What leads to better visitor outcomes in live interpretation?

Stern, Marc J. ; Powell, Robert B.

Introduction

Live interpretive programs can serve multiple purposes (Ham, 2013). These include enhancing the experiences and the enjoyment of visitors to special places (Moscardo, 1999; Stern et al., 2011), increasing visitors' knowledge and understanding of natural and cultural resources and places (Ham, 1992; Tilden, 1957), fostering a sense of appreciation or other attitudes toward those resources (Powell et al., 2009), and promoting stewardship behaviors, both on-site and after visitors leave the site of the interpretation (Ham, 2009).

While volumes have been published outlining what might be considered best practices for producing such outcomes, a recent review of the empirical literature suggests that the linkage between these best practices and visitor outcomes have only circumstantial support, despite strong theoretical grounding (Skibins et al., 2012). This is largely due to a lack of comparative studies, which can empirically isolate which practices are the ones most likely causing desired outcomes. Most research studies have evaluated the outcomes of single programs rather than mixtures of programs with varying characteristics. While findings of positive outcomes across multiple studies suggest the broad efficacy of interpretation in general, no study has yet isolated the influence of different interpretive practices and approaches upon visitor outcomes.

This study aims to close this gap in the literature through a comparative study of live interpretive programs across the National Park Service (NPS), by identifying which practices and approaches most consistently lead to more positive outcomes, including visitor satisfaction, enhancement of visitor experience and appreciation of the park unit and its resources, and intentions to change behaviors resulting from program attendance.

Hypothesized best practices for interpretation

Skibins et al. (2012) identified consensus-based best practices of the field in a recent review article. Many of these practices stem from Freeman Tilden's (1957) original six principles first identified in 1957. The principles generally highlight the importance of making communication relevant to the audience; of telling holistic stories; of practicing the art of revelation based on information rather than information dissemination; of provoking the audience to want to do something, whether it be to reflect more deeply, learn more, or act upon new information; and of tailoring interpretation to different audiences. Many others have expanded upon those original best practices to provide insights into how to best craft stories; how to organize content; how to make interpretation relevant, engaging, and entertaining; and how to achieve particular outcomes (see Skibins et al., 2012, for a summary of this work). We drew upon this broad body of literature to develop many of the key program characteristics of interest in this study (see Table 3 for full list).

The role of the interpreter

In addition to characteristics of programs, the characteristics of the interpreters and their delivery styles also likely influence program outcomes. Passion on behalf of the interpreter, for example, has long been recognized as an important element of successful interpretive programs (e.g., Beck and Cable, 2002; Ham & Weiler 2002; Ward & Wilkinson, 2006). We supplement this concept with additional theories from education and communication to further explore the impact of the interpreter on visitor outcomes in addition to the content and format of the program.

The concepts of immediacy, credibility, and clarity have been studied extensively in the communications and education fields (Finn et al., 2009). Immediacy behaviors are those that tend to enhance the familiarity and reduce psychological distance between the communicator and his or her audience (Mehrabian, 1969). Such behaviors might include friendly physical gestures, small talk, calling people by name, or the sharing of personal information (Myers et al., 1998). These behaviors may also be related to "affinity-seeking," or the process through which communicators attempt to get listeners to like them (McCroskey et al., 1986). Studies suggest that such behaviors can enhance the openness of audiences (most studies involve students and their teachers) to the content of lessons (Finn et al., 2009). Others have also assumed that general likeability may be an important factor in audience response (Ward & Wilkinson, 2006).

Credibility refers to audience members' perceptions of the believability or legitimacy of the communicator. Credibility has been found to be important in predicting the responses of message recipients in multiple fields (e.g., Ajzen 1992; Rogers 1995; Stern 2008). Within the education and communications fields, Finn and others (2009) suggest that this credibility is composed of three dimensions: competence, trustworthiness, and caring. Competence can be related to the apparent knowledge, confidence, and eloquence of the communicator. Trustworthiness can be based on multiple factors, including the interpreter's appearance, performance, degree of comfort and/or authority, title or position, and/or personal interactions with the audience. Caring is primarily related to the sincerity with which the interpreter communicates as well as his or her interactions with the audience.

Clarity is not only related to eloquence, but also to the consistency, or "fidelity" of the communicative experience (Chesebro & Wanzer, 2006). Finn and others' review (2009) found that lessons taught with any combination of these characteristics (clarity, credibility, and immediacy) tend to be more effective for learners than those exhibiting only one of them.

Interpreters also have the ability to assume particular roles as communicators. These range from friend to authority figure to the "walking encyclopedia" that Enos Mills warned future nature guides against becoming nearly 100 years ago (Mills, 1920). Each of these identities may be differentially appropriate in different situations and with different audiences (Wallace & Gaudry, 2005). Other items of interest include any apparent bias, misinformation, or false assumptions about the audience made by the interpreter, which could detrimentally influence audience responses.

Interpreters' planning processes and psychological states might also influence the quality of their programs (see Stern at al., this issue). As noted above, interpretation can be used for many purposes, ranging from teaching to entertainment to persuasion. Interpreters' intentions may drive, at least to some extent, audience responses to their programs (Ham, 2013).

Methods

Selection of sites

We aimed to select park units that reflected the diversity of locations, types, and resources of the U.S. NPS system. Criteria for selecting park units for the study included annual visitation numbers, park location (region of the country and distance from population centers), programming focus, number of programs offered to the public, and willingness to participate in the study. In order to ensure adequate visitor attendance at interpretive programs, we only considered parks that received at least 35,000 annual recreation visits. Parks were categorized as urban, urban-proximate, or remote based on their proximity to metropolitan centers. Metropolitan areas were defined as having an urban core of at least 50,000 residents. Urban parks were located within the limits of these metropolitan areas. Urban-proximate parks were located outside these cores, but within a 60-mile radius of these areas. As such, they were typically in rural or suburban areas. Remote parks were located at least 60 miles from any metropolitan area. Parks were placed into one of three categories based on their primary resource base: predominantly cultural, predominantly natural, or a mix of the two. We aimed to have our selection of units mirror the makeup of the NPS system and also allow us to observe at least 10 programs in each park (or within nearby clusters of parks in cases such as Aztec Ruins and Navajo National Monuments) in five days or less. Twenty-four park units were selected for inclusion in the study (Table 1).

We observed programs in 14 predominantly culturally focused park units, seven predominantly nature-focused park units, and three park units with a mixed focus. This roughly mirrors the distribution of these different types of park units throughout the NPS, where roughly 30% of park units are predominantly nature-focused and roughly 60% are predominantly culturally focused. (1) We visited 11 remote park units, five urban-proximate parks, and eight urban park units. This variability provides a reasonable sample from which to make generalizations to the broader population of live interpretive programs across the NPS. Park units were organized for logistical purposes by geographic region into six clusters. Teams of two researchers collected data from each park unit. One team of researchers sampled Great Smoky Mountains National Park and the mid-Atlantic, Washington D.C., and California locations. The other team sampled the Southwest, Midwest, and South Dakota locations.

Sampling and data collection

Individual live interpretive programs served as the unit of analysis for this study. Programs were selected within each park based on variability (with regard to subject matter--natural vs. cultural--and types of delivery--guided walks vs. campfire programs vs. hands-on activities, etc.) and their time and location to maximize the number of programs observed at each park unit. Regular programs were selected over children's programs whenever possible, as adult respondents were the targets of visitor surveys. We attempted to attend 488 scheduled programs, of which only 376 occurred. From these 376 programs, we collected 3,603 surveys from visitors (Table 2). Data from 312 programs were used in the analyses contained within this paper (see "Interpretive program sample development and data cleaning" below for more detail).

Throughout the research, the same procedure was followed for observing all programs. Upon arrival at the program site, a brief interview was conducted with the interpreter. Interview questions included interpreters' intended programmatic outcomes, questions about program development, and others about the preparation and the level of enthusiasm of the interpreter. The interviews also collected basic background information about the interpreter, which included age, gender, and interpretation experience. These interviews were conducted on all but 15 programs. In those cases, time did not allow for the interviews to take place. Basic information about the program itself was recorded by the observer, including time, location, type, topic focus, and size and age breakdown of the audience.

At the end of the program we asked visitors over the age of 15 to complete a short survey regarding their opinions of the program and its influence on them. For programs with fewer than 50 participants, we attempted a census of all eligible attendees. In programs that were particularly large (more than 50 attendees), the researchers employed systematic sampling whenever possible--for example, selecting every nth row to complete surveys at Ford's Theatre. In these cases, the researchers chose the sample interval in attempt to target at least 20 respondents.

During each program, researchers maintained an unobtrusive presence within the group, acting simply as another member of the audience. The researchers completed observation sheets during and immediately following each program.

Throughout the duration of all field work, researchers would periodically attend programs together to ensure reliability and consistency in scoring each variable. Occasional check-ins were also completed between team members to ensure that observation techniques were consistent, to clarify questions about scoring certain variables, and to add variables that were deemed relevant to the research. No new variables were added after the first week of fieldwork.

Measurement

Dependent variables: outcomes

The dependent variables in the study were composed of retrospective assessments provided by program attendees on surveys administered immediately following their programs. (2) While interpretation may produce multiple outcomes, we focused primarily on visitor satisfaction and shifts in knowledge, attitudes, and behavioral intentions relevant to the park experience.

Overall satisfaction with the program was measured on a scale from 0 to 10, with 0=Terrible and 10=Excellent. An additional battery of survey items provided response prompts for the following question: "To what degree did the program you just attended influence any of the following for you?" Response categories were composed of a five-point Likert-type scale, with answer choices: Not at all (1), A little (2), Somewhat (3), A moderate amount (4), and A great deal (5). The survey items included:

* Made me think deeply

* Made me reflect on my own life

* Enhanced my appreciation for this park

* Enhanced my appreciation for the National Park Service

* Made me more likely to avoid harming park resources

* Increased my knowledge about the program's topic

* Made my visit to this park more enjoyable

* Made my visit to this park more meaningful

* Changed the way I will behave while I'm in this park

* Changed the way I will behave after I leave this park

* Made me want to tell others about what I learned

* Made me care more about this park's resources

* Made me care more about protecting places like this

These items were developed based on key literature (e.g., Ham, 1992; Moscardo, 1957; Tilden, 1957; Ward & Wilkinson, 2006) and extensive input from NPS staff. This input included interviews and focus groups with the NPS National Education Council; a focus group and associated surveys conducted with NPS interpreters at the National Association for Interpretation (NAI) National Workshop in Las Vegas, November 2010; and two surveys conducted in 2010 and 2011 with NPS superintendents and supervisors of interpretation, respectively (see Stern & Powell, 2011). The resulting responses were analyzed to reduce the items into fewer latent factors reflecting the key outcomes of programs for visitors (see Results section).

Independent variables: predictors

Our primary independent, or predictor, variables of interest included both interpreter characteristics and the interpretive practices employed during a program. These practices were primarily drawn from an extensive literature review aimed at identifying best practices in the field (Skibins et al., 2012) as well as characteristics identified by interpretive experts within the NPS and ranked highly by interpretive staff in surveys (Stern and Powell, 2011). Additional items emerged as potentially important in pilot tests (e.g., consistency of tone and quality throughout a program) and were also measured.

Program characteristics were based in theory found in key texts within the interpretation literature (Table 3). A subset of these characteristics, however, were based primarily within the field of social psychology and relate to programs that explicitly aim to influence the behavior of participants. In short, the Theory of Planned Behavior (Ajzen, 1991) suggests that people base their behaviors upon three types of evaluations they make about the likely outcomes of performing that behavior: the benefits vs. the costs of the expected outcomes of the behavior (behavioral beliefs), what they perceive their peers might think about the behavior (normative beliefs), and the degree of control and/or ability they feel with regard to carrying out the behavior (control beliefs). We translated the theory into observable characteristics that would theoretically address these evaluations (see "Behavioral theory elements," Table 3).

Interpreter characteristics, meanwhile, focused upon the appearance, identity, and overall styles of the interpreters themselves, drawn largely from the communications and education literature, though many of these factors are also referenced in the interpretation literature (Table 4). Citations are provided where characteristics were drawn from the literature. Additional insights and examples can be found in a companion article in this same issue (Stern et al., this issue).

We also collected details pertaining to the experience level and demographics of the interpreter, their intended outcomes for their programs, and their level of excitement about the particular program they were about to deliver. In addition, we tracked information on the context for the program including location (e.g., indoors vs. outdoors), type of program, its focus (natural vs. cultural/historical vs. both), and other unexpected circumstances that could impact program outcomes (e.g., weather). In addition, we estimated the number of attendees at each program and the ratio of youth (ages 15 and under) to adults. Each of these contextual variables is examined in another article within this issue (Powell and Stern, this issue).

Pilot testing

Extensive pilot testing aided instrument development and refinement and enhanced the reliability of measurement across the research team. Prior to the field research, we observed video-recorded interpretive programs from an undergraduate interpretation class. These programs were used to develop consistent measurement of each relevant characteristic. Programs were viewed repeatedly and scores were compared among team members on each characteristic. These exercises were also used to refine the scoring of several variables.

From this testing, a preliminary assessment sheet was developed. These assessment sheets were further pilot tested at Great Smoky Mountains National Park in May of 2011, where the research team observed three live interpretive programs. Extensive discussion allowed us to further refine definitions and observation techniques for each of the characteristics under study. For each measure, we aimed to maximize the number of points in each scale to differentiate practices/attributes and enhance variability in the findings. However, existing definitions from the literature and results of pilot-testing limited most scales to four or fewer points. Pilot testing revealed that the middle-points on larger scales for many variables were not easily differentiated in a consistent manner by the research team. As a result, the scoring for each item varies to maximize the potential range of scores while maintaining inter-rater reliability. Binary scores were used in cases where the most appropriate measure was to indicate presence or absence.

Reliability and calibration

We built a calibration phase into the research design to ensure that each researcher's scores of each observed characteristic were consistent and reliable and therefore could be interpreted similarly. This involved three steps. First, immediately upon the completion of the field research and data entry, we carefully examined differences in the average scores of each variable between each member of the research team using a one-way ANOVA with posthoc tests. We identified all statistically significant differences between the mean scores for observations by different members of the research team. Second, through detailed examination of field notes and group discussions, we determined whether any of these differences might be attributed to systematic differences in observation techniques as opposed to differences in the unique sets of programs observed by each researcher. Two types of systematic differences emerged. In the first case, one researcher was systematically higher or lower than the other three on a particular measurement scale. In these cases, scoring procedures were reviewed, consensus definitions were refined, and that one researcher re-coded the variable based on these definitions and their qualitative program notes. Variables that were re-coded in this manner included comfort of the interpreter, passion, apparent knowledge, sincerity, provocation, holistic story, and appropriateness for the audience. In the second case, a researcher had misinterpreted the response scale (scoring values) of the variable being coded. Again, a consensus definition was clarified and re-coding of that variable took place. These variables included cognitive engagement, clear theme, and central message. In one case, a variable was removed due to inconsistent interpretation of its definition in the field: place-based messaging.

Data entry and cleaning

Post-program surveys and program audits were coded and entered into Microsoft Access Database and Microsoft Excel to facilitate data entry. Data were then transferred to SPSS for screening and analysis. The visitor survey data were first screened for missing values and any surveys missing more than 50% of the items per factor were removed. A total of 118 respondents were removed as a result. Data were then screened for univariate and multivariate outliers on outcome variables following Tabachnick and Fidell (2007) using Mahalanobis Distance (MAH) and studentized deleted residuals (SDRESID). A total of 58 cases were removed for exceeding +/- 3 standard deviations, or the criterion Mahalanobis Distance value. This reduced our sample to 3,427 individual surveys from 376 interpretive programs.

Interpretive program sample development and data cleaning

Because the interpretive program is the unit of analysis in this study, we aggregated individual data at the program level by calculating the mean score of each visitor outcome for each program. To do so, we first needed to determine how many completed surveys within a particular program would serve as a viable reflection of the quality of that program and its impacts on visitors. Prior research suggests that programs with particularly small numbers of attendees may be inherently different than programs with larger numbers of attendees (Forist, 2003; McManus, 1987, 1988; Moscardo, 1999). In particular, programs with fewer than five attendees may have a high likelihood of serving only a single cohesive group (e.g., a single family). Meanwhile, programs with five or more have a higher likelihood of being composed of multiple groups. Moreover, a greater number of survey responses enhances the reliability of the research findings. Based on this rationale, we separated programs with fewer than five attendees from those with five or more attendees, and analyzed them separately.

For groups with five or more attendees, we included in the analysis all programs with 10 or more respondents to the surveys. We only included those programs with fewer than 10 respondents if the number of respondents represented at least half of the eligible respondents at the program (those over the age of 15). This yielded a total of 272 programs with five or more attendees for analysis.

For programs with fewer than five attendees (n = 45), we only included those in which all eligible respondents (those over the age of 15) completed a survey. If a census was not achieved, the program was dropped from further analysis. This resulted in the removal of five of these smaller programs, leaving 40 in the sample for further analysis.

Results

Index development: Dependent variables

Before conducting further analyses, we conducted exploratory and confirmatory factor analyses to explore the relationships between items and form factors made up of multiple items that represent a concept. The items that vary together as part of a factor can be combined to create scales or composite indexes that represent coherent concepts for use in subsequent analyses (DeVellis, 2003). Following procedures outlined by DeVellis (2003) we conducted exploratory and confirmatory factor analysis on dependent variables using the individual respondent data. Exploratory factor analyses and reliability analyses revealed the presence of two latent factors. Confirmatory factor analysis (CFA), which is a form of structural equation modeling, further refined the structure of these two factors. The resulting CFA model confirmed two factors while also providing a more parsimonious solution. Model fit statistics were all within the acceptable range (S-B [chi square] = 338.41; CFI = .96; RMSEA = .08). We labeled the resulting factors Visitor Experience and Appreciation and Behavioral Intentions (Table 5).

These factors form two of the three outcomes employed in this study. The first factor reflects an overall assessment of the impact of the program on the individual's experience, attitudes, and knowledge. Taken as a whole, it may be the best reflection of the first two elements of the classic statement from an old NPS manual quoted by Tilden (1957), "Through interpretation, understanding; through understanding, appreciation; through appreciation, protection." The Behavioral intentions factor relates to the third part of the classic quote, actually influencing the behavior of visitors in some way. The third outcome, satisfaction, was measured through a single survey item: "On a scale of 0 to 10, 10 being the best, please rate your overall level of satisfaction with the program you just attended."

Composite indexes were created for each of the factors by equally weighting each item and taking the average of all items within the index. Table 6 shows the individual items that comprise each resulting index, as well as Cronbach's alpha scores for each. Cronbach's alpha is a measure of internal consistency of each index and can range from 0 to 1. Cronbach's alpha scores above 0.7 are considered acceptable for developing indexes (DeVellis, 2003). Higher Cronbach's alpha scores indicate greater internal consistency of the index. Both indexes were found to be highly reliable.

Index development: Independent variables

To explore the relationships between the individual program characteristics, we conducted exploratory factor analyses and reliability analyses on program observations. We did not conduct confirmatory factor in this case because program characteristics are formative variables that are observed and represent a specific practice or attribute that is thought to directly influence a dependent variable. This is opposed to reflective indicators, which are thought to represent a broader concept and are not directly observed (see Kline, 2005; Diamantopoulis & Siguaw, 2006; Jarvis et al., 2003; Padsokoff et al., 2003, for further explanation). Exploratory factor analyses and reliability analyses on program level data revealed the presence of four latent factors: two interpreter characteristics and two program characteristics. We have named the two resulting interpreter characteristics factors confidence and authentic emotion and charisma. We labeled the two resulting program characteristics factors organization and connection. The items making up each factor are included in Table 6.

The confidence factor generally reflects the notion that the interpreter appears in control of the program and is comfortable with what they are presenting. We use the term authentic emotion and charisma to denote a special sort of identity that the interpreter exudes to his or her audience. Interpreters scoring high on this factor showed apparent and obvious passion and care for what they were interpreting and were generally likeable. Organization reflects many of the best practices taught by the National Park Service's Interpretive Development Program in addition to the writings of Sam Ham (e.g., Ham, 1992). Meanwhile, Connection strongly reflects the core elements of Tilden's classic core principles (Tilden, 1957).

While the factor analyses revealed that confidence, authentic emotion and charisma, organization, and connection are separate constructs, they are also moderately correlated with each other (r ranges from .357 to .623). This suggests that when an interpreter scores highly on any one of these indexes, he or she is likely to score highly on the others as well.

Visitor characteristics

All descriptive statistics reported below are calculated only from the 312 programs that met our sampling criteria. More than half of the respondents to the surveys were female (56.4%). The ages of respondents ranged from 16 to 88, with a mean of 45 and a median of 46. Eighty-seven percent of respondents described themselves as White and not of Hispanic descent. Roughly 7% described themselves as Hispanic (3.6%) or Asian (3.6%). Only 34 respondents (1.1%) described themselves as Black and not of Hispanic descent; 15 respondents identified themselves as Native American and 25 respondents identified themselves as "other." Twenty-five respondents marked more than one category. Roughly 5% were from a country other than the United States. For comparison, a 2009 survey of U.S. residents conducted by the National Park Service estimated that roughly 78% of all visitors to National Park units were White; roughly 9% were Hispanic; roughly 7% were African American; roughly 3% were Asian; and roughly 1% were Native American (Taylor et al. 2010). Less than 5% of survey respondents attended the program alone. More than half (50.8%) were visiting with children. Most (59.1%) had been in the park less than one full day when they attended the program, and 37.4% had attended a ranger-led program in the same park prior to the one they were attending on the day they were surveyed.

Descriptive statistics: Outcomes

Table 7 displays the means and standard deviations of each outcome variable for programs with five or more attendees and for smaller programs. While satisfaction and visitor experience and appreciation consistently scored highly, items associated with behavioral intentions were more evenly distributed. Visitor satisfaction scores ranged from 5 to 10 on the 0 to 10 scale and 95% of respondents scored above the midpoint on the visitor experience and appreciation index. Meanwhile, 43% percent of respondents scored above the midpoint on the behavioral change index. There were no statistically significant differences in visitor outcome scores between larger programs and programs with fewer than five attendees.

Descriptive statistics: Program types and attendees

We attempted to investigate 488 programs. Only 376 programs actually occurred. Programs were cancelled for a range of reasons including weather, no visitor attendees, or failure of the interpreter to appear. Data from 312 programs were used for analyses in this paper. Advertised program lengths for these programs ranged from 15 minutes to four hours. Actual program lengths ranged from 10 minutes to three and a half hours. The average program length was just over 48 minutes. One-hundred and ninety-eight (64%) of the programs focused primarily on cultural heritage; 74 (24%) had a primary focus on the natural environment. Thirty-three (11%) had a dual focus and others had neither central focus (for example, general orientation talks). Programs included guided tours, talks, demonstrations, hands-on activities, and multi-media presentations. Guided tours and stationary talks made up over 80% of the programs we observed. Seventy-two percent of programs took place outdoors; 20% took place indoors; and others used both indoor and outdoor settings. The breakdowns of program lengths and types were roughly similar for programs in the two different size classes described above.

The number of attendees at each program ranged from one person to approximately 600 people. The median number of attendees was 17. Only 17% of the programs had no children in their audiences. Forty programs (13%) ended with fewer attendees than they had begun with. Forty-eight programs (15%) were at least 20% shorter than advertised; 53 programs (17%) were at least 20% longer than advertised. Thirteen (4%) of the programs experienced notably bad weather. No significant differences were noted in program length or weather-related variables when comparing small (fewer than five attendees) with larger programs.

Descriptive statistics: Interpreter characteristics

Two-hundred and seventy-one (87%) of the observed interpreters were park rangers; 37 were volunteers, and five were concessionaires. Sixty-four percent were male. Nineteen percent were under the age of 25; 23% were between the ages of 25 and 34; 24% were between the ages of 35 and 50; and 34% were over 50 years old. The interpreters averaged 9.6 years of experience in the NPS and 7.1 years in interpretation at their current park unit. Nearly one quarter of the interpreters (24.7%) had presented the program we observed at least 100 times before. More than one-third (36.0%) had presented the program at least 50 times before. Nearly one-third (32.6%) had presented the program 10 or fewer times. For seven interpreters, this was their first time presenting the program we observed.

We asked interpreters prior to their programs to indicate their intended visitor outcomes for that program (Table 8). The most commonly noted intended outcome was providing the audience with new knowledge. Most (90%) noted more than one intended outcome. We also asked interpreters how their programs were developed (Table 9). Most reported developing their own programs with little guidance beyond a suggested topic.

We asked a subset of interpreters (n = 188) about their level of excitement about the program they were about to present. The level of excitement averaged 7.81 on a 10-point scale, with responses ranging from 2 to 10 on the scale. Seven percent ranked their level of excitement below the midpoint (5) on the scale; 4% selected the midpoint; and 89% rated their level of excitement above the midpoint.

Descriptive statistics: Interpreter delivery styles

Tables 10 and 11 display descriptive statistics of each of the interpreter delivery styles observed in the study. Table 10 contains ordinal variables (variables that are measured on an increasing scale). Table 11 contains binary and categorical variables, or those in which the presence or absence of the characteristics is the essential feature being measured. Means comparisons, chi-square tests, and effect size calculations revealed few meaningful differences between the two size classes of programs. Interpreters typically scored slightly lower on the confidence index in smaller groups (t = 2.0; p = 0.042; Cohen's d=0.38). We also more commonly observed the "friend" identity in smaller groups ([chi square] = 8.0; p = 0.005).

Descriptive statistics: Program characteristics

Tables 12 and 13 display descriptive statistics for each of the program characteristics observed in the study. Table 12 displays ordinal variables, while Table 13 displays categorical variables. No statistically significant differences were observed between the two size classes of programs.

Which practices and approaches most consistently lead to more positive outcomes for visitors?

Interpreter and program characteristics

Table 14 displays (in rank order) correlations between all ordinal independent variables (program and interpreter characteristics) and visitor outcomes for programs with five or more attendees. Statistical significance is displayed in two ways within the table. A single asterisk indicates that the correlation is statistically significant at p < 0.05. A double asterisk indicates that the correlation is statistically significant at p < 0.01. As such, the stronger relationships are those with two asterisks. These are bolded and italicized for ease of interpretation. Cells with no asterisks represent no statistically significant relationships between the variables.

Behavioral theory elements were observed in 42 programs overall, including 31 with five or more attendees. Only one behavioral theory element showed a statistically significant correlation with the behavior change index, "costs of action" (r = .597, p < .001). This suggests that programs that explicitly addressed the costs of undertaking a potential behavior were generally more successful at influencing behavior change intentions than others.

T-tests and ANOVAs were performed to examine the relationships of categorical variables upon visitor outcomes. These variables included fact-based messaging, unexpected positive and negative circumstances, pace, bias, impatience, inequitable treatment of the audience, questionable information, use of props, and interpreter identities. Tables 15 and 16 summarize only the statistically significant relationships observed in the data. To facilitate interpretation of the t-tests, we calculated Cohen's d for each of the statistically significant associations. Cohen's d is an effect size measure that provides an assessment of the meaningfulness of the difference between groups. Cohen (1988) suggested that even statistically significant differences may not be meaningful in a practical sense. They may rather be an artifact of large sample sizes. Cohen posited that meaningful differences begin at d = 0.2. Differences near 0.2 may be considered small, while those approaching 0.5 are considered medium and 0.8 large.

Programs in which the interpreter outwardly expressed impatience with the audience received lower satisfaction and visitor experience and appreciation scores than others, as did programs with an unexpected negative occurrence. Programs in which the interpreter employed the "friend" identity manifested higher satisfaction scores than others. Meanwhile, programs in which the interpreter employed the "walking encyclopedia" identity yielded lower behavioral intention scores than others. Paces that felt too fast or too slow resulted in lower satisfaction scores. A too-slow pace was related to lower visitor experience and appreciation scores, and a too-fast pace was associated with weaker behavioral intentions. No statistically significant differences were observed for smaller programs (fewer than five attendees).

Program attrition and outcomes

Program attrition (people leaving a program before it was completed) was related to both satisfaction and visitor experience and appreciation for programs with five or more attendees (see Table 17), suggesting that program attrition may serve as another reasonable indicator of program quality. Thirty-six of programs with five or more attendees experienced attrition. The best predictors of program attrition for programs with five or more attendees included interpreters' lack of responsiveness to the audience, inaudibility, false assumptions about the audience, the identity of the walking encyclopedia, inappropriate logistics, the use of props, slow pace, lack of interpreter confidence, a lack of organization of the program, and an unexpected negative circumstance (see Tables 17 and 18). (3) No other interpreter or program characteristics exhibited any statistically significant relationship with program attrition at p < 0.05.

Relationship between interpreter and program characteristics and outcomes in programs with fewer than five attendees

Fewer statistically significant correlations (p < 0.05) were observed in programs with fewer than five attendees. In rank order, they included:

Correlated with Satisfaction:

* Connection index: r = .492, p = .001

* Organization index: r = .420, p = .007

* Appropriate for the audience: .337, p = .033

* Humor quality: r = .323, p = .045

Correlated with Visitor experience and appreciation:

* Connection index: r = .438, p = .005

* Organization index: r = .368, p = .020

* Appropriate for the audience: .348, p = .028

Correlated with Behavioral intentions:

* Novelty: r = .408, p = .009

Thus, a subset of the variables that predicted positive outcomes in larger programs predicted similar outcomes in smaller programs. Because only four programs within this sample experienced attrition, no additional analyses were conducted pertaining to attrition.

Interpreters' background, excitement, and intentions

For the smaller program sample (those with fewer than five attendees), no statistically significant relationships were observed between interpreter backgrounds, level of excitement, program origin, or intended outcomes and visitor outcomes. Some differences were noted, however, in the larger sample.

For larger group sizes (five or more attendees), program outcomes were not related to the age, gender, or experience of interpreters, nor their degree of autonomy in program development. The interpreters' degree of excitement about the program was positively correlated with visitor satisfaction (r = .186; p = 0.013) and visitor experience and appreciation (r = .153; p = 0.041). Interpreters expressing higher degrees of excitement also exhibited higher levels of confidence (r = .324, p < .001) and authentic emotion and charisma (r = .475; p < .001). Volunteers tended to achieve lower degrees of visitor satisfaction than did park rangers (means: 8.70 vs. 8.98; t = -2.4; p = .019; Cohen's d = 0.42).

We examined the relationships between interpreters' intended outcomes and visitor-reported outcomes by conducting independent samples t-tests, which compare the means of two groups. In these cases, groups were defined by the presence of an intended outcome or not. Table 19 summarizes only the statistically significant relationships between interpreters' intended outcomes and visitor survey responses. Cohen's d statistics are also provided as effect size estimates. Visitor experience and appreciation was the most sensitive to interpreters' intended outcomes, with five different desired outcomes related to more positive visitor responses. Satisfaction was related to a subset of these items. Only one intention was negatively related to visitor outcomes. Interpreters who were aiming to increase visitors' knowledge as a primary outcome of their program generally achieved lower visitor experience and appreciation scores. Two intended outcomes were positively related to reported behavioral intentions by visitors: increasing the audience's level of concern and changing visitors' behaviors.

Discussion

The study sought to determine which practices and approaches most consistently lead to more positive outcomes for live interpretive programs' attendees. In this manuscript, we have limited our analyses to bivariate relationships between practices and outcomes rather than employing multivariate statistics. We did this for two reasons. First, we wished to examine the individual relationship of each observed practice and interpreter characteristic with visitor outcomes. Second, multivariate analyses are used to provide the most parsimonious statistical model of observed phenomenon. In multivariate processes, certain observed characteristics may be removed from the best explanatory model if they explain a similar portion of the variance as another variable, despite being an important part of influencing a particular outcome (Byrne, 2006). As a result, the multivariate approach may lead to misinterpretation of the importance (or lack thereof) of particular practices and program characteristics. If one were to focus only on the variables contained in the multivariate statistical model, at the expense of others that covaried with those same variables, there would be a danger of inappropriately assuming that practices not in the model are unimportant. In a companion piece, we use structural equation modeling to develop more parsimonious causal models (see Powell and Stern, this issue). These multivariate analyses help to illuminate the inter-relationships of different interpreter and program characteristics and their roles in influencing outcomes. However, they do not negate the bivariate relationships shared in this article.

Understanding outcomes

Live interpretive programs across the NPS generally seem to produce consistently high levels of satisfaction in their attendees. Eighty-five percent of the analyzed sample rated the program as an 8 or better on the 0 to 10 satisfaction scale. Such satisfaction skewness is common in customer satisfaction surveys, and the modal response is typically the most positive response allowed by the scale (Peterson & Wilson, 1992). The mode in our case was a 9 out of 10. Prior research suggests that satisfaction assessments may be influenced by social desirability bias or acquiescence (Peterson & Wilson, 1992). In our case, such social factors might include some degree of gratitude or sympathy toward the interpreter regardless of the program quality, leading respondents to check a positive response. High satisfaction scores might also be attributed in part to what is known as assimilation effects (Sherif & Hovland, 1961). In the context of tourism, this means that expectations are often a stronger driver of satisfaction ratings than the quality of the actual experience (del Bosque & San Martin, 2008). In other words, if visitors strongly expect an experience to be positive, they have a high tendency to rate it as such regardless of its specific qualities. This may of course be the case with visitors to national parks. Still, the particularly high satisfaction values observed in this study suggest that few visitors were dissatisfied with their interpretive experiences. Visitor experience and appreciation also showed similar trends.

Despite the skewness of the data, we observed significant statistical relationships between certain program characteristics and visitor outcomes. The positively skewed dependent variables, however, suggest that our findings do not necessarily identify the practices that separate good programs from bad programs. Rather, the findings illuminate which characteristics most commonly move programs along a scale from good to better from a visitor's standpoint (see Stern et al., this issue).

The behavioral intentions outcome was centered closer to the midpoint of the five-point scale. This is likely due to widely varying baselines in terms of visitors' behaviors prior to programs (some visitors wrote on the survey cards things like "I already respect the parks"). For example, if a visitor is a major park supporter and an environmentally sensitive visitor, we might expect them to report no change, despite experiencing what may have been an outstanding program. Meanwhile, an inexperienced visitor to the same program might have reported a great deal of change. As such, we might expect muted results regarding program and interpreter characteristics' associations with the behavioral intentions outcome. This may in part explain the smaller number of independent variables associated with intentions to change behaviors. Other authors have also expressed concern when measuring intentions and behavior change, especially in nature-based settings (see Beaumont, 2001; Powell et al, 2008).

What leads to better outcomes?

Interpreters who expressed that a primary goal of their program was to increase the knowledge of the audience about their program's topic achieved lower visitor experience and appreciation scores than others. Those aiming to change their audience's attitudes, appreciation, understanding, and/or desire to learn achieved more positive attitudinal outcomes. Interpreters who explicitly aimed to increase their audience members' levels of concern or change their behavior were more likely to achieve more positive post-program behavioral intentions than others.

The best predictors of positive outcomes varied somewhat for different outcomes. In programs with at least five attendees, the outcomes Satisfaction and visitor experience and appreciation were correlated with a similar list of program and interpreter characteristics, including: confidence, authentic emotion and charisma, appropriateness for the audience, organization, connection, humor quality, consistency, a clear message, responsiveness, verbal engagement, audibility, and appropriate logistics and pace. Multisensory engagement and fact-based messaging (negative relationship) were additionally related to satisfaction.

Behavioral theory suggests that interpretation (and other communication/ educational experiences) should not be expected to change behavior unless a specific behavior is explicitly targeted and communication is designed to address attitudes relevant to that behavior (e.g., Ajzen, 1991; Ham et al., 2007). Programs in which the interpreter explicitly targeted behavior change as an intended outcome (7%) were more successful at doing so. Programs of this nature that explicitly addressed the costs of taking that action were the most successful, supporting Ajzen's (1991) emphasis on both ability and trade-offs in predicting behavior. Moreover, confidence, authentic emotion and charisma, a clear message, verbal engagement, and appropriate logistics showed the strongest statistically significant correlations with the behavioral intentions outcome. These items mirror theoretical constructs from multiple disciplines known to be predictive of behavior change, including credibility and trust in the communicator (Rogers, 1995; Stern, 2008), empowerment of the message recipient and verbal engagement (Ajzen, 1991; Stern, 2008), and the elimination of distraction and clear orientation to place (Moscardo, 1999). For a broader discussion of behavior change and interpretation see Ham et al. (2007) and Ham (2009).

Figure 1. Best practices for live interpretive programs observed in
the study.

1. Confidence

* Comfort, eloquence, apparent knowledge

2. Authentic emotion and charisma

* Passion, sincerity, charisma

3. Appropriateness for audience

4. Organization

* Quality of introduction, appropriate sequence, effective
transitions, holistic story, clear theme, link between introduction
and conclusion

5. Connection

* Links to intangibles and universal concepts, cognitive
engagement, relevance to audience, affective messaging, provocation

6. Consistency

7. Clear message

8. Responsiveness

9. Audibility

10. Appropriate logistics

11. Verbal engagement

12. Multisensory engagement

13. Appropriate pace

14. Avoid focusing on knowledge gain as the program's central goal
and communicating solely factual information

15. Avoid making uncertain assumptions about the audience

A smaller subset of interpreter and program characteristics were correlated with outcomes for smaller programs (those with fewer than five attendees). Connection, organization, and appropriateness for the audience were each correlated with satisfaction and visitor experience and appreciation. Humor quality was additionally correlated with satisfaction. Only novelty was correlated with post-program behavioral intentions for these smaller programs.

Implications for live interpretation

The study carries implications for both the practice of live interpretation as well as future research pertaining to best practices. Figure 1 provides a list of the program characteristics most strongly associated with the outcomes measured in this study. These "best practices" cut across multiple contexts (see Powell & Stern, this issue) and constitute elements of interpretation that could inform interpretive training both within the National Park Service and beyond. While humor quality also was positively related to outcomes, we don't list it as a best practice, as not all programs should necessarily be funny.

Although each of the practices listed in Figure 1 was statistically correlated with better outcomes, variability within the sample suggests that the entire suite of best practices is not a necessary precursor to a high-quality program. Rather, each of these practices in various combinations was found to enhance outcomes across a majority of programs in which they were practiced. A wide range of diverse approaches led to positive visitor outcomes. As such, we recommend maintaining the freedom for interpreters to be creative and innovative in their presentations. This is further supported by correlations between interpreters' own excitement about a program and positive visitor outcomes.

While many of the "best practices" in Figure 1 speak to specific interpretive techniques, some, at first glance, appear to exist outside of the famous "interpretive equation" used in NPS trainings (Lacome, 2013). The interpretive equation is presented as a "foundation" for NPS interpretive training and as a tool for identifying "the elements of successful interpretation" and the relationships between them. In its simplest form, the equation states that an interpretive opportunity (IO: "one that provides a favorable set of circumstances for a meaningful moment of connection between audience and resource," p. 5) is brought about by knowledge of the resource (KR), knowledge of the audience (KA), and appropriate techniques (AT).

The Interpretive Equation: KR + KA x AT = IO

Many of the "best practices," in particular confidence, authentic emotion and charisma, and avoiding a focus on knowledge gain, do not clearly constitute "knowledge of the resource," "knowledge of the audience," or "appropriate techniques" directly. They are rather the observable manifestations of internal states specific to individual interpreters during their programs. Their significance speaks to the importance of the appropriate translation of the interpretive equation into action. While knowledge of the resource is critical, it should not necessarily be the focus of communications within an interpretive setting. Rather, knowledge of the resource may play a more important role in enhancing the confidence of the interpreter and allowing his or her own positive emotions and connections to the resource to show through. Presenters who are more familiar with their topics generally experience less anxiety (Daly et al. 1989). When coupled with knowledge of the audience and appropriate techniques, feelings of self-confidence and freedom to express oneself might be instrumental in moving from good, or adequate, visitor outcomes toward more powerful ones. This also suggests that the general organizational culture in which the interpreter finds herself is likely important as well. More supportive and empowering cultures may lead to better performance (Pearce & Sims, 2002; Rafferty & Griffin, 2006). The particular roles of interpreter characteristics vs. program characteristics are examined in greater detail in a companion article within this issue (Powell & Stern, this issue).

Implications for future interpretive research

This research suggests that certain interpretive practices are statistically linked to desired outcomes across a range of contexts. Without the ability to compare a large sample of programs, this identification would not have been possible. We thus urge others to undertake similar forms of research and to learn from our shortcomings. Even comparative research of just a few programs can shed additional light on what practices and approaches are linked to more positive visitor outcomes (see Ballantyne & Packer 2002, for example).

Our limitations and shortcomings were many in this effort, including both controllable and uncontrollable factors. Those most relevant to future research involve the selection and measurement of the key independent and dependent variables of the study. The treatment (an interpretive program in a national park setting) is a complex phenomenon that is influenced by an interaction between the resource and its qualities, the social environment, including the makeup of social groups, the characteristics of the interpreter and the individual attendees, and the topic and characteristics of the program (Powell et al, 2009). This research focuses on the relationships between visitor outcomes and selected interpreter and program characteristics. As such, other potential influences are not accounted for.

Our experience revealed that it required considerable and iterative training, feedback, and adjustment for our team to produce consistent and reliable monitoring results. This is a well-known challenge in any research using a team of human observers, who have a tendency to cling to their own personal biases or sometimes idiosyncratic interpretations of similar events (Jacobs et al., 2012). In an ideal situation, additional pilot testing and assimilation of the team toward consistent definitions could take place and programs would be consistently observed in pairs, rather than by individuals.

Our selection of dependent variables was quite challenging due to the wide diversity of program content and formats included in this study. Visitor survey items were designed to be rather general in their content so as to be appropriate and relevant to all programs. The general nature of outcome measures may have also contributed to a "ceiling effect," which describes the phenomenon when individuals (in this case, NPS visitors) come into an experience with already high scores on the outcomes considered (in this case the specific attitudes and intentions measured in the study). As such, some respondents would report little to no change for an outcome measure because their attitudes or intentions may already be at the high end of the spectrum for the outcome in question. In these cases, the survey items may not be sensitive enough to detect the influence of a program. We urge future researchers to develop more sensitive dependent variables, and, if possible, include a control group. In particular, other researchers have found that multiple measures of satisfaction with both positive and negative wording can produce more variability (Peterson & Wilson, 1992). A rigorous approach to control group sampling might involve a similar design as our own (see endnotes) with a larger sample of non-participants. Alternatively, researchers might consider comparison groups exposed to similar interpretation with the exception of only a few variables (or ideally one experimental variable) at a time.

Conclusions

Overall, our analysis suggests that Tilden (1957), writing over 50 years ago, was right about a lot of things. Programs that are relevant to the audience, tell holistic stories, provoke the audience to reflect, and move beyond facts into the realm of revelation tend to produce better visitor outcomes than programs that are fact-based and detached from the audiences' lives. It also suggests that more recent interpretive texts and training programs include numerous ideas that can enhance the interpretive experience, including the passion of the interpreter (e.g., Beck & Cable, 2002; Ward & Wilkinson, 2006), the organization of the material (e.g., Ham, 1992; Larsen, 2003), the importance of a central message (e.g., Ham, 1992; Jacobson, 1999), the connection of tangible objects to intangible meanings and universal concepts (NPS, 2003), and multiple forms of engagement and responsiveness (Beck & Cable, 2002; Knudson et al., 2003; Lewis, 2005; Moscardo, 1999). The study also revealed some factors that appear less regularly in existing training programs, but are certainly not surprising. In essence, the study revealed the importance of the sincerity, passion, confidence, and delivery style of individual interpreters, as much as the planning and content of the program itself. We echo Tilden (1957) in believing that "interpretation is an art ... and that any art is in some degree teachable." We hope that the results of this study can contribute to the learning process of the committed individuals around the world who care deeply enough about our world to call themselves "interpreters."

Literature Cited

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Decision-Making Processes, 50(2), 179-211.

Ajzen, I. (1992). Persuasive communication theory in social psychology: a historical perspective. In: Manfredo, M. (Ed.). Influencing Human Behavior. Theory and Applications in Recreation, Tourism, and Natural Resources Management. Champaign, IL: Sagamore Publishing, Inc.

Babbie, E. (2007). The Practice of Social Research. Belmont, CA: Wadsworth.

Ballantyne, R. and J. Packer. (2002). Nature-based excursions: School students' perceptions of learning in natural environments. International Research in Geographical and Environmental Education 11, no. 3: 218-236.

Beck, L., & Cable, T. T. (2002). Interpretation for the 21st century: Fifteen guiding principles for interpreting nature and culture (2nd ed.). Champaign: Sagamore.

Beaumont, N. (2001). Ecotourism and the conservation ethic: Recruiting the uninitiated or preaching to the converted? Journal of Sustainable Tourism, 9(4):317-341.

Brochu, L., & Merriman, T. (2002). Personal interpretation: Connecting your audience to heritage resources. Fort Collins, CO: InterpPress.

Chesebro, J. L., & Wanzer, M. B. (2006). Instructional message variables. Handbook of instructional communication: Rhetorical and relational perspectives (pp. 89-116). Boston: Allyn & Bacon.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (second ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.

Daly, J.A., Vangelisti, A.L., Neel, H.L., & Cavanough, P.D. 1989. Pre-performance concerns associated with public speaking anxiety. Communication Quarterly, 37(1); 39-53.

del Bosque, I. R., & San Martin, H. (2008). Tourist satisfaction: A cognitive-affective model. Annals of Tourism Research, 35(2): 551-73.

DeVellis, R.F. 2003. Scale development: Theory and applications applied social research methods. 2nd ed. Thousand Oaks, CA: Sage Publishing.

Diamantopoulos, A., & Winklhofer, H.M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 38(2); 269-277.

Eppley Institute for Parks and Public Lands. (2012). "About interpretation." http://idp. eppley.org/about-interpretation. Last accessed July 16, 2012.

Finn, A. N., Schrodt, P., Witt, P. L., Elledge, N., Jernberg, K. A., & Larson, L. M. (2009). A meta-analytical review of teacher credibility and its associations with teacher behaviors and student outcomes. Communication Education, 58(4), 516-537.

Forist, B. (2003). Visitor Use and Evaluation of Interpretive Media. A Report on Visitors to the National Park System. National Park Service Visitor Services Project. http://nature.nps.gov/socialscience/ docs/Visitor_Use_and_Evaluation.pdf. Accessed May 24, 2012.

Frauman, E., & Norman, W.C. (2003). Managing visitors via "mindful" information services: One approach in addressing sustainability. Journal of Park and Recreation Administration, 21(4): 87-104.

Ham, S. H. (1992). Environmental interpretation: A practical guide for people with big ideas and small budgets. Golden: Fulcrum.

Ham, S. (2009). From interpretation to protection: is there a theoretical basis? Journal of Interpretation Research 14(2): 49-57.

Ham S. H. (2013). Interpretation: Making a difference on purpose. Golden, CO: Fulcrum.

Ham, S. H., Brown, T. J., Curtis, J., Weiler, B., Hughes, M., & Poll, M. (2007). Promoting persuasion in protected areas: A guide for managers. Developing strategic communication to influence visitor behavior. Southport, Queensland, Australia: Sustainable Tourism Cooperative Research Centre.

Ham, S H. & Weiler, B.M. (2002). Tour guide training: a model for sustainable capacity building in developing countries. Journal of Sustainable Tourism 10(1): 52-69.

Interpretive Development Program. (2008). National Park Service: Professional Standards for Learning and Performance. U.S. Department of the Interior.

Jacobs, W.J., Sisco, M., Hill, D., Malter, F., and Figueredo, A.J. 2012. On the practice of theory-based evaluation: information, norms, and adherence. Evaluation and Program Planning 35, no. 3: 354-369.

Jacobson, S. K. (1999). Communication skills for conservation professionals. Washington, DC: Island Press.

Jarvis, C.B., MacKenzie, S.B., & Podsakoff, P.M. (2003). A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research, 30(2); 199-218.

Kline, R.B. (2005). Principles and practice of structural equation modeling. New York: The Guilford Press.

Knapp, D., & Benton, G. M. (2004). Elements to successful interpretation: A multiple case study of five national parks. Journal of Interpretation Research, 9(2), 9-25.

Knapp, D. & Yang, L. (2002). A phenomenological analysis of long-term recollections of an interpretive program. Journal of Interpretation Research, 7(2), 7-17.

Knudson, D. M., Cable, T. T., & Beck, L. (2003). Interpretation of cultural and natural resources (2nd ed.). State College: Venture Publishing.

Lacome, B. (2013). The Interpretive Equation. http://idp.eppley.org/IDP/sites/default/files/ EquationEssay.pdf. Accessed April 6, 2013.

Larsen, D. L. (2003). Meaningful Interpretation: How to Connect Hearts and Minds to Places, Objects, and Other Resources. Eastern National.

Lewis, W. J. (2005). Interpreting for park visitors (9th ed.). Fort Washington: Eastern National.

Madin, E. M. P., & Fenton, D. M. (2004). Environmental interpretation in the Great Barrier Reef Marine Park: An assessment of programme effectiveness. Journal of Sustainable Tourism, 12(2), 121-137.

McCroskey, J.C., Richmond, V.P., & Stewart, R.A. (1986). One on one: The foundations of interpersonal communication. Englewood Cliffs, NJ: Prentice-Hall.

McManus, P.M. (1987). It's the company you keep ... The social determination of learning-related behaviour in a science museum. The International Journal of Museum management and Curatorship, 6: 263-270.

McManus, P.M. (1988). Good companions. More on the social determination of learning-related behaviour in a science museum. The International Journal of Museum Management and Curatorship 7: 37-44.

Mehrabian, A. (1969). Attitudes inferred from non-immediacy of verbal communications. Journal of Verbal Learning & Verbal Behavior, 6, 294-295.

Mills, E. 1920. Adventures of a Nature Guide. New York: Doubleday, Page & Company.

Moscardo, G. (1999). Making visitors mindful: Principles for creating quality sustainable visitor experiences through effective communication. Champaign: Sagamore.

National Association for Interpretation (NAI). 2012. Mission, Vision and Core Values. http://www.interpnet.com/about_nai/mission.shtml. Accessed May 24, 2012.

National Park Service 2003. Module 101: Fulfilling the NPS mission: The process of interpretation. Washington, DC: National Park Service.

Pearce, C.L. & Sims, H.P. (2002). Vertical versus shared leadership as predictors of the effectiveness of change management teams: an examination of aversive, directive, transactional, transformational, and empowering leader behaviors. Group dynamics: theory, research, and practice, 6(2), 172-197.

Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., & Podsakoff, N.P. (2007). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5) 879-903.

Powell, R.B., Skibins, J.C. & Stern, M.J. (2010) Linking interpretation best practices with outcomes: A review of literature. Clemson University and U.S. National Park Service, National Education Council.

Powell, R.B., Kellert, S.R., & Ham, S.H. (2008). Antarctic tourists: Ambassadors or consumers? Polar Record, 44(230): 233-241.

Powell, R. B., Kellert, S. R., & Ham, S. H. (2009). Interactional theory and the sustainable nature-based tourism experience. Society & Natural Resources, 22(8), 761-776.

Rafferty, A.E. & Griffin, M.A., (2006). Refining individualized consideration: distinguishing developmental leadership and supportive leadership. Journal of Occupational and Organizational Psychology, 79(1), 37-61.

Regnier, K., Gross, M., & Zimmerman, R. (1992). The Interpreter's Guidebook: Techniques for Programs and Presentations. Interpreter's Handbook Series. Stevens Point: UWSP Foundation Press.

Rogers, E. M. (1995). Diffusion of Innovations (Fourth Edition ed.). New York: Free Press.

Skibins, J.C., Powell, R.B., & Stern, M.J. (2012). Exploring empirical support for interpretation's best practices. Journal of Interpretation Research.

Sharpe, G.W. (1976). Interpreting the environment. New York: John Wiley & Sons.

Sherif, M., and C. Hovland (1961). Social Judgment: Assimilation and Contrast Effects in Communication and Attitude Change. New Haven: Yale University Express.

Stern, M.J. (2008). Coercion, voluntary compliance, and protest: the role of trust and legitimacy in combating local opposition to protected areas. Environmental Conservation, 35(3): 200-210.

Stern, M.J. & Powell, R.B. (2011). The Views of National Park Service Superintendents and Interpretation and Education Supervisors on Interpretation and Education in the National Park Service. Submitted to the National Education Council of the U.S. National Park Service. October, 2011.

Stern, M.J., R.B. Powell, and K.S. Hockett (2011). Why do they come? Understanding interpretive program attendance at Great Smoky Mountains National Park. Journal of Interpretation Research 16(2): 35-52.

Tabachnick, B.G. and L.S. Fidell. (2007). Using multivariate statistics (5th ed.) Needham Heights, MA: Allyn & Bacon.

Taylor, P.A., Grandjean, B.D., Gramann, J.H. (2011). National Park Service Comprehensive Survey of the American Public 2008-2009: Racial and Ethnic Diversity of National Park System Visitors and Non-Visitors. Natural Resource Report NPS/NRSS/SSD/NRR-2011/432. Wyoming Survey & Analysis Center, University of Wyoming, Laramie, Wyoming. Online: http://nature.nps.gov/ socialscience/docs/CompSurvey2008_2009RaceEthnicity.pdf.

Tilden, F. (1957). Interpreting our heritage (3rd ed.). Chapel Hill: The University of North Carolina Press.

Veverka, J. A. (1998). Interpretive master planning: the essential planning guide for interpretive centers, parks, self guided trails, historic sites, zoos, exhibits and programs (2nd ed.). Tustin: Acorn Naturalists.

Wallace, G. N., & Gaudry, C. J. (2002). An evaluation of the "authority of the resource" interpretive technique by rangers in eight wilderness/backcountry areas. Journal of Interpretation Research, 7, 43-68.

Wearing, S., & Wearing, B. (2001). Conceptualizing the selves of tourism. Leisure Studies, 7, 11-23.

Ward, C.W., & Wilkinson, A. E. (2006). Conducting meaningful interpretation: A field guide for success. Golden: Fulcrum.

Notes

(1.) Based on a review of web pages of all park units at the time of the research (www.nps.gov).

(2.) Our original research design also included administering shorter pre-experience surveys at different, but similar programs across the parks in our sample. These surveys contained two batteries of survey items that could be compared to the post-experience surveys to create a control group against which to compare outcomes. Unfortunately, an insufficient number of these surveys were administered at most parks to create a reliable control group. As a result, we did not include these data in further analyses.

(3.) Our field observations suggest that the association between the use of props and increased attrition may be influenced by cases in which not all visitors were able to engage with the prop(s). This may have motivated their departure.

Marc J. Stern

Department of Forest Resources and Environmental Conservation, Virginia Tech

Robert B. Powell

Department of Parks, Recreation and Tourism Management and School of Agricultural

and Forest Environmental Sciences, Clemson University

Table 1. Park units included in the study.

                                                          Annual
                               Resource                   Recreation
Park Unit                      Focus      Park Location   Visits (a)

Aztec Ruins National           Cultural   Remote          37,437
  Monument
Badlands National Park         Natural    Remote          977,778
Bryce Canyon National Park     Natural    Remote          1,285,492
Chaco Culture National         Cultural   Remote          34,226
  Historical Park
Ford's Theater National        Cultural   Urban           662,298
  Historic Site
Fort McHenry National          Cultural   Urban           611,582
  Monument and Historic
  Shrine
Gettysburg National Military   Cultural   Urban-          1,031,554
  Park                                    Proximate
Grand Canyon National Park     Natural    Remote          4,388,386
                                          Urban-
Great Smoky Mountains          Mix        Proximate       9,463,538
  National Park
Harpers Ferry National         Cultural   Urban-          268,822
  Historical Park                         Proximate
Independence National          Cultural   Urban           3,751,007
  Historical Park
Jefferson National Expansion   Cultural   Urban           2,436,110
  Memorial
Jewel Cave National Monument   Natural    Remote          103,462
Lincoln Home National          Cultural   Urban           354,125
  Historic Site
Manassas National              Cultural   Urban-          612,490
  Battlefield Park                        Proximate
Mesa Verde National Park       Mix        Remote          559,712
Mount Rushmore National        Cultural   Remote          2,331,237
  Memorial
National Mall                  Cultural   Urban           1,363,389
Navajo National Monument       Mix        Remote          90,696
Point Reyes National           Natural    Urban-          2,067,271
  Seashore                                Proximate
San Francisco Maritime         Cultural   Urban           4,130,970
  National Historical Park

Ulysses S. Grant National      Cultural   Urban           39,967
  Historic Site
Wind Cave National Park        Natural    Remote          577,141
Yosemite National Park         Natural    Remote          3,901,408

(a) Annual visitation from 2010 (http://www.nature.nps.gov/stats/)

Table 2. Programs observed and total number of surveys collected.

Park unit                           Programs    Programs   Surveys
                                    attempted   observed   collected

Aztec Ruins National Monument           4           2            4
Badlands National Park                 22          19          157
Bryce Canyon National Park             12          12          133
Chaco Culture National Historical       9           8           85
  Park
Ford's Theater National Historic       20          20          519
  Site
Fort McHenry National Monument         23          14          133
  and Historic Shrine
Gettysburg National Military Park      26          21          206

Grand Canyon National Park             30          30          384
Great Smoky Mountains National         19          14           96
  Park
Harpers Ferry National Historical      21          15          100
  Park
Independence National Historical       36          22          156
  Park
Jefferson National Expansion           22          16          146
  Memorial
Jewel Cave National Monument           20          20          190
Lincoln Home National Historic         18          14           89

  Site
Manassas National Battlefield          20          17           88
  Park
Mesa Verde National Park               14          14          301
Mount Rushmore National Memorial       23          19          171
National Mall                          47          22           65
Navajo National Monument                8           3           23
Point Reyes National Seashore          12           9           34
San Francisco Maritime National        20          16           69
Historical Park
Ulysses S. Grant National              15           9           40
  Historic Site
Wind Cave National Park                18          18          215
Yosemite National Park                 29          22          199

Totals                                488         376        3,603

Park unit                            Used in analyses

                                    Programs   Surveys

Aztec Ruins National Monument           2           4
Badlands National Park                 14         118
Bryce Canyon National Park             12         127
Chaco Culture National Historical       7          70
  Park
Ford's Theater National Historic       18         448
  Site
Fort McHenry National Monument         11         113
  and Historic Shrine
Gettysburg National Military Park      18         186
Grand Canyon National Park             28         363
Great Smoky Mountains National         12          86
  Park
Harpers Ferry National Historical      12          79
  Park
Independence National Historical       17         122
  Park
Jefferson National Expansion           14         135
  Memorial
Jewel Cave National Monument           18         177
Lincoln Home National Historic         10          72
  Site
Manassas National Battlefield          15          80
  Park
Mesa Verde National Park               14         290
Mount Rushmore National Memorial        9         101
National Mall                          16          49
Navajo National Monument                3          23
Point Reyes National Seashore           8          32
San Francisco Maritime National        14          64
Historical Park
Ulysses S. Grant National               8          36
  Historic Site
Wind Cave National Park                13         175
Yosemite National Park                 19         172

Totals                                312       3,122

Table 3. Program characteristics observed in the study, their
definitions, and operationalization.

Program characteristic Definition              Scoring

Introduction quality   Degree to which the     3 = Oriented audience
(Brochu and Merriman,  introduction captured   and captured
2002; Ham, 1992;       the audience's          attention
Jacobson, 1999)        attention and
                       oriented (or pre/       2 = Minimally
                       disposed) the           oriented audience;
                       audience to the         did not necessarily
                       program's content       capture attention
                       and/or message.
                                               1 = Poorly executed

Appropriate            Degree to which basic   4 = Well planned and
logistics (Jacobson,   audience and program    appropriate
1999; Knudson et       needs were met (i.e.,
al., 2003)             restrooms, weather,     3 = Audience/program
                       technology,             needs mostly
                       accessibility, shade,   addressed
                       etc.).
                                               2 = Needs marginally
                                               addressed

                                               1 = Needs not met

Appropriate for        Degree to which the     5 = Very appropriate
audience (Beck and     program aligned with
Cable, 2002;           audience's ages,        4 = Appropriate
Jacobson, 1999;        cultures, and level
Knudson et al.,        of knowledge,           3 = Moderately
2003)                  interest, and           appropriate
                       experience.
                                               2 = Only slightly
                                               appropriate

                                               1 = Not appropriate

Appropriate sequence   Degree to which the     4 = Enhanced
(Beck and Cable,       program followed a      messaging
2002; Ham, 1992;       logical sequence.
Jacobson, 1999;                                3 = Appropriate
Larsen, 2003)
                                               2 = Choppy

                                               1 = Detracted from
                                               messaging

Transitions            Degree to which         4 = Enhanced
(Beck and Cable,       program used            messaging and were
2002; Brochu and       appropriate             smooth
Merriman, 2002; Ham,   transitions that kept
1992; Jacobson,        the audience engaged    3 = Appropriate
1999; Larsen, 2003)    and did not detract
                       from the program's      2 = Forced or
                       sequence.               irrelevant

                                               1 = Detracted from
                                               messaging or not
                                               present

Links to intangible    Communication           5 = Extensively
meanings and           connected tangible      developed; powerful
universal concepts     resources to            concepts
(NPS Module 101;       intangible meanings
Beck and Cable,        and universal           4 = Well developed
2002; Brochu and       concepts.
Merriman, 2002; Ham,   Intangibles: stories,   3 = Present but weak
1992; Knudson, et      ideas, meanings, or
al., 2003; Larsen,     significance that       2 = Difficult to
2003; Lewis, 2005;     tangible resources      detect or slightly
Moscardo, 1999;        represent               used
Tilden, 1957; Ward     Universals: concepts
and Wilkinson, 2006)   that most audience      1 = Clearly not
                       members may identify    present
                       with

Multisensory (Beck     Degree to which the     3 = Explicit/
and Cable, 2002;       program intentionally   purposeful inclusion
Knudson et al.,        and actively engaged    of two sense beyond
2003; Lewis, 2005;     more than just basic    sight and sound
Moscardo, 1999;        sight and sound.
Tilden, 1957;                                  2 = Actively
Veverka, 1998; Ward                            incorporated a sense
and Wilkinson, 2006)                           beyond passive use of
                                               sight and sound, or
                                               actively focused upon
                                               either of these and
                                               senses as a vehicle
                                               for conveying the
                                               message (e.g., "close
                                               your eyes and sound
                                               listen")

                                               1 = Primarily a talk
                                               in which the ranger
                                               did not explicitly
                                               use multiple sense
                                               beyond passive use of
                                               sight (scenery/
                                               objects) and sound
                                               (words)

Physical engagement    Degree to which the     4 = Central
(Beck and Cable,       program physically      programming element
2002; Knudson, et      engaged audience
al., 2003; Lewis,      members in a            3 = Occurred multiple
2005; Moscardo,        participatory           times
1999; NPS Module       experience; i.e.,
101; Sharpe, 1976;     through touching or     2 = Minimal effort to
Tilden, 1957)          interacting with        engage
                       resource.
                                               1 = No efforts

Verbal engagement      Degree to which the     5 = Central
(Knudson, et al.,      program verbally        programming element
2003; Moscardo,        engaged audience
1999; Sharpe, 1976;    members in a            4 = Occurred multiple
Tilden, 1957;          participatory           times
Veverka, 1998)         experience; i.e.,
                       dialogue (a two-way     3 = Modestly engaged
                       discussion).
                                               2 = Minimal effort to
                                               engage

                                               1 = No efforts

Cognitive engagement   Degree to which the     5 = Central
(Knudson, et al.,      program cognitively     programming element
2003; Moscardo,        engaged audience
1999; Sharpe, 1976;    members in a            4 = Occurred multiple
Tilden, 1957;          participatory           times
Veverka, 1998)         experience beyond
                       simply listening;       3 = Modestly engaged
                       i.e. calls to imagine
                       something, reflect,     2 = Minimal effort to
                       etc.                    engage

                                               1 = No efforts

Multiple activities    Degree to which the     4 = 2+ primary
(Knapp and Benton,     program consisted of    activities included
2004; Moscardo,        a variety of
1999; Ward and         activities and          3 = 2+ secondary
Wilkinson, 2006)       opportunities for       activities included
                       direct audience
                       involvement (not        2 = One secondary
                       including dialogue).    activity included

                                               1 = One activity only

Props (Jacobson,       A visual aide beyond    1 = Prop(s) used
1999; Knapp and        a screen-based
Benton, 2005; Ham,     slideshow.              0 = Not used
1992; Ward and
Wilkinson, 2006)

Relevance to           Degree to which the     5 = Major focus of
audience (Beck and     program explicitly      messaging
Cable, 2002; Brochu    communicated the
and Merriman, 2002;    relevance of the        4 = Well developed
Ham, 1992; Jacobson,   subject to the lives    efforts
1999; Knapp and        of the audience.
Benton, 2004; Lewis,                           3 = Moderate efforts
2005; Moscardo,
1999; NPS Module                               2 = Minimal efforts
101; Sharpe, 1976;
Tilden, 1957;                                  1 = No efforts
Veverka, 1998)

Affective messaging    Degree to which the     5 = Central
(Jacobson, 1999;       program communicated    programming element
Lewis, 2005; Madin     emotion (in terms of
and Fenton, 2004;      quantity, not           4 = Frequent and
Tilden, 1957; Ward     quality).               repeated messages
and Wilkinson, 2006)
                                               3 = Occasional
                                               messages

                                               2 = Minimal effort to
                                               include messages

                                               1 = Messages absent

Fact-based messaging   Degree to which the     1 = Messaging was
(Frauman and Norman,   program communicated    solely fact-based
2003; Jacobson,        factual information.
1999; Lewis, 2005;                             0 = Messaging was not
Tilden, 1957; Ward                             solely fact-based
and Wilkinson, 2006)                           (incorporated
                                               affective messaging)

Surprise (Beck and     Degree to which the     3 = Major element
Cable, 2002;           program used the
Moscardo, 1999)        element of surprise     2 = Minor element
                       in communication.
                       This could include      1 = Not used
                       "aha" moments or
                       unexpected or
                       contrasting messages.

Novelty (Beck and      Degree to which the     3 = Major element
Cable, 2002; Frauman   program presented
and Norman, 2003;      novel ideas,            2 = Minor element
Knapp and Benton,      techniques, or
2004; Moscardo,        viewpoints as an        1 = Not used
1999)                  element of
                       communication; i.e.,
                       using a device not
                       usually associated
                       with or related to
                       resource.

Provocation (Beck      Degree to which the     4 = Powerful and
and Cable, 2002;       program explicitly      explicit inclusion
Brochu and Merriman,   provoked participants
2002; Knudson, et      to personally reflect   3 = Occasional
al., 2003; Tilden,     on content and its      inclusion
1957)                  deeper meanings.
                                               2 = Isolated or vague
                                               inclusion

                                               1 = No attempt made

Multiple viewpoints    Degree to which the     3 = Multiple
(Beck and Cable,       program explicitly      viewpoints developed;
2002; Brochu and       acknowledged multiple   none given clear
Merriman, 2002;        perspectives or         priority
Tilden, 1957)          uncertainty within a
                       theme or message.       2 = Primarily one
                       (Primarily for          viewpoint, with some
                       controversial           focus on others
                       messaging; when an
                       argument is made, was   1 = No effort
                       a relevant counter-
                       argument provided?)     NA = not applicable

Holistic               Degree to which the     5 = Holistic story
storytelling (Beck     program aimed to        used throughout; all
and Cable, 2002;       present a holistic      messaging tied to
Larsen, 2003;          story (with             story
Tilden, 1957)          characters and a
                       plot) as opposed to     4 = Holistic story
                       disconnected pieces     present; some info
                       of information.         did not relate to
                                               story

                                               3 = Equal mix of
                                               storytelling and
                                               factual information,
                                               no single, holistic
                                               story

                                               2 = Factual
                                               information primarily
                                               used; some stories
                                               used to create
                                               relevance.

                                               1 = Facts and
                                               information
                                               primarily; no attempt
                                               at storytelling.

Place-based            Degree to which the     5 = Central focus of
messaging (Beck &      program emphasized      messaging
Cable, 2002;           the connection
Knudson, et al.,       between the visitor     4 = Well-developed
2003; Lewis, 2005;     and the site/           connection through
Moscardo, 1999; NPS    resource.               repetition and
Module 101; Sharpe,                            engagement
1976)
                                               3 = Moderately
                                               emphasized through
                                               repetition or
                                               engagement

                                               2 = Slightly
                                               developed verbally

                                               1 = Not developed

Introduction and       Degree to which         4 = Intro and
conclusion linkage     program connected       conclusion were
(Beck and Cable,       conclusion back to      linked in a cohesive
2002; Brochu and       the introduction in     way that enhanced
Merriman, 2002;        an organized or         messaging
Larsen, 2003)          cohesive way (i.e.,
                       program "came full      3 = Intro and
                       circle.")               conclusion were
                                               linked, but didn't
                                               necessarily enhance
                                               messaging

                                               2 = Intro and
                                               conclusion were
                                               weakly linked

                                               1 = Intro and
                                               conclusion were
                                               disconnected from
                                               each other

Clear theme (Beck      Degree to which the     4 = Theme is clearly
and Cable, 2002;       program had a clearly   developed and
Brochu and Merriman,   communicated            communicated
2002; Ham, 1992;       theme(s). A theme is
Jacobson, 1999;        defined as a single     3 = Easy to detect,
Knudson, Cable, and    sentence (not           but not well
Beck, 2003; Larsen,    necessarily             developed
2003; Lewis, 2005;     explicitly stated)
Moscardo, 1999;        that links tangibles,   2 = Difficult to
Sharpe, 1976;          intangibles, and        detect, present but
Veverka, 1998; Ward    universals to           at least somewhat
and Wilkinson, 2006)   organize and develop    ambiguous
                       ideas.
                                               1 = Unclear/not
                                               present

Central message        Degree to which         4 = Clearly
(Beck and Cable,       program's message(s)    communicated and well
2002; Brochu and       was clearly             developed
Merriman, 2002; Ham,   communicated; i.e.,
1992; Jacobson,        the "so what?"          3 = Easy to detect,
1999)                  element of the          but not well
                       program.                developed

                                               2 = Difficult to
                                               detect, ambiguous

                                               1 = Unclear/not
                                               present

Consistency (Beck      Degree to which the     3 = Consistent
and Cable, 2002;       program's tone and
Ham, 1992)             quality were            2 = Some shift in
                       consistent throughout   either tone or
                       the program             quality during the
                                               program

                                               1 = Shift in both
                                               tone and quality

Pace (Jacobson,        Degree to which the     Categorical:
1999)                  pace of the program
                       allowed for clarity     Too fast
                       and did not detract
                       from the program.       Too slow

                                               Just fine

Quality of the         Degree to which the     3 = Contextually
resource               resource where          iconic or grandiose
                       program took place is
                       awe-inspiring or        2 = Pleasant but not
                       particularly iconic.    iconic

                                               1 = Unimpressive/
                                               generic

Unexpected negative    Were there any          1 = Yes
circumstance           unexpected
                       interruptions or        0 = No
                       emergencies during
                       the program, such as
                       a sudden change in
                       weather, medical
                       emergency, technical
                       difficulties, or
                       hazardous conditions
                       that detracted from
                       the quality of the
                       program?

Unexpected positive    Was there an            1 = Yes
circumstance           unexpected experience
                       that occurred during    0 = No
                       the program, such as
                       seeing charismatic
                       wildlife or other
                       unique phenomena that
                       added significantly
                       to the quality of the
                       experience?

Behavioral theory elements

The following were only measured for programs in which a behavioral
change was expressed by the interpreter as a desired program outcome.

Benefits of action     Degree to which the     4 = Explicitly/
(Ajzen, 1991; Ham      program emphasized      purposefully
et. al., 2007;         the potential           emphasized
Jacobson, 1999;        benefits resulting
Knudson, et al.,       from performing a       3 = Mentioned a
2003; Moscardo,        particular action(s).   moderate amount
1999; Peake et. al,
2009)                                          2 = Explained a
                                               little

                                               1 = No mention

                                               NA = not applicable

Costs of action        Degree to which the     4 = Explicitly/
(Ajzen, 1991; Ham      program emphasized      purposefully
et. al., 2007;         the potential costs     emphasized
Jacobson, 1999;        resulting from
Knudson, et al.,       performing a            3 = Mentioned a
2003; Moscardo,        particular action(s).   moderate amount
1999; Peake et. al,
2009)                                          2 = Explained a
                                               little

                                               1 = No mention

                                               NA

Norms of action        Degree to which the     4 = Explicitly/
(Ajzen, 1991; Ham      program emphasized      purposefully
et. al., 2007;         the social              emphasized
Jacobson, 1999;        acceptability of
Knudson, et al.,       performing a            3 = Mentioned a
2003; Moscardo,        particular behavior     moderate amount
1999)                  or desired action.
                                               2 = Explained a
                                               little

                                               1 = No mention

                                               NA

Ease of action         Degree to which the     4 = Explicitly/
(Ajzen, 1991; Ham      program communicated    purposefully
et. al., 2007;         the ease (or            emphasized
Jacobson, 1999;        difficulty) of
Knudson, et al.,       performing a            3 = Mentioned a
2003; Moscardo,        particular behavior     moderate amount
1999; Tilden, 1957)    or desired action.
                                               2 = Explained a
                                               little

                                               1 = No mention

                                               NA

Demonstrates action    Degree to which the     4 = Majority of
(Ajzen, 1991; Beck     program provided        audience engaged
and Cable, 2002;       examples of, or
Knudson, et al.,       opportunities for,      3 = Demonstration by
2003; Moscardo,        performing a desired    ranger or small
1999; Sharpe, 1976;    action.                 proportion of
Widner Ward and                                audience
Wilkinson, 2006)
                                               2 = Verbal
                                               description

                                               1 = No mention/
                                               demonstration

                                               NA

Table 4. Interpreter characteristics observed in the study, their
definitions, and operationalization.

Interpreter            Definition              Scoring
characteristic

Professional           The extent to which     0 = Interpreter
appearance             the interpreter         appears disheveled or
                       appears properly        unkempt and is not
                       dressed and groomed.    professionally
                                               dressed

                                               1 = Interpreter
                                               appears well-groomed
                                               and is professionally
                                               dressed

Comfort of the         Degree to which the     1 = Interpreter seems
interpreter            interpreter             scared, nervous, or
(Lewis 2008;           presenting the          unable to lead the
Moscardo, 1999; Ward   program seems           program
and Wilkinson, 2006)   comfortable with the    2 = Interpreter seems
                       audience and capable    nervous and struggles
                       of successfully         with much of the
                       presenting the          program
                       program without
                       apparent signs of       3 = Interpreter seems
                       nervousness or self-    comfortable, but
                       doubt.                  might become
                                               uncomfortable at
                                               times

                                               4 = Interpreter is
                                               not nervous and
                                               handles the program
                                               with ease

Responsiveness         The extent to which     NA = Not able to
(Jacobson, 1999;       the interpreter         observe (e.g., large
Knudson et al.,        interacts with the      programs in dark
2003; Lewis, 2008)     audience, collects      theatres)
                       information about
                       their interests and     1 = Interpreter is
                       backgrounds, and        aloof or averse to
                       responds to their       the visitors'
                       specific questions      presence
                       and requests or non-
                       verbal cues.            2 = Interpreter is
                                               somewhat responsive
                                               to visitors'
                                               questions/body
                                               language

                                               3 = Interpreter was
                                               very responsive to
                                               the audience

Inequity               The presence of         1 = Interpreter did
(Ham and Weiler,       unequal attention       not pay equal
2002)                  devoted to certain      attention to all
                       attendees and not       audience members.
                       others through
                       greater interaction     0 = No inequity
                       or attentiveness.       issues.

Humor quality          How funny is the        1 = Not funny at all
(Ham and Weiler,       interpreter overall?
2002; Knapp and        Does the audience       2 = A little funny
Yang, 2002; Regnier    react positively to
et al., 1992)          the interpreter's use   3 = Moderately funny
                       of humor and seem to
                       enjoy it?               4 = Hilarious

Humor quantity         The extent to which     1 = Interpreter
                       the interpreter         attempts no humor
                       attempts to use         throughout the
                       humor, sarcasm, or      presentation
                       jokes to share the
                       topic with the          2 = Interpreter
                       visitor, regardless     rarely uses humor
                       of their success.
                                               3 = Interpreter uses
                                               an equal mix of humor
                                               and non-humor to
                                               convey the message

                                               4 = Interpreter is
                                               mostly trying to be
                                               humorous

                                               5 = Interpreter uses
                                               humor as the primary
                                               vehicle to convey
                                               their message

Sarcasm                The degree to which     1 = Not at all
                       the interpreter used
                       sarcasm (the use of     2 = Done to some
                       mocking,                extent
                       contemptuous, or
                       ironic language or      3 = A central feature
                       tone) or self-          of the delivery style
                       deprecation that was
                       not meant to be
                       serious, as a part of
                       presenting their
                       program.

Charisma (Ward and     A general sense of      1 = Not likeable/
Wilkinson, 2006)       the overall             found interpreter
                       likeability/charisma    irritating
                       of the interpreter,
                       commonly recognized     2 = Somewhat off-
                       by seemingly genuine    putting
                       interaction with the
                       visitors, including     3 = Neither liked or
                       smiling, looking        disliked interpreter
                       people in the eye,
                       and having an overall   4 = More or less
                       appealing presence.     liked interpreter

                                               5 = Found interpreter
                                               very likeable/
                                               charismatic

Sincerity              The degree to which     1 = Interpreter
(Ham, 2009)            the interpreter seems   seemed to only be
                       genuinely invested in   going through the
                       the messages he or      motions, with no real
                       she is communicating,   emotional connection
                       as opposed to           or sincerity
                       reciting information,
                       and seems sincere in    2 = Interpreter
                       the emotional           seemed somewhat
                       connection they may     connected through the
                       exude to the message    words they used,
                       and/or the resource.    though their
                       In other words, the     mannerisms or
                       extent to which the     intonation didn't
                       interpretation was      corroborate their
                       delivered through       words
                       authentic emotive
                       communication.          3 = Interpreter
                                               seemed mostly sincere
                                               with authentic
                                               emotive communication
                                               for most of the
                                               program

                                               4 = Communication was
                                               clearly sincere and
                                               authentic throughout
                                               the program, as
                                               evidenced by words,
                                               gestures, intonation,
                                               or other mannerisms

Passion                The interpreter's       1 = Interpreter seems
(Beck and Cable,       apparent level of       completely detached/
2002; Ham and          enthusiasm for the      disinterested from
Weiler, 2002;          material, as opposed    the program
Moscardo, 1999)        to a bored or
                       apathetic attitude      2 = Low levels of
                       toward it. The          passions
                       overall vigor with
                       which the material is   3 = Interpreter shows
                       presented.              moderate levels or
                                               sporadic instances of
                                               passion

                                               4 = Pretty high
                                               levels of passion
                                               overall

                                               5 = Interpreter seems
                                               extremely passionate
                                               about the program

Personal sharing       The degree to which     1 = Interpreter did
(Jacobson, 1999;       the interpreter         not share any
Myers et al., 1998)    shared personal         personal information
                       insights or             about themselves with
                       experiences, answered   the audience
                       questions about
                       themselves for the      2 = Interpreter
                       audience, or provided   shared minimal
                       their own opinion on    personal information
                       topics or events        or viewpoints
                       relevant to the
                       program.                3 = Interpreter
                                               shared a large amount
                                               of personal
                                               information and
                                               perspective

                                               4 = Interpreter's
                                               personal life/point
                                               of view is explicitly
                                               the central focus of
                                               the experience (used
                                               themselves as the
                                               primary framework for
                                               the program)

Apparent knowledge     The degree to which     1 = Interpreter seems
(Ham and Weiler,       the interpreter         not at all
2002; Lewis, 2008;     appears to know the     knowledgeable (unsure
Ward and Wilkinson,    information involved    of facts or has a
2006)                  in the program, the     hard time recalling
                       answers to visitors     the information
                       questions, and has      intended for the
                       local knowledge of      program)
                       the area and its
                       resources.              2 = Interpreter seems
                                               somewhat
                                               knowledgeable, but
                                               appears to forget a
                                               few things or leave
                                               out important details

                                               3 = Interpreter
                                               appears more or less
                                               knowledgeable without
                                               any major hiccups or
                                               uncertainty
                                               throughout the
                                               program.

                                               4 = Interpreter's
                                               presentation of facts
                                               and information
                                               during the program is
                                               flawless

Audibility             The extent to which     1 = Interpreter could
                       the interpreter can     not be heard by the
                       clearly be heard and    audience during the
                       understood by the       majority of the
                       audience.               program

                                               2 = Interpreter could
                                               be clearly heard for

                                               the majority of the
                                               program, but wasn't
                                               audible during some
                                               parts

                                               3 = Interpreter could
                                               be clearly heard
                                               throughout the entire
                                               program

Eloquence              The extent to which     1 = Interpreter
(Lewis, 2008)          the interpreter spoke   stumbled on their
                       clearly and             speech throughout
                       articulately, and did   their entire program
                       not mumble or           and was hard to
                       frequently use filler   understand
                       words such as "um" or
                       "like."

                                               2 = Interpreter had
                                               some minor issues
                                               with mumbling or
                                               unclear speech

                                               3 = Interpreter had
                                               no such issues during
                                               the program

                                               4 = Interpreter was
                                               exceptionally
                                               eloquent

Impatience             Did the interpreter     1 = Interpreter was
                       show any explicit       explicitly impatient
                       impatience toward       with the audience
                       audience members?

                                               0 = No issues noted

Formality              The degree to which     1 = Interpreter was
                       the interpreter was     extremely casual
                       very formal and
                       official vs. casual     2 = More casual than
                       and laid back about     formal
                       the presentation.
                                               3 = Interpreter was
                                               neither explicitly
                                               casual nor formal

                                               4 = More formal than
                                               casual

                                               5 = Interpreter was
                                               entirely formal

False assumption of    At any point during     1 = No problem with
the audience           the program, did the    false assumptions
                       interpreter make
                       assumptions of the      2 = Some minor false
                       audience's attitudes    assumptions that

                       or knowledge that       likely did not
                       could have easily       detract from the
                       been false?             quality of the
                                               program

                                               3 = Obvious false
                                               assumptions that made
                                               the experience less
                                               enjoyable or
                                               meaningful

Character acting       The degree to which     0 = Interpreter does
                       role playing or         no character role
                       character acting is     playing during the
                       incorporated into the   program, he/she is
                       program, either to      simply leading the
                       add authenticity or     program
                       to help tell a story.
                                               1 = Interpreter acts
                                               like one or more
                                               characters during
                                               parts of the program

                                               2 = Interpreter is in
                                               full costume or does
                                               not break character
                                               at any point during
                                               the program

Primary identity       Friend: outwardly       1 = primary identity;
(Ham and Weiler,       friendly, casual,       0 = not
2002; Ham, 2002;       approachable, mingles
Knapp and Yang,        informally
2002; Larsen, 2003;
Mills, 1920; Wallace   Authority figure:       1 = primary identity;
and Gaudry, 2005)      emphasizes own role     0 = not
                       as a park ranger and
                       focuses on rules,
                       regulations, and/or
                       authority to
                       communicate

                       Walking encyclopedia:   1 = primary identity;
                       Focused on conveying    0 = not
                       a large volume of
                       facts

Questionable           Obvious factual         1 = present
information            inaccuracy (incorrect
                       or inaccurate           0 = not present
                       information) or false
                       attribution
                       (unfounded claims
                       about others, e.g.,
                       "the native people
                       were happy to hand
                       over their land so a
                       National Park could
                       be formed.")

Bias                   Did the interpreter     1 = yes
                       share any apparent
                       bias or strong          0 = no
                       opinion with
                       potential effects on
                       relationships with
                       audience members?

Table 5. Outcome indexes developed through confirmatory factor
analyses.

OUTCOME INDEXES

Program outcome: Visitor Experience and Appreciation (Cronbach's
[alpha] = 0.89)

To what extent did the program you just attended influence any of the
following for you?

* Made my visit to this park more enjoyable

* Made my visit to this park more meaningful

* Enhanced my appreciation for this park

* Increased my knowledge about the program's topic

* Enhanced my appreciation for the National Park Service

Program outcome: Behavioral intentions (Cronbach's [alpha] = 0.94)

To what extent did the program you just attended influence any of
the following for you?

* Changed the way I will behave while I'm in this park

* Changed the way I will behave after I leave this park

Table 6. Independent variable indexes developed through exploratory
factor analyses.

INDEPENDENT VARIABLE INDEXES

Interpreter characteristic: Confidence (Cronbach's [alpha] = 0 .70)

* Comfort of the Interpreter

* Apparent knowledge

* Eloquence

Interpreter characteristic: Authentic emotion and charisma
(Cronbach's [alpha] = 0.85)

* Passion

* Charisma

* Sincerity

Program characteristic: Organization (Cronbach's [alpha] = 0.82)

* Quality of the introduction

* Appropriate sequence

* Effective transitions

* Holistic story

* Clarity of theme

* Link between introduction and conclusion
Program characteristic: Connection (Cronbach's [alpha] = 0.88)

* Links to intangible meanings and universal concepts

* Cognitive engagement

* Relevance to audience

* Affective messaging

* Provocation

Table 7. Means and standard deviations of outcome variables measured
in visitor surveys.

Variable (Scale)                     Means (with standard deviations)

                                     Five or more      Fewer than
                                       attendees      five attendees

Satisfaction (0 to 10)                8.96 (0.68)      9.02 (0.89)

Visitor experience and                4.41 (0.32)      4.57 (.042)
  appreciation (1 to 5)

* Made my visit to this park more     4.55 (0.30)      4.70 (0.43)
  enjoyable (1 to 5)
* Made my visit to this park more     4.49 (0.32)      4.69 (0.45)
  meaningful (1 to 5)
* Enhanced my appreciation for        4.36 (0.37)      4.51 (0.51)
  this park (1 to 5)
* Increased my knowledge about the    4.45 (0.34)      4.62 (0.47)
  program's topic (1 to 5)
* Enhanced my appreciation for the    4.27 (0.36)      4.38 (0.58)
  National Park Service (1 to 5)

Behavioral intentions (1 to 5)        2.92 (0.64)      3.02 (0.98)

* Changed the way I will behave       2.92 (0.67)      3.08 (0.97)
  while I'm in this park (1 to 5)
* Changed the way I will behave       2.92 (0.61)      2.97 (1.04)
  after I leave this park (1 to 5)

Table 8. Intended outcomes expressed by interpreters immediately
prior to their programs.

I want my audience to ...                                 Proportion
                                                          expressing
                                                             each
                                                           outcome

Have an increased knowledge of the program topic            79.5%
Have an increased appreciation for this park                56.4%
Have an increased understanding of the park's resources     39.1%
Want to learn more about the program topic                  24.8%
Be entertained                                              15.6%
Have an increased appreciation of the NPS                   14.1%
Have an increased concern for a specific topic              11.5%
Change their attitudes toward something                     10.6%
Change a certain behavior in the future                      7.0%
Develop and practice a new skill                             3.5%

Table 9. How interpretive programs were developed.

Program development                                      Proportion
                                                         expressing
                                                            each

Program provided for ranger with full script planned         <1%
  out
Program provided for ranger with some freedom to             14%
  inject own style

Program topic provided, little restrictions on               20%
  information or style to be presented
General topic suggested, but wrote own script and            53%
  selected information
Interpreter selected and developed entire program free       13%
  of restrictions

Table 10. Means and standard deviations of ordinal interpreter
delivery styles.

Variable (Scale)                   Means (with standard deviations)

                                   Five or more      Fewer than
                                     attendees      five attendees

Confidence index (1 to 4)           3.28 (0.49)      3.12 (0.41)

* Comfort of the interpreter        3.49 (0.60)      3.25 (0.63)
  (1 to 4)
* Apparent knowledge (1 to 4)       3.45 (0.63)      3.40 (0.59)
* Eloquence (1 to 4)                2.99 (0.65)      2.83 (0.50)

Authentic emotion and charisma      3.57 (0.85)      3.46 (0.70)
index (1 to 5)

* Passion (1 to 5)                  3.23 (1.02)      3.08 (1.04)
* Charisma (1 to 5)                 3.82 (0.86)      3.68 (0.69)
* Sincerity (1 to 4)                2.93 (0.77)      2.88 (0.65)
Responsiveness (1 to 3) (a)         2.81 (0.41)      2.82 (0.45)
Humor quality (1 to 4)              2.08 (0.73)      1.92 (0.58)
Humor quantity (1 to 5)             2.08 (0.72)      1.85 (0.53)
Personal sharing (1 to 4)           1.68 (0.72)      1.79 (0.73)
Audibility (1 to 3)                 2.86 (0.36)      2.85 (0.36)
Formality (1 to 5)                  3.21 (0.86)      3.00 (0.68)
Sarcasm (1 to 3)                    1.23 (0.46)      1.15 (0.36)
False assumptions of audience (1 to 1.17 (0.40)      1.08 (0.27)

(a) Responsiveness was not observable in every case. For larger
programs, n = 245.

Table 11. Descriptive statistics of interpreter delivery styles
(categorical variables).

Interpreter delivery style                   % of programs in which
                                             delivery style occurred

                                             Five or     Fewer than
                                               more        five
                                             attendees   attendees

Professional appearance of the interpreter     98.2        100.0

Inequitable treatment of audience               2.9          2.5
Impatience                                      1.8          2.5
Primary identity: Friend                       18.0         37.5
Primary identity: Authority                     4.4          2.5
Primary identity: Walking encyclopedia         76.8         67.5
Character acting: partial                       2.6          2.5
Character acting: complete                      2.9          0.0
Interpreter bias                                3.3          7.5
Questionable information                        9.9          2.5

Table 12. Means and standard deviations of ordinal program
characteristics.

Variable (Scale)                     Means (with standard deviations)

                                     Five or more      Fewer than
                                       attendees      five attendees

Organization index (1 to 5)           3.34 (0.71)      3.14 (0.65)

* Quality of introduction (1 to 3)    2.13 (0.45)      1.93 (0.42)
* Appropriate sequence (1 to 4)       2.79 (0.69)      2.70 (0.69)
* Transitions (1 to 4)                2.72 (0.76)      2.55 (0.71)
* Holistic story (1 to 5)             2.78 (1.01)      2.78 (0.77)
* Conclusion linked to intro          2.63 (0.86)      2.48 (0.75)
  (1 to 4)
* Clear theme (1 to 4)                2.82 (0.86)      2.58 (0.90)

Connection index (1 to 5)             2.77 (0.78)      2.74 (0.55)

* Links to intangible meanings and    2.88 (0.94)      3.00 (0.80)
  universal concepts (1 to 5)
* Cognitive engagement (1 to 5)       2.85 (0.94)      2.78 (0.83)
* Relevance to audience (1 to 5)      2.86 (0.86)      2.70 (0.69)
* Affective messaging (1 to 5)        2.43 (0.95)      2.38 (0.71)
* Provocation (1 to 4)                2.24 (0.72)      2.25 (0.67)
Clear message (1 to 4)                2.20 (0.94)      2.00 (0.85)
Appropriate logistics (1 to 4)        3.11 (0.93)      3.15 (0.89)
Appropriate for the audience          3.93 (0.70)      4.15 (0.83)
  (1 to 5)
Multisensory (1 to 3)                 2.39 (0.51)      2.35 (0.48)
Physical engagement (1 to 4)          1.42 (0.69)      1.50 (0.75)
Verbal engagement (1 to 5)            2.51 (1.02)      2.68 (0.80)
Surprise (1 to 3)                     1.10 (0.31)      1.03 (0.16)
Novelty (1 to 3)                      1.18 (0.43)      1.10 (0.30)
Consistency (1 to 3)                  2.88 (0.37)      2.88 (0.34)
Resource quality (1 to 3)             2.37 (0.70)      2.13 (0.69)
Multiple viewpoints (1 to 3) (a)      2.63 (0.51)      2.61 (0.50)

Behavioral theory elements (b)

Benefits of action (1 to 4)           2.52 (0.63)      2.80 (0.45)
Costs of action (1 to 3)              1.97 (0.75)      2.40 (0.89)
Norms of action (1 to 3)              1.48 (0.57)      1.40 (0.55)
Ease of action (1 to 3)               1.81 (0.65)      1.20 (0.45)
Demonstrates action (1 to 4)          2.13 (0.96)      2.20 (1.30)

(a) Multiple viewpoints were not appropriate or relevant in every
case (e.g., a talk on butterfly life cycles). We only observed this
variable where it seemed potentially relevant (n = 94 for larger
programs; n = 22 for smaller programs).

(b) These variables are explicitly associated with behavioral
change theory. As such, they were only observed on a small subset
of cases within the sample where specific behaviors were discussed
by the interpreter (n = 31 for larger programs; n = 5 for smaller
programs).

Table 13. Descriptive statistics of program characteristics
(categorical variables).

Program characteristics            % of programs w program
                                   characteristic was observed

                                   Five or more   Fewer than
                                    attendees        five
                                                   attendees

Fact-based messaging                  26.8%          25.0%
Use of props                          30.5%          27.5%
Pace too fast                          6.2%           5.0%
Pace too slow                          9.2%           5.0%
Pace just right                       84.6%          90.0%
Unexpected positive circumstance       1.8%           2.5%
Unexpected negative circumstance      15.8%          10.0%

Table 14. Pearson correlations between ordinal independent variables
and visitor outcomes for programs with five or more attendees.

Variable                    Satisfaction     Visitor      Behavioral
                                            experience    intentions

                                               and
                                           appreciation

Interpreter style:            .479 **        .277 **        .174 **
  Confidence index
Interpreter style:            .423 **        .303 **        .182 **
  Authentic emotion and
  charisma index
Program characteristic:       .381 **        .378 **        .153 *
  Approp. for audience
Program characteristic:       .362 **        .219 **        .132 *
  Organization index
Program characteristic:       .342 **        .259 **        .124 *
  Connection index
Interpreter style: Humor      .288 **        .233 **        .155 *
  quality
Program characteristic:       .271 **        .281 **        .034
  Consistency
Program characteristic:       .255 **        .281 **        .187 **
  Clear message
Interpreter style:            .241 **         245 **        .061
  Responsiveness
Program characteristic:       .234 **        .240 **        .162 **
  Verbal engagement
Program characteristic:       .216 **        .115           .141 *
  Multisensory engagement
Interpreter style:            .197 **        .134 *         .104
  Audibility
Interpreter style: False     -.172 **       -.197 **       -.088
  assumption of audience
Program characteristic:        170 **        .245 **        .165 **
  Appropriate logistics
Program characteristic:       .150 *         .151 *         .127 *
  Surprise
Program characteristic:       .145 *         .024           .014
  Novelty
Interpreter style: Humor      .144 *         .097           .062
  quantity
Program characteristic:       .074           .120 *         .061
  Physical engagement
Interpreter style:           -.069          -.155 *        -.023
  Formality
Interpreter style:            .105           .053          -.114
  Sarcasm
Program characteristic:       .077           .068           .065
  Quality of the resource
Interpreter style:            .035           .048           .112
  Personal sharing
Program characteristic:       .031           .157           .128
  Multiple points of view

Table 15. Statistically significant t-tests results,
comparing the means of visitor outcome scores for
selected categorical variables for programs with five
or more attendees.

Observed
category                  Satisfaction

                  Mean                    Cohen's
                  diff.     t       P       d

Impatience        -0.36   -2.2    0.031   0.68
"Friend"           0.23    2.3    0.023   0.36

"Walking
encyclopedia"

Fact-based        -0.34   -3.9   <0.001   0.50
messaging

Unexpected neg.   -0.29   -2.8    0.006   0.45
circumstance

Observed          Visitor experience and
category                  appreciation

                  Mean                    Cohen's
                  diff.     t      P        d

Impatience        -0.47   -3.3    0.001   1.28
"Friend"

"Walking
encyclopedia"

Fact-based        -0.12   -2.6    0.011   0.36
messaging

Unexpected neg.   -0.19   -3.6   <0.001   0.60
circumstance

Observed
category          Behavioral intentions

                  Mean                    Cohen's
                  diff.     t      P        d

Impatience
"Friend"

"Walking          -0.20   -2.2    0.031   0.32
encyclopedia"

Fact-based
messaging

Unexpected neg.
circumstance

The following categorical variables yielded no
statistically significant differences in visitor
outcomes: Inequitable treatment of the audience,

questionable information, "Authority" identity,
unexpected positive circumstances, use of props.

Table 16. One-way ANOVA comparing outcome variables for programs of
different pace with five or more attendees. Items not sharing the
same superscript are statistically different from one another.

                                 Means

                                Visitor
Pace          Satisfaction   experience and   Behavioral intentions
                             appreciation

Too fast      [8.62.sup.A]   [4.27.sup.AB]        [2.56.sup.A]
Too slow      [8.43.sup.A]    [4.23.sup.A]        [2.84.sup.AB]
Appropriate   [9.03.sup.B]    [4.44.sup.B]        [2.96.sup.B]

Statistics    F = 12.9;      F = 6.9,         F = 3.2, p = 0.042
              p < 0.001      p = 0.001        Cohen's d
              Cohen's d      Cohen's d        (appropriate pace vs.
              (appropriate   (appropriate     others): 0.34
              pace vs.       pace vs.
              others):       others): 0.57
              0.78

Table 17. Independent samples t-tests comparing means of
characteristics for programs that experienced attrition (people left
the program early) vs. those that did not.

                         Program                             Cohen's
Characteristic          attrition?   Means    t       p         d

Responsiveness of the      Yes       2.62    -2.4    0.020     0.46
  interpreter               No       2.83
Audibility                 Yes       2.72    -2.3    0.025     0.49
                            No       2.91
False assumption of        Yes       1.31     2.4    0.020     0.50
  the audience              No       1.11
Appropriate logistics      Yes       2.44    -5.0   <0.001     0.86
                            No       3.23
Confidence                 Yes       3.08    -2.8    0.006     0.46
                            No       3.32
Organization               Yes       3.09    -2.2    0.031     0.32
                            No       3.36

                         Program
Outcomes                attrition?   Means    T       p

Satisfaction               Yes       8.49    -3.9   <.001     0.79
                            No       9.04
Visitor experience         Yes       4.26    -2.6    .014     0.51
  and appreciation          No       4.44
Behavioral intentions      Yes       2.73    -1.8    .070     0.34
                            No       2.95

Table 18. Chi-square tests comparing programs that experience attrition
vs. those that did not.

                                     Pearson              Relation to
Characteristic                     [chi square]    P       attrition
                                    statistic

Interpreter identity: walking          3.6        .058   More attrition
  encyclopedia
Use of props                           12.4       .001   More attrition
Slow pace                              5.8        .026   More attrition
Unexpected negative occurrence         8.9        .006   More attrition

Table 19. Statistically significant t-tests results, comparing the
means of visitor outcome scores for interpreters who expressed
different intended outcomes for their interpretive programs.

                                         Satisfaction
                                 Mean                   Cohen's
Intended outcome                 diff.    t      P         d

Increased knowledge
Increase desire to learn         0.20    2.2   0.029     0.30
Change attitude                  0.18    2.0   0.048     0.31
Increase appreciation for Park   0.22    2.7   0.007     0.34
Increase understanding of
  resource
Increase level of concern
Change visitor behavior

                                 Visitor experience and
                                         appreciation
                                 Mean                   Cohen's
Intended outcome                 diff.    t       P         d

Increased knowledge              -0.12   2.4    0.019    0.37
Increase desire to learn          0.14   3.2    0.002    0.46
Change attitude                   0.16   4.3   <0.001    0.45
Increase appreciation for Park    0.09   2.2    0.028    0.28
Increase understanding of         0.08   2.1    0.040    0.26
  resource
Increase level of concern
Change visitor behavior

                                 Behavioral intentions
                                 Mean                   Cohen's
Intended outcome                 diff.    t      P         d

Increased knowledge
Increase desire to learn
Change attitude
Increase appreciation for Park
Increase understanding of
  resource
Increase level of concern        0.27    2.2   0.032     0.41
Change visitor behavior          0.41    2.7   0.008     0.66