Research Guide
Outdoor Education Meta-analyses

Meta-Analytic Research on the Outcomes of Outdoor Education

James Neill
Last updated:
05 February, 2008

James T. Neill
University of New Hampshire

Paper presented to the 6th Biennial Coalition for Education in the Outdoors Research Symposium, Bradford Woods, IN, 11-13 January, 2002

Updated January, 2008
 

Abstract

The impacts of outdoor education programs on participants remains a topic of much interest to practitioners and researchers. Hundreds of empirical pre-post studies have been conducted in an attempt to better understand the impacts. Typically, psychosocial constructs such as self-concept are measured before and after outdoor education programs. In order to summarize the results of such studies, five relevant meta-analyses have been conducted (Bunting & Donley, 2002; Cason & Gillis, 1994; Hans, 2000; Marsh, P. E., 1999; Marsh, H. W., Neill, & Richards, 1997). These research reviews show that education programs have small-moderate impacts on constructs such as self-concept, locus of control, and teamwork. Impressively, the effects appeared to be retained over time. The most effective programs were conducted by particular organizations, involved adult participants, and were longer. Comparisons to meta-analyses for other types of interventions revealed that outdoor education program impacts were similar to other self-concept change programs and other psychological interventions, but not as powerful as individual psychotherapy. Recommendations that emerged from this research review were that future research should seek to understand more about why different programs and different participants achieve different outcomes and for effect sizes to be used by outdoor education programs to monitor educational effectiveness.
 

What are the effects of outdoor education?

Are people any different after participating in challenging adventure activities? Which programs or participants have the most dramatic effects? What are the effects of outdoor education? Such questions have attracted animated discussion since modern forms of outdoor education started appearing over 50 years ago.

Typically, the effects of outdoor education have been in the form of testimonial support, anecdotal examples, and passionate rhetoric. For example, the promotional material of outdoor education organizations invariably communicates highly positive claims and examples, outdoor education instructors’ verbiage with clients exhibits highly optimistic and positive hopes, and a surprising amount of so-called academic work has a clear tendency to uncritically promote the view that outdoor education programs are good things for people.

For most outdoor education advocates the ready-at-hand, ad hoc evidence is sufficient to support their enthusiastic claims about outdoor education. Indeed, the smile when a participant conquers a challenge and the glowing words spoken during a debrief are often compelling. But is it enough to let such evidence speak for itself?

A list of claims about the effects of outdoor education would be very long. Yet for most advocates of outdoor education, what they have personally seen, heard and felt provides sufficient evidence for many wider claims.

Such a critique of outdoor educators may seem harsh. But professional outdoor educators should know more about research results, as should all stakeholders in outdoor education programs and related fields of endeavor. Indeed, so should researchers of outdoor education! To date, however, there has been a lack of high quality research reviews that have been written for both researchers and practitioners. Thus the current article undertakes to compare and summarize the findings of empirical outcome research on outdoor education and related programs for both general audiences. A challenging aim for the article is to provide maximum information about the empirical outcome findings in outdoor education in a minimum of space.
Initially this paper overviews two ways of measuring outcomes, then describes the use of meta-analysis as a method for objectively summarizing pre-post empirical research results, and finally provides recommendations for future outdoor education research and evaluation studies.
 

Ways of researching outcomes

There have been two main objective ways that outdoor education outcomes have been researched.

Post-program Surveys

One of the more common has been to directly ask participants at the end of a program something like “How valuable was this program for your personal growth?”. On face value this approach has appeal. Indeed, it appears to be a prevalent method. In a recent survey of the practices of over 100 Australian outdoor educators approximately 80% reported that their program used written end-of-program surveys, or ‘happy sheets’ as they are colloquially known (Richards, Neill, & Butters, 1997). However, such evaluations are notoriously problematic indicators of program effectiveness. For example, a study by Neill, Richards and Badenoch (1997) found little relationship between staff and participants’ ratings of amount of personal growth experienced by participants (correlations between .2 and .4). Similarly low correlations were found between participants’ end-of-program self-reports and their pre-post changes on reliable, psychometric questionnaires.

End-of-program reports of the impact of the value of an experience are one valuable source of evidence. However, post-hoc reports are vulnerable to many potential distortions. For example, scores can be inflated by post-group euphoria (Marsh, Richards, & Barnes, 1986), post-purchase rationalization (or post-commitment justification), group think, and other problems related to difficulties in validity remembering, understanding, and then accurately reporting on what happened during an experience and resultant effect on one’s psyche.

Disparities between different ways of assessing intervention programs is a broader phenomenon. For example, in a review of management training research, it was claimed that the “reaction of trainees bore no relationship to the effectiveness of the training” (Bright, cited in Lawson, 1997, p. 7).

To date very little research has explored the relationships between what stakeholders, such as staff and participants, say about a program’s value and other indicators of the program effectiveness. Thus, the validity of post-experience ratings remains an area for further investigation.

Pre-post Testing

A second major approach to examining the effectiveness of outdoor education has been to gather participants’ self-perceptions (or observer perceptions) before and after a program and then test to see if there are any differences. Hundreds of empirical evaluation studies have utilized the basic pre-post design, or variations of it. Such studies have used a wide variety of dependent measures (e.g., self-esteem, locus of control, etc.) and have been conducted on several different types of programs. The validity of this approach depends on important factors such as the quality of measurement tool, the use of control or comparison groups, and whether followup testing is conducted (Campbell, & Stanley, 1963).

The variety in the methods and measures of outdoor education pre-post outcome research makes the task of reviewing the literature formidable. There have been several notable reviews of the empirical research (e.g., Barrett & Greenaway, 1995; Burton, 1981; Crompton & Sellar, 1981; Ewert, 1982; Gillis, 1992; Gillis & Thomsen, 1996; Godfrey, 1974; Iida, 1975; Reddrop, 1997; Richards, 1977; Shore, 1977), however they have all faced difficulties in presenting a systematic overview of the outcome research. Perhaps an indicator of this is that there is only one review which has been published in a peer-reviewed journal (Crompton & Sellar, 1981) and it is over 20 years old.

One of the main difficulties faced by these reviews has been to effectively synthesize and systematically summarize the empirical findings from the variety of empirical studies. Cason and Gillis (1994) described the problem that “the number of research endeavors in adventure programming has grown so large that qualitative literature reviews (like annotated bibliographies) are no longer sufficient tools for understanding the strengths and weaknesses of trying to measure change that results from adventure programming Cason and Gillis” (1994, p. 40).

Since the 1980’s, a new, systematic review method appeared in the form of meta-analysis. Meta-analysis provides a method for entering the empirical results reported by different studies into a central database which is then analyzed and summarized. Well conducted meta-analyses allows for more objective appraisal of the outcome evidence, which can lead to clearer understanding about the nature of outdoor education’s impacts on participants. To date, five meta-analytic reviews of outdoor education and related outcome literature have been conducted, and it is these five reviews which the current article summarizes.
 

Understanding meta-analysis

Meta-analysis is a statistical technique for accumulating and representing the research results reported in various studies. Meta-analysis pools the findings about a single research question from many different sources and analyzes the overall effects.

Meta-analysis results are reported in terms of effect sizes. An ES, as used in this article, is a measure of how much difference existed between ratings at two different points in time (e.g., before and after an outdoor education program). In essence, the effect size indicates the quantified amount of change.

For general interpretation, an effect size of 0 means no change, a negative effect size means a reduction for a measured outcome, and a positive effect size means an increase in a measured outcome. It can also be noted that effect sizes are proportional. Thus, an effect size of .40 represents twice as much change as an effect size of .20.

Various experts have offered suggestions about the meaning of different effect sizes. For example, it has been suggested that an effect size of .50 is a change of practical significance while .25 is educationally significant (Wolf, 1986). Cohen (1977) suggested that .2 was small, .5 moderate, and .8 large. Other authors have warned against global guides to interpreting effect sizes since who and what is being measured needs to be identified before a genuine assessment about the value of a particular effect size is made.

In practice, a small effect size can be very impressive if, for example, the outcome is difficult to change (e.g., a personality construct) or if the outcome is very valuable (e.g., an increase in life expectancy). On the other hand, a large effect size does not necessarily mean that there is any practical value in the change, particularly if it isn’t related to the aims of the intervention (e.g., political orientation).

Meta-analysis as a review method is not without its critics. Criticisms tend to fall into two categories (Bangert-Drowns & Rudner, 1991). Some researchers complain that meta-analysis obscures important information by averaging numerical representations across studies. Other researchers argue that research is best reviewed by a reflective expert who sifts through research and develops kernels of insight from the array of argumentation in a field. Thus, there are advantages and disadvantages to both traditional style and meta-analytic reviewing. An ideal reviewing method, then might involve making use of both meta-analytic methods and traditional review methods.
 

Meta-analytic outdoor education research

Five meta-analyses of the effects of outdoor education and related programs have been conducted (Bunting & Donley, 2002; Cason & Gillis, 1994; Hans, 2000; Marsh, P. E., 1999; Marsh, H. W., Neill, & Richards, 1997; see Table 1). The studies vary considerably in their scope and focus. The major study (Hattie, Marsh, H. W., Neill, & Richards, 1997) included 96 studies, representing over 12, 000 participants, included outdoor education programs results for both youth and adults, coded for pre-program effects, program effects and post-program effects. Cason and Gillis (1994) conducted the first outdoor education meta-analysis, based on pre-post outcomes reported in outdoor education research with adolescents (43 studies). Most of the studies in the Cason and Gillis (1994) study were included in the Hattie et al. (1997), hence the results of the two studies are generally consonant.

The other three outdoor education meta-analyses reviewed the pre-post outcomes for particular types of programs. Marsh, P. E. (1999; 22 studies) focused on the impacts of American camping programs (22 studies). Hans (2000; 24 studies) focused on the effects of adventure programming on locus of control outcomes. Bunting and Donley (2002; 15 studies) focused on the effects of ropes challenge course programs on teamwork, self-concept and self-esteem outcomes.

Table 1

Overall results from five meta-analyses related to outdoor education

Study Focus d N studies N
effects
N participants
Cason & Gillis (1994) Adventure programming for adolescents .31 43 147 2,291
Hattie et al. (1997) Adventure education and Outward Bound programs .34 96 1,728 12,057
Hans (2000) Adventure programming locus of control outcomes .38 24 30 1,632
Marsh, P. E. (1999) Camping programs .20 22 37 na
Bunting & Donley (2002) Ropes Challenge Courses .55 15 na na
Gillis & Speelman (2008) Ropes challenge courses .43 44 390 2,796

Overall Findings

The five outdoor education meta-analyses reported that outdoor education programs has small-moderate effects (see Table 1). The lowest average effect size was reported for American camp programs (.20; Marsh, P. E., 1999) and the highest average effect size was reported for ropes challenge courses (.55; Bunting & Donley, 2002). For the other studies, Hans (2000) reported an effect size of .38, Hattie et al. (1997) reported .34 and Cason and Gillis (1994) reported .31. The findings for the larger studies and the ones focused on outdoor education programs (i.e, Cason & Gillis; Hattie et al., 1997) suggest overall effect sizes between .3 and .4. By broader educational and psychological standards this is be considered a small-moderate amount of change (Cohen, 1977).

Camping programs’ seemingly low effect size of .20 may not be unexpected, given that not all camps aim to enhance campers’ personal and/or social development. Indeed, camping programs that had a personal development philosophy had an effect size of .41. It should also be noted that Bunting and Donley’s (2002) relatively high overall effect size of .55 for ropes challenge courses was based on only 15 studies, and these studies exhibited a wide range in effect sizes. Further undermining the reliability of this study is issue that a strong negative relationship was found between quality of study and size of outcome; in other words, poorer quality studies tended to have higher ES, a finding also reported by Cason and Gillis (1994).

In summary, a reasonable consensus amongst the five outdoor education meta-analyses is that outdoor education programs have a small-moderate impact for typically measured outcomes such as self-esteem, behavior problems, and teamwork. In an attempt to be more precise, Hattie et al. (1997) reported that their results suggested that 65% of participants were better off for having participated in outdoor education programs.

Do the Effects of Outdoor Education Programs Last?

Hattie et al., (1997) examined the long-term effects of outdoor education programs and reported an additional effect size of .17 for assessments up to 18 months after the end of the program. This finding means that the average program effect size of .34 is followed by an average ongoing improvement of .17, suggesting a moderate overall effect size of .51. This evidence suggests that, on average, outdoor education participants experience additional growth on returning to their home environments. Despite wide effect size variations between different studies, Hattie et al. (1997) noted that the additional .17 is impressive because longitudinal education and training research generally shows a loss over time of immediate program benefits. However, it should also be noted that much of the research on which these long-term results were based was Outward Bound expedition-style programs, thus the generalizability of the finding is limited.

What are the Main Outcomes of Outdoor Education Programs?

While overall outcomes are useful as a general guide, detailed information about the effects is needed. Hattie et al. (1997) grouped the results into six broad categories – leadership, self-concept, academic, personality, interpersonal, and adventuresome (see Table 2).

Table 2

Effect sizes (ESs) by outcome category reported by Hattie et al (1997)

Category

Description

Program ES (N)

Follow-up ES (N)

Leadership

Conscientiousness, Decision making, Leadership, Organisational ability, Time management, Values, Goals

 

.38* (222)

.15 (58)

Self-concept

Physical ability/appearance, Peer relations, General, Academic, Confidence, Self-efficacy, Family, Self-understanding, Well-being, Independence

.28* (271)

.23* (149)

Academic

Mathematics, Reading, Grade Point Average (GPA), Problem solving

.46* (30)

.21 (9)

Personality

Femininity/Masculinity, Achievement motivation, Emotional stability, Aggression, Assertiveness, Locus of control, Maturity, Neurosis reduction

.37* (235)

.14 (76)

Inter-personal

Cooperation, Interpersonal communication, Social competence, Behavior, Relating skills, Recidivism

.32* (176)

.17 (36)

Adventure-some

Challengeness, Flexibility, Physical fitness, Environmental awareness

.38* (69)

-.06 (19)

Note. * p < .05

The pre-post program effect sizes for each of Hattie et al.’s (1997) six categories were statistically significant. Interestingly, the smallest program effect size was for self-concept (.28), however the largest follow-up effect size was for self-concept (.23). This might indicate a sleeper effect whereby self-concept changes in particular are begun during a program and then continue to unfold afterwards.

The effect sizes reported by Cason and Gillis (1994) for programs with adolescents were placed into seven categories – behavioral assessment by others, self-concept, academic, attitude, locus of control, clinical scales, and school attendance (see Table 3). There was a very high effect size (1.05) for studies using clinical measurement scales. This finding could be due to the fact that clinical scales are mainly used for researching specialized and intensive outdoor education programs for participants who have emotional, behavioral or psychological difficulties. Such programs tend to utilize a more highly disciplined and individualized approach than is usually the case in programs with an personal development, educational or recreational emphasis. High effect size for therapeutic and/or behavioral problem clients may also be due to the clients having a greater room for growth than average outdoor education program participants. The results for the other six categories of outcomes reported by Cason and Gillis (1994) ranged between .30 (locus of control) and .61 (academic).

Table 3

Effect sizes (ESs) by outcome category reported by Cason and Gillis (1994)

Category

Program ES (N)

Behavioral assessment by others

.40 (23)

Self-concept

.34 (23)

Academic

.61 (10)

Attitude surveys

.46 (19)

Locus of control

.30 (13)

Clinical scales

1.05 (12)

School attendance

.47 (9)

Hans (2000) focused on a single outcome, locus of control, and this study largely corroborated the findings of Cason and Gillis (1994) and Hattie et al. (1997) for small to moderate shifts towards an internal locus of control. Marsh, P. E. (1999) focused on outcomes for self-constructs (mostly self-esteem and self-concept). Bunting and Donley (2002) reported results for three main outcome categories, team development (ES = .58), self-concept (ES = .23) and self-esteem (ES = .16).

Roughly, the small-moderate average effect size finding applies to different outcomes categories, with the notable exception of clinical scales in the Cason and Gillis (1994) study. The largest followup effects appear to be in the area of self-concept. The overall effects seem to suggest enhancement of self-related constructs, action-orientation, and coping behaviors.

Are All Outdoor Education Programs Effective?

According to Hattie et al. (1997, p.77), outdoor education programs “can obtain notable outcomes and have particularly strong, lasting effects. It is clear, however, that adventure programs are not inherently good. There is a great deal of variability in outcomes between different studies, different programs, and different individuals”.

This theme of wide variability in the effect sizes was also evident in the Cason and Gillis (1994, p. 46) meta-analysis which stated that “when one considers the large range of effect sizes collected, -1.48 to 4.26, and the wide variation in the findings, a singular summary effect size...cannot adequately reflect the effectiveness of adventure programming”.

Other research reviews have come to similar conclusions about variability in outcomes. Burton (1981), for example, reviewed 161 outdoor education studies and reported that 38% had “mostly positive” outcomes, 26% had “some positive” outcomes, 34% had “non-significant” outcomes and 2% had “negative” outcomes.

This variability in outcomes between programs is a striking feature in the outdoor education research and warns against accepting false comfort from the small-moderate average effect sizes. Hattie et al. (1997, p.70) emphasized that “only some adventure programs are effective, and then on only some outcomes, and it is probable that only parts of the programs are influencing these outcomes”.

Moderators of Outdoor Education Outcomes?

Meta-analytic methods of summarizing research can be used to consider the effect of moderator (or process) variables on outcomes. The meta-analytic studies have examined the influence of program length, participant gender, participant age, participant type (e.g., delinquent, management, gender mix of group (all male, all female, co-educational), nature of program (e.g. sailing, wilderness), organization running the program and quality of the study.

Hattie et al. (1997) found that 36% of the variance in outcomes could be explained by the abovementioned factors. Outcomes were most influenced by the organization running the program (Outward Bound Australia, ES = .47; all other outdoor education program, ES = .17), program length (programs greater than 20 days, ES = .41; less than 20 days, ES = .23), and age of participants (adults, ES = .38; students, ES = .21).

Cason and Gillis (1994) found that the quality of the study had an impact on the outcomes, with lower quality studies reporting higher effect sizes, as did Bunting and Donley (2002). This should warn researchers to be wary of using instrumentation and designs which are below par since they may inflate results. This concern becomes even more important in the light of how few programs use high quality assessment tools. According to a recent survey only 7% of outdoor educators reported that their programs utilized standard tests as part of their evaluation strategy (Richards, Neill, & Butters, 1997).

Hattie et al (1997) and Cason and Gillis (1994) reported a weak but positive relationship between program length and the size of outcome, and Bunting and Donley (2002) reported a correlation of .28 between the length of ropes course programs and outcomes. Hans (2000) did not report any effects for program length, and length was not used by Marsh, P. E. (1999). Overall, outdoor education program length explains a small but consistent amount of variance in participant outcomes.

With regard to the effect of participant age, stronger outcomes were found by Cason and Gillis (1994) for younger rather than older adolescents. This negative correlation between age and outcome was also found for the camping studies (Marsh, P. E., 1999). This is consistent with other research suggesting that older adolescence is a period during which self-concept is somewhat resilient to change (e.g. see Hattie, 1992). However, the strongest outcomes of all were found for adult-aged participants (Hattie et al., 1997), although Bunting and Donley (2002) found linear effects of age on ES, from school-age, through college-age to corporate client-age.

Hans (2000) found two significant influences on locus of control outcomes. Firstly, programs that had therapeutic goals as their primary purpose had higher effect sizes than programs with goals such as recreation. This finding seemed to be consistent with Marsh, P. E.’s (1999) finding that camps with a self-development philosophy achieved a reasonably high ES, whereas for camps without such a philosophy, participant changes were negligible. These findings lend credibility to the possibility that organizational mission and culture is a critical factor underlying program effects and this could explain the strong effect sizes that have been reported for Outward Bound programs in Australia (Hattie et al, 1997).

The second significant moderator from the Hans (2000) study was that residential and semi-residential programs were more effective than sessional programs which utilized adventure activities but did not take participants away from their familiar environments overnight.

In summary, the main influences on empirical outdoor education research outcomes are the outdoor education organization running the program, the age of participants, and the length of program. Other moderators of note were quality of study and whether the program was residential. Overall, it appears that the results of outdoor education program research show that there are small-moderate average effects which vary considerably from participant to participant and from program to program.

 

Outdoor education outcomes compared to other intervention programs

Research on many educational, training and therapeutic intervention programs can provide useful comparison with outdoor education outcomes. The most famous meta-analytic study (based on 475 studies) reported an average effect size of .70 for psychotherapy (Smith, Glass & Miller, 1980), noticeably greater than the average outdoor education effect sizes, although it should be noted that psychotherapy tends to be conducted with a different clientele than is normally the case for outdoor education programs.

Another useful comparison is Lipsey and Wilson’s (1993) meta-analyses on psychological treatments, a study which represented over one million participants. The overall effect size was moderate (.47), with 83% of the reviewed meta-analyses finding effect sizes over .20. Lipsey and Wilson stated that “little in conventional reviews and past discussions of these treatment areas, either individually or collectively…prepares a reviewer for the rather stunning discovery that meta-analysis shows nearly every treatment examined to have positive effects” (p. 1192). The Lipsey and Wilson (1993) study provides outdoor education programs with a useful comparative benchmark, particularly for outdoor education programs that seek recognition for effecting psychological changes.

Another useful comparison can be made with non-outdoor education programs designed to enhance self-concept. Hattie, J. A. (1992) conducted a meta-analysis of 89 self-concept enhancement studies, and found an average effect size of .37. This effect size represents an average of outcomes for self-concept change programs such as psychology courses, behavior therapy, cognitive therapy, and client-centered counseling. An effect size .37 for non-outdoor education self-concept program changes is similar to outdoor education programs. As with the outdoor education meta-analyses, however, the Hattie, J. A. (1992) study of non-outdoor education program effects found large variations depending on the type of intervention program.

 

Recommendations

On the basis of the research reviewed, three main recommendations emerge for future research and evaluation in outdoor education. Firstly, future research should describe the program’s methods in more detail. Secondly, investigation is needed of the influence of individual differences on outcomes. Thirdly, effect size benchmarks should be used in program evaluation studies.
 

Future Research Should Describe the Program's Methods in More Detail

All five outdoor education meta-analyses were consistent in calling for future research to provide more details about the educational methods used in the program being investigated. The lack of methods information in the primary empirical studies limited the extent to which meta-analysis could analyze process variables. A lack of more detailed information about program features such as program philosophy, type of activities, instructional techniques and facilitation style, instructor experience, program difficulty, weather, environment, group sizes, activity sequencing, and methods for dealing with behavior problems was noted.

Given the wide variation of outcomes between different types of programs it is critical that future research must undertake to document in much greater detail the nature of the programs being conducted.

An argument can be made that future empirical studies be treated as mixed mode studies with indepth qualitative descriptions of the program process. Other intervention research areas, such as psychology, provide limited information because they use standardized treatment regimens. However, outdoor education programs often lack reliable documentation. outdoor education programming methods needs to receive greater attention in future research.
 

Investigation is needed of the influence of individual differences on outcomes

Meta-analytic findings to date have provided little insight into the phenomenon of individual differences influencing outdoor education participant outcomes. Research reviews have struggled to identify and provide models of relevant participant characteristics and processes that significantly influence the nature of their experience and learning. The lack of definitive evidence about causes of individual variations in outcomes may signal a need for different lines of investigation. Two recent studies have illustrated some possible approaches.

Fry and Heubeck (1998) examined the effects of personality on the mood states of participants in Outward Bound programs. In this study, extroverts and introverts had noticeably different mood patterns during the different phases of the program. Neill and Heubeck (1998) found that school students who reported using productive coping skills (such as ‘focusing on the positive’, ‘working hard’ and ‘solving the problem) experienced greater mental health benefits from outdoor education programs than those who reported using non-productive coping skills (such as ‘self-blame’, ‘worrying’ and ‘ignoring the problem’).

These two studies suggest that research on individual differences, such as personality or coping styles, may enlighten a path towards developing more effective, individualized programming in outdoor education.
 

Use Effect Size Benchmarks in Program Evaluation

The reporting of effect sizes has utility beyond meta-analysis. Reporting of effect sizes is now considered a standard part of empirical reporting (American Psychological Association, 2001). Effect sizes allow for comparative quantification of program effectiveness. Program evaluation results can be compared with internal benchmarks (e.g., compare this program to last year’s program) or external benchmarks (e.g., compare this program with typical effects for outdoor education programs or for other types of intervention programs).

An important contribution of such comparisons can be to provide feedback about a program’s relative strengths and weaknesses. This can be valuable diagnostic information to help identify areas for possible improvement, and to identify particular strengths of the program.

For more information, see Why Use Effect Sizes instead of Significant Testing in Program Evaluation?
 

Conclusion

After five decades of modern day outdoor education, the empirical outcome research has been synthesized. On average, outdoor education programs appeared to have small to moderate effects on participants’ self-perceptions of personal qualities and capabilities. This finding is similar to the average outcome for psychological training and other types of educational self-concept change programs. These results provide scientific support that outdoor education programs, on average, provide legitimate and effective educational training. It would seem that outdoor education methods have something genuine to offer education, training, and psychology. A particularly impressive strength is that outdoor education programs seem capable of triggering an ongoing cycle of positive change within participants.

However, readers should be warned against the comfort of average results. Some outdoor education programs are much less effective and some outdoor education programs are more effective. Outdoor education is not a panacea. Many programs which claim to deliver personal development through the outdoors may be ineffective and almost all programs could be substantially improved if the outcomes of psychological interventions and psychotherapy are accepted as reasonable benchmarks.

Unfortunately, most outdoor education programs do not conduct systematic, rigorous evaluations of program effectiveness, nor are the methodologies of various outdoor education programs well described and available for public dissemination. For the programs that conduct some evaluation, there are disturbingly few that use standardized testing instruments and too many rely on end-of-program written surveys as indicators of their effectiveness.

Further outdoor education meta-analyses would also be useful. The Hattie et al. (1997) study could be updated. In addition, meta-analytic studies of specific aspects of outdoor education programming, such as adventure therapy, would be useful. For example, a meta-analysis of adventure therapy programs could be conducted, as could a meta-analysis examining the influence of different facilitation techniques on participant outcomes.

Outdoor education scholars might also consider other innovative research reviewing methods. For example, a central electronic database of primary outdoor education research data could be developed to allow powerful secondary and cumulative analyses.
It is no longer sufficient for outdoor education advocates to simply believe in the benefits of outdoor education, or to accept anecdotal evidence at face value. Instead, all available tools be used in order to understand how what effects outdoor education programs have on participants and how the best possible learning experiences can be provided.
 

References

American Psychological Association (2001).  Publication Manual of the American Psychological Association (5th ed.).  Washington, DC: American Psychological Association.

Bangert-Drowns, R. L., & Rudner, L. M. (1991).  Meta-analysis in Educational Research.  ERIC Digest.  Retrieved February 17, 2002, from the Educational Resources Information Center Web Site: http://www.ed.gov/databases/ERIC_Digests/ed339748.html.

Barrett, J., & Greenaway, R. (1995). Why Adventure? The role of outdoor adventure in young people's personal and social development. Foundation for Outdoor Adventure: Coventry, UK.

Bunting, C. J., & Donley, J. P. (2002).  Ten Years of Challenge Course Research: A Review of Affective Outcome Studies.  Poster presented at the 6th Coalition for the Education in the Outdoors Research Symposium, January 11-13, Bradford Woods, IN.

Burton, L. M. (1981). A critical analysis and review of the research on Outward Bound and related programs. Dissertation Abstracts International, 42, 1581B.

Campbell, D. T., & Stanley, J. C. (1963). Experimental and Quasi-Experimental Designs for Research.  Houghton Mifflin: Boston.

Cason D., & Gillis, H. L. (1994). A meta-analysis of outdoor adventure programming with adolescents. Journal of Experiential Education, 17(1), 40-47.

Cohen, J. (1977). Statistical power analysis for behavioral sciences (Rev. ed.). New York: Academic Press.

Crompton, J. L., & Sellar, C. (1981). Do outdoor education experiences contribute to positive development in the affective domain? Journal of Environmental Education, 12(4), 21-29.

Ewert, A. (1982). Outdoor adventure and self-concept: A research analysis. Eugene: University of Oregon, Centre of Leisure Studies.

Fry, S. K. & Heubeck, B. (1998). The effects of personality and situational variables on mood states during Outward Bound wilderness courses: An exploration. Personality and Individual Differences, 25, 649-659.

Gillis, H. L. (1992). Therapeutic uses of adventure-challenge-outdoor-wilderness: Theory and research.  Paper presented to the Coalition for Education in the Outdoors Symposium, Bradford Woods, Indiana University, Martinsville, IN.

Gillis, H. L., & Thomsen, D. (1996). A research update (1992 – 1995) of adventure therapy: Challenge activities and ropes courses, wilderness expeditions, and residential camping.  Paper presented to the Coalition for Education in the Outdoors Symposium, Bradford Woods, Indiana University, Martinsville, IN.

Godfrey, R. (1974, April). A review of research and evaluation literature on Outward Bound and related educational programs. Paper presented at the Conference on Experiential Education, Estes Park, CO.

Hans, T. (2000).  A meta-analysis of the effects of adventure programming on locus of control.  Journal of Contemporary Psychotherapy, 30(1),33-60.  Penultimate draft (.2MB).

Hattie, J. (1992). Self-concept. NJ: Lawrence Erlbaum.

Hattie, J., Marsh, H. W., Neill, J. T. & Richards, G. E. (1997). Adventure Education and Outward Bound: Out-of-class experiences that have a lasting effect. Review of Educational Research, 67, 43-87.

Iida, M. (1975). Adventure-oriented programs – A review of research.  In B. van der Smissen (Ed.) Proceedings from the national research workshop, Pennsylvania State University, December 3 – 6.  Penn State HPER Series No. 11.

Lawson, M. (1997, November 24). Wilderness training yet to prove its worth.. The Australian Financial Review (p.7).

Lipsey, M. W., & Wilson, D. B. (1993). The efficacy of psychological, educational, and behavioral treatment. American Psychologist, 48, 1181-1201.

Marsh, P. E. (1999). What does camp do for kids? A meta-analysis of the influence of organized camping experience on the self constructs of youth.  Unpublished Master of Science thesis, Department of Recreation and Park Administration, Indiana University, IN.

Marsh, H. W., Richards, G. E., & Barnes, J. (1986). Multidimensional self-concepts: The effect of participation in an Outward Bound program. Journal of Personality and Social Psychology, 50, 195-204.

McKenzie, M. D. (2000). How are Adventure Education Program Outcomes Achieved?: A review of the literature. Australian Journal of Outdoor Education, 5(1), 19-28.

Neill, J. T., & Heubeck, B. (1997). Adolescent coping styles and outdoor education: Searching for the mechanisms of change. Paper presented to the 1st International Adventure Therapy Conference, University of Western Australia, Perth, W.A., Australia, July 1-5, 1997

Neill, J. T., & Richards, G. E. (1998). Does outdoor education really work? A summary of recent meta-analyses. Australian Journal of Outdoor Education, 3(1), 2-9. 

Neill, J. T., Richards, G. E., & Badenoch, A. (1997). A research evaluation of short-term camps for youth at risk. Canberra, ACT, Australia: National Outdoor Education & Leadership Services.

Reddrop, S. (1997). Outdoor programs for young offenders in detention: An Overview. Hobart, Australia: National Clearing House for Youth Studies.

Richards, G. E. (1977). Some educational implications and contributions of Outward Bound. Sydney: Australian Outward Bound Foundation.

Richards, G. E., Neill, J. T., & Butters, C. (1997). Summary statistical report of attendees at the 10th National Outdoor Education Conference., Colloroy, NSW, Australia, January, 1997. Canberra, ACT, Australia: National Outdoor Education & Leadership Services.

Shore, A. (1977). Outward Bound: A reference volume. Greenwich, CT: Topp Litho.

Smith, M. L., Glass, G. V., & Miller, T. I. (1980).  The benefits of psychotherapy. Baltimore: Johns Hopkins University Press.

Wolf, F. M. (1986).  Meta-analysis: Quantitative Methods for Research Synthesis. Beverly Hills, CA: Sage.