**3. Study of Forest SWB within the Upper Yasu Watershed (Study 2)**

A second study was conducted on the upper watershed, a subsection within the first study site [42]. The purpose of this second survey was to identify patterns of responses among populations that have more intense ties with forests. The upper watershed areas have higher ratios of forested areas, whose residents historically experience active forestry activities. The city in the upper watershed includes 81% of the all the six watershed cities [39]. Considering the entire watershed survey, it was found that residents in forested areas were less likely to derive forest SWB from forest-related activities, which is another motivation for this study.

Questionnaires were mailed from January through April 2018 to 6559 households in all postal codes of two upper watershed areas of the study area. Similar to the previous questionnaire, SWB, relationships (social capital), nature (forest-related activities), and other aspects of everyday life were questioned. A total of 1457 questionnaires were returned, with a response rate of 17.2%. The average age of the respondents was 59, consisting of 39% of female respondents while the average age of citizens of the upper watershed city was 46 and the ratio of female population was 50% [40].

In this study, forest SWB was measured using five items. Similar to study 1, the first item asked about personal affective evaluation towards local forests. The other four items assessed satisfaction, fulfillment (eudaimonia), positive affect, and negative affect, based on the Organization for Economic Co-operation and Development (OECD) guidelines for the measurement of SWB [43].

Forest satisfaction: Respondents were asked about their satisfaction using an 11 point scale, which ranged from "completely dissatisfied" (0) to "completely satisfied" (10) ("How satisfied are you with your current relationships with forests?"). Forest fulfillment (eudaimonia): Forest fulfillment was assessed on an 11-point scale ranging from "do not feel at all" (0) to "feel strongly" (10) ("How much worth, fulfilment, or sense of accomplishment do you feel regarding your relationships with forests?"). This question concerns eudaimonia, a concept first indicated by Aristotle in his Nicomachean Ethics, in which he asserts that people are happy not by feeling pleasure (hedonia), but by leading virtuous lives [44,45].

Positive and negative affect (feelings): Next, respondents were asked "How much have you experienced the following feelings during your experiences regarding forests?" For this item, respondents were provided the following list of feelings: forward-looking, backward-looking, pleasant, not pleasant, happy, sad, fearful, joyful, angry, satisfied, proud, shameful, awe, and respect. For each feeling, respondents were asked to choose one option on a five-point scale: 1 = "very rare," 2 = "rare," 3 = "sometimes," 4 = "frequently," and 5 = "very frequently." Based on the results of a factor analysis, by choosing items with factor loadings greater than 0.5, the authors determined the scores of the positive and negative affects by adding the scores of the five items for the positive affect (i.e., forward-looking, pleasant, happy, joyful, and satisfied), and the scores of two items for the negative affect (i.e., backward-looking and unpleasant).

Figure 3 presents the distributions of responses to the forest SWB items. The values in brackets indicate the average scores. Similar to the previous assessment, the responses are distributed widely, indicating sufficient variation in responses.

**Figure 3.** Forest SWB in the upper watershed study [42]. The authors gave permission to use this chart.

Figure 4 presents the distribution of forest SWB evaluated by respondents considering their affect (feelings). The values in brackets indicate the average. In general, respondents reported a higher frequency for positive affect than for negative affect. Nearly all indicators for the positive affect, such as "Forward-looking," "Pleasant," and "Happy," have higher scores than the mid-point (3.0) between 1 and 5; except for "Satisfied" [2.9] and "Proud" [2.6]. All the indicators for the negative affect, such as "Backward-looking," "Not Pleasant," "Sad," "Fearful," "Angry," and "Shameful," have lower scores than the mid-point (3.0). Although the number of responses in the second survey was 1457, the number of responses to complete the emotion items ranged from 647 to 842, because respondents who did not engage with forests were instructed not to answer these questions.

The validity of these measures was examined [42]. These measures are based on the manual for measuring SWB by the OECD [43]. While these measures concentrate on a specific domain, that is, relationships with forests, this treatment is justified by studies in psychology, in which a domain-specific SWB, such as the job and marriage of a participant, are measured [43]. Construct validity was verified using confirmatory factor analysis; the results obtained were satisfactory [42].

Ordinary least squares (OLS) regressions were performed in STATA to identify significant associations (*p* < 0.10) between forest SWB, demographic factors, and (i) natural, (ii) man-made, and (iii) social capitals, with positive and negative coefficients indicating the direction of these associations (Table 5). In the written explanation of the results to follow,

the numbers for each variable in Table 5 will be used in text to refer to their corresponding variables.


**Table 5.** Summary results of regression analyses in the upper watershed study: explanatory variables with positive and negative statistically significant coefficients \* (*p* < 0.10) [42].

\* Variables indicating built or manufactured capital, such as hospitals, are omitted from this table for simplicity. \*\* As the dependent variable was reversed, "positive" indicates that an increase in the explanatory variable leads to a decrease in the negative affect, and vice versa.

 study **Figure 4.** Forest SWB (positive and negative affect) in the upper watershed study [42]. The authors gave permission to use this chart.

Considering the groups of variables, certain demographic variables have positive or negative correlations with the indicators.


Next, we consider the variables indicating respondent behavior regarding forests.


Unexpectedly, the authors did not identify a relationship between the physical presence of the forests and forest SWB.

• The forest ratios of the respective postal areas where the respondents resided did not correlate with the forest SWB.

The adjusted R<sup>2</sup> values for models with dependent variables 2–6 (forest SWB) were 0.103, 0.116, 0.151, 0.160, and 0.107, respectively. All F-statistic *p*-values of the corresponding OLS were less than 0.0001. The adjusted R<sup>2</sup> values for models of general SWB were larger than those for the forest SWB (0.415).
