Destination image refers to the set of perceptions, impressions, and evaluations that tourists form toward a place before, during, and after a visit [4]. Gunn treated destination image as a central component of tourism-region design [4], while Baloglu and McCleary emphasized that image formation reflects the joint influence of personal factors and stimulus factors [1]. Recent studies further suggest that emotional responses to destination imagery [5], social media affordances [6], and user-generated visual content [7] also shape how tourism destinations are perceived. In agrotourism, destination image typically includes not only natural scenery but also food, agricultural experience, service encounters, environmental cleanliness, and value-for-money perceptions.

Prior work in tourism and agritourism has treated destination image as a multidimensional concept. In the literature in Taiwan region, destination image has been organized into such dimensions as landscape, activities, quality, facilities, service, and price [8][9]. Later work adapted these into more operational groupings such as product, quality, service, and price image [10]. This four-part structure is particularly useful for agrotourism management because it translates a broad destination-image concept into concrete service domains that can be monitored and improved.

Revisit intention is usually defined as a visitor’s willingness to return to a destination or to recommend it to others after the visit [11]. Gitelson and Crompton described repeat visitation as a core manifestation of tourism loyalty [11], and Kozak argued that revisit behavior is shaped by satisfaction, prior experience, and travel-related constraints [12]. In empirical studies, revisit and recommendation intention are often used as concise behavioral indicators of favorable post-visit evaluation [2][13][14]. Recent reviews also confirm that revisit intention remains a central loyalty indicator in tourism and hospitality research [15]. In the present article, these two items are treated as behavioral-intention attributes within the descriptive IPA framework rather than as dependent variables in a causal model.

Importance-Performance Analysis (IPA) was introduced by Martilla and James as a simple but powerful managerial tool for diagnosing service strengths and weaknesses [16]. By plotting the perceived importance of each attribute against its perceived performance, IPA helps identify attributes that should be maintained, monitored, or improved. In tourism research, IPA has been widely used because it turns diffuse visitor evaluations into a clear resource-allocation logic. For agrotourism destinations, the main value of IPA lies in showing where managers should concentrate limited attention and investment to improve the overall visitor experience.

This article is a manuscript redevelopment and secondary analysis of an archived survey study originally conducted by the authors on Hualien agrotourism. The archived Chinese report indicates that questionnaires were administered on site at multiple farms, agricultural sightseeing areas, and agricultural-experience venues in Hualien. The target population comprised visitors who had participated in agrotourism activities in Hualien and were willing to complete the questionnaire. A total of 141 valid questionnaires were retained for analysis.

Several procedural details requested by reviewers cannot be fully reconstructed from the archived report alone. Specifically, the exact survey period, the exact number of participating sites, and the step-by-step recruitment log were not preserved in the manuscript file available for redevelopment. For this reason, the present article describes the sample conservatively as an on-site convenience sample and does not claim statistical representativeness. These archival limitations are acknowledged explicitly so that the empirical scope of the study remains transparent.

Transparency statement.This article is an English-language manuscript redevelopment and secondary analysis of an archived survey study prepared by the same authors.The present version reorganizes the preserved tables and analytical outputs for journal submission and does not involve additional data collection.

The questionnaire contained three parts. First, tourist characteristics were measured through seven background variables: gender, age, education, occupation, income, place of residence, and prior travel experience in Hualien. These background variables were adapted from domestic tourism studies on visitor socioeconomic attributes [17][18]. Second, destination image was assessed through 24 adopted items organized into four dimensions: product image (7 items), quality image (5 items), service image (8 items), and price image (4 items). These items were adapted from earlier destination-image studies [8]-[10]. Third, behavioral intention was measured with two items—revisit intention and recommendation intention—adapted from [2][13][14].

All substantive items were evaluated on a 7-point Likert scale twice: once for perceived importance and once for perceived performance (reported in the original study as satisfaction). To improve traceability across Methods and Results, the present article labels the 24 destination-image items as DI1 - DI24 and the two behavioral-intention items as BI1 - BI2. Because the archived report preserved the analytical outputs but not the original raw dataset, the wording and item grouping reported here follow the archived instrument rather than a newly reconstructed questionnaire.

Notably, the archived report states that destination image contained 23 items, but the reproduced descriptive tables enumerate 24 destination-image items (DI1 - DI24) plus two behavioral-intention items (BI1 - BI2). Because the item-level tables and IPA output are internally consistent with 26 total attributes, the present article follows the 24-item destination-image count. Table 1 summarizes the construct definitions, item codes, and adopted sources.

Table 1. Construct definitions, item codes, and adopted sources.

The archived report preserved only aggregate measurement diagnostics: Cronbach’s alpha values above 0.80 for the full scales, KMO = 0.867, and Bartlett’s test of sphericity with chi-square = 16316.575, df = 1326, p < 0.001. These results indicate acceptable internal consistency and factorability at the overall scale level. However, dimension-specific alpha coefficients, factor loadings, and other validity statistics were not preserved. Accordingly, the four-dimension structure is treated in this article as an adopted framework derived from the original instrument rather than a factor structure independently reconfirmed in the current manuscript.

For the IPA, the archived report used an overall importance mean of 5.8434 and an overall performance mean of 5.805 as reference values. Items were then grouped into four conventional categories: keep up the good work, possible overkill, low priority, and concentrate here. The archived graphical output appears to have been generated from a standardized spreadsheet rather than from the reproduced raw means alone. Because that spreadsheet is not available, the present manuscript reports the archived quadrant assignment, while also disclosing the raw item means for the two archived priority items so that readers can judge their placement more carefully.

The archived descriptive paragraph also reports an overall performance mean of 5.22, whereas the IPA section reports 5.805. Because the reproduced item-level means range from 5.53 to 6.00 and are therefore consistent with the higher value, the present article uses the archived IPA crosshair value of 5.805 while flagging the discrepancy as an archival inconsistency. The archived reliability and validity information reported in this study is summarized in Table 2.

Table 2. Archived reliability and validity information.

The sample was dominated by male respondents (55.3%). The largest age group was 21 - 30 years (45.4%). Most respondents had a university or junior-college education (64.5%), and the largest occupational group was the service sector (27.0%). Monthly income most commonly fell between NT$30,000 and NT$50,000 (40.4%). Respondents from eastern Taiwan region and the outlying islands accounted for the largest residential group (41.4%). Overall, the sample mainly represents relatively young adult visitors with moderate spending capacity and a substantial regional connection to eastern Taiwan region. The sample profile is summarized in Table 3.

Table 3. Summary of sample characteristics.

At the overall scale level, the archived report indicated Cronbach’s alpha values above 0.80 for the combined importance and performance instruments, suggesting good internal consistency. The KMO value of 0.867 and Bartlett’s test result (chi-square = 16316.575, df = 1326, p < 0.001) further indicated that the item set had sufficient intercorrelation for factor-analytic treatment in the original study design. Nevertheless, because the archived material does not preserve dimension-level diagnostics, the present article does not claim new confirmation of the four-dimensional structure.

The overall importance mean was 5.84, indicating that respondents generally regarded Hualien’s agrotourism attributes as important. The highest importance scores were DI8 (“helps me feel physically and mentally relaxed”) and BI1 (“I am willing to plan time to revisit this destination”), both at 6.02. BI2 (“I would recommend Hualien to others”) followed closely at 6.01. These results suggest that relaxation and favorable future behavioral intention are central to visitors’ evaluation framework.

Most performance means fell between 5.53 and 6.00, indicating moderate to high satisfaction overall. The highest performance scores were DI1 (“the scenery is beautiful”) and BI2 (“I would recommend Hualien to others”), both at 6.00, followed by DI8 (“helps me feel physically and mentally relaxed”) and BI1 (“I am willing to revisit”), both at 5.99. The lower-performing items were DI2 (“transportation is convenient,” 5.53) and DI24 (“fees for experience activities are reasonable,” 5.69). Thus, Hualien’s agrotourism appears strongest in scenery, relaxation, and positive word of mouth, while accessibility and perceptions of activity pricing remain comparatively weaker. The complete item-level importance and performance means, together with the archived IPA category assigned to each attribute, are reported in Table 4.

Table 4. Item-level importance and performance means with archived IPA classification.

Note. DI20 was not explicitly assigned to a quadrant in the archived narrative even though its mean values place it close to the overall crosshair. The two archived “concentrate here” items (DI1 and DI15) are retained for transparency, but their placement should be interpreted cautiously.

Using the archived IPA classification, ten attributes were placed in the “keep up the good work” quadrant: delicious food (DI6), distinctive food characteristics (DI7), relaxation (DI8), comfortable dining environment (DI10), positive staff attitude (DI14), rapid problem solving (DI16), staff empathy (DI17), creative activity design (DI19), revisit intention (BI1), and recommendation intention (BI2). Six items fell into “possible overkill”: clean dining environment (DI11), quality public facilities (DI12), ease of obtaining tourism information (DI18), reasonable dining prices (DI22), reasonable agricultural-product prices (DI23), and reasonable activity fees (DI24). Seven items were classified as “low priority”: convenient transportation (DI2), attractive folk-cultural activities (DI3), attractive agrotourism experiences (DI4), distinctive agricultural specialty products (DI5), learning new knowledge (DI9), professional interpretive ability (DI13), and reasonable ticket prices (DI21).

The archived graphical output identified two “concentrate here” items: scenic beauty (DI1) and staff appearance/clean dress (DI15). To make this placement transparent, Table 5 reports their reproduced raw means. DI1 had an importance mean of 5.81 and a performance mean of 6.00, while DI15 had an importance mean of 5.73 and a performance mean of 5.84. Because both items lie close to the overall reference values and because the archived plot was based on a standardized spreadsheet that is no longer available, these two attributes are best interpreted as expectation-sensitive managerial watch points rather than as straightforward low-performing failures. This interpretation is especially important for scenic beauty, whose absolute performance mean was high even though the archived IPA figure still marked it as a priority area.

Table 5. Item-level means for the two archived “concentrate here” attributes.

Note. The overall IPA crosshair values were importance = 5.8434 and performance = 5.805. Both archived priority items are located close to these reference values, so their quadrant placement should be interpreted cautiously.