Evaluation of Recreation Motivation and Activity Involvement in Affecting Place Attachment by Hikers

Evaluation of Recreation Motivation and Activity Involvement in Affecting Place Attachment by Hikers

Yu-Jen Chiang, ¹⁾ Dar-Hsiung Wang ^2,3)

【 Summary】

Owing to the rich forest resources in mountain areas, there are various popular hiking activi- ties in Taiwan. Determining how to provide suitable trails for hikers has become a noteworthy issue. The effects of recreation motivation and of activity involvement on place attachment were verified in several previous empirical studies. However, relationships among them have not yet been simultaneously examined. Therefore, the purpose of this study was to investigate relation- ships among hikers’ recreation motivation, activity involvement, and place attachment (comprised of place dependence and place identity). We conducted a questionnaire survey and selected hik- ers on the trail of Guanzihling Dadong Mt. in Tainan City as sampling subjects. In total, 364 valid questionnaires were obtained. Statistical software of structural equation modeling Amos 18 was used to perform a confirmatory factor analysis to estimate parameters of the model, identify the relationship of the overall model, and test the fitness. Results showed that activity involvement was a stronger predictor of the 2 dimensions (place dependence and place identity) of place attachment compared to that of recreation motivation. In terms of the overall model, recreation motivation, activity involvement, and place dependence were all antecedents that had an effect on place iden- tity. Additionally, activity involvement played a mediating role between recreation motivation and place attachment (both dimensions of place dependence and place identity).

Key words: hiking, recreation motivation, activity involvement, place attachment.

Chiang YJ, Wang DH. 2016. Evaluation of recreation motivation and activity involvement in affect- ing place attachment by hikers. Taiwan J For Sci 31(1):1-17.

1)

Department of Cultural Resources and Leisure Industries, National Taitung Univ., 369, Sec. 2, University Rd., Taitung 95092, Taiwan. 國立臺東大學文化資源與休閒產業學系，95092臺東市大學 路二段369號。

2)

Division of Forest Management, Taiwan Forestry Research Institute, 53 Nanhai Rd., Taipei 10066, Taiwan. 林業試驗所森林經營組，10066台北市南海路53號。

3)

Corresponding author, e-mail:dhwang@tfri.gov.tw 通訊作者。

Received March 2015, Accepted August 2015. 2015年3月送審 2015年8月通過。

研究報告

登山健行者遊憩動機與活動涉入對地方依附影響之研究

江昱仁 ¹⁾ 汪大雄 ^2,3)

摘 要

登山健行者的遊憩動機與活動涉入個別對地方依附之影響，已有一些前人研究證實其重要性與實 務應用性，但三者同時就其間關係進行驗證與探討與釐清地方依附之地方依賴和地方認同構面兩者關 係，則尚屬少見。本研究選定台南市關仔嶺大棟山為研究地點，旨在瞭解登山健行者遊憩動機、活動 涉入、地方依附(含地方依賴與地方認同)之因果關係。本研究採問卷調查法，有效問卷為364份。利用 線性結構方程模式統計軟體Amos 18，進行驗證性因素分析、模式之參數校估及整體模式之因果模式 與配適程度驗證。研究結果發現，相對於遊憩動機，活動涉入對地方依附之地方依賴與地方認同更具 預測影響力；遊憩動機、活動涉入及地方依賴三者皆為地方認同之前行變數。就整體模式而言，活動 涉入為影響遊憩動機與地方依附(地方依賴與地方認同構面)間關係的重要因子。

關鍵詞：登山健行、遊憩動機、遊憩涉入、地方依附。

江昱仁、汪大雄。 2016。登山健行者遊憩動機與活動涉入對地方依附影響之研究。台灣林業科學

31(1):1-17。

INTRODUCTION

Taiwan is a mountainous island with various terrains and abundant ecological re- sources, which provide diverse hiking sites.

According to data from the Tourism Bureau’s 2013 Survey of Travel by R.O.C Citizens, among recreational activities preferred by tourists, “hiking on forest trails, mountain climbing, and camping” accounted for 11.8%

and ranked second highest following the ac- tivity of scenic sightseeing (Taiwan Tourism Bureau 2014). To respond to public preferenc- es for hiking activities, in 2001, the Taiwan Forestry Bureau began to establish a national hiking trail system and designed for over 60 trails around Taiwan (Lin and Ong 2007). It is clear that hiking activities are quite popular in Taiwan. With the popularity of hiking activi- ties, understanding relationships between hik- ers’ experiences and nature settings is worthy of attention in trail management.

Conceptualized as place bonding, place

attachment has received increasing attention

in research and practical management over

the past 4 decades (Lewicka 2011). In a natu-

ral setting context, place attachment can be

interpreted as nature bonding, and it refers

to emotional and functional connections de-

veloped between individuals and the natural

environment. Recently, a lot of studies have

investigated understanding of the relation-

ship of place attachment with its antecedents

(Bricker and Kerstetter 2000, Vaske and Ko-

brin 2001, Kyle et al. 2004c, Hou et al. 2005),

and the influential role of place attachment in

outcome variables (Stedman 2002, Lin 2008,

Tsaur and Sun 2009). However, even though

studies of place attachment have turned to-

ward understanding the formative process of

place attachment and the relationship with

other variables (López-Mosquera and Sán-

chez 2013), the formative process through which place attachment develops is still being debated in existing studies (Lewicka 2011).

In this study, we tried to understand questions regarding why people are driven to a specific natural setting, and how psycho- logical bonds subsequently developed with place. Two antecedents, recreation motivation and activity involvement, were applied to in- vestigate their influence on place attachment in this study.

Weiner (2000) suggested that the forma- tion of attachments involves motivational pro- cesses. Kyle et al. (2004c) and Luo and Deng (2008) showed similar results of motivation having a partial effect on place attachment.

Therefore, we argued that place attachment might not be directly influenced by motiva- tion, but rather by other experience factors after visiting a specific place. From the view- point of an enduring association with an ac- tivity, we suggested that activity involvement should be studied as having a congruent role between motivation and place attachment.

Kyle et al. (2003) showed that both at- traction and self-expression of the activity involvement construct are related to place identity, whereas self-expression was the only construct related to place dependence in an investigation of hikers on the Appala- chian Trail in the USA. Moreover, Gross and Brown (2008) also showed that involvement had a positive and significant effect on place attachment in a tourism context. These find- ings indicated that there is a relationship be- tween activity involvement and place attach- ment. In addition, Kyle et al. (2006) showed that motivation had a significant effect on place attachment with activity involvement.

Other work also showed that motivation can be recognized as an antecedents to activity involvement (Funk et al. 2004, Chen 2010).

These findings also indicated that there is a

relationship between activity involvement and place attachment.

As already noted, existing studies dem- onstrated the relationship between two of the recreation motivations, activity involvement and place attachment, in the recreationists’

psychological context. However, the effects of motivation and activity involvement on place attachment have not yet been simulta- neously examined in a “hiker-trail” context.

Furthermore, while many studies used place dependence and place identity as dimensions for measuring place attachment (Williams et al. 1992, Moore and Graefe 1994, Bricker and Kerstetter 2000, Gross and Brown 2008), few empirical studies investigated the relationship between these 2 dimensions (Dai et al. 2008).

Thus, we suggested that the relationship be- tween place dependence and place identity should be included as part of the overall test model to obtain a more-complete understand- ing of the formative process of place attach- ment.

Overall, the purpose of this study was to further investigate relationships among hikers’

recreation motivation, activity involvement, and place attachment (both place dependence and place identity) using a structural equation model. It wasw expected that the results of this study can be helpful to more completely understand the formative process of place attachment, and provide reference for trail management agencies and hiking associa- tions to promote hiking activities, and for trail planning and management.

Theoretical background and research hypotheses

The purpose of this study was to in- vestigate relationships among recreation motivation, activity involvement, and place attachment. This proposed conceptual model

“motivation → activity involvement → place

attachment” was constructed on the concept of motivation and attitudes.

Kyle et al. (2004c) applied “the expec- tancy-value model of motivation theory”

as a theoretical basis for understanding the relationship between motivation and place attachment. They suggested that individuals are motivated to interact with natural settings in pursuit of their specific benefits, involve- ment as well as attachment, which may occur through interactions between these natural settings and people. However, their study provided less insight in explaining why indi- viduals engage in a certain activity in a spe- cific natural setting even when many various environments or activities could also satisfy their specific needs or provide similar desired outcomes.

Kyle et al. (2006) proposed that there is a relationship between motivation and activity involvement based on the expectancy-value model and the process by which ego attitudes are activated. Their work applied the concept of attitude to strengthen the understanding of the process driven by motivation. Many previous studies also suggested that attitude is a critical construct toward an object which is determined by motivation (Gnoth 1997, Hsu et al. 2010), and relationships between their motivations and specific behaviors are mediated by attitudes within specific contexts (Baloglu 2000).

Kyle et al. (2003) demonstrated that both activity involvement and place attachment are attitudinal constructs reflecting the similarity of ego and self-identity. McIntyre and Pigram (1992) noted that leisure involvements can enhance an individual’s commitment to a specific place. Additionally, Herna’ndez et al. (2007) and Thomson et al. (2005) demon- strated that the concept of place attachment could be similar to psychological commit- ment or attitudinal loyalty. Thus, according to

the 3 studies mentioned above, we assumed that activity involvement is an essential ante- cedent in the formation of place attachment.

Consequently, this proposed conceptual model “motivation → activity involvement

→ place attachment” was constructed on the basis of motivation and attitude concepts. Ad- ditionally, this study presents the following research hypotheses.

Recreation motivation and activity in- volvement

Funk et al. (2004) pointed out that con- sumers’ consumption motivations for profes- sional baseball teams should be the anteced- ent of activity involvement. Kyle et al. (2006) found a positive correlation between hiking motivation and activity involvement. As a re- sult, this study proposed a hypothesis 1(H1):

hikers’ recreation motivation has a significant effect on activity involvement.

Recreation motivation and place attach- ment

Kyle et al. (2004c) found that recreation motivation of park users was correlated with place dependence and place identity of place attachment. In an investigation of the relation- ship between recreationists’ motivation and place attachment at the Nanjenshan Ecologi- cal Protection Area, Kenting National Park, Chiang et al. (2008) reported that recreation- ists’ motivation was significantly correlated with place dependence and place identity.

Based on those studies, this study proposed 2 hypotheses, H2: hikers’ recreation motivation has a significant effect on place dependence;

and H3: hikers’ recreation motivation has a significant effect on place identity.

Activity involvement and place attachment

Kyle et al. (2003) investigated the rela-

tionship between these 2 variables in a study

of hikers on Appalachian Mountains Trails, and showed that activity involvement is the antecedent of place attachment. Dai et al.

(2008) also showed that activity involve- ment has a significant effect on both the place dependence and place identity of hikers.

According to the above studies, this study proposed 2 hypotheses, H4: hikers’ activity involvement has a significant effect on place dependence; and H5: hikers’ activity involve- ment has a significant effect on place identity.

Place dependence and place identity Moore and Graefe (1994) suggested that place dependence was formed in a short period of time, and place identity cannot be developed until a user continuously visits a place many times and participates in an ac- tivity for a long period of time. Moore and Graefe (1994) investigated railway users’ at- tachment to the recreational facilities around the railway, and found that users’ place de- pendence had an effect on place identity. Dai et al. (2012) also found that leisure divers’

place dependence had an effect on place iden- tity. Therefore, this study proposed research

hypothesis H6: hikers’ place dependence has a significant effect on place identity.

According to the proposed research hy- potheses based on the aforementioned litera- ture, the conceptual model of the study to be verified is shown in Figure 1.

MATERIALS AND METHODS Site description

The trail of Guanzihling Dadong Mt. (1 trail of the national trail system), located in Paiho District, Tainan City, Taiwan, belongs to Taiwan Forestry Bureau, and is near the Guanzihling Hot Spring Area. The elevation of Dadong Mt. is 1241 m, and it is the highest mountain in Tainan as well as being known as one of the “little hundred mountains” in Taiwan (Taiwan Forestry Bureau 2013). The mountain trail runs for about 6 km along the slope. It is a popular mountain site for hikers in southern Taiwan, as well as a hiking trail suitable for the public (Taiwan Forestry Bu- reau 2013). Because of the high accessibility of Dadong Mt. (a suburban mountain), it is very frequently visited by hikers, and there

Fig. 1. Conceptual model of the study.

are more interactions between hikers and the trail environment. Therefore, we selected Guanzihling Dadong Mt. trails as the study area.

Questionnaire design

In this study, a recreation motivation questionnaire containing 5 dimensions (com- prised of health promotion, natural experi- ence, self-learning and growth, escape, relax- ation, and interpersonal relationships) with 20 items was developed by referring to the scales of Kyle et al. (2004c, 2006).

To measure activity involvement, McIn- tyre and Pigram (1992) developed a 3- dimen- sional measurement of activity involvement comprised of attraction, self-expression, and centrality to lifestyle. This study developed 15 items based on 3 dimensions including attraction, centrality, and self-expression by referring to the scale of Kyle et al. (2004a).

We applied place dependence and place identity as dimensions for measuring place at- tachment. These 2 dimensions were verified in previous studies (Williams et al. 1992, Moore and Graefe 1994, Bricker and Kerstetter 2000, Gross and Brown 2008). This study referred to research items (8 items totally) developed by Kyle et al. (2004b). The items mentioned above were all amended accord- ing to the characteristics of hiking activities.

A Likert 5-point scale was used for scoring. Additionally, the questionnaire was amended according to the suggestions of 3 relevant experts and scholars.

Sample selection and data collection We selected hikers on Guanzihling Da- dong Mt. trails as the sample. A convenience sampling approach was adopted to distribute the questionnaires to subjects at the pavilion square near the triangulation point of Dadong Mt. trails. In total, 380 questionnaires were

distributed. After invalid questionnaires were removed, there were 364 valid questionnaires.

Data processing

We used the statistical software SPSS 17.0 (SPSS, Chicago, IL, USA) to perform the descriptive analysis, reliability analysis, and explorative factor analysis (EFA) of samples. In addition, we also used Structural Equation Modeling (SEM) statistical software Amos 18 to carry on the confirmatory factor analysis (CFA), estimate the parameters of the model, test the research hypotheses, verify the relationship of the overall model, and fit it finally.

RESULTS AND DISCUSSION Respondents’ profile

Results of the descriptive analysis of the demographic statistics and recreational characteristics of respondents were shown in Table 1. The majority of respondents were male, married, aged 51~60 yr (with 74% over the age of 41), were engaged in commerce/service, earned New Taiwan (NT$) 20,001~40,000 (followed by NT$

40,000~60,000 per month), lived in Tainan City, visited the study site 1 or 2 times per month with nearly 40% visiting the study site more than 3 times per month, hiked with their friends, families, or relatives, and spent more than 4 h to complete the roundtrip hike, followed by 3~4 h (64% totally). The longest experience with hiking was > 10 yr. The main source of information obtained by respon- dents came from friends and family.

Instrument validation

To test whether internal consistencies in

the scales used in this study existed, a reli-

ability analysis and an EFA were performed

prior to the CFA. Items with poor reliability

Table 1. Demographic and characteristic profile of respondents

Variable Group N %

Gender Male 214 58.8

Female 150 41.2

Marital status Married 285 78.3

Unmarried 79 21.7

Age (yr) ≦ 20 7 1.9

21~30 26 7.1

31~40 61 16.8

41~50 98 26.9

51~60 136 37.4

≥ 61 36 9.9

Education level Junior high school or below 41 11.3

high school 102 28.0

University or college 195 53.6

Graduate school 26 7.1

Occupation Student 12 3.3

Office worker or teacher 69 19.0

Service industry 100 27.5

Laborer 93 25.5

Housekeeper 38 10.4

Retiree 30 8.2

Others 22 6.0

Monthly income (NT$) ≥ 20,000 68 18.7

20,001~40,000 128 35.2

40,001~60,000 84 23.1

60,001~80,000 57 15.7

80,001~100,000 15 4.1

≥ 100,001 12 3.3

Residence Chiayi 15 4.1

Tainan 312 85.7

Kaohsiung/Pingtung 31 8.5

Other areas 6 1.6

Visitations per month (times) ≦ < 1 109 29.9

1 or 2 110 30.2

3 or 4 90 24.7

5 or 6 23 6.3

7 or 8 12 3.3

≥ 9 20 5.5

Companion By oneself 27 7.4

Family 121 33.2

Friends 184 50.5

Association 32 8.8

Traffic vehicle Car 342 94.0

con’t

Traffic vehicle Motorcycle 17 4.7

Bus 5 1.4

Time spent hiking (h) < 2 67 18.4

2~3 68 18.7

3~4 117 32.1

> 4 112 30.8

Hiking experience (yr) < 1 66 18.1

1~3 81 22.3

4~6 80 22.0

7~9 42 11.5

≥ 10 95 26.1

Information source Informed by friends/relatives 269 73.9

Internet 15 4.1

Notice of hiking association 30 8.2

Hiking supply store 13 3.6

Mass media 24 6.6

Other sources 13 3.6

and low factor loading were removed. In the recreation motivation scale, after 3 items (“To learn more about nature”, “To reduce built-up tension”, and “To be with people who have the same interests”) were removed, Cron- bach’s α values of the dimensions of motiva- tion in “learning and exploration”, “pursuit of nature”, “escape and relaxation”, “social bonding”, and “promotion of health” were 0.89, 0.86, 0.76, 0.79, and 0.8, respectively.

Cronbach’s α values of the dimensions of activity involvement of “attraction”, “central- ity”, and “self-expression” were 0.94, 0.90, and 0.94, respectively. Cronbach’s α value of the place dependence scale was 0.91, and that of the place identity scale was 0.93. Hair et al. (1998) indicated that when a reliability analysis is used to inspect consistency, the generally agreed lower limit for Cronbach’s α value is 0.70. The reliability of various di- mensions in this study were between 0.76 and 0.94, which were all acceptable.

This study used a CFA to test factor structure, convergent validity, and discrimi-

nant validity of various dimensions. With respect to recommended values used in the CFA model, Jöreskog and Sörbom (1989) suggested deleting items with excessively high residuals or an excessively low factor loading, and only those with a standardized factor loading of > 0.45 were retained. In addition, the squared multiple correlations (SMCs) of items were at least 0.20. Based on the standards mentioned above, we removed 3 items of activity involvement: “I really enjoy hiking”, “hiking is one of the most enjoyable things to me”, and “Hiking says a lot about who I am”. Factor loadings of the remaining items ranged 0.57~0.95, their SMC values ranged 0.32~0.90, and t-values of the estimated parameters were all > 1.96 (p <

0.05). Those values all met the determination criteria suggested by Jöreskog and Sörbom (1989), and are shown in Table 2.

Regarding the composite reliability (CR)

and average variance extracted (AVE) in the

CFA, the recommended values of Bagozzi and

Yi (1988) were > 0.60 and > 0.50, respectively.

Table 2. Confirmatory factor analysis of measurement model

Variables Measured items SFL t-value SMC CR AVE

Recreation Learning/exploration

motivation To expand/explore knowledge 0.75 — 0.56

To experience new and different things 0.85 16.52* 0.72 0.89 0.67

To satisfy my curiosity 0.83 15.22* 0.69

To experience the unknown 0.84 15.42* 0.71

Nature seeking

To be close to nature 0.74 — 0.55

To enjoy the tranquility 0.80 14.49* 0.64 0.86 0.61

To enjoy the natural scenery 0.84 14.55* 0.70

To enjoy the natural atmosphere 0.73 13.07* 0.53

Escape/relaxation

To get away from routine affairs 0.57 — 0.32

To get away from crowded areas 0.91 8.79* 0.83 0.78 0.54

To experience privacy 0.69 9.81* 0.48

Social bonding

To bring family/friends closer relationships 0.74 — 0.55

To be with my family/friends 0.83 11.13* 0.69 0.79 0.56

To share my skill and knowledge with others 0.66 10.83* 0.43 Health

To be physically active 0.87 — 0.76

To obtain health 0.91 20.59* 0.84 0.89 0.73

To get exercise 0.77 17.44* 0.59

Recreation Attraction

involvement Hiking is pleasurable 0.70 — 0.48

Hiking interests me 0.89 16.97* 0.79 0.91 0.71

Hiking is one of the most satisfying things that I do 0.95 16.34* 0.90

Hiking is important to me 0.81 14.54* 0.65

Self expression

Others can see in a way that I want them to see me when I go hiking 0.71 — 0.51

When I participate in hiking I can really be myself 0.95 15.54* 0.90 0.88 0.71 You can learn a lot about a person by watching them hiking 0.85 15.51* 0.72 Centrality

I have many friends who are in some way connected with hiking 0.84 — 0.70

I enjoy discussing hiking with my friends 0.81 18.82* 0.65 0.94 0.77 A lot of my life is organized around hiking activities 0.91 22.81* 0.83

Hiking plays a central role in my life 0.93 23.50* 0.86 A lot of my life is organized around hiking 0.88 21.50* 0.77 Place Place dependence

dependence I enjoy hiking on this trail more than any other trail 0.91 — 0.84 I get more satisfaction from this trail than any other trail 0.94 31.82* 0.89 0.92 0.75 Hiking here is more important than hiking any other place 0.93 29.75* 0.86

I would not substitute any other trail for the activities I participate in here 0.65 14.82* 0.42 Place Place identity

identity Hiking on this trail means a lot to me 0.87 — 0.76

I am very attached to this trail 0.84 21.07* 0.71 0.92 0.75

I identify strongly with this trail 0.92 23.82* 0.84

I have a special connection to this trail and the people who hike here 0.84 20.57* 0.70

SD, standard deviation; SFL, standardized factor loading; SE, standard error; SMC, square multiple correlations; AVE, average variance extracted; CR, composite reliability.

* p < 0.05.

Table 3. Examination of the discriminant validity (implied correlation of dimensions)

Dimension n Correlation value

A B C D E F G H I J

A. Learning/exploration 4 0.82 B. Nature seeking 4 0.30 0.78 C. Escape/relaxation 3 0.25 0.24 0.74 D. Social bonding 3 0.44 0.34 0.21 0.75 E. Health 3 0.12 0.28 0.10 0.25 0.85 F. Attraction 4 0.22 0.25 0.19 0.35 0.27 0.84 G. Self expression 3 0.37 0.25 0.20 0.36 0.12 0.40 0.84 H. Centrality 5 0.29 0.19 0.20 0.37 0.16 0.64 0.52 0.88 I. Place dependence 4 0.22 0.10 0.10 0.19 .0.06 0.25 0.29 0.38 0.87 J. Place identity 4 0.31* 0.26 0.23 0.37 0.20 0.52 0.44 0.52 0.53 0.87 Note 1: Values on the diagonal (boldface) denote the square root of the average value extracted of di-

mensions. Values under the diagonal are standardized correlation coefficients.

* denotes p < 0.05.

The CR values of various dimensions in this study ranged 0.78~0.94, and AVE values ranged 0.54~0.77, which both met the recom- mended values of Bagozzi and Yi (1988).

Therefore, the convergent validity of the measurement model in this study should be acceptable.

In testing the discriminant validity, Hair et al. (2006) and Fornell and Larcker (1981) indicated that the AVE for each pair of con- structs should be greater than the square of the correlation between each of those 2 con- structs. As shown in Table 3, the square root of each dimension was greater than the cor- relation coefficient of each dimension. The results showed that the judgment criteria were met, thus verifying that the discriminant va- lidity of the measurement model in this study was acceptable.

Evaluation of the overall model

In an analysis of overall structure model, in order to simplify the items, means of the 5 dimensions of recreation motivation and 3 dimensions of recreation involvement were calculated and used as input data of observed

variables. Regarding the dimensions of place dependence and place identity, the input data of the original 4 observed variables were con- tinuously used for analysis.

On the test of offending estimates, Hair et al. (1998) suggested that the determina- tion of non-offending estimates should meet the criteria that there be no negative error variance (EV), or oversizes standard error (SE), and the standardized factor loading (SFC) should not be greater than or close to 1. As shown in Table 4, the SFC in this study ranged 0.58~0.95, and the EV was not nega- tive and reached statistical significance, sug- gesting that there was no offending estimate in the estimation of the model in this study.

A test of the fit of the overall model could be performed.

After confirmation that there was no of-

fending estimate, we then tested the fitness

of the overall model. Measurement indices

for the test of fitness on the hypothesis in this

study included the Chi-square value, Chi-

square degree of freedom ratio (normed Chi-

square), goodness of fit index (GFI), adjusted

goodness of fit index (AGFI), root mean

Table 4. Offending estimates of latent variables

Latent variable Observed variable SFL EV SE SMC CR AVE

Recreation RM1. Learning/exploration 0.76* 0.22* 0.02 0.58

motivation RM2. Nature seeking 0.75* 0.13* 0.01 0.56

RM3. Escape/relaxation 0.61* 0.33* 0.03 0.37 0.83 0.51

RM4. Social bonding 0.82* 0.14* 0.02 0.67

RM5. Health 0.58* 0.19* 0.02 0.34

Recreation AI1. Attraction 0.83* 0.13* 0.01 0.70

involvement AI2. Self expression 0.75* 0.26* 0.02 0.56 0.88 0.72

AI3. Centrality 0.95* 0.06* 0.02 0.91

Place PD1. I enjoy hiking here more than any other trail 0.91* 0.11* 0.01 0.84 dependence PD2. I get more satisfaction out 0.94* 0.07* 0.01 0.89

of visiting this trail than from visiting any other trail

PD3. Hiking here is more important than hiking any other place 0.93* 0.11* 0.01 0.86 0.92 0.75 PD4. I would not substitute any other trail for the type 0.65* 0.54* 0.04 0.42

of activities I do here

Place PI1. This trail means a lot to me 0.87* 0.15* 0.02 0.76 identity PI2. I am very attached to this trail 0.84* 0.17* 0.02 0.71

PI3. I identify strongly with this trail 0.92* 0.11* 0.01 0.84 0.92 0.75 PI4. I have a special connection to this trail and 0.84* 0.21* 0.02 0.70

people who hike here

SFL, standardized factor loading; EV, error variance; SE, standard error; SMC, square multiple correlations; AVE, average variance extracted; CR, composite reliability.

* P < 0.05.

square error of approximation (RMSEA), normed fit index (NFI), Tucker-Lewis index (TLI), comparative fit index (CFI), parsimo- nious normed fit index (PNFI), parsimonious comparative fit index (PCFI), and root mean square residual (RMR). Measurement indices and standards for the test of fitness, as well as the result of the fitness on the overall model in this study were given in Table 5. Most of the results met the standards required by vari- ous indices. The AFGI value was 0.87. Al- though it did not meet the standard of 0.90, it was still acceptable. Overall, the fitness of the research model was acceptable.

Tests of hypotheses

In terms of the tests of the research hy- potheses, as shown in Fig. 2, the findings of this study confirm that recreation motivation has a positive effect on activity involvement

(β = 0.81, t = 13.09, p < 0.001). Thus, H1 was accepted. This result is consistent with those of other studies, such as those by Iwasaki Havitz (2004), Kyle et al. (2006), and Lin and Lee (2010). It was also confirmed that recre- ation motivation positively and significantly influenced place identity (β = 0.22, t = 3.11, p

< 0.01) in this study. Thus, H3 was accepted.

This research result is in line with those of Kyle et al. (2004c) and Chiang et al. (2008).

The prediction of a relationship between ac- tivity involvement and place dependence (β

= 0.69, t = 6.61, p < 0.001) was confirmed in this study. Thus, H4 was accepted. This find- ing is in line with a few previous empirical studies (Kyle et al. 2003, Dai et al. 2008).

Similarly, findings of this study confirmed

that activity involvement had a significant

influence on place identity (β = 0.38, t = 4.70,

p < 0.001). Thus, H5 was accepted. This

result is consistent with those of the studies by Bricker and Kerstetter (2000), Kyle et al.

(2003), Kyle and Mowen (2005), and Dai et al. (2012). The effect of place dependence on place identity was confirmed by this study (β

= 0.39, t = 7.19, p < 0.001). Thus, H6 was ac- cepted. This result is consistent with those of studies by Vaske and Kobrin (2001) and Dai et al. (2012).

Surprisingly, hypothesis 2, the effect of recreation motivation on place dependence, was not supported in this study (β = 0.00, t = 0.03, p > 0.05). Consequently, recreation mo- tivation had an influence on place identity but not on place dependence. Overall, recreation motivation was a stronger predictor of place identity than it was of place dependence. This result is inconsistent with those of several existing studies showing that motivation has a significant influence on the 2 dimensions (place dependence and place identity) of place attachment (Kyle et al. 2004c, Anderson and Fulton 2008, Chiang et al. 2008). Although our findings are not consistent with those of several previous studies, Kyle et al. (2004c) pointed out that compared to place identity, the effect of motivation on place dependence was less significant in their study. Low and Altman (1992) suggested that effects/emotions toward symbols of specific places could be stronger than actual places in attachment for- mation. Similarly, Lewicka (2011) also dem- onstrated that physical attributes may facilitate

social contacts and thus indirectly influence place attachment. Furthermore, our findings showed that the factor loading of social bond- ing was the highest among the 5 dimensions of recreation motivation in the CFA (Table 4).

This result infers that interpersonal relation- ships may facilitate individuals attaching to a specific place, and social ties or social identi- fication may directly drive an internal process in the formation of place identity directly.

From the above discussion, it can be inferred that respondents would extend the meaning, value, social interaction, and affec- tion of trail hiking activities at Dadong Mt., and thus, have a strong sense of identity with it. As indicated by Dai et al. (2012), if an indi- vidual feels that he/she belongs to a space and perceives his/her own importance to it, then he/she will develop a sense of identity for the space and develop a sense of belonging to the group within it.

From aspects of the direct effect, results showed that recreation motivation had a di- rect effect on recreation involvement, and the path coefficient was the highest one (β = 0.81, p < 0.001) in the overall structural model (Fig. 2). Clearly, activity involvement can be recognized as the variable most influenced by recreation motivation compared to the others.

Additionally, our study found that the ef- fects of recreation motivation on place depen- dence (β = 0.00, p > 0.05) and place identity (β

= 0.22, p < 0.01) were weaker than the effect Table 5. Summary table of indices of the goodness of fit of the overall model

Absolute fit measurement Incremental fit Parsimonious fit

measurement measurement

Χ

(p value) X

/df GFI AGFI RMSEA NFI TLI CFI PNFI PCFI RMR

Casual model 285.63 (p = 0.00) 2.92 (df = 98) 0.90 0.87 0.07 0.90 0.95 0.96 0.77 0.78 0.02

Threshold value p < 0.05 < 3 > 0.9 > 0.9 < 0.08 > 0.9 > 0.9 > 0.9 > 0.5 > 0.5 < 0.1

df, degree of freedom; GFI, goodness of fit index; AGFI, adjusted goodness of fit index; RMSEA, root mean square er-

ror of approximation; NFI, normed fit index; TLI, Tucker-Lewis index; CFI, comparative fit index; PNFI, parsimonious

normed fit index; PCFI, parsimonious comparative fit index; RMR, root mean square residual.

of activity involvement on place dependence (β = 0.69, p < 0.001) and place identity (β = 0.38, p < 0.001) (Fig. 2). Thus, the findings indicated that respondents’ activity involve- ment was a stronger predictor of place at- tachment (both place dependence and place identity) than that of recreation motivation.

In addition to the results of direct effects, indirect effects among the studied variables were also investigated. We found that the path coefficient of “recreation motivation → activity involvement → place dependence”

(0.81×0.69 = 0.56) was higher than that of

“recreation motivation → place dependence”

(0.00). Thus, in the absence of a significant direct influence of recreation motivation on place dependence, it can be stated that activ- ity involvement fully mediates the influence of recreation motivation on place depen- dence. Additionally, since the path coefficient of “recreation motivation → place identity”

(0.22) was lower than that of “recreation motivation → activity involvement → place identity” (0.81×0.38 = 0.31), and similar to the path coefficient of “recreation motivation

→ activity involvement → place dependence

→ place identity” (0.81×0.69×0.39 = 0.22), activity involvement may play an intermedi- ary role between recreation motivation and place identity. To determine the role media- tion played by activity involvement in this study, a test was conducted on the structural model without the activity involvement vari- able. The goal was to understand changes in the value of the effect of recreation motiva- tion on place identity. In the structural model without the activity involvement variable, the value of the effect of recreation motivation on place identity was 0.74. This value was the same as the effect of recreation motivation on place identity to the structural model with the activity involvement variable, namely 0.74.

Furthermore, in the structural model without

the activity involvement variable, the path coefficient of “recreation motivation → place dependence” was 0.56 (p < 0.001). This value was greater than that of “recreation motiva- tion → place dependence” in the structural model involving the activity involvement variable, namely 0.00 (p > 0.05). This means that the influence of place dependence be- tween recreation motivation and place iden- tity was mostly absorbed by activity involve- ment. Thus, it can be concluded that activity involvement played a mediating role between recreation motivation and place dependence, and also mediated the effect of recreation mo- tivation on place identity.

Overall, since the path of “recreation motivation → activity involvement → place dependence → place identity” (0.81×0.69×

0.39 = 0.22) was found in this study, our findings confirmed that when individuals are driven to participate in and learn the benefits provided from various recreation activities in a specific place, over time they may become more deeply involved in activities which meet their needs, and then place dependence may form in a short period of time, and place identity develops later over a longer period of time (Moore and Graefe 1994, Dai et al.

2008).

In summary, our findings show that place attachment is not always influenced directly from the recreation motivation, but rather from activity involvement. Clearly, respon- dents’ activity involvement is a stronger pre- dictor of place attachment (both place depen- dence and place identity) than their recreation motivation.

Furthermore, this study suggests that activ-

ity involvement play a mediating role between

recreation motivation and place attachment

(both dimensions of place dependence and

place identity). Hence, the relationship of “rec-

reation motivation → recreation involvement

→ place dependence → place identity” is verified in hiking extent, thus, potentially contributed to the academic literature on place attachment and the application of the theoretical frameworks.

CONCLUSIONS

Along with an increasing awareness of the importance of place attachment, it is also

important to understand that relationships of place attachment with its antecedents. This study developed a conceptual model that incorporated recreation motivation and activ- ity involvement as critical factors in order to strengthen place attachment among hikers on Taiwanese trails.

The findings of this study also have prac- tical implications for managers. Owing to the relationship model “recreation motivation → Fig. 2. Estimated structural model.

Note: RM1..RM5, AI1..AI3, PI1..PI4, and PD1..PD4 are observed variables associated with

corresponding latent variables.

recreation involvement → place dependence

→ place identity” being verified in this study, this result can be provided as a reference for management agencies to develop trail and hiker management strategies. In order to in- crease hikers’ recreation involvement, and dependence and identity with Dadong Mt.

trails, management agencies can focus on providing environmental, social, and man- agement contextual settings to meet the mo- tivations of hikers. When hikers spend more time and efforts in a place and engage in its relevant activities, they will develop a sense of identity with the place. In other words, the power of hikers’ place attachment can also be used to effectively manage and maintain local resources and provide them to users for con- tinuous use.

Despite its contributions, this study has certain limitations. First, since this study fo- cused on respondents hiking on the trail of Guanzihling Dadong Mt, our findings are lim- ited to this setting, and might may not reflect the situation on other trails around Taiwan.

We suggest that follow-up studies choose different suburban mountain types assites to investigate differences. Second, our findings present only relationships among recreation motivation, activity involvement, and place attachment. We suggest that follow-up studies expand this conceptual model presented here with other antecedents such as image and experience to enrich our understanding of the formative process of place attachment.

LITERATURE CITED

Anderson DH, Fulton DC. 2008. Experience preferences as mediators of the wildlife related recreation participation: place attachment rela- tionship. Hum Dimens Wildl 13:73-88.

Bagozzi RP, Yi Y. 1988. On the evaluation of structural equation models. J Acad Market Sci

16(1):76-94.

Baloglu S. 2000. A path-analytical model of visitation intention involving information sources, socio-psychological motivations and destinations images. In: Woodside A, Crouch G, Mazanec J, Oppermann M, Sakai M, edi- tors. Consumer Psychology of Tourism Hospi- tality and Leisure. New York: CABI Publish- ing. p 63-90.

Bricker KS, Kerstetter DL. 2000. Level of specialization and place attachment: an ex- ploratory study of whitewater recreationists.

Leisure Sci 22:233-57.

Chen FM. 2010. Examining the relationships among recreation motivation, recreation in- volvement, travel behavior-example of moun- tain hiking. Rev Leis Sport Health 1(2):100- 19. [in Chinese with English summary].

Chiang YJ, Tsai CF, Hwung SH. 2008. Tour- ists’ motivation and place attachment of ecot- ourism destination- a case study of Nanjenshan ecological reserve, Kenting National Park. Leis Ind Res 6(3):47-70. [in Chinese with English summary].

Dai DD, Huang WS, Lee HC, Chen KY.

2012. A study on relationships between satis- faction of setting attributes and place attach- ment in recreational scuba diving. J Outdoor Recreat Stud 25(1):25-55. [in Chinese with English summary].

Dai YD, Chen KY, Lee SH. 2008. A study on causal relationships among recreation involve- ment, place dependence, and place identity: a case of recreation bikers at Tong-Fon bikeway green corridor. J Outdoor Recreat Stud 21(4):21- 57. [in Chinese with English summary].

Fornell C, Larcker D. 1981. Evaluating stru- ctural equation models with unobservable variables and measurement error. J Mark Res 18(1):39-50.

Funk DC, Ridinger LL, Moorman AM.

2004. Exploring origins of involvement: un-

derstanding the relationship between consumer

motives and involvement with professional sport teams. Leisure Sci 26:35-61.

Gnoth J. 1997. Tourism motivation and expec- tation formation. Ann Tourism Res 24(2):283- 304.

Gross MJ, Brown G. 2008. An empirical structural model of tourists and places: pro- gressing involvement and place attachment into tourism. Tourism Manage 29(6):1141-51.

Hair JF, Anderson RE, Tatham RL, Black WC. 1998. Multivariate data analysis (5

ed.).

Upper Saddle River, NJ: Pearson-Prentice Hall. 118 p.

Hair JF, Black WC, Babin BJ, Anderson RE, Tatham RL. 2006. Multivariate data anal- ysis.(6

ed.). Upper Saddle River, NJ: Pearson- Prentice Hall. 778 p.

Herna’ndez B, Hidalgo MC, Salazar-Laplace ME, Hess S. 2007. Place attachment and place identity in natives and non-natives. J Environ Psychol 27:310-19.

Hou JS, Lin CH, Morais DB. 2005. Ante- cedents of attachment to a cultural tourism destination: the case of Hakka and non-Hakka Taiwanese visitors to Pei-Pu, Taiwan. J Travel Res 44(2):221-33.

Hsu CHC, Cai LA, Li M. 2010. Expectation, motivation, and attitude: a tourist behavioral model. J Travel Res 49(3):282-96.

Iwasaki Y, Havitz ME. 2004. Examining relationships between leisure involvement, psychological commitment and loyalty to a recreation agency. J Leisure Res 36(1):45-72.

Jöreskog KG, Sörbom D. 1989. LISREL 7: a guide to the program and application. Chicago, IL: Scientific Software International. 420 p.

Kyle GT, Absher JD, Hammitt W, Cavin JC. 2006. An examination of the motivation- involvement relationship. J Leisure Sci 28:467- 85.

Kyle GT, Graefe AR, Manning R, Bacon J. 2003. An examination of the relationship between leisure activity involvement and place

attachment among hikers along the Appala- chian Trail. J Leisure Res 35(3):249-73.

Kyle GT, Graefe AR, Manning R, Bacon J. 2004a. Effect of activity involvement and place attachment on recreationists’ perceptions of setting density. J Leisure Res 36(2):209-31.

Kyle G, Graefe A, Manning R, Bacon J.

2004b. Effects of place attachment on users’

perceptions of social and environmental condi- tions in a natural setting. J Environ Psychol 24:213-25.

Kyle GT, Mowen AJ, Tarrant M. 2004c. Lin- king place preferences with place meaning: An examination of the relationship between place motivation and place attachment. J Environ Psychol 24:439-54.

Kyle GT, Mowen AJ. 2005. An examination of the leisure involvement-agency commitment relationship. J Travel Res 37(3):342-63.

Lewicka M. 2011. Place attachment: How far have we come in the last 40 years? J Environ Psychol 31(3):207-30.

Lin CH. 2008. Examining the role of place attachment in tourist’s loyalty to a hot springs destination: the case of Gu-Guan. J Outdoor Recreat Stud 21(3):1-22. [in Chinese with English summary].

Lin HZ, Ong LS. 2007. Development of na- tional trail system. Taiwan For J 33(1):28-38.

Lin WR, Lee MJ. 2010. The Influence of windsurfers’ recreation motivation on activity involvement. J Sport Leis Hosp Res 5(3):152- 70. [in Chinese with English summary].

López-Mosquera N, Sánchez M. 2013. Di- rect and indirect effects of received benefits and place attachment in willingness to pay and loyalty in suburban natural areas. J Environ Psychol 34:27-35.

Low SM, Altman I. 1992. Place attachment:

a conceptual inquiry. In: Altman I, Low SM, editors. Place Attachment. New York: Plenum Press. p 1-12.

Luo Y, Deng J. 2008. The new environmental

paradigm and nature-based tourism motiva- tion. J Travel Res 46(2):392-402.

McIntyre N, Pigram JJ. 1992. Recreation specialization reexamined: the case of vehicle- based campers. Leisure Sci 14:3-15.

Moore RL, Graefe AR. 1994. Place attach- ments to recreation settings: the case of rail- trail users. Leisure Sci 17:17-31.

Stedman R. 2002. Toward a social psychology of place: predicting behavior from place-based cognitions, attitude, and identity. Environ Be- hav 34(5):561-81.

Taiwan Forestry Bureau. 2013. National trail system. Available at http://recreation.forest.

gov.tw/RT/RT_index.aspx. Accessed 2013 Dec 11.

Taiwan Tourism Bureau. 2014. National travelling investigation 2013. Available at http://admin.taiwan.net.tw/statistics/market.

aspx?no=133. Accessed 2014 July 10.

Thomson M, MacInnis DJ, Park CW. 2005.

The ties that bind: measuring the strength of consumers’ emotional attachments to brands. J Consum Res 15(1):77-91.

Tsaur SH, Sun CY. 2009. Antecedents and consequences of place attachment. J Geogr Sci (55):43-63. [in Chinese with English sum- mary].

Vaske JJ, Kobrin KC. 2001. Place attachment and environmentally responsible behavior. J Environ Educ 32(4):16-21.

Weiner B. 2000. Intrapersonal and interper- sonal theories of motivation from an attributio- nal perspective. Educ Psychol Rev 12(1):1-14.

Williams DR, Patterson ME, Poggenbuck

JH, Watson AE. 1992. Beyond the commodity

metaphor: examining emotional and symbolic

attachment to place. Leisure Sci 14:29-46.