Q-method is a pile-sorting method that systematically explores and analyses different perspectives on a particular topic based on the degree of agreement and disagreement among respondents across the same set of statements about that topic (Brown, 1980; Watts and Stenner, 2012). The Q-method combines quantitative analysis and qualitative analysis (ethnographic inquiry), which is based on factor analysis to find explanatory models, and qualitatively interprets these models based on the participants’ arrangement of relationships between statements to uncover any common patterns of engagement, orientation, or forms of understanding (Zabala and Pascual,2016). The advantage of the Q-method over other traditional survey-based methods is that it provides a systematic means of identifying and exploring the different perspectives on a topic represented in a selected group of participants, while avoiding the researcher’s bias against clustering participants (Brown,1996). Thus, the Q-method is particularly suitable for studying highly controversial phenomena and as a powerful method for examining subjectivity or perceptions about a particular topic (Cross, 2005). Q-method has been used for various research topics related to natural resource management, such as farmers’ motivations for crop diversification choices (Nordhagen et al.,2017; Buckwell et al., 2020), Indonesian perceptions of pulpwood plantations (Pirard et al., 2016), blueberry growers’ attitudes towards deforestation in Georgia, USA (Upadhaya and Dwivedi, 2019), heterogeneity in the motivations for adopting silvopastures (Zabala et al.,2017), etc. However, it has not been applied to issues related to WEP collection and consumption. For this study, the use of the Q-method was valuable because it helped the researchers to identify the diversity of potential perspectives on WEP collection and consumption and the groups that share similar patterns of thinking about behavioral explanations.

The study area is located in the Samaba Terraces in Honghe County, one of the ten areas of the “Honghe Hani Rice Terraces System”, and the study villages include Gata Village, Chaoyang Village, and Baohua Village. Honghe County is a national key county for poverty alleviation and development, which is located in the Ailao Mountain and on the south bank of the Red River, with 97% mountainous terrain of three types: low mountain valley area, semi-mountainous area, and mountainous area. It has a subtropical monsoon climate with abundant sunshine, an average annual temperature of 20 ℃, and average annual precipitation of 1340 mm (Liu et al., 2014; Zhang et al., 2017). The economy is mainly based on agriculture. Gata Village is 7 km from the government of Baohua town and 40 km from Honghe County, with a total land area of 11.8 km², an altitude of 1700-1800 m, an average annual temperature of 16 ℃, and average annual precipitation of 3200 mm. Crops grown in the village include rice, corn, and wheat. There are 2 km² of arable land and more than 6.5 km² of forest land, with agricultural income accounting for more than 95% of the total. Chaoyang Village is 3.70 km from the town, with an altitude of 1720 m. The annual average temperature is 14.3 ℃ and the annual precipitation is 1485 mm. Baohua Village is the seat of the township government, with an average altitude of 1830 m, an average annual temperature of 13.7 ℃ and an annual rainfall of 1470 mm, and there is a farmer product market in the village.

Conducting research using Q-method involves a series of steps. The first step is to create a Q sample, i.e., to develop a list of statements that includes as many ideas as possible about the research topic. The main challenge in developing the Q sample was to make them short and easily be understood by the participants. The statements in the Q sample was not facts, but rather views and opinions in a broad sense. Most of the statements in our study were obtained through focus group discussions with local farmers and statements by researchers, while a few statements were obtained by collating the text of declarations submitted to the FAO and scientific literature published in academic journals. Initially, we selected 30 statements, and this list was refined by considering that the size of the Q sample should be manageable, a long Q sample requires more time for analysis, and the same topic can be discussed from different perspectives in several statements if it is considered important. Therefore, by removing overly similar statements and piloting them in the field, 23 key statements were finally identified, covering five themes (Table 1). Based on the literature and previous field surveys, these themes were deemed the most relevant for explaining farmers’ perspectives on the collection and consumption of WEPs. The themes covered were (A) knowledge; (B) economics; (C) availability; (D) socioeconomic and political; and (E) cultural services.

The second step is to recruit participants and form a P sample. Participants were recruited purposefully, not randomly, to capture different perspectives (Dziopa and Ahern, 2011). For this study, we selected three villages in the core area of the heritage site. Due to the lack of census information, a strict random sampling was not possible, so participants were recruited for the sessions through local leaders. Participants were randomly selected and their composition was examined to ensure roughly representative distributions of age, economic status, market, and natural exposure. We ended up recruiting 34 farmer participants, and we first explained our study to the participants and then obtained their verbal informed consent. Four farmers discontinued the survey midway through due to the long duration of the survey. Ultimately, we ended up with a P sample of 30 WEP collectors and eaters, which is an acceptable and appropriate number of respondents for a Q-method study (Watts and Stenner, 2012).

The third step was to ask participants to sort Q statements using their subjectivity. In our study, the low literacy level and language barrier of local villagers were challenges. Since interviewing only literate people may bias the results, all of our work was conducted verbally even though some villagers were able to read and understand the written content. Three university students who were fluent in the local language were recruited for the study, and each of the three students was trained to interact one-on-one with the interviewees. The trained local university students instructed the participants to read the statements one by one and to classify the statements into two categories: agree and disagree. Each statement was then ranked in three groups (agree, disagree, and neutral), placed in a quasi-normal distribution grid, with the most agreeable and least agreeable statements at each end of the grid, and then the other statements were ranked and placed in the grid (Fig. 1). The person in charge of the study was on-site and always available to clarify questions. Throughout the Q-sorts process, participants could move statements between columns and re-sort them within columns. As statements were sorted, we asked participants to justify some choices to ensure that they fully understood the rules and made choices that reflected their true thoughts. Each interview lasted between 40 minutes and 1 hour. Finally, we administered a short survey to participants to collect basic demographic information, frequency of using wild plants, and other relevant information.

We used PQMethod software (available at http://schmolck.org/qmethod/downpqmac.htm) to analyze the collected data. We conducted the Principal Component Analysis (PCA) to extract the factors, and eight of the 23 components had eigenvalues greater than 1, with a maximum eigenvalue of 10.42. Since factors with eigenvalues less than one did not capture enough variation to explain even the views of individual participants, following Watts and Stenner (2012), we excluded factors with eigenvalues below 1.00. We selected four top factors for the explanation, and these four factors explained 61% of the variability among respondents, a percentage consistent with other Q studies (Watts and Stenner, 2012; Lynch et al.,2014). Seven respondents were not labeled because they had relatively high loadings on two or three factors, meaning that they were characterized by more than one viewpoint at the same time. Next, factor loadings, i.e., correlation coefficients between the factors and Q-sort, were calculated to show how much each Q-sort loaded on a particular factor: i.e., how much does that individual agree with this viewpoint? (Sandbrook et al., 2011). For clarity, we rotated the remaining factors using the Varimax module (QVARIMAX) to maximize the differences between the factors and make them as distinct as possible (Watts and Stenner, 2012). The rotated factors were re-expressed by ranking the factor z scores as statement scores (i.e., from -3 to 3). Thus, each group is represented by an “ideal” Q-sort, which shows how a hypothetical respondent with 100% loading on that factor would sort the statements (Watts and Stenner, 2012).

The Q-method analysis yields three main results: overall factor characteristics, factor loadings of respondents, and stated factor and z scores. The z scores are “the weighted average of the respondents most closely related to the factor for a given statement” ( Zabal and Pascual, 2016) and are used for the main results of the interpretation. The z-scores of the statements indicate how the prototypical respondents for each factor would rank the statements. The interpretation of the factors is based on the stated results, i.e., their salience in each factor and their unique position compared to other factors. Statements with significant differences in scores among the factors are differentiating statements, and they represent statements where there is significant disagreement among the factors. Statements with the smallest differences in scores between factors are consensus statements, and they indicate a common view and shared understanding of the issue.

The availability of WEPs is influenced by a number of factors. Our respondents generally believed that fewer species could be collected for WEPs, suggesting that WEP-related biodiversity is declining, which affects WEP availability. Changes in availability are the most common explanation affecting WEP use (Sõukand, 2016; Wijaya,2019). In our case sites, the availability of some commonly used plants did not decrease (e.g., fiddlehead, dulcamara, and herba houttuyniae are very commonly used in this area), but the variation in availability varied widely among the plants. Changes in plant collection were not conclusively proven in our study; some farmers suggested that changes in forest structure, quarrying activities, etc., had reduced WEPs, but others believed that changes in market conditions had increased WEP consumption. Some villagers suggested that changes in knowledge have led to many wild plants no longer being consumed and collected. The loss of traditional knowledge (Wijaya,2019) can also affect the availability of WEPs. Regarding the collection of WEPs, some interviewees noted changes during their lifetime, such as “when I was a child, I had more knowledge about WEPs and knew a lot of WEPs, especially the elderly collected more WEPs, and then this knowledge was gradually forgotten.” The traditional knowledge of the diversity and uses of some wild plants is being lost, indirectly affecting the availability of WEPs. On the other hand, there is a lack of knowledge of how to cook these traditional foods, and there is little innovation in cooking or processing methods. Therefore, closing the knowledge gap will potentially improve farmers’ knowledge, and the collection and consumption of WEPs.

Changes in livelihoods and lifestyles can also affect WEP collection and consumption. More and more people in the local area are engaging in gardening activities, and many farmers have started vegetable gardens in front of their gardens. Many people in this study felt that there was no conflict between the consumption of WEPs and the consumption of their home-grown vegetables, but WEP collection is still an inefficient task. The decrease in people’s efforts for WEP collection partly reflects the lower economic value of WEPs and their high substitutability. Also, people’s tastes are beginning to change, especially among young people who have less contact with nature, leading to a decrease in interest in some WEPs. The increasing number of fruits and vegetables traded in the market can also affect the use of WEPs. Although WEPs are still common in local restaurants of all types, there is a lack of species richness. Therefore, future research needs to investigate the market potential and economic value of WEPs in detail.

Pawera et al. (2020) studied the barriers, motivations, and reasons for the changes in WEP consumption and concluded that the main motivation for people to consume WEPs is free and “uncontaminated” natural foods, the main drivers of change are socioeconomic factors and changes in the market, and the barriers are decreasing availability and limited knowledge of WEP nutritional value. However, none of these factors were reflected in our study. In the Pawera et al. (2020) study, local markets were considered more convenient and increasingly common for purchasing and cooking ingredients, but this was not corroborated in our study. In some studies, laboring in the fields or forests is a characteristic of masculine identity, and thus, WEP gathering activities are usually performed by men (Acosta-Naranjo et al., 2020). In our study, even though Hani men are more involved in manual labor, some women will often participate in WEP collection.

Participants generally rejected the idea that WEPs are hunger food or for the poor. This suggests that locals have moved away from reliance on WEPs as a basic food and toward higher levels of need, such as cultural needs. Leisure is not the most important reason for the Hani to collect WEPs. WEP collection is an important activity in the lives of Hani people, and some respondents believe that WEP collection and consumption can bring them pleasure, but those who collect WEPs as recreation are still in the minority. The local people’s interest in WEP collection and consumption may come from the influence of traditional food culture. WEP collection is a local tradition and has developed a strong culture, as local people still enjoy eating WEPs and often engage in collection activities in their daily lives. The persistence of this strong culture has slowed the change in dietary structure (Pawera et al.,2020), which in turn has sustained WEP collection and consumption.

Although speculative, this study argues that a focus on local WEP-related cultural services and increased market development for WEPs would benefit WEP conservation and resource utilization. The fundamental role of WEPs is as a food that provides provisioning services. WEPs represent a traditional food that is often richer in micronutrients than cultivated crops (Hunter et al., 2019). In rural areas, WEPs offer the potential to alleviate micronutrient deficiencies in some cases. Some studies have linked the decline in wild food consumption to the perception of wild foods as food for the poor or as a reserve food in times of famine (Duguma, 2020). In the Hani Terraces, most respondents did not consider WEPs as “food only for the poor” and many denied that the use of wild plants was related to food scarcity, although in some areas WEPs were often associated with famine food (Muller and Almedom, 2008). In contrast, most respondents in this study had a positive attitude, associating WEPs with good taste, nutrition, and healthy, and some even argued that WEPs are a food eaten by the rich because they usually found in restaurants and the poor do not have the time to gather them. The health benefits provided by food are classified as cultural services in some studies (e.g., Reyes-García et al., 2015) and the same view is adopted in this paper. WEPs can provide rich cultural services (Wijaya,2019) such as entertainment (Pouta et al., 2006), health, and deliciousness. Contrary to Wijaya’s study, the idea that “limited knowledge of its nutritional and health benefits can influence WEP consumption” was not reflected in this study. Specifically, the statement “I would eat more wild foods if I knew about their nutritional and health benefits” was not supported, although there was strong agreement with the statement that “WEP consumption is beneficial to health.” The disagreement was strongest for the statement “WEPs plants are associated with lower social status.” These responses indicate that local people generally agree that WEPs are healthy and nutritious.

The use of WEPs tends to persist in indigenous communities, and in this study, the Hani people considered the consumption of WEPs to be a traditional local custom. People still enjoy and value the traditional foods of their people. Even so, the impact of cultural change on behavior toward WEPs was not assessed in this paper. WEPs are still consumed in many parts of the world, not only in subsistence-oriented economies but often in rural and even urban areas of developed countries. This is partly attributed to the supply of cultural services (Reyes-García et al., 2015), including recreation or the need for innovative foods, and the greater the cultural services of the WEP species, the greater biological benefits of the species are less necessary, and changes in demand for cultural services are thought to be the main cause of reduced WEP consumption (Reyes-García et al., 2015). In contrast to Wijaya’s (2009) study, respondents in this study perceived WEPs as usually tasty, which may be related to plant resource utilization and food culture.

In the Hani Terraces, WEPs are a part of the local people’s life, so local people do not pay much attention to the fact that WEPs are free and do not consider some plants as a source of household income, and only some of them believe that WEP consumption can save household expenses and bring indirect benefits. Many studies argue that the free availability of WEPs in nature leads to their low economic value and reduces their visibility and promotion (Ngome et al.,2017). In our study, there was a large variation in local perceptions of the economic value of WEPs. Many farmers believed that the economic value of WEPs was high and that increased marketability would increase their chances of consuming WEPs. However, even with the high economic value, most farmers would not participate in the sale of WEPs but would collect WEPs out of a desire to be self-sufficient. Therefore, farmers were neutral regarding indications that the government should promote the benefits of WEPs and increase the purchase price of WEPs. As the economy develops and more WEPs are being sold or purchased in the market, some people have acquired the knowledge of WEP domestication, and increased marketability will probably increase WEP collection and consumption.

In terms of methodology, this is the first time that the Q-method has been applied to the topic of WEP collection and consumption, especially in the context of the global trend of generally decreasing WEP consumption. This method has strengths and limitations. During the interview, participants generally reported good results. Some interviewees had difficulty prioritizing the statements, but this did not necessarily create analytical difficulties, as the Q-method aims to force participants to make trade-offs and prioritize (Nordhagen et al., 2017). Also, some new statements emerged during the study and these comments could only be analyzed as ranked interviews. While studies using the Q-method typically produce strong differences of opinion (Hiedanpää et al.,2020), the relatively broad overlapping of opinions in this study is not an uncommon phenomenon (e.g., Andrew et al., 2020). Finally, naming groups under the Q-method may give a simplistic and false impression: in fact, the factors represent not groups but a continuum along which people can expect to change, each with a different degree of association with the group (Fairweather and Klonsky,2009). In our preliminary study, specific factors influencing WEP availability were not collected and specific changes in WEP consumption were not measured in the field, and the incorporation of such factors could be enhanced in future studies. Overall, the method proved to be very useful for gaining a clearer understanding of farmers’ motivations or perceptions.