when the setting was evaluated as a social experience to 50% when it was evaluated as wilderness. Similarly, Mt. Evans is located 70 miles west of Denver and attracts approximately a quarter of a million people each year. Depending on the evaluation context, norm prevalence at Mt. Evans ranged from 43% (wildlife viewers’ evaluations of

other viewers) to 62% (wildlife viewers evaluations of hunters). As predicted by hypothesis 1, norm prevalence was higher in backcountry (mean ⫽ 81.2%, SD ⫽ 15.9) compared to frontcountry (60.7% ⫾ 13.8%) evaluation contexts (Table 3). This difference was statistically significant, F(1, 54) ⫽ 26.8, P ⬍ 0.001. Type of Activity Hypothesis 2 predicted there would be no relationship between type of activity and norm prevalence. Support for this prediction is evident in Table 2, where 22 of the nonconsumptive evaluation contexts were above the median, and 21 were below. For the consumptive activities, eight fell above the median and five below. As suggested by hypothesis 2, no statistical difference was observed between consumptive (76.3% ⫾ 15.3%) and nonconsumptive (68.7% ⫾ 18.4%) activities, F(1, 54) ⫽ 1.6, P ⫽ 0.215 (Table 3). Type of Encounter Type of encounter influenced the percent reporting an encounter norm. In Gwaii Haanas National Park Reserve in British Columbia, for example, 86% of the kayakers gave a norm for meeting motorboaters, while only 34% of the motorboaters specified a norm for kayakers (Table 2). On the Poudre River, a popular day-use frontcountry setting in Colorado, 77% of the kayakers reported a norm for meeting rafters, while 39% of the rafters had a norm for kayakers. Hypothesis 3 predicted that, in conflict situations, respondents would be more likely to report an encounter norm, and the data supported this prediction (Table

410


3). Norm prevalence was higher in the conflict evaluation contexts (82.9% ⫾ 11.9%) when compared to the no-conflict contexts (68.1% ⫾ 18.0%), F(1, 54) ⫽ 5.0, P ⫽ 0.030. Question Response Format The response format used by respondents influenced norm prevalence. For example, the Colorado bow elk hunter surveys as well as the studies conducted in Alaska all used the two-category implicit response option, and norm prevalence in all of these studies was 78% or higher. In contrast, all but one of the studies using the three-category response format reported norm prevalence levels of 71% or lower (Table 2). Norm prevalence varied statistically with question response format [F(2, 53) ⫽ 26.1, P ⬍ 0.001]. As predicted by hypothesis 4, the average norm prevalence was highest for the two-category implicit format (86.8% ⫾ 8.3%), and lowest for the three-category response format (52.0% ⫾ 16.2%). In evaluation contexts where response categories encouraged respondents to specify a number, more people were likely to do so (Table 3). Predicting Norm Prevalence In summary, the data supported the four hypotheses. As anticipated, norm prevalence varied by two of the experiential variables—type of resource (backcountry versus frontcountry) and type of encounter (no conflict versus conflict), and the methodological variable— question response format (two-category implicit, twocategory explicit, and three-category). In addition, there was no relationship between norm prevalence and activity type (consumptive versus nonconsumptive). The relative importance of these predictor variables in explaining norm prevalence is examined in Table 4. In terms of the bivariate correlations, type of resource (r ⫽ ⫺0.58, P ⬍ 0.001), type of encounter (r ⫽ 0.29, P ⫽ 0.030), and question response format (r ⫽ ⫺0.70, P ⬍ 0.001) were significantly related to norm prevalence. Consistent with these findings and the results in Table 3, the regression model indicated that type of resource (␤ ⫽ ⫺0.36, P ⬍ 0.001), type of encounter (␤ ⫽ 0.19, p ⫽ 0.023), and question response format (␤ ⫽ ⫺0.56, P ⬍ 0.001) significantly influenced norm prevalence.5 Taken together, these three variables ex5As

a check on multicollinearity, the zero-order correlations between each pair of independent variables were also calculated. The correlations ranged from ⫺0.009 to ⫺0.449, with an average correlation of 0.206; thus, collinearity was not considered to be a problem. In addition, a saturated model indicated that none of the higher-order interactions among the independent variables significantly influenced norm prevalence.

Table 4.

Predicting norm prevalence Dependent variable: Norm prevalencea

Independent variablesb resourcec

Type of Type of activityd Type of encountere Question response formatf Adjusted R 2 aF(3,

P

Standardized regression coefficient

P

⫺0.58 ⫺0.17 0.29

0.001 0.215 0.030

⫺0.36 n.s. 0.19

0.001 n.s. 0.023

⫺0.70

0.001

⫺0.56 0.64

0.001

Zero-order correlation

52) ⫽ 33.7, P ⬍ 0.001.

bNone

of the higher order interactions among the independent variables significantly influenced norm prevalence. cVariable

coded as 0 ⫽ backcountry, 1 ⫽ frontcountry.

dVariable

coded as 0 ⫽ consumptive, 1 ⫽ nonconsumptive.

eVariable

coded as 0 ⫽ no conflict, 1 ⫽ conflict.

fVariable coded as 1 ⫽ 2-category implicit, 2 ⫽ 2-category explicit, 3 ⫽ 3-category.

plained 64% of the variance in norm prevalence. Type of activity was not significant in either the zero-order correlations or the regression model. The regression model in Table 4 describes the relative influence of variables on norm prevalence. The equation, however, can also be used to predict, prior to data collection, the degree of norm prevalence likely to be observed in a given study. An estimate of norm prevalence can be obtained by using the three significant predictor variables, the unstandardized regression coefficients, and the associated code values (Table 5). As illustrated by the first example in Table 5, if a researcher uses the two-category implicit response format in a backcountry setting where no conflict is evident, norm prevalence is predicted to be 89%, with a 95% confidence interval of ⫾10.77%. In an identical setting (i.e., backcountry, no conflict), use of the three-category response format yields a predicted norm prevalence of only 60% (⫾10.77%). In such a situation, use of the three-category response format obviously has some trade-offs. Example 3 in Table 5 illustrates the predictive model when using the three-category response format in a frontcountry area where conflict exists. Under this scenario, the predicted norm prevalence is 57%. By simply modifying the response format from three to two categories (implicit), norm prevalence increases to 85% (example 4). These findings highlight the extent to which norm prevalence can be influenced by question response format. The decision criteria researchers and managers should consider when choosing one response format over another are discussed below.


Table 5.

411

Three variable regression model predicting norm prevalencea Example 1

Independent variable Constant Question response format

Unstandardized regression coefficient 102.66 ⫺14.08

Example 2

Coefficient⫻ Code values

1 1 ⫽ 2 category implicit 2 ⫽ 2 category explicit 3 ⫽ 3 category Type of resource ⫺12.91 0 ⫽ Backcountry 1 ⫽ Frontcountry Type of encounter 9.71 0 ⫽ No conflict 1 ⫽ Potential conflict Dependent variable ⫽ predicted norm prevalence (%) Minimum norm prevalence (predicted ⫺ standard error) (%) Maximum norm prevalence (predicted ⫹ standard error) (%)

Example 3

Example 4

Study value

study value

Study value

study value

Study value

study value

Study value

Coefficient⫻ study value

1 1

102.66 ⫺14.08

1 3

102.66 ⫺42.24

1 3

102.66 ⫺42.24

1 1

102.66 ⫺14.08

0

0

0

0

1

⫺12.91

1

⫺12.91

0

0

0

0

1

9.71

1

9.71

89 78 99

Coefficient⫻

60 50 71

Coefficient⫻

57 46 68

85 75 96

aR 2 ⫽ 0.64. Standard error of estimate ⫽ 10.77. The examples depict the use of the: 1. 2-category implicit response format in a backcountry setting where the types of encounters are not in conflict. 2. 3-category response format in a backcountry setting where the types of encounters are not in conflict. 3. 3-category response format in a frontcountry setting where the types of encounters are conflicting. 4. 2-category implicit response format in a frontcountry setting where the types of encounters are conflicting.

Discussion The current debate regarding norm existence can be traced to the findings from one investigation that found a relatively low percentage of individuals willing to specify an encounter norm (Roggenbuck and others 1991). To place this debate on a more empirical basis, this comparative analysis summarized studies that have examined the prevalence of encounter norms and evaluated variables influencing norm prevalence. Across the 56 evaluation contexts examined here, norm prevalence averaged 70%. This finding indicates that, on average, nearly three quarters of all respondents are willing to provide a numerical response when asked. Interestingly, the study that generated the debate (Roggenbuck and others 1991) ranked last in terms of norm prevalence among the 56 evaluation contexts examined. While the norm prevalence percentages shown here indicate that most respondents are capable of specifying a norm, it is equally clear that norm prevalence varies considerably depending on the type of resource, the type of encounter, and the question response format. Type of Resource Consistent with previous speculations (Shelby and Vaske 1991, Shelby and others 1996, Vaske and others 1996), backcountry visitors were more likely to specify an encounter norm when compared to frontcountry respondents. In part, these findings can be attributed to the difficulty of the task and the importance of the impact. As the number of visitors to an area increases, the task of stating a precise number is more difficult (Shelby and others 1987). In addition, in situations where visitors expect numerous others to be present,

the relative importance of encounters decreases (Whittaker 1992). These observations suggest that researchers should work with managers to identify which impacts are important to the setting being studied and measure not only encounter norms, but other normative issues as well (e.g., waiting time to use or gain access to facilities, safety norms). Type of Activity As predicted, type of activity (consumptive versus nonconsumptive) had no influence on norm prevalence. Such findings may suggest the consumptive/ nonconsumptive categorization of activities is too broad to differentiate participants. For example, within the consumptive category, a variety of both hunter and angler types can be identified. While encounters may be important for the experienced fly angler who is seeking solitude, the presence of others may be expected and tolerated for opening-day angling experiences on popular rivers. Similarly, Manfredo and Larson (1993) identified four types of wildlife viewing experiences, ranging from high involvement to occasional participation, differences that are masked by the nonconsumptive categorization. This suggests that, relative to norm prevalence, there may be an interaction effect between specific types of activities and level of specialization within that activity. While the data available for this analysis did not allow examination of specialization, the topic should be explored in future investigations. From a more general perspective, the findings here as well as those reported in other analyses of crowding (Shelby and others 1989) and satisfaction (Vaske and others 1982) can be explained in terms of the degree of measurement correspondence among the concepts. In

412


the satisfaction comparative analysis (Vaske and others 1982), both activity type (consumptive versus nonconsumptive) and satisfaction (overall evaluation) were measured at a general level and a relationship was observed. In both the crowding paper (Shelby and others 1989) and the norm prevalence analysis described here, type of activity was measured at a general level, and crowding/norm prevalence were specific measures. For these latter two analyses no relationship was observed. Although the sample sizes available for this analysis did not allow for examination of differences among more specific categorizations of activities, the topic should be explored in future research. For example, differences may be apparent between traditional activities and more recently popular activities. Type of Encounter Consistent with past research, our findings highlight the existence of asymmetrical relationships among activities. In conflict situations, encounter norms were more prevalent. The traditional conflict model (Jacob and Schreyer 1980) suggests that conflicts occur when the physical presence of an individual or group interferes with the goals of another individual or group. This paper suggests the need to broaden the focus of conflict research. Conflicts can also arise between groups who do not share the same norms (Ruddell and Gramann 1994) and/or values (Vaske and others 1995b). These situations can be labeled either ‘‘norm conflicts’’ or ‘‘value conflicts’’ (Williams 1993). Hunters versus wildlife viewers represents an example of conflicts in norms/ values. For example, Vaske and others (1995b) show that to the extent conflict exists on Mt. Evans in Colorado, much of the problem stems from differences in social values held by the hunting and nonhunting publics. More research is needed to examine the relative importance of goal interference versus norms/ values in influencing norm prevalence. Question Response Format The data reported here show clear response effects for the two- versus three-category response options. As noted earlier, both formats have advantages and disadvantages (Hall and others 1996). The three-category option offers a more complete response set, but the two-category approach may provide more managerially relevant information because more respondents are likely to give a number (norm) that can be directly translated into a management standard. Encouraging individuals to be specific may be acceptable if respondents can reasonably be expected to make the evaluation (e.g., encounters in backcountry settings). The regression model presented in Table 5 can assist in this

judgment call by allowing researchers and managers to assess the trade-offs between different response formats given the situation being studied. Researchers have used other strategies for presenting normative options that involve some form of multiple-choice format where respondents do not have to come up with a number on their own. For example, several investigations in high-density resources have asked about an acceptable time in sight of others (Hall and others 1996, Whittaker and Shelby 1988). Rather than specifying a number, respondents circled a number on a scale (see Donnelly and others 1992, for a review of other measurement options). In the Hall and others (1996) study, time in sight of other boaters (scale ⫽ circle a number) showed the highest percentage giving a norm, while number of encounters with others (scale ⫽ respondent specifies a number) showed fewer giving a norm. Other researchers have used a visual approach to setting encounter norm standards in both frontcountry (Manning and others 1996) and backcountry (Vaske and others 1995a) settings. In a study at Arches National Park, Utah (Manning and others 1996), for example, respondents were asked to rate the acceptability of a series of photographs depicting different numbers of people at Delicate Arch. The findings suggest that pictures showing varying numbers of people helped respondents evaluate encounters in a high-density situation. Moreover, this response format avoids the problem of forcing individuals to identify a discrete number of acceptable encounters. All of the above, however, are complex measurement issues that warrant careful consideration in future work.

Conclusions At the individual level, one can argue that a norm exists when a person can specify a value for his/her personal norm. At the aggregate level, however, the question of norm existence is less clear. If prevalence is 100%, it would seem a social norm does exist; if prevalence is 0%, a social norm does not exist. All of the evaluation contexts examined here, however, fell somewhere between these two extremes. Rather than becoming mired in decision rules about the existence of social norms, we think the important goal is to understand norms that are relevant in the specific evaluation context under investigation. Understanding social norms has typically involved examination of four characteristics: the optimal situation, tolerable range, amount of agreement, and intensity of the norm. This paper suggests the importance of an additional characteristic, norm prevalence. Prevalence


413

Jacob, G. R., and R. Schreyer. 1980. Conflict in outdoor recreation: A theoretical perspective. Journal of Leisure Research 12:368–380.

should be viewed as an initial step when evaluating norm data. When prevalence is low, either there are measurement problems or the impact in question is not relevant in that particular evaluation context. In either case, judgments regarding the other characteristics of social norms should be made cautiously. When prevalence is high, researchers and managers can be more confident in assessing other norm characteristics and using the data to set standards for what is acceptable. By specifying the conditions under which people will give an encounter norm, this paper helps to clarify one component of the normative debate. Future analyses concerned with other conceptual and methodological issues surrounding the debate would further enhance our understanding of norms.

Kuentzel, W. 1990. Motive uniformity across recreation settings: A meta-analysis of REP scales. Paper presented at the annual National Recreation and Parks Association meeting, Phoenix, Arizona.

Literature Cited

Kuss, F. R., A. R. Graefe, and J. J. Vaske. 1990. Visitor impact management: A review of research. National Parks and Conservation Association, Washington, DC, 256 pp.

Adelman, B. J., T. A. Heberlein, and T. M. Bonnicksen. 1982. Social psychological explanations for the persistence of a conflict between paddling canoeists and motorcraft users in the Boundary Waters Canoe Area. Leisure Sciences 5:45–62. Alaska Division of Parks and Recreation. 1993. Kenai River carrying capacity study. Soldotna, Alaska. Dawes, R. M., and T. Smith. 1985. Attitude and opinion measurement. Pages 509–566 in G. Lindzey and E. Aronson (eds.), Handbook of social psychology, Vol. 2. Random House, New York. Donnelly, M. P., J. J. Vaske, and B. Shelby. 1992. Measuring backcountry standards in visitor surveys. Pages 38–52. in B. Shelby, G. Stankey, and B. Shindler (Eds.), Defining wilderness quality: The role of standards in wilderness management—a workshop proceedings. General Technical Report PNW-GTR-305. USDA Forest Service, Portland, Oregon. Eagly, A. H., and S. Chaiken. 1993. The psychology of attitudes. Harcourt, Brace, Jovanovich, New York. Fulton, D. C., M. J. Manfredo, J. J. Vaske, L. Johnson, J. George, and R. Kahn. 1995. Crowding and satisfaction among Colorado elk bowhunters. HDNRU Report No. 5. Colorado State University, Fort Collins. Gilljam, M., and D. Granberg. 1993. Should we take Don’t Know for an answer? Public Opinion Quarterly 57:348–357. Gramann, J. H., and R. J. Burdge. 1981. The effects of recreational goals on conflict perception: The case of water skiers and fishermen. Journal of Leisure Research 13:15–27. Hall, T., and B. Shelby. 1996. Who cares about encounters? Differences between those with and without norms. Leisure Sciences 18:7–22. Hall, T., B. Shelby, and D. Rolloff. 1996. Effect of varied question format on boaters’ norms. Leisure Sciences 18:193– 204. Homans, G. C. 1950. The human group. Brace and World, New York. Jackson, E. L., and R. Wong. 1982. Perceived conflict between urban cross-country skiers and snowmobilers in Alberta. Journal of Leisure Research 14:47–62.

Johnson, R. L., B. Shelby, and N. S. Bregenzer. 1990. Economic values and product shift on the Rogue River: A study of non-commercial whitewater recreation. Water Resources Institute, Corvallis, Oregon. Knopp, T. B., and J. D. Tyger. 1973. A study of conflict in recreational land use: Snowmobiling versus ski touring. Journal of Leisure Research 5:6–17. Krosnick, J. A., and H. Schuman. 1988. Attitude intensity, importance and certainty and susceptibility to response effects. Journal of Personality and Social Psychology 54:940–952.

Lime, D. W. 1975. Backcountry river recreation: problems and opportunities. Naturalist 26:1–6. Lucas, R. C. 1964a. The recreational capacity of the QueticoSuperior Area. Res. Pap. LS-15. US Department of Agriculture, Forest Service, Lake States Forest Experiment Station, St. Paul, Minnesota. Lucas, R. C. 1964b. Wilderness perception and use: The example of the Boundary Waters Canoe Area. Natural Resources Journal 3:394–411. Manfredo, M. J., and R. Larson. 1993. Managing for wildlife viewing recreation experiences: A case study in Colorado. Wildlife Society Bulletin 21:226–236. Manfredo, M. J., B. L. Driver, and M. A. Tarrant. 1996. Measuring leisure motivation: A meta-analysis of the recreation experience preference scales. Journal of Leisure Research 28:188–213. Manning, R. E., D. W. Lime, W. A. Freimund, and D. G. Pitt. 1996. Crowding norms at frontcountry sites: A visual approach to setting standards of quality. Leisure Sciences 18:39– 59. McCay, R., and G. Moeller. 1976. Compatibility of Ohio trail users. Res. Note NE-225. US Department of Agriculture, Forest Service, Northeastern Forest Experiment Station, Upper Darby, Pennsylvania. McDonald, C. D. 1996. Normative perspectives on outdoor recreation behavior: Introductory comments. Leisure Sciences 18:1–6. Nielsen, J. M., and B. Shelby. 1977. River running in the Grand Canyon: How much and what kind of use. Pages 168–177 in Proceedings: River Recreation Management and Research Symposium. Gen. Tech. Rep. NC-28. USDA Forest Service, St. Paul, Minnesota. Puttkamer, A. 1994. A managerial and theoretical approach to the management of the Mt. Shasta Wilderness. Masters thesis. Colorado State University, Fort Collins. Roggenbuck, J. W., D. R. Williams, S. P. Bange, and D. J. Dean. 1991. River float trip encounter norms: Questioning the use of the social norms concept. Journal of Leisure Research 23:133–153.

414


Rolloff, D. B., T. E. Hall, and B. Shelby. 1995. Whitewater recreation on the Clackamus River. USFS Pacific Northwest Experiment Research Paper. USDA Forest Service, Portland, Oregon. Ruddell, E. J., and J. H. Gramann. 1994. Goal orientation, norms, and noise-induced conflict among recreation area users. Leisure Sciences 16:93–104. Schreyer, R., and M. L. Nielsen. 1978. Westwater and Desolation Canyons: Whitewater river recreation. Institute for the Study of Outdoor Recreation and Tourism. Utah State University, Logan. Schuman, H., and G. Kalton. 1985. Survey methods. Pages 635–698 in G. Lindzey and E. Aronson (eds.), The handbook of social psychology, 3rd ed. Random House, New York. Shelby, B. 1980. Contrasting recreational experiences: Motors and oars in the Grand Canyon. Journal of Soil and Water Conservation 35:129–131. Shelby, B. 1981. Encounter norms in backcountry settings: Studies of three rivers. Journal of Leisure Research 13:129–138. Shelby, B., and R. B. Colvin. 1979. Determining use levels for the Rogue River. Report WRRI-63. Water Resources Research Institute, Oregon State University, Corvallis, Oregon. Shelby, B., and R. B. Colvin. 1981. Carrying capacity for the Illinois River. Report WRRI-72. Water Resources Research Institute, Oregon State University, Corvallis, Oregon. Shelby, B., and B. Shindler. 1992. Rogue River, user study: Assessing displacement and product shift on the Wild Rogue. USFS Pacific Northwest Experiment Research Paper. USDA Forest Service, Portland, Oregon. Shelby, B., and K. Stein. 1984. Recreational use and carrying capacity of the Klamath River. Report WRRI-92. Water Resources Research Institute, Oregon State University, Corvallis, Oregon. Shelby, B., and J. J. Vaske. 1991. Using normative data to develop evaluative standards for resource management: A comment on three recent papers. Journal of Leisure Research 23:173–187. Shelby, B., and M. Wing. 1992. Impacts of whitewater boating on the White Salmon River. USFS Pacific Northwest Experiment Research Paper. USDA Forest Service, Portland, Oregon. Shelby, B., D. Whittaker, R. Speaker, and E. E. Starkey. 1987. Social and ecological impacts of recreation use on the Deschutes River Scenic Waterway. Report to the Oregon Legislature. Oregon State University, Corvallis, Oregon. Shelby, B. B., J. J. Vaske, and T. A. Heberlein. 1989. Comparative analysis of crowding in multiple locations: Results from fifteen years of research. Leisure Sciences 11:269–291.

of Agriculture, Forest Service, Intermountain Research Station, Ogden, Utah. Vaske, J. J., and M. P. Donnelly. 1993. Poudre River recreation study. Unpublished manuscript. Colorado State University, Fort Collins. Vaske, J. J., M. P. Donnelly, T. A. Heberlein, and B. Shelby. 1982. Differences in reported satisfaction ratings by consumptive and nonconsumptive recreationists. Journal of Leisure Research 14:195–206. Vaske, J. J., B. Shelby, A. R. Graefe, and T. A. Heberlein. 1986. Backcountry encounter norms: Theory, method and empirical evidence. Journal of Leisure Research 18:137–153. Vaske, J. J., M. P. Donnelly, and B. Shelby. 1993. Establishing management standards: Selected examples of the normative approach. Environmental Management 17:629–643. Vaske, J. J., M. P. Donnelly, W. A. Freimund, and T. Miller. 1995a. The 1995 Gwaii Haanas visitor survey. HDNRU Rep. No. 24. Colorado State University, Human Dimensions in Natural Resources Unit, Fort Collins. Vaske, J. J., M. P. Donnelly, K. Wittmann, and S. Laidlaw. 1995b. Interpersonal versus social-values conflict. Leisure Sciences 17:205–222. Vaske, J. J., M. P. Donnelly, and J. P. Petruzzi. 1996. Country of origin, encounter norms and crowding in a frontcountry setting. Leisure Sciences 18:161–176. Watson, A. E., D. R. Williams, and J. J. Daigle. 1991. Sources of conflict between hikers and mountain bike riders in the Rattlesnake NRA. Journal of Park and Recreation Administration 9:59–71. Watson, A. E., M. J. Niccolucci, and D. R. Williams. 1994. The nature of conflict between hikers and recreational stock users in the John Muir Wilderness. Journal of Leisure Research 26:372–385. Whittaker, D. 1992. Selecting indicators: Which impacts matter more? Pages 13–22 in B. Shelby, G. Stankey, and B. Shindler (eds.), Defining wilderness quality: The role of standards in wilderness management—a workshop proceedings. General Technical Report PNW-GTR-305. USDA Forest Service, Portland, Oregon. Whittaker, D. 1996. Kanektok, Goodnews and Togiak rivers: User survey findings and implications. Report prepared for Togiak National Wildlife Refuge, Dillingham, Alaska. Whittaker, D., and B. Shelby. 1988. Types of norms for recreation impacts: Extending the social norms concept. Journal of Leisure Research 20:261–273.

Shelby, B. B., J. J. Vaske, and M. P. Donnelly. 1996. Norms, standards and natural resources. Leisure Sciences 18:103–123.

Whittaker, D., L. Anderson, and J. Mosby. 1990. Susitna Basin recreation user survey results. Report to Alaska Department of Natural Resources, Anchorage, Alaska.

Stankey, G. H. 1973. Visitor perception of wilderness recreation carrying capacity. Res. Pap. INT-142. US Department

Williams, D. R. 1993. Conflict in the great outdoors. Parks and Recreation September: 28–34, 122.